From 3618a97c20e386aa49a39c1f6af55ce7414ffc5c Mon Sep 17 00:00:00 2001 From: Aaron Ji <127167174+DresAaron@users.noreply.github.com> Date: Mon, 23 Sep 2024 13:45:09 +0800 Subject: [PATCH 01/79] feat: extend api params for Jina Embeddings V3 (#8657) --- .../model_providers/jina/jina.yaml | 43 +++++++++++++++++++ .../jina/text_embedding/text_embedding.py | 42 ++++++++++++++---- 2 files changed, 76 insertions(+), 9 deletions(-) diff --git a/api/core/model_runtime/model_providers/jina/jina.yaml b/api/core/model_runtime/model_providers/jina/jina.yaml index 23e18ad75f6885..9c70d6ff33ba41 100644 --- a/api/core/model_runtime/model_providers/jina/jina.yaml +++ b/api/core/model_runtime/model_providers/jina/jina.yaml @@ -67,3 +67,46 @@ model_credential_schema: required: false type: text-input default: '8192' + - variable: task + label: + zh_Hans: 下游任务 + en_US: Downstream task + placeholder: + zh_Hans: 选择将使用向量模型的下游任务。模型将返回针对该任务优化的向量。 + en_US: Select the downstream task for which the embeddings will be used. The model will return the optimized embeddings for that task. + required: false + type: select + options: + - value: retrieval.query + label: + en_US: retrieval.query + - value: retrieval.passage + label: + en_US: retrieval.passage + - value: separation + label: + en_US: separation + - value: classification + label: + en_US: classification + - value: text-matching + label: + en_US: text-matching + - variable: dimensions + label: + zh_Hans: 输出维度 + en_US: Output dimensions + placeholder: + zh_Hans: 输入您的输出维度 + en_US: Enter output dimensions + required: false + type: text-input + - variable: late_chunking + label: + zh_Hans: 后期分块 + en_US: Late chunking + placeholder: + zh_Hans: 应用后期分块技术来利用模型的长上下文功能来生成上下文块向量化。 + en_US: Apply the late chunking technique to leverage the model's long-context capabilities for generating contextual chunk embeddings. + required: false + type: switch diff --git a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py index ceb79567d54ca2..6c96699ea2e3f1 100644 --- a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py @@ -27,6 +27,38 @@ class JinaTextEmbeddingModel(TextEmbeddingModel): api_base: str = "https://api.jina.ai/v1" + def _to_payload(self, model: str, texts: list[str], credentials: dict) -> dict: + """ + Parse model credentials + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :return: parsed credentials + """ + + def transform_jina_input_text(model, text): + if model == "jina-clip-v1": + return {"text": text} + return text + + data = {"model": model, "input": [transform_jina_input_text(model, text) for text in texts]} + + task = credentials.get("task") + dimensions = credentials.get("dimensions") + late_chunking = credentials.get("late_chunking") + + if task is not None: + data["task"] = task + + if dimensions is not None: + data["dimensions"] = int(dimensions) + + if late_chunking is not None: + data["late_chunking"] = late_chunking + + return data + def _invoke( self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None ) -> TextEmbeddingResult: @@ -49,15 +81,7 @@ def _invoke( url = base_url + "/embeddings" headers = {"Authorization": "Bearer " + api_key, "Content-Type": "application/json"} - def transform_jina_input_text(model, text): - if model == "jina-clip-v1": - return {"text": text} - return text - - data = {"model": model, "input": [transform_jina_input_text(model, text) for text in texts]} - - if model == "jina-embeddings-v3": - data["task"] = "text-matching" + data = self._to_payload(model=model, texts=texts, credentials=credentials) try: response = post(url, headers=headers, data=dumps(data)) From 4c7beb9d7b5e566209600ea0975677c7d46a6926 Mon Sep 17 00:00:00 2001 From: haike-1213 <2100797950@qq.com> Date: Mon, 23 Sep 2024 07:23:52 +0000 Subject: [PATCH 02/79] fix: Assignment exception (#8663) Co-authored-by: fum --- api/core/app/apps/base_app_generate_response_converter.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/api/core/app/apps/base_app_generate_response_converter.py b/api/core/app/apps/base_app_generate_response_converter.py index c6855ac85494d6..62e79ec444a48a 100644 --- a/api/core/app/apps/base_app_generate_response_converter.py +++ b/api/core/app/apps/base_app_generate_response_converter.py @@ -75,10 +75,10 @@ def _get_simple_metadata(cls, metadata: dict[str, Any]): :return: """ # show_retrieve_source + updated_resources = [] if "retriever_resources" in metadata: - metadata["retriever_resources"] = [] for resource in metadata["retriever_resources"]: - metadata["retriever_resources"].append( + updated_resources.append( { "segment_id": resource["segment_id"], "position": resource["position"], @@ -87,6 +87,7 @@ def _get_simple_metadata(cls, metadata: dict[str, Any]): "content": resource["content"], } ) + metadata["retriever_resources"] = updated_resources # show annotation reply if "annotation_reply" in metadata: From 86f90fd9ff4905e5c1b85cd445466cfb3f2c457c Mon Sep 17 00:00:00 2001 From: Bowen Liang Date: Mon, 23 Sep 2024 15:28:57 +0800 Subject: [PATCH 03/79] chore: skip PLR6201 linter rule (#8666) --- api/core/tools/provider/builtin/firecrawl/tools/crawl.py | 4 ++-- api/core/tools/provider/builtin/firecrawl/tools/scrape.py | 4 ++-- api/pyproject.toml | 1 - 3 files changed, 4 insertions(+), 5 deletions(-) diff --git a/api/core/tools/provider/builtin/firecrawl/tools/crawl.py b/api/core/tools/provider/builtin/firecrawl/tools/crawl.py index 9675b8eb913351..15ab510c6c889c 100644 --- a/api/core/tools/provider/builtin/firecrawl/tools/crawl.py +++ b/api/core/tools/provider/builtin/firecrawl/tools/crawl.py @@ -35,10 +35,10 @@ def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMe scrapeOptions["excludeTags"] = get_array_params(tool_parameters, "excludeTags") scrapeOptions["onlyMainContent"] = tool_parameters.get("onlyMainContent", False) scrapeOptions["waitFor"] = tool_parameters.get("waitFor", 0) - scrapeOptions = {k: v for k, v in scrapeOptions.items() if v not in {None, ""}} + scrapeOptions = {k: v for k, v in scrapeOptions.items() if v not in (None, "")} payload["scrapeOptions"] = scrapeOptions or None - payload = {k: v for k, v in payload.items() if v not in {None, ""}} + payload = {k: v for k, v in payload.items() if v not in (None, "")} crawl_result = app.crawl_url(url=tool_parameters["url"], wait=wait_for_results, **payload) diff --git a/api/core/tools/provider/builtin/firecrawl/tools/scrape.py b/api/core/tools/provider/builtin/firecrawl/tools/scrape.py index 538b4a1fcbf056..f00a9b31ce8c2c 100644 --- a/api/core/tools/provider/builtin/firecrawl/tools/scrape.py +++ b/api/core/tools/provider/builtin/firecrawl/tools/scrape.py @@ -29,10 +29,10 @@ def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> list[ToolInv extract["schema"] = get_json_params(tool_parameters, "schema") extract["systemPrompt"] = tool_parameters.get("systemPrompt") extract["prompt"] = tool_parameters.get("prompt") - extract = {k: v for k, v in extract.items() if v not in {None, ""}} + extract = {k: v for k, v in extract.items() if v not in (None, "")} payload["extract"] = extract or None - payload = {k: v for k, v in payload.items() if v not in {None, ""}} + payload = {k: v for k, v in payload.items() if v not in (None, "")} crawl_result = app.scrape_url(url=tool_parameters["url"], **payload) markdown_result = crawl_result.get("data", {}).get("markdown", "") diff --git a/api/pyproject.toml b/api/pyproject.toml index 506f379aaf116f..066b4772a99184 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -28,7 +28,6 @@ select = [ "PLR0402", # manual-from-import "PLR1711", # useless-return "PLR1714", # repeated-equality-comparison - "PLR6201", # literal-membership "RUF019", # unnecessary-key-check "RUF100", # unused-noqa "RUF101", # redirected-noqa From b37954b9661067977f45524dce0bdb8863309b9b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Mon, 23 Sep 2024 15:33:06 +0800 Subject: [PATCH 04/79] fix: png avatar upload as jpeg (#8665) --- .../components/base/app-icon-picker/index.tsx | 2 +- .../components/base/app-icon-picker/utils.ts | 21 ++++++++++++++++++- 2 files changed, 21 insertions(+), 2 deletions(-) diff --git a/web/app/components/base/app-icon-picker/index.tsx b/web/app/components/base/app-icon-picker/index.tsx index 825481547501a8..ba375abdd9c04a 100644 --- a/web/app/components/base/app-icon-picker/index.tsx +++ b/web/app/components/base/app-icon-picker/index.tsx @@ -88,7 +88,7 @@ const AppIconPicker: FC = ({ if (!imageCropInfo) return setUploading(true) - const blob = await getCroppedImg(imageCropInfo.tempUrl, imageCropInfo.croppedAreaPixels) + const blob = await getCroppedImg(imageCropInfo.tempUrl, imageCropInfo.croppedAreaPixels, imageCropInfo.fileName) const file = new File([blob], imageCropInfo.fileName, { type: blob.type }) handleLocalFileUpload(file) } diff --git a/web/app/components/base/app-icon-picker/utils.ts b/web/app/components/base/app-icon-picker/utils.ts index 0c90e96febda83..14c9ae3f289d7d 100644 --- a/web/app/components/base/app-icon-picker/utils.ts +++ b/web/app/components/base/app-icon-picker/utils.ts @@ -11,6 +11,23 @@ export function getRadianAngle(degreeValue: number) { return (degreeValue * Math.PI) / 180 } +export function getMimeType(fileName: string): string { + const extension = fileName.split('.').pop()?.toLowerCase() + switch (extension) { + case 'png': + return 'image/png' + case 'jpg': + case 'jpeg': + return 'image/jpeg' + case 'gif': + return 'image/gif' + case 'webp': + return 'image/webp' + default: + return 'image/jpeg' + } +} + /** * Returns the new bounding area of a rotated rectangle. */ @@ -31,12 +48,14 @@ export function rotateSize(width: number, height: number, rotation: number) { export default async function getCroppedImg( imageSrc: string, pixelCrop: { x: number; y: number; width: number; height: number }, + fileName: string, rotation = 0, flip = { horizontal: false, vertical: false }, ): Promise { const image = await createImage(imageSrc) const canvas = document.createElement('canvas') const ctx = canvas.getContext('2d') + const mimeType = getMimeType(fileName) if (!ctx) throw new Error('Could not create a canvas context') @@ -93,6 +112,6 @@ export default async function getCroppedImg( resolve(file) else reject(new Error('Could not create a blob')) - }, 'image/jpeg') + }, mimeType) }) } From c66cecaa555722b7f82e069f58c3874bf39c771f Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Mon, 23 Sep 2024 16:18:55 +0800 Subject: [PATCH 05/79] add Qwen model translate (#8674) --- .../tongyi/llm/farui-plus.yaml | 4 +++ .../tongyi/llm/qwen-coder-turbo-0919.yaml | 4 +++ .../tongyi/llm/qwen-coder-turbo-latest.yaml | 4 +++ .../tongyi/llm/qwen-coder-turbo.yaml | 4 +++ .../model_providers/tongyi/llm/qwen-long.yaml | 4 +++ .../tongyi/llm/qwen-math-plus-0816.yaml | 4 +++ .../tongyi/llm/qwen-math-plus-0919.yaml | 4 +++ .../tongyi/llm/qwen-math-plus-latest.yaml | 4 +++ .../tongyi/llm/qwen-math-plus.yaml | 4 +++ .../tongyi/llm/qwen-math-turbo-0919.yaml | 4 +++ .../tongyi/llm/qwen-math-turbo-latest.yaml | 4 +++ .../tongyi/llm/qwen-math-turbo.yaml | 4 +++ .../tongyi/llm/qwen-max-0107.yaml | 4 +++ .../tongyi/llm/qwen-max-0403.yaml | 4 +++ .../tongyi/llm/qwen-max-0428.yaml | 4 +++ .../tongyi/llm/qwen-max-0919.yaml | 4 +++ .../tongyi/llm/qwen-max-latest.yaml | 4 +++ .../tongyi/llm/qwen-max-longcontext.yaml | 4 +++ .../model_providers/tongyi/llm/qwen-max.yaml | 4 +++ .../tongyi/llm/qwen-plus-0206.yaml | 4 +++ .../tongyi/llm/qwen-plus-0624.yaml | 4 +++ .../tongyi/llm/qwen-plus-0723.yaml | 4 +++ .../tongyi/llm/qwen-plus-0806.yaml | 4 +++ .../tongyi/llm/qwen-plus-0919.yaml | 4 +++ .../tongyi/llm/qwen-plus-chat.yaml | 4 +++ .../tongyi/llm/qwen-plus-latest.yaml | 4 +++ .../model_providers/tongyi/llm/qwen-plus.yaml | 4 +++ .../tongyi/llm/qwen-turbo-0206.yaml | 4 +++ .../tongyi/llm/qwen-turbo-0624.yaml | 4 +++ .../tongyi/llm/qwen-turbo-0919.yaml | 4 +++ .../tongyi/llm/qwen-turbo-chat.yaml | 4 +++ .../tongyi/llm/qwen-turbo-latest.yaml | 4 +++ .../tongyi/llm/qwen-turbo.yaml | 4 +++ .../tongyi/llm/qwen-vl-max-0809.yaml | 30 +++++++++++++++++++ .../tongyi/llm/qwen-vl-max.yaml | 30 +++++++++++++++++++ .../tongyi/llm/qwen-vl-plus-0201.yaml | 30 +++++++++++++++++++ .../tongyi/llm/qwen-vl-plus-0809.yaml | 30 +++++++++++++++++++ .../tongyi/llm/qwen-vl-plus.yaml | 30 +++++++++++++++++++ .../tongyi/llm/qwen2-math-1.5b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2-math-72b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2-math-7b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-0.5b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-1.5b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-14b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-32b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-3b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-72b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-7b-instruct.yaml | 4 +++ .../tongyi/llm/qwen2.5-coder-7b-instruct.yaml | 4 +++ 49 files changed, 326 insertions(+) diff --git a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml index aad07f56736e52..e5de586c1caf16 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml index ebba565d572aec..6ab39cde2d052d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml index 361e2c2373d652..be6d9a0e07272e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml index f4032a4dd316d8..d2aca4f514ae09 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml index dbe7d024a50f17..a59a3350f62dda 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml @@ -63,6 +63,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -70,6 +71,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml index 89d1302abef70d..cab7233c98d691 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml index 032b3c970d7f46..f82fba0c012aa8 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml index 31dd9f69725760..e2fb6e0e554d46 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml index 1a51d57f7814ba..8803e747e5eb7e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml index 1894eea417a43d..0dc5a066f03001 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml index b8365618b0499f..2ac0e4692a701b 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml index 8d346d691e6b2f..9a7f1312e97246 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml index c0ad12b85e5469..c0eef375570d40 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml index b00fb44d29fa72..c12444bd7b57b1 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml index 1848dcc07d1853..173c55b6b9796f 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml index 238882bb121898..692a38140dbed0 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml index 9d7d3c2fcbf67e..afd7fb4b77696d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml index a7bdc42f7347e5..d02ba7af18318a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml index 57888406afa0a3..c6a64dc507497e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml index 1e0b81661731a1..1111298c3712ac 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml index f70c373922fb3a..ef8dd083ade5dd 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml index c6007e9164ca2c..87a4417df59b8a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml index 2f53c433366988..967f258fa91626 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml index 90b54ca52e2d44..9d44852ac9cd5e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml index 59e8851240b0ac..df9448ae04ddf2 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml index 2a821dbcfe0011..32ccb8d6159034 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml index 626884f4b29542..f3fce30209e48f 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml index 844fced77a684a..bf976b518a9114 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml index 0152f75579e5c9..060e7fb4c9d092 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml index 19c6c8d293f4e3..97cd34929b4f4d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml index f557f311ef9c6d..8d77ba7a2a698a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml index be2475847ef22c..4458c706aad40e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml index 90f13dc19f5ae3..33f05967c2a462 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml @@ -62,6 +62,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -69,6 +70,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml index 41d45966e9d628..12e9e0dd569f93 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml @@ -9,6 +9,15 @@ model_properties: mode: chat context_size: 32000 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. - name: top_p use_template: top_p type: float @@ -50,6 +59,27 @@ parameter_rules: en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. - name: response_format use_template: response_format + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: enable_search + type: boolean + default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. + - name: response_format + use_template: response_format pricing: input: '0.02' output: '0.02' diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml index 78d0509374131e..b811fdece48f5a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml @@ -9,6 +9,15 @@ model_properties: mode: chat context_size: 32000 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. - name: top_p use_template: top_p type: float @@ -50,6 +59,27 @@ parameter_rules: en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. - name: response_format use_template: response_format + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: enable_search + type: boolean + default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. + - name: response_format + use_template: response_format pricing: input: '0.02' output: '0.02' diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml index 8944388b1ee5a2..188dea389a7e21 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml @@ -9,6 +9,15 @@ model_properties: mode: chat context_size: 8000 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. - name: top_p use_template: top_p type: float @@ -50,6 +59,27 @@ parameter_rules: en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. - name: response_format use_template: response_format + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: enable_search + type: boolean + default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. + - name: response_format + use_template: response_format pricing: input: '0.02' output: '0.02' diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml index 869e0ea71c01b1..bc623e2f032110 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml @@ -9,6 +9,15 @@ model_properties: mode: chat context_size: 32768 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. - name: top_p use_template: top_p type: float @@ -50,6 +59,27 @@ parameter_rules: en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. - name: response_format use_template: response_format + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: enable_search + type: boolean + default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. + - name: response_format + use_template: response_format pricing: input: '0.008' output: '0.008' diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml index da11bacc646c87..8977e12e4f2d90 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml @@ -9,6 +9,15 @@ model_properties: mode: chat context_size: 8000 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. - name: top_p use_template: top_p type: float @@ -50,6 +59,27 @@ parameter_rules: en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. - name: response_format use_template: response_format + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: enable_search + type: boolean + default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. + - name: response_format + use_template: response_format pricing: input: '0.008' output: '0.008' diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml index cfe4b5a6662a37..de237842af7d8e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml index e541c197b0ff4f..1fda35abaf1fec 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml index ba4514e3d6eadb..06fd33c5f47400 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml index e5596041af6f21..ebf809955337b4 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml index 4004c59417107a..e9bc99339de0b1 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml index d8f53666ced415..3ed85dade82817 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml index 890f7e6e4e5d96..328519c16822ff 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml index 6d3d2dd5bb5e61..d1ed3c2a73edf1 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml index 17d0eb5b351d85..0e88c24aa83475 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml index 435b3f90a2be53..35313cd1f70947 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml index 435b3f90a2be53..35313cd1f70947 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml @@ -60,6 +60,7 @@ parameter_rules: type: float default: 1.1 label: + zh_Hans: 重复惩罚 en_US: Repetition penalty help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 @@ -67,6 +68,9 @@ parameter_rules: - name: enable_search type: boolean default: false + label: + zh_Hans: 联网搜索 + en_US: Web Search help: zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. From 3554a803e70152e1e5064e2163d5dc13d389264d Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Mon, 23 Sep 2024 16:19:42 +0800 Subject: [PATCH 06/79] add zhipuai web search (#8668) --- .../model_providers/zhipuai/llm/glm-4-0520.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm-4-air.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm-4-airx.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm-4-flash.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm_3_turbo.yaml | 9 +++++++++ .../model_runtime/model_providers/zhipuai/llm/glm_4.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm_4_long.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm_4_plus.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm_4v.yaml | 9 +++++++++ .../model_providers/zhipuai/llm/glm_4v_plus.yaml | 9 +++++++++ 10 files changed, 90 insertions(+) diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-0520.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-0520.yaml index b1f9b7485cd9dd..7fcf6922023fbf 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-0520.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-0520.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.1' output: '0.1' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-air.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-air.yaml index 4e7d5fd3cc9775..fcd7c7768c7179 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-air.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-air.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.001' output: '0.001' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-airx.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-airx.yaml index 14f17db5d6c06a..c9ae5abf196360 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-airx.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-airx.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.01' output: '0.01' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-flash.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-flash.yaml index 3361474d737df6..98c4f72c7237f7 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-flash.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm-4-flash.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0' output: '0' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_3_turbo.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_3_turbo.yaml index bf0135d1985481..0b5391ce2f83fd 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_3_turbo.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_3_turbo.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 8192 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.001' output: '0.001' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4.yaml index ab4b32dd826b00..62f453fb775b9b 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.1' output: '0.1' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_long.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_long.yaml index d1b01731f54632..350b080c3fc11b 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_long.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_long.yaml @@ -49,6 +49,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.001' output: '0.001' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_plus.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_plus.yaml index 9ede308f18d55a..2d7ebd71cf26e1 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_plus.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4_plus.yaml @@ -46,6 +46,15 @@ parameter_rules: default: 1024 min: 1 max: 4095 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.05' output: '0.05' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v.yaml index 28286580a7ec1b..3a1120ff375c19 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v.yaml @@ -44,6 +44,15 @@ parameter_rules: default: 1024 min: 1 max: 1024 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.05' output: '0.05' diff --git a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v_plus.yaml b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v_plus.yaml index 4c5fa2403413ea..14b9623e5a8c44 100644 --- a/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v_plus.yaml +++ b/api/core/model_runtime/model_providers/zhipuai/llm/glm_4v_plus.yaml @@ -44,6 +44,15 @@ parameter_rules: default: 1024 min: 1 max: 1024 + - name: web_search + type: boolean + label: + zh_Hans: 联网搜索 + en_US: Web Search + default: false + help: + zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 + en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. pricing: input: '0.01' output: '0.01' From a126d535cf146ce5473974bfbe500d209a5b1693 Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Mon, 23 Sep 2024 16:39:46 +0800 Subject: [PATCH 07/79] add Spark Max-32K (#8676) --- .../model_providers/spark/llm/_client.py | 3 +- .../model_providers/spark/llm/_position.yaml | 1 + .../spark/llm/spark-max-32k.yaml | 33 +++++++++++++++++++ 3 files changed, 36 insertions(+), 1 deletion(-) create mode 100644 api/core/model_runtime/model_providers/spark/llm/spark-max-32k.yaml diff --git a/api/core/model_runtime/model_providers/spark/llm/_client.py b/api/core/model_runtime/model_providers/spark/llm/_client.py index b99a657e715edb..48911f657a52e3 100644 --- a/api/core/model_runtime/model_providers/spark/llm/_client.py +++ b/api/core/model_runtime/model_providers/spark/llm/_client.py @@ -25,6 +25,7 @@ def __init__(self, model: str, app_id: str, api_key: str, api_secret: str, api_d "spark-pro": {"version": "v3.1", "chat_domain": "generalv3"}, "spark-pro-128k": {"version": "pro-128k", "chat_domain": "pro-128k"}, "spark-max": {"version": "v3.5", "chat_domain": "generalv3.5"}, + "spark-max-32k": {"version": "max-32k", "chat_domain": "max-32k"}, "spark-4.0-ultra": {"version": "v4.0", "chat_domain": "4.0Ultra"}, } @@ -32,7 +33,7 @@ def __init__(self, model: str, app_id: str, api_key: str, api_secret: str, api_d self.chat_domain = model_api_configs[model]["chat_domain"] - if model == "spark-pro-128k": + if model in ["spark-pro-128k", "spark-max-32k"]: self.api_base = f"wss://{domain}/{endpoint}/{api_version}" else: self.api_base = f"wss://{domain}/{api_version}/{endpoint}" diff --git a/api/core/model_runtime/model_providers/spark/llm/_position.yaml b/api/core/model_runtime/model_providers/spark/llm/_position.yaml index 458397f2aaf1c6..73f39cb1197b48 100644 --- a/api/core/model_runtime/model_providers/spark/llm/_position.yaml +++ b/api/core/model_runtime/model_providers/spark/llm/_position.yaml @@ -1,3 +1,4 @@ +- spark-max-32k - spark-4.0-ultra - spark-max - spark-pro-128k diff --git a/api/core/model_runtime/model_providers/spark/llm/spark-max-32k.yaml b/api/core/model_runtime/model_providers/spark/llm/spark-max-32k.yaml new file mode 100644 index 00000000000000..1a1ab6844c69c5 --- /dev/null +++ b/api/core/model_runtime/model_providers/spark/llm/spark-max-32k.yaml @@ -0,0 +1,33 @@ +model: spark-max-32k +label: + en_US: Spark Max-32K +model_type: llm +model_properties: + mode: chat +parameter_rules: + - name: temperature + use_template: temperature + default: 0.5 + help: + zh_Hans: 核采样阈值。用于决定结果随机性，取值越高随机性越强即相同的问题得到的不同答案的可能性越高。 + en_US: Kernel sampling threshold. Used to determine the randomness of the results. The higher the value, the stronger the randomness, that is, the higher the possibility of getting different answers to the same question. + - name: max_tokens + use_template: max_tokens + default: 4096 + min: 1 + max: 8192 + help: + zh_Hans: 模型回答的tokens的最大长度。 + en_US: Maximum length of tokens for the model response. + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + default: 4 + min: 1 + max: 6 + help: + zh_Hans: 从 k 个候选中随机选择一个（非等概率）。 + en_US: Randomly select one from k candidates (non-equal probability). + required: false From c7eacd1aacf7c67e9ec10c885d6f46bee1333fc2 Mon Sep 17 00:00:00 2001 From: Nam Vu Date: Mon, 23 Sep 2024 17:40:40 +0700 Subject: [PATCH 08/79] chore: Optimize I18nObject class for better performance and readability (#8681) --- api/core/tools/entities/common_entities.py | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) diff --git a/api/core/tools/entities/common_entities.py b/api/core/tools/entities/common_entities.py index b52119fdc4759b..924e6fc0cf9f17 100644 --- a/api/core/tools/entities/common_entities.py +++ b/api/core/tools/entities/common_entities.py @@ -1,6 +1,6 @@ from typing import Optional -from pydantic import BaseModel +from pydantic import BaseModel, Field class I18nObject(BaseModel): @@ -8,19 +8,16 @@ class I18nObject(BaseModel): Model class for i18n object. """ - zh_Hans: Optional[str] = None - pt_BR: Optional[str] = None - ja_JP: Optional[str] = None en_US: str + zh_Hans: Optional[str] = Field(default=None) + pt_BR: Optional[str] = Field(default=None) + ja_JP: Optional[str] = Field(default=None) def __init__(self, **data): super().__init__(**data) - if not self.zh_Hans: - self.zh_Hans = self.en_US - if not self.pt_BR: - self.pt_BR = self.en_US - if not self.ja_JP: - self.ja_JP = self.en_US + self.zh_Hans = self.zh_Hans or self.en_US + self.pt_BR = self.pt_BR or self.en_US + self.ja_JP = self.ja_JP or self.en_US def to_dict(self) -> dict: return {"zh_Hans": self.zh_Hans, "en_US": self.en_US, "pt_BR": self.pt_BR, "ja_JP": self.ja_JP} From 11d09a92d0cd954aa81e6cbcfad961e6d8ddb815 Mon Sep 17 00:00:00 2001 From: Hash Brown Date: Mon, 23 Sep 2024 18:44:09 +0800 Subject: [PATCH 09/79] fix: send message error when last sent message not succeeded (#8682) --- .../debug/debug-with-single-model/index.tsx | 13 ++++--------- .../base/chat/chat-with-history/chat-wrapper.tsx | 13 ++++--------- .../base/chat/embedded-chatbot/chat-wrapper.tsx | 13 ++++--------- web/app/components/base/chat/types.ts | 2 +- web/app/components/base/chat/utils.ts | 10 ++++++++++ .../panel/debug-and-preview/chat-wrapper.tsx | 13 ++++--------- 6 files changed, 27 insertions(+), 37 deletions(-) diff --git a/web/app/components/app/configuration/debug/debug-with-single-model/index.tsx b/web/app/components/app/configuration/debug/debug-with-single-model/index.tsx index 5faef46d988e0d..2de500a3a6211b 100644 --- a/web/app/components/app/configuration/debug/debug-with-single-model/index.tsx +++ b/web/app/components/app/configuration/debug/debug-with-single-model/index.tsx @@ -22,6 +22,7 @@ import { import Avatar from '@/app/components/base/avatar' import { useAppContext } from '@/context/app-context' import { ModelFeatureEnum } from '@/app/components/header/account-setting/model-provider-page/declarations' +import { getLastAnswer } from '@/app/components/base/chat/utils' type DebugWithSingleModelProps = { checkCanSend?: () => boolean @@ -83,17 +84,11 @@ const DebugWithSingleModel = forwardRef { diff --git a/web/app/components/base/chat/chat-with-history/chat-wrapper.tsx b/web/app/components/base/chat/chat-with-history/chat-wrapper.tsx index 225bbac714084c..9d7b360f38d066 100644 --- a/web/app/components/base/chat/chat-with-history/chat-wrapper.tsx +++ b/web/app/components/base/chat/chat-with-history/chat-wrapper.tsx @@ -6,6 +6,7 @@ import type { OnSend, } from '../types' import { useChat } from '../chat/hooks' +import { getLastAnswer } from '../utils' import { useChatWithHistoryContext } from './context' import Header from './header' import ConfigPanel from './config-panel' @@ -67,17 +68,11 @@ const ChatWrapper = () => { }, []) const doSend: OnSend = useCallback((message, files, last_answer) => { - const lastAnswer = chatListRef.current.at(-1) - const data: any = { query: message, inputs: currentConversationId ? currentConversationItem?.inputs : newConversationInputs, conversation_id: currentConversationId, - parent_message_id: last_answer?.id || (lastAnswer - ? lastAnswer.isOpeningStatement - ? null - : lastAnswer.id - : null), + parent_message_id: last_answer?.id || getLastAnswer(chatListRef.current)?.id || null, } if (appConfig?.file_upload?.image.enabled && files?.length) @@ -111,13 +106,13 @@ const ChatWrapper = () => { const prevMessages = chatList.slice(0, index) const question = prevMessages.pop() - const lastAnswer = prevMessages.at(-1) + const lastAnswer = getLastAnswer(prevMessages) if (!question) return handleUpdateChatList(prevMessages) - doSend(question.content, question.message_files, (!lastAnswer || lastAnswer.isOpeningStatement) ? undefined : lastAnswer) + doSend(question.content, question.message_files, lastAnswer) }, [chatList, handleUpdateChatList, doSend]) const chatNode = useMemo(() => { diff --git a/web/app/components/base/chat/embedded-chatbot/chat-wrapper.tsx b/web/app/components/base/chat/embedded-chatbot/chat-wrapper.tsx index ed2f24274dcd7b..b97c940eece7d4 100644 --- a/web/app/components/base/chat/embedded-chatbot/chat-wrapper.tsx +++ b/web/app/components/base/chat/embedded-chatbot/chat-wrapper.tsx @@ -6,6 +6,7 @@ import type { OnSend, } from '../types' import { useChat } from '../chat/hooks' +import { getLastAnswer } from '../utils' import { useEmbeddedChatbotContext } from './context' import ConfigPanel from './config-panel' import { isDify } from './utils' @@ -69,17 +70,11 @@ const ChatWrapper = () => { }, []) const doSend: OnSend = useCallback((message, files, last_answer) => { - const lastAnswer = chatListRef.current.at(-1) - const data: any = { query: message, inputs: currentConversationId ? currentConversationItem?.inputs : newConversationInputs, conversation_id: currentConversationId, - parent_message_id: last_answer?.id || (lastAnswer - ? lastAnswer.isOpeningStatement - ? null - : lastAnswer.id - : null), + parent_message_id: last_answer?.id || getLastAnswer(chatListRef.current)?.id || null, } if (appConfig?.file_upload?.image.enabled && files?.length) @@ -113,13 +108,13 @@ const ChatWrapper = () => { const prevMessages = chatList.slice(0, index) const question = prevMessages.pop() - const lastAnswer = prevMessages.at(-1) + const lastAnswer = getLastAnswer(prevMessages) if (!question) return handleUpdateChatList(prevMessages) - doSend(question.content, question.message_files, (!lastAnswer || lastAnswer.isOpeningStatement) ? undefined : lastAnswer) + doSend(question.content, question.message_files, lastAnswer) }, [chatList, handleUpdateChatList, doSend]) const chatNode = useMemo(() => { diff --git a/web/app/components/base/chat/types.ts b/web/app/components/base/chat/types.ts index 489dbb44cf67e7..0bc50518ebb814 100644 --- a/web/app/components/base/chat/types.ts +++ b/web/app/components/base/chat/types.ts @@ -63,7 +63,7 @@ export type ChatItem = IChatItem & { conversationId?: string } -export type OnSend = (message: string, files?: VisionFile[], last_answer?: ChatItem) => void +export type OnSend = (message: string, files?: VisionFile[], last_answer?: ChatItem | null) => void export type OnRegenerate = (chatItem: ChatItem) => void diff --git a/web/app/components/base/chat/utils.ts b/web/app/components/base/chat/utils.ts index e851c4c4633fb5..305df5995d128b 100644 --- a/web/app/components/base/chat/utils.ts +++ b/web/app/components/base/chat/utils.ts @@ -19,6 +19,15 @@ function getProcessedInputsFromUrlParams(): Record { return inputs } +function getLastAnswer(chatList: ChatItem[]) { + for (let i = chatList.length - 1; i >= 0; i--) { + const item = chatList[i] + if (item.isAnswer && !item.isOpeningStatement) + return item + } + return null +} + function appendQAToChatList(chatList: ChatItem[], item: any) { // we append answer first and then question since will reverse the whole chatList later chatList.push({ @@ -71,5 +80,6 @@ function getPrevChatList(fetchedMessages: any[]) { export { getProcessedInputsFromUrlParams, + getLastAnswer, getPrevChatList, } diff --git a/web/app/components/workflow/panel/debug-and-preview/chat-wrapper.tsx b/web/app/components/workflow/panel/debug-and-preview/chat-wrapper.tsx index 86519af60357d7..230b2d7fa0c65c 100644 --- a/web/app/components/workflow/panel/debug-and-preview/chat-wrapper.tsx +++ b/web/app/components/workflow/panel/debug-and-preview/chat-wrapper.tsx @@ -25,6 +25,7 @@ import { stopChatMessageResponding, } from '@/service/debug' import { useStore as useAppStore } from '@/app/components/app/store' +import { getLastAnswer } from '@/app/components/base/chat/utils' type ChatWrapperProps = { showConversationVariableModal: boolean @@ -76,19 +77,13 @@ const ChatWrapper = forwardRef(({ showConv ) const doSend = useCallback((query, files, last_answer) => { - const lastAnswer = chatListRef.current.at(-1) - handleSend( { query, files, inputs: workflowStore.getState().inputs, conversation_id: conversationId, - parent_message_id: last_answer?.id || (lastAnswer - ? lastAnswer.isOpeningStatement - ? null - : lastAnswer.id - : null), + parent_message_id: last_answer?.id || getLastAnswer(chatListRef.current)?.id || null, }, { onGetSuggestedQuestions: (messageId, getAbortController) => fetchSuggestedQuestions(appDetail!.id, messageId, getAbortController), @@ -103,13 +98,13 @@ const ChatWrapper = forwardRef(({ showConv const prevMessages = chatList.slice(0, index) const question = prevMessages.pop() - const lastAnswer = prevMessages.at(-1) + const lastAnswer = getLastAnswer(prevMessages) if (!question) return handleUpdateChatList(prevMessages) - doSend(question.content, question.message_files, (!lastAnswer || lastAnswer.isOpeningStatement) ? undefined : lastAnswer) + doSend(question.content, question.message_files, lastAnswer) }, [chatList, handleUpdateChatList, doSend]) useImperativeHandle(ref, () => { From 52da5b16e7002fc15343d2df73838b3f8220cccd Mon Sep 17 00:00:00 2001 From: Likename Haojie Date: Mon, 23 Sep 2024 18:44:24 +0800 Subject: [PATCH 10/79] fixbug tts(stream) not work on ios safari(17.1+) (#8645) Co-authored-by: crazywoola <427733928@qq.com> --- web/app/components/base/audio-btn/audio.ts | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/web/app/components/base/audio-btn/audio.ts b/web/app/components/base/audio-btn/audio.ts index a61fd085d44c45..baf675d0be92d2 100644 --- a/web/app/components/base/audio-btn/audio.ts +++ b/web/app/components/base/audio-btn/audio.ts @@ -12,7 +12,7 @@ export default class AudioPlayer { mediaSource: MediaSource | null audio: HTMLAudioElement audioContext: AudioContext - sourceBuffer?: SourceBuffer + sourceBuffer?: any cacheBuffers: ArrayBuffer[] = [] pauseTimer: number | null = null msgId: string | undefined @@ -33,7 +33,7 @@ export default class AudioPlayer { this.callback = callback // Compatible with iphone ios17 ManagedMediaSource - const MediaSource = window.MediaSource || window.ManagedMediaSource + const MediaSource = window.ManagedMediaSource || window.MediaSource if (!MediaSource) { Toast.notify({ message: 'Your browser does not support audio streaming, if you are using an iPhone, please update to iOS 17.1 or later.', @@ -43,6 +43,10 @@ export default class AudioPlayer { this.mediaSource = MediaSource ? new MediaSource() : null this.audio = new Audio() this.setCallback(callback) + if (!window.MediaSource) { // if use ManagedMediaSource + this.audio.disableRemotePlayback = true + this.audio.controls = true + } this.audio.src = this.mediaSource ? URL.createObjectURL(this.mediaSource) : '' this.audio.autoplay = true From 4f69adc8ab0b67d45ff5f578502c521086f0ba13 Mon Sep 17 00:00:00 2001 From: Vikey Chen Date: Mon, 23 Sep 2024 18:45:10 +0800 Subject: [PATCH 11/79] fix: document_create_args_validate (#8569) --- api/services/dataset_service.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/api/services/dataset_service.py b/api/services/dataset_service.py index 30c010ef29f623..e96f06ed40fa4f 100644 --- a/api/services/dataset_service.py +++ b/api/services/dataset_service.py @@ -1100,8 +1100,8 @@ def document_create_args_validate(cls, args: dict): DocumentService.data_source_args_validate(args) DocumentService.process_rule_args_validate(args) else: - if ("data_source" not in args and not args["data_source"]) and ( - "process_rule" not in args and not args["process_rule"] + if ("data_source" not in args or not args["data_source"]) and ( + "process_rule" not in args or not args["process_rule"] ): raise ValueError("Data source or Process rule is required") else: From d7aada38a111028ade07b597ff2f9e201d57e35b Mon Sep 17 00:00:00 2001 From: ice yao Date: Mon, 23 Sep 2024 19:57:21 +0800 Subject: [PATCH 12/79] Add nomic embedding model provider (#8640) --- .../model_providers/nomic/__init__.py | 0 .../nomic/_assets/icon_l_en.svg | 13 ++ .../nomic/_assets/icon_s_en.png | Bin 0 -> 25814 bytes .../model_providers/nomic/_common.py | 28 +++ .../model_providers/nomic/nomic.py | 26 +++ .../model_providers/nomic/nomic.yaml | 29 +++ .../nomic/text_embedding/__init__.py | 0 .../text_embedding/nomic-embed-text-v1.5.yaml | 8 + .../text_embedding/nomic-embed-text-v1.yaml | 8 + .../nomic/text_embedding/text_embedding.py | 170 ++++++++++++++++++ api/poetry.lock | 78 +++++++- api/pyproject.toml | 2 + .../model_runtime/__mock/nomic_embeddings.py | 59 ++++++ .../model_runtime/nomic/__init__.py | 0 .../model_runtime/nomic/test_embeddings.py | 62 +++++++ .../model_runtime/nomic/test_provider.py | 22 +++ dev/pytest/pytest_model_runtime.sh | 3 +- 17 files changed, 506 insertions(+), 2 deletions(-) create mode 100644 api/core/model_runtime/model_providers/nomic/__init__.py create mode 100644 api/core/model_runtime/model_providers/nomic/_assets/icon_l_en.svg create mode 100644 api/core/model_runtime/model_providers/nomic/_assets/icon_s_en.png create mode 100644 api/core/model_runtime/model_providers/nomic/_common.py create mode 100644 api/core/model_runtime/model_providers/nomic/nomic.py create mode 100644 api/core/model_runtime/model_providers/nomic/nomic.yaml create mode 100644 api/core/model_runtime/model_providers/nomic/text_embedding/__init__.py create mode 100644 api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.5.yaml create mode 100644 api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.yaml create mode 100644 api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py create mode 100644 api/tests/integration_tests/model_runtime/__mock/nomic_embeddings.py create mode 100644 api/tests/integration_tests/model_runtime/nomic/__init__.py create mode 100644 api/tests/integration_tests/model_runtime/nomic/test_embeddings.py create mode 100644 api/tests/integration_tests/model_runtime/nomic/test_provider.py diff --git a/api/core/model_runtime/model_providers/nomic/__init__.py b/api/core/model_runtime/model_providers/nomic/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/nomic/_assets/icon_l_en.svg b/api/core/model_runtime/model_providers/nomic/_assets/icon_l_en.svg new file mode 100644 index 00000000000000..6c4a1058ab9c70 --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/_assets/icon_l_en.svg @@ -0,0 +1,13 @@ + + + + + + + + + + + + + diff --git a/api/core/model_runtime/model_providers/nomic/_assets/icon_s_en.png b/api/core/model_runtime/model_providers/nomic/_assets/icon_s_en.png new file mode 100644 index 0000000000000000000000000000000000000000..3eba3b82bc1e3fcb6b27cd9038786e19b7c2adcf GIT binary patch literal 25814 zcmce8c{o*H`?qc#G9K{_ZDQ;ou+Q{u_24=Hr{Y2>&m?^_GR7 z`1k~x9X4&V0E2h$-VF~A9~c;DY;3HotYk14=gysriHW&(?OIh;)$`}i-@JKq`SRtr zZ{IdGHPzJAE zO=o9k5n*8~E32fWq>mpzDk>?l+3YiC&gkjsMMg#<2$GhTW^QgiciucYo&M<2qoJXp zg9i^vN=T@xsumU&E?&I2y1M$ri4!MJo@{S#-??+=yLayn95}#Yv0PnU@87?F<;s=$ z^XD5G8Ch6ZJbwIm=B!yGBO^zT9t{l*6%rDvudmnD)g2rh%*@P`m6c^OnGgg;M@Rel z`Ptdog@lBxSh2#!#>U;<-O|!hO-=32ojV^sd=M28xpe8$g9i^98XEfg`lO_!UcP)e zIy$PUsX1%*>_dkRJ$(4^#*G{Q{PRyjLPAPPijR-a_3PJ%hlkg!S<}|mrlhP~P*8C8 z?AdMGw#CNAmXwrqc6R#u`l_p|8yFarm6Zhq1UNZ4xwyEjTenV9Qu683r!6fl$;ruK zVPQHtI@`Bz-?eL3U|?WkVxpj+U`|fX#fum7^Yb@u+}P34v1-+-O`A5=*4DPRwq|5x zWM^lanwoCdut8ZxB`z+GLZJi)2XEfIIVdRT`0?YrckfP5PtVHAl2=ezzkdDfIdiP7 zttG_8U%YrBCMxRh?>|po{`KqEMMXs@ist3ztz5bC!i5XX&CR!N-I_CZE)2uHy}f(( z>{+^W>5?T&jvP56ARyr7<;Bm>udS_ZW@aWVuD=o3;-@29?6&g#{@wqZ_n*4oWq;^LcUAUq_vZa67j00{( zd|LNnJnF6&mE3uB#kT~0Q8V0OxAN+uAK4+ox=lZ58BR05T)cVjw9F2LrT!LD zX>kp=a?J==Mp5R|VYe1mOgKBFzV5g3KS^{{i%N47S~8fTH^;{476;!ilsPtej^)uL+RE-Z8X_a|`auUSr+RpEgcUro*5gR1BbyM%C}bPd9}i7dkv3LBv3 z(p(1E!tg>iTd=R;p7JESJNb<{GF)oe*j9|it;WKH=_kjPkBsQD40UI0f=Z;hWiHlG zpVsf=Lk2-*PinRsMtF0Kc4auSs_uW@M~~8tbGwEaERGP*VhvFAktA%*YW^oC=c4S2 zZwYULG+M@(g>aJn5=d6-^~EG#`o0;LFh!-&J=s+6(x|QWQ3nUST%lvItvfQox4cP# zcEum*QTfq7^TCY=vXX&O-laDOCLuQEF;%uuZKPU9-kM9;3)AT2d@7pm{Pn)<>t4iJQBZ%qb3N6Msewul zEKq$os%n>8D3b^O@^j4SWO8rD&x1ckjemdaLoX8b<$U_$BWl}!yZ)xFH57FT{p|Yc z+fPMjewNk5KBg{Oa%gRfujzTo&H`9>2Q(?^3*KGlZ{N!1GcZxQ4V@j?xkrbAY5((X zTwZQ*9Hy;(46d!N#yZ_mxHhO+h%G&8jikM})(_hZ&ieU00)h0F*>6R9%7;2e_ctPG z<}B_LWbLf)3#&D}>57>GbQc-t=3qe|j@!Z-bAS7U*5fuEn5Kp+LvSU+=~;C zX|{jU+Y^nx2jx0ljq9mLfOKr{O}KqO{=1?#+Aa&JJ#lexqneEny4+<@o%n7SdQ2I1MPc+FP%?dHra9eNp-Zmn zq)+^}JwGj(PGe(>H&1_KeouW>NrG<(w2sB4=a?Mu;%r}3dT{1rr=A5jV8_$YCw!|q zjs9B4r-LDB>sVbH>60xdu{#c*ki2yh58iI`M_DU-qpYA-_CGI9(s!*c%|#_T*(usZ;=8vxIv&WuB6P+b&$+uFBXf)4;ngAU`8F^0l?~mWreN6K zz4!)fc^di-_EZubKX-dkvCh5wNNr-0r;J?g4J_zs_-4KFEy-bvy-H&TD5k(N=a#=2 z{0!}{54ak84>67*Uc8POXs)I#Pg+W!5JC@ZS>U!(+`i!=_Ne;9tIy8+A1=L^eh-1n z7TJZOk+RqI-$fZZz4SVP-BJG7P^GT?WnZXyD_j2is|cxa2@lUQ?8jAAxSMw23R1i1 z_cyKnqB%iV81V6j_3rL=_Ald^hMCT@MlY$9x+`xy<;yWxcjXgO+c#RqSlIcVzRCjH zF8f<>weB9XR<=oJ&!T}ig|Sul+5YJB`8ndHaM$47D;A+)`+hiXWmbt(UDhm#ezgJm zQnp5ZrNaF^q1BjCX5rZqxM$3*aou-&726h9!Fu>2Ft{#&Sg}ZdJD?(Tvi{kddfJ6! z$b{uD-L%a^9L|ClQr*nbt)eEC~WNy$DhBk zt$#IaqjD_Nhvjlr_4`EzJQ0?p)xT72!yet>w3d%j?fM=O$h31N+=hV5W~Ry#iU|(UMXmH8tAD_4k{L zS(AEKtBlU-qrza@SZx1oEMr5Bv$46GMlDyv+9J~*pQmXSxQA7BZ@+l($2V0`+v&fL zuG{{!D`;+F9<|fMOK<1m1A2Nc(2`91zJv6#p*YozA>yHCWyx98NAt|4}zKp8mjo_X~d%Rv&P5xgw|i!K{Y1&LS5pCpPuFmJ70=tFn@hk2yZw&$5t7kl!1`j4)h`*iE0(z}13xRM3un?z|X zypGMxxB*-L^WmiTgc;;_K2X1`0w zg2N9z)3w`%!t?If*lTnhR}2>ylg*dc*bH5ZhRy%& z{xc}0_R6LU*zpfw#wDay>h{_}&*170+jm5}DU4oE+lxL_8C_(iacHrx?6~+<`(F?o zeSFE6m9;Lm>ef(cP|VFUmyonmeVG?H_D#a7Pd7kM1In@2GTbeIsa;V22wHby?hw?+ z`JS;5{WdB3QI{1xIq&NitCX*NnKQ;*im0*kcS3ujp(jAxC9vQ%% z`Q9A4D&;p?(mDOUvJY?7j^50cNU^{v2<5 zFqm_)o&BWwd3Sc1KrkvY1VX&P)z1VHZ;2qygG8#Xi3-ty&X{VyZ|>%PSj+& zc)0&Q#j}XChg&zY4=S_>!k&e!Yb$Sq&|-d`t-Y4uaW-3V2$^4uC@ zSxQ?ILKRlq42^~t#}vRi>cNlRMsk`5eeUjU@zz&BBv;ARn9&s(f-XD@9?*B!?mARF?xYpsc0Y!0bMLkd8x#qeZek4Hd_P{91k9*y@i zSzPse`IsHzV)r`PT7a@6$7XzZ%S}$Ve(R<4+F!?mX&A+dzPhh1yd`JO*xlZ*p4dMN zeK>~)jo!&?WMK*U5sLfP9T$_gy7G05Iqy()sxFdt4>3P(XmG4_0lloNhk3;^3;R)# z9&movW@v){GW%eg#Q67!-=2>iH?**|Kc9x8(F-@@FBieD_zwk@c?ZsJ{&*Gf>$nY|j3H0o_sz~4%3m!Un)BDuV$XQ;P=%`X=s*~o##mn-zLsD=J8B; zei0uh+FgqWT-g9GWIyn>h9<+pPH3%1-hBJGcq?@CG!z07a{SEvC20ES zti9^3lu{?d%ZTKv-bQ8A8oFp?>&-cqsU{gBzCjJuZD@Bb7;t@0F0~Od{2{dS4D<1D zhwTk)zBG3wK#7z^c3Y7zlZ$Wv5058q+9#aa0Nr-&d+Vdq?5`6H;PB?x`nmA0=cd+k z)zHt!0+WhhT0PtgFp$;RwiK-R&RG+=x9YoFDN%mtLxoZFa;>m;{EYnNsLM#wUVehu zZlvc#lB|17@A3`Md7}f~oHdyXB@+n7JK1zN{_?ujgH^Yj6EOKhNGltg+NlKo)p#{x zLC9)Iw3V{J&Ka4wYJWkPB+r*_Z&#PN7r#=2At*k(iaejTAyQ zvEFp)Oid5o9GwP#op6$+Ja|!9)_HY;qf%I{9{x#kbP2Qje0Jq&LN<}uKT_N=Qe++o zRAhfqmr$bwbDl?asu`h1#WvU9CYf4aZYg-_pt_pPu4e{Z?~&5Cg*?CSy&TV5YMm5U z038&X8|p*nIL|;n(jH)N~v%%v_KAr z87Vw?xac!J>pgRjo4Uq3zN;v0J{p#sbTg_nDTuFZ}mg#4Zv zTx@?S#6;={cyYFO=@0)>FH)BI@Z+%iUvsGYOAkI}IziP=I2hCfYh}E;mPlB@KBS1+ zyG)T#$#d`0yV>@N-tpZW1S+YT8oA?YQ!QC_k#b~(pa1T zF!0OPG=#8+b1Wx%wbL5X9@^P{nI>I)CBUX1B1Hzh%SfAE?nB?1X+p^L-;X)-d5_;A zMXLZCS9jxVY=-QTln7M^pdE^%b4Zt&%c_dDr|=$6X1(y|Aa1iYa1+?o?i#)?)nr02 z9qbEM((hfSO2!!eIoRltsmE^q z7woQul5!}759}TzJW?L?lU8YbTv%z)upAdeAvM!{LA=l!8%Rov87v7#4(jqo;9qwr z~Q5xB&MZaqO7kR@_>O{n{daQ&xR zxGP+Tw~l-vRHKHN-`E*9EBO4o^DFT=HUc2oQ+ON6QPmDzuBYy0M(<6*PCWRWwE>#^ z@{F8FLnbm?x>iI(bzYqC9QgF>qnq>KsOMjl`Hv)H@-)O;8>IY%7WRc!SMx)+HQU(8 zdy;xJG+k=<1Eo>*Y&a>GH!uJBpGD2bM(ds!$Ey_2w}I9YYv%JdbjMR6-UKAq%b4WF zoJ)AZD+9LJ<)Ezdl-c#FaZQseZL4=Ye;&Mn?1QbE{WxYF3j3uKiWg~f%^|yOis#-~ zT*G8oF9;2TmQuKt8*7<{KfSAE^RFrmlts)_@#W2HXY1ANOWiEsCbBRj@6Rap;MC~z zx_ymX1$1N=-n#W?^s=hrElTq?#DmXA=Y_v4p2(#ie$H}$CVOH@mKlhoUaufXJ_K8! z^ZAGrlLKm@FnSFbZRs^5*NZ+)$NUH)qkMFp`pc^D*4YM?>XO7q`N*F2`@QJV2a|o5 zoDj*Gk7Kr=lc|R?aw9T;UF?8<8~j-F$X`|-o$Pses_sq>tOk_Qk@2mG!L+*j&OCy} zRq0>%dw{YMnRxLsx$ab}Yl4yMY+w3~1Eg>5fZ_$_{gA~qn(g5vFmEHFonpB3<$J=g zBUvCWJpWl%)!aQzf`*NHsLo!5nZG@!!Q49IefvEwy zY!7L$)!2L~?kApB&w+l(6MZ7XO5XqplZrTQG|1JFLtD}?1NUJGsbc)+Fpps$i6Z>XE1xP$-~S@6et_~aim z1O-p`N)i&B!+d19TX=T6qB-NkhbAJXFTr!A4IVjwMck)BwL9PMoa3d5>TYOpMaiy` zuV&ml5u~<_x6%bHPBjb-irZPG)c5PV#XP}@3%FN;xn|5~q!e=Cn24Ck7y z5YZ;w7QB$hi*5_OIg9%Lxr^%=@^0yiCNvyEpV`7SAnLw6G~bAuj|GYe;Gft7?Wn z{McV7kSB>*sKPI`jnK2&K8x#^k2KyHAbh#3gfQ_Hot>{o6dhQb6h(&~{?(pX9%O-`SfS9+^Jro0!G86s`Ib@qe^TRn* zk<|$FZD>(tJ=LAb{aEN5L0V9xm+Uyd(R1Dgao7vt&K)v%NV~Ol9=|UU6C;@1mX@v^ zDm<^whLa;!rPRY$c?7DVVItKHQwzMDKJr{ZyQXh3NqhJNqt>7PJ+*O;)bBP#9Rm zEY1O*^*(85x7Mvp!=sb)vAQ+xmkGCKU@}75wNK_Z#n2W2x#eL;>)~fSeb7)9(cK1* z=aW*uf!#4ZcKP`Xp-iQ|q2Z{`Kb*Hy#!g0jSl?4%+sMn%0v(jGL&byAnteB7j`>?3x@}|fh9s*H){*Clyo;TiJugL;8_~kf zjdBv;DZZ03+`K^zh)B&1dL`_bLYRE8sE)8Bk;VCa;{&}GHiHT)p%gBPNY2F(UvNi% zZrw&AMY^NFvQ1jm_3Gmt*qQ*GOe9YIvUR^)8iGS`O0jio0|4c}_l#9bM;;py+ zqQ-fA`nW*7n}VyIa7pe0aPR!#gGvdWja`~xYKF$#4a#uO z{o%0%@c8y#Ze#=nY`+d2m*zhGJ2a*jvdvqK5lxi%4N2IpHt=XpZig5Ht_HtyP$x{#Qcm_vPZEs-V(Fn<2jI(e1V*i9pUn2uDHTu7UG6%nzcf%c^#15bvP>W+iKj_yS_!QB1{LMql5jg zXGsY1F9$g!7x@CpXxzzDULOTpyiE`F+M~t*y+Fne$czI2a#FZ;d{vF#;Fl&6EZuv2%-xhu(4L6Sa$ZMWM$vE4uO+51X2ZsiRinbjWW=@ zfe}TlZ(0wx;xTTEvQ8^AT3Im!+Nnki?S{1fJ7hIAM5%{Y)1K=baZJD!Z2|CA=F~G2dRRo~woRandwiaVQH$IoBmPS}kM}g^ShQgTZm$>v2RWsKemW4g)HnRup~ubNm5L%Fyp~UjzdM{xjsDm=Nh>J zl(t$wAUBIf6ZDRcjKs7BH|9ul_w&!u8P--O@cVjHJt|Q&Mu1-$>?49ysE#J%AQG1t z2*|Oa|7=!cL^2=Ot%K_PIgA_!(O7Mxe?ZOyg_)!%nyjKIt`BTb4Kb##mmVgHm*-~` z2l-c#^6bc%AcY=z3i~D2UTvh__K*gf#TAg7Z$s6iErU!`J!aFw0DY&h1oXC^dJ^8% z^IZ}*uMm$T&qB6+pC9pqXdwktOCD7*y?&aI-xknd-ksq#Q>*~=;Q&aaN6fDn z*noE$aQ2US(X%hm@U?-`r7h=lD{7t~e7!UKoFCupL>#+=!yz_L%!`okbpyQddS9y@X~@sDVI>2u5a$4}5`A%OH)Oe-2%JT-(%{cUKjAOp01>nx#o zzU|Ef;NQD<4r%7#tm|HHiJF{)E8^jvO>G>=c*0{^8-eI+)$oZRJ{4v+j;ySaScBJ8 zwdZdHEG^R^5&}Sp<1LbyrkD_(@8rEXFzs@;gW6uE=+CHSgkF!Jt5p@nw}Cd3JUAR= zJj-xkP$M@M#cMC%4FkzjE~|fg0@J&*F`)hl&IOi%&zA*)aP$O@*K?l0 zwI>oTaT*H+77Nm`R58+c)CW>@YBQe%K_8Ia!74cBHxtn-a6K5-tiqrEPc znoVH1`Q{)zFUI)W*YZ6C%H5AtD2!gFe%eIrat6!0BObA69GC6P6_yu`8b|Js#%2ZS2>4POrG?r-4^#~iw3q#k4sFaP5MaJmSf*S*60U>kj>hF^a*+3aQb3TnYi)eOG@r2swzJ~~qiMiA!z(Ffe zt2vr1%N^lqyrGq?*3%dbB{hA66ojIAP?^4 z(O`on%W$ReY7M03hPI41ykJp?#Zg>}e5*yzV{w+*BwzMW9N;S%ESiT*2s!wD$BimX zAa*`s&n9>zp(M@|D*)X!8ncSBYh)ecu)1StE>t=*1lj55I74k()M9i~tnpz8ik{iqxfcZ_^=I3U51!{hbToyWlqt9cQ z-PuE{90=GR;|K7@TuRCeG0kNl%yqf4{hN;Ql;%g|o^*aADMm5i$`#aaE}tga=7xoT z&2n3qOQ!EiV9d>5^aS|#0lUMHMi3}?*>T`Hi~~%3dk77iQE>G2#Z?*tzIQHNTt~!_ zanSD7pGg8ZZhp?WE`d4u1nOs5FZm_3?IMwSKrtZXxL-)%IuOkWD^;&o5AXluG1d)7 z#!dO6#E>$xYLP?07;f+fF382PfAKS_wp0&~kjUv25f*Jknw9ckKFX#kT58`KZIM4=jRa-PJ2=ZG1 z3bm05C5I!QaT!c5Uv+|*dzF=X9C#M!?Sh)9p*>57&*Bw4^f6e_>28AJzy?MEQM@if zCLRste&7W~ZKPIQv*svnC|g0H3O8t`v7SFLG*vBwdHRZIN`?ig@&A!#e!iA4lVe+UAKkp@+>`)KnlK zeByfOgKDFRRKbtRQ(bqYs|_ZrQzj#?6CMUij5lP3SO=;)e-1}dXL4f5ri~XHHCr$J-)W&Ipmp~8>dOj2` z&4V82n1hFa(7{Xnxzr=ogZK#~uC2>hM0OdrL5e^mN4+`AXno_H>XL}zI~Q8q2$$H0 z95&LQSpfR0pIX_vV7PR}nBdV(mf=Kby6BCArYI(jW&yZN#hM2)+IW$29|DQ*fRFD| zV`Kqaw}oEfR)D6qx$>4ZKfbYPlO*m`AanNU=6{N!4U{li*ioy|U?B4cpP8D66iZz(`SqEMmZ5rf!>36_K{~Q06sVUqp?s}KmvHT9NL$$T_MC|`PiFa zA4hIa;QH5gcxx$LKt^7o-Hai3%_iHrcp*uYd}Jv-U*-S5Q=UW`nxaBwEl=N1&4riizL zR`gLHjwqP@De4)Uw6D&LZ@lE4k2zXau0D~h2NA_^)()r>_})8EUSuG8@?_5jlo#NH zVka%~B8|2hsJjSbpB$-3GJID_)W;S#u^U1XBNyewG(jU8RKNiTKR-?g3mU8F?bHx_ zNYv<>os@u4@279nA_D-uPAD}VEHHT3`M;aqK!#l+{FqUa<;LXDZs-v&7Tf{1TDrYN zyY>Z>yWYai17vy~h%S`H)!Un`Qk>AYlMvlITy!Qe@A9Xohym)6q~p zfOGNEHULG1wpGt0U-|A+>QCNVz+K}OC<5ERe5TP>0ah!pU*|`TXokO_YvI91G7g(5 zLvDJNRi#zlc1oT^!0X@1KGt9tUG93^gV2?{R}%)(_$A;CJ0aLMu|#H1 z+lf1bB39rHj)lLX4e?QcfwP6bnZ^1Aie+jKO=vo};iO6mqUx$A5k(MlaZUPiLiZM# zZ}>i|8(ihaPaGK39$;Pn0IAYZ^c0v2&G|~UW2~X?K*_p*iz-awG_;79sdsoaedYO& z3sh=_tw}i9WNu5M_PK~2U;} z#8{wvm=>q$O;j=9jKf8`eTOJO#TP(m_AL$OEAQr4c4#LXC4blc)WsG}crMc~XyHP( zF|w$QJP!#+sWE;*d`)u$2_C=+gceZpMxbOl6yDa(uMj8o`eW7&9ex#}WCb;(7DG%? zOcPN2smt}fe<^;Ap9K^(3stvh&Xi+ZWIm2yz;yN6^hfrMS2hyo1M8uD0Si|?0D1uv zLTE|sNRXm$>ZoKy2XfVrG-u@ProDHQj4GJ#c7s+LPL3r~frl%w9m9*W0Us=%K8X9z z90?&%fyp~CyCw|E2gE~btKaNq#PDXDn+n#Lf>;U`==b8x-djJ*D0lk!pVj-l=ySl} zj%kDI(LKPUPXW<`k8a}qjtYFbtYv;#@(5Ai-`@!B=lS~?%yB6_3Eb{bMZY-;ARCAy ztz2pau#`2x+Am=26C+1w(3j9P367VcC4%jd6cHIvxXLFmqeIgnU8jqN7l{sDGM1q5 z{M-gLMjSAQ9N0HOMd$UEvD2jLS~I$V>QsR0tRa}I?(EtyA@zeiN!kQCPEixIawy8A z=+ITV;@2MWA)>W+>M~Xg!U4GDDzyXYk?tN_7o;~toU}^n0W2$_9vHC_keY%8t*^YD zpdvlBYCPX5ipWR4eLx6W=preu77w4SXk%}oU4tXn&=deUwh-MCGkKJ@9Gs(pM&*5y z8hp|jETz)xh+z4k?!hvzD*fgSM5HiA6?jT-W(FJMgH_;Y5D#1A$TlFEW}wbFc+5Fh-+3|0kWy`HVMgk1(9wKEHZ&eugQ1zcocqZS6#c*wSfq{a?P z6-FL&zir(E|V*w8X%K(%|L>N70?E3xtv7Vwi}}@>SQx9Lzz8 z=wQn%a7NubHu7ef-`V@Mpsj?<1rJYdVtL|t63nG_vMbYLD_OpOy88+|A`@6u`#rN)p^79ij*tj&p*et*y=X~i zx7Vs#OVaH+GVTGl1D7KCcS7m))OsF}@?%w*IvjZL>TQGsFuL_nh4jRM))voM^`;en z&b5o+t&u(p0Pjly?=KN3E4yjd=`d1lMbs#uHhgL&_CDzAW#5-%W_yGz1IQ4s@Z5b% zx!uhY*ffPgUJkMDOF?DqCbOV;Xn#%uPT0X-ItPs5{RuEOk)U0h_TRB$&?f|AZ3)`Om#2@xeC@`p zB0kW)$4jh@Jt-Xdmv2=nZEpyX?dMZ0%u+RS)$2O~)qv`wixKy+Co52MKxlAdkN>{h`vtph#l3BkQbmI1J^1GFRy7;xm@F2ggH zp`_+Z76l)nF}b>LGGwKPJ|`TCkLR5)-ro%k^9h%Z?vT|2oU}Ky0|bw0dG?y~R_#&M zdEf$i5X}T7oCeM&ztkwT3Wl3PSlB@ z>8~PkQ5n?TWuR^4U=68ewD;Fz(uB|k7;rfd+Cy;Gu@Onr?>)>rjk|9SSxjZU&fp>A z&#PIPhYIeR@?wAq)9DCU&=D+X13`8Lck|7|2);qbU@hbQyZX6;zQDd%OhahhWVg&W z0^m!sc))fD`6qdp~2K5s{SRtGbnUXoKqe40fpNC z*|5Eh{k=l##u46tDmsT3`9P`hEdSf9$LD_c<$vGDyDjpm#TCRka5Nmh)&hk+i@!sm z=SZflpu3{iWp+vwgGzW?JKNObA@}nVax+yloo^ew|I-%tTwp|5*j(k{J>d?IWYlEJK6d!}t|qyzglR3G<$>Vg+~IMuJy&kY|`l*cfncT#$0;WJ@9U z>=&QRp=}fKJSzQLU4H>8*8wVR(dXfxY99*@4-$tDk{YOr@I{r|Gk1Whx}Xrwc%w}U z3-m2MnCeLH7T3b*?_l@*pnqAqoRF0UdVueuN+>WAP$)ZWhd85mdy1HZ(_LkzhLg`h z6VGq6IjKI2;psHe8sbBIu z0fcS~>F}MhS1FVE&$}E7co%rsK9D`UJ8xP05S1|2+Np$57f_ z--Cn*&OHCFOt!I$K!$0hoN~_)pZi$S45jjKqm)VQ042{9UEjSq>lkC1c{+lQVBska zjw3y)1t$FNnX;LQvjfu?1G$dQ@ncltEM+XYwVltrrm0Uhn4oK?+COk!224bvQKs_*{ z4BWI(LpO^NjrC9MY-I+<1k+a$2x3`4)}`|Kyx=8=hG|q_duXUPeT94M|6|7@EC3G% z(|5dq7|+l&4U{ee*9oR}1X(Vz_qD9RL!HcWksSa_f-?;%zfBRkt0$*WljWG~@B*h%tV?1^N-5<`}|R zu>G8N(Wl`n1HlRQUqdK?@i+1A*^~=aVRt0kGnstV1kjt+Wlxfnr^Y0PRhoRD<#d(WO(fA9o!O57uh8p;YhZG3lQw2nkUVa`hIf1nT7XkNX zr<-P2Rq?|sw3vrZg5ZkHU*?6tAXc-eBfvm7ORCg^=RI5fF2lP{FRzoIi_QdP(QRliy?FL6kQv)lQRZ7hs zEf4Eh(q{02^ zzjOfC_4H1vk68Q$=m&ovG^HPbba8-}E*M^*13&hAuB!xIDon)%Bg6)@?W;kK<9(Y*94<^5e;js4;Kfrq)e>y|}vGeBK*bEWH=XYhQH%+6E;my%xwEP^7Xq=wo zuV6CZRZy?u?*gP^1Z+P}kZ1e0kP=_)d)XP0|C$1NnwO6V$i(($Yv89!3;ra-es5Xa znKw?`v)`Mez}RY!EWkcm{>iuZ+Snen{J~u6Kcxpjiaa&z;q14c$Q%Dh5$AbB%wqh@ zt=K=12k@iew~IgBRp(!usrkqPr0_4F1pOUPM)JQVpD2K-k&eW7WoIx7u%vfuc_-I$ z;5*-TZUO!&+Vriu{B}_#m^H5W$7@p*uWhG*TFQSqXt>u-@%MT=NPI9h<+z*B1L9vV z;CHpSZptaplT63YhPzsf*Qgo1*=-{LlHOtNl8-bw#{A<;?{MMDY@t zo;)2`;0RxQ`V)Uvuxj%GePH41A+vrUO;Mh~g1~QBaZriB&vJT@1h1Qtum8!X!cNfP zh=bYFqewM4cO5bS-y;6?#vnxn7-WQjJ^#-2ZZtl?5)ewWCAOAe)GYF z)7@?-)X@EG!Sr1kkTglA61dwrE&R~U>W|ZE6Q}Rvhw?{T#sQQBJi!x$-Jeptuq17uL4Tp{K&Yz>4G_fvG*c}4lq0nr_0$55*1Xboi69^ zh_UR@O8|C3@*W2QRI_R)j#=+@xsDSo|FHxMVKiNUIWirPf;5zVpwTlTC4KtNbHD{E zn^)5{w|OlTEV$fw`nF5GO20R6AUcVvhq5k>E#UFi-t79}U)zF+Y^KEhsn=WqZlq2z zoOK%07C9CKQ0!D52d$#hv?lQR_*5E;QAD8GL)}1;fVT?3n~5X|Ouzu>PEFSva0l<< zye58Iy8`kXnJEHOkXP5CIzQH%H}}ueePe%Rq!jG(>LDM$_}ITT$@gKG7e0FG^bBngf5q4 zs*t_<%V3wQ;#0f4DyJJOc^S|CL)4{1DMJ3pu|;T~5HSOI?usau)1l#HyeR}eA_foC z(2keKmZM>kL&j3XytPQt&2PJq9?|I!0ztG29GaNJ;H)6RhvNV+|2e=!C<|UlwX=7V z&s!ixn*PuXjPq$@o7HocyQpvL*|ja9Lv}bA)pKFMm{2MUr>`G;`WFr5Uv||XQ@Sg`glG7gmLF9jDzA)t(YXauu2WMd!0=r~MKm7AR zgm?gApa?(%2l)UR(0%=E6Ivq8q=3VXKXSJ~)A^@A0K`{)XPU}-QRV*9bPgGY=L}KB z{yw>a7z+~*fH_0JoNwXxiGZ-oSyc<0A-TVTUB%)GGPr7l-4d#qfrho^2XYSn>!GiY z4tKh~t|L7ZFH^q1`&~yS^qNEYz`Ee>0tmM!Oa?dO;wwaI<)^<0ghR)?`-b~O^i?DN za=9I)O8n;mXl9U>zX4iR>FNYE@%QDE7xFbh1ssik+aZ5we=&AYsK%beh`3-?TG(Nx zC4+~k&0+YVqIW!DJ%Oy?HUD`~rSqQ$slm930PS;txW=gG4yvIUGN3TMxa$vn>QNJ@|kh{G!T4 z#ILPnc*%hkol8r=XN~`PurKmI54vf;3(3IH>Ka?J0O`4x|4n{yuQD!15cR1*Y7JY% zdF!*IBlxzGag*5K1}w|RpGU$Ff=3r<;=d0b^G|;x&Wp1ovs3btq|l`kgxszGuPMeO z4N&_Nr14LZCCvQ}mhjx3?CJ3SjS0@&^sfg&qSGG$Jyx2Q%nJ}paK}$vd3yZDs#-uk zSP$6)D;pCc)Z}Rij(~jB?l1gn#e!a(RN6`Fzw!`mX5akR10bLW`QMCW`DQHst6B;T zDiS0f0EsOF6073bStW)E%-=+IRB#Tc5!f-h(8^7JU7C6V`1te(fCY`PT{^F=)#}&3 zM0#Wzn31+c(h-h5;0Iw&;lLZK6n-7rQqc98RvI09%vSvb)+w@b)5#LpCUTQ;Hxm)} z3M{lO+KhuoZ%!NSs%Pvm6~3=)o)vV>p6|txB9@(SQfpgdP}0f2YqAKu3oc-#<50h4 zo}LqQSN+aM@VhjTc0#C15-1Fvp$6l_bJ84(9(3O!_D>Ody?bw)PdKAP$8VRdU<=#8 z4xabfQSE`yOT_%Wilv!mL+y)k8@)FK39=!k&tnZrS0 zuT8+O_k^gAR(P+9l|!5b#dP^AF}Y>oeR&=6k&Zc;*6X3h!vMa8UUfFf1mD+%e)33j zr1S4mm<&IrsLl*Fl>0{jTqgYz_v_f)bUVv6i4<=wP%GuZ9T7&8Nw&KtufLWAMQ}9W zED6*XIJwz(Pg`hgC>wM>qDWLF7(=Ha?Pq zp5?@?90CX7pyi9MRn42owa8XVv~w#T(vOe_mzZt|$GH2--m;qft?PBtc|J;ngCsL2K89pe6wOSRTAL|0NQ|mY7s&oVXqFbsP#&6$9KQmhUK6E-M92|66G6 zEBJjU!AaFFv`|Tx8LV_f_R%$8PG;gZkyqP#XMXsSF=h&4_bD%ALNOu#?!J)Bwgl7P zyV;&A&S2V~eeRq|ON7eAYOvAuTasl>;!gf5KYZcdX^=7{T9{Q@V1~_S!RRvXkMz05 zitDA&Z|wNqlZS?o9=UVde^&JhUNaDN9udRAyDh}!3;8iwva!dV^G$krkp3C)jfrH9 zq~tYUS!iUkNTT+(RNy%4NgSwmz}j6C%5nwYyQm{D`&Hw9n>O(*>QHI-n`TxQq;hdM zvV3u7UEbda$|U(;H%-|vV8#hB{*{}+{0hEf_c&Z+}rpI@t4Px z_J#Xa*v8@v*g_x%+{VCfEp;-&E8pE%Jj()Vh6dELq^{i#$F%$CgoewKl0^Ur*)!G(X# zD7!QGYIc?v-SeyK*CH>@PW_(6JKvd|O8H%I j#hd$5Bz3DL**g^>vTwt4UwW3_V z9g*2R>O21oCZBnk?{}ph__eriGFJ8I<;}{+$~)xER=s-Z0Cez~87+G?fIZxBZ>{9Z z-cXO2v*0_c{@Tq}5a(1|?(f1t-_T*p!7ZS?oh{(QQBZ3Ypjhk3bd8#M+^w)!-@?8Q zx`v(2uJ?D|iC>g*TuKMUI!5@(mobju<+%wzhOfz#%t=L`l?M(cq33a-SMsFoeZR;`@M#t*)#b`jq7@u+jXr)bhU8noYUdfC4+(5|EITW|A#6K<7A_< zMN8STjYJ}3Q#x5O)+LqLO-!G48J9_oq*+~D<}hw;YE?t`*v6tr%0x4U#v!6wng|(& z&QOf4=`_TgL&MCz{)zqheLm+rpXc*F?{oXU-*?)hna(B5r>}uTes#IS+xpwSROg31 zOoLCkjvGU@i0iH~Y|p*etrRW8SjVTR9C1RYlp|3-B~K^z8e;J*)Gb{av4``)a<(dY z^L1+Ciz|)-u2SLrM<0Xb1v{H{b48pd4S?&BDbIq>&Quiy9 zxEaxyTGi#ZC1=MNf^3(#ENigg^TaODWDsvC&Hmy+|9RIAgb~y9&||e^U+5xE$ztqr z!fsvu66j8&L@+2|JqWf5_pO%ce=hLk5vvGe+7kok@=^rdw~!mXiczWv0x}dZ?r2Kc zu0BqPz%(J@;Qq!^e8}g=;7P2qqj@@=&Q9Runa9op4`lI6a$hcN*5fY#M9Wh(fjy+- z!X16tjR0`(5sE#?A)jr%V!sI1zm2?otqRx&BQ$+#M~YiuWEpVyT8S_7UELZzd%wQm zwb+_W^hUk}R?;RGc!*PUJ5KTLj1R?s>jfX)oH@mGDTYc*@h<-{1dUnm5UV@T+iz8jT>j`~`_2;7NvKD4LkZ=HaF^=O+pbeLIBk+F<^y`X(X2&pT^fdvME4;m5cJ9mrWO(yMw5T28?}RWT3{D*_2)NHjWQj zVR`hOFw(8j@)G_1?D-1Nz4!KgWO=f1K@1mCHMUbHGiHfTk^unywv@c4GQ!xzR|gh_ zW?RJ*=$dNl*~sj$;h}Mwc@DdHT>F{V?lP9vNFR! zJEKpLFEI^H2U0ee689*)bb7+~bD46bkuQvth!Jv5Z4Mc)^Ow zHI7+{^Mk=Njw+FR#CgqcLdhnOITq#)tndf&H@E3!GgybP3wEIA+L@0tk);B?b!=Ix zw7gailx6Q*>z36;YJ->^Uy`*%synCt*|dtx2fY}dB3GbmHp_64YoqW4*N;Nxsh-8p{~bnJJUYSp)SZ{xD=Au)hy9vq(sbd!`!dQlMcj(4<=<&jVmPYjcDH$22;RV`KF{;)-h%9}ACh=h`Y=U2n0 z7aOK6V4Iza!(P(XN@z!9_al<4lZHPCQkTJ|^ z=W+^HPlQM{akHN={{93IH0WXQB)I8G@`J^C?fY8_Ny2y)L{ekrGf=7~t^Rn4K5e#! z47xqdMwX+v|J) literal 0 HcmV?d00001 diff --git a/api/core/model_runtime/model_providers/nomic/_common.py b/api/core/model_runtime/model_providers/nomic/_common.py new file mode 100644 index 00000000000000..406577dcd7e701 --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/_common.py @@ -0,0 +1,28 @@ +from core.model_runtime.errors.invoke import ( + InvokeAuthorizationError, + InvokeBadRequestError, + InvokeConnectionError, + InvokeError, + InvokeRateLimitError, + InvokeServerUnavailableError, +) + + +class _CommonNomic: + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + """ + Map model invoke error to unified error + The key is the error type thrown to the caller + The value is the error type thrown by the model, + which needs to be converted into a unified error type for the caller. + + :return: Invoke error mapping + """ + return { + InvokeConnectionError: [InvokeConnectionError], + InvokeServerUnavailableError: [InvokeServerUnavailableError], + InvokeRateLimitError: [InvokeRateLimitError], + InvokeAuthorizationError: [InvokeAuthorizationError], + InvokeBadRequestError: [KeyError, InvokeBadRequestError], + } diff --git a/api/core/model_runtime/model_providers/nomic/nomic.py b/api/core/model_runtime/model_providers/nomic/nomic.py new file mode 100644 index 00000000000000..d4e5da2e98ec97 --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/nomic.py @@ -0,0 +1,26 @@ +import logging + +from core.model_runtime.entities.model_entities import ModelType +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.model_provider import ModelProvider + +logger = logging.getLogger(__name__) + + +class NomicAtlasProvider(ModelProvider): + def validate_provider_credentials(self, credentials: dict) -> None: + """ + Validate provider credentials + + if validate failed, raise exception + + :param credentials: provider credentials, credentials form defined in `provider_credential_schema`. + """ + try: + model_instance = self.get_model_instance(ModelType.TEXT_EMBEDDING) + model_instance.validate_credentials(model="nomic-embed-text-v1.5", credentials=credentials) + except CredentialsValidateFailedError as ex: + raise ex + except Exception as ex: + logger.exception(f"{self.get_provider_schema().provider} credentials validate failed") + raise ex diff --git a/api/core/model_runtime/model_providers/nomic/nomic.yaml b/api/core/model_runtime/model_providers/nomic/nomic.yaml new file mode 100644 index 00000000000000..60dcf1facb475d --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/nomic.yaml @@ -0,0 +1,29 @@ +provider: nomic +label: + zh_Hans: Nomic Atlas + en_US: Nomic Atlas +icon_small: + en_US: icon_s_en.png +icon_large: + en_US: icon_l_en.svg +background: "#EFF1FE" +help: + title: + en_US: Get your API key from Nomic Atlas + zh_Hans: 从Nomic Atlas获取 API Key + url: + en_US: https://atlas.nomic.ai/data +supported_model_types: + - text-embedding +configurate_methods: + - predefined-model +provider_credential_schema: + credential_form_schemas: + - variable: nomic_api_key + label: + en_US: API Key + type: secret-input + required: true + placeholder: + zh_Hans: 在此输入您的 API Key + en_US: Enter your API Key diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/__init__.py b/api/core/model_runtime/model_providers/nomic/text_embedding/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.5.yaml b/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.5.yaml new file mode 100644 index 00000000000000..111452df579f8f --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.5.yaml @@ -0,0 +1,8 @@ +model: nomic-embed-text-v1.5 +model_type: text-embedding +model_properties: + context_size: 8192 +pricing: + input: "0.1" + unit: "0.000001" + currency: USD diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.yaml b/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.yaml new file mode 100644 index 00000000000000..ac59f106ed2928 --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/nomic-embed-text-v1.yaml @@ -0,0 +1,8 @@ +model: nomic-embed-text-v1 +model_type: text-embedding +model_properties: + context_size: 8192 +pricing: + input: "0.1" + unit: "0.000001" + currency: USD diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py new file mode 100644 index 00000000000000..6cccff6d4601cf --- /dev/null +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py @@ -0,0 +1,170 @@ +import time +from functools import wraps +from typing import Optional + +from nomic import embed +from nomic import login as nomic_login + +from core.model_runtime.entities.model_entities import PriceType +from core.model_runtime.entities.text_embedding_entities import ( + EmbeddingUsage, + TextEmbeddingResult, +) +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.text_embedding_model import ( + TextEmbeddingModel, +) +from core.model_runtime.model_providers.nomic._common import _CommonNomic + + +def nomic_login_required(func): + @wraps(func) + def wrapper(*args, **kwargs): + try: + if not kwargs.get("credentials"): + raise ValueError("missing credentials parameters") + credentials = kwargs.get("credentials") + if "nomic_api_key" not in credentials: + raise ValueError("missing nomic_api_key in credentials parameters") + # nomic login + nomic_login(credentials["nomic_api_key"]) + except Exception as ex: + raise CredentialsValidateFailedError(str(ex)) + return func(*args, **kwargs) + + return wrapper + + +class NomicTextEmbeddingModel(_CommonNomic, TextEmbeddingModel): + """ + Model class for nomic text embedding model. + """ + + def _invoke( + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :return: embeddings result + """ + embeddings, prompt_tokens, total_tokens = self.embed_text( + model=model, + credentials=credentials, + texts=texts, + ) + + # calc usage + usage = self._calc_response_usage( + model=model, credentials=credentials, tokens=prompt_tokens, total_tokens=total_tokens + ) + return TextEmbeddingResult(embeddings=embeddings, usage=usage, model=model) + + def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :return: + """ + if len(texts) == 0: + return 0 + + _, prompt_tokens, _ = self.embed_text( + model=model, + credentials=credentials, + texts=texts, + ) + return prompt_tokens + + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ + try: + # call embedding model + self.embed_text(model=model, credentials=credentials, texts=["ping"]) + except Exception as ex: + raise CredentialsValidateFailedError(str(ex)) + + @nomic_login_required + def embed_text(self, model: str, credentials: dict, texts: list[str]) -> tuple[list[list[float]], int, int]: + """Call out to Nomic's embedding endpoint. + + Args: + model: The model to use for embedding. + texts: The list of texts to embed. + + Returns: + List of embeddings, one for each text, and tokens usage. + """ + embeddings: list[list[float]] = [] + prompt_tokens = 0 + total_tokens = 0 + + response = embed.text( + model=model, + texts=texts, + ) + + if not (response and "embeddings" in response): + raise ValueError("Embedding data is missing in the response.") + + if not (response and "usage" in response): + raise ValueError("Response usage is missing.") + + if "prompt_tokens" not in response["usage"]: + raise ValueError("Response usage does not contain prompt tokens.") + + if "total_tokens" not in response["usage"]: + raise ValueError("Response usage does not contain total tokens.") + + embeddings = [list(map(float, e)) for e in response["embeddings"]] + total_tokens = response["usage"]["total_tokens"] + prompt_tokens = response["usage"]["prompt_tokens"] + return embeddings, prompt_tokens, total_tokens + + def _calc_response_usage(self, model: str, credentials: dict, tokens: int, total_tokens: int) -> EmbeddingUsage: + """ + Calculate response usage + + :param model: model name + :param credentials: model credentials + :param tokens: prompt tokens + :param total_tokens: total tokens + :return: usage + """ + # get input price info + input_price_info = self.get_price( + model=model, + credentials=credentials, + price_type=PriceType.INPUT, + tokens=tokens, + ) + + # transform usage + usage = EmbeddingUsage( + tokens=tokens, + total_tokens=total_tokens, + unit_price=input_price_info.unit_price, + price_unit=input_price_info.unit, + total_price=input_price_info.total_amount, + currency=input_price_info.currency, + latency=time.perf_counter() - self.started_at, + ) + + return usage diff --git a/api/poetry.lock b/api/poetry.lock index 78816683d8d24d..184cdb9e81e57d 100644 --- a/api/poetry.lock +++ b/api/poetry.lock @@ -4135,6 +4135,20 @@ files = [ {file = "joblib-1.4.2.tar.gz", hash = "sha256:2382c5816b2636fbd20a09e0f4e9dad4736765fdfb7dca582943b9c1366b3f0e"}, ] +[[package]] +name = "jsonlines" +version = "4.0.0" +description = "Library with helpers for the jsonlines file format" +optional = false +python-versions = ">=3.8" +files = [ + {file = "jsonlines-4.0.0-py3-none-any.whl", hash = "sha256:185b334ff2ca5a91362993f42e83588a360cf95ce4b71a73548502bda52a7c55"}, + {file = "jsonlines-4.0.0.tar.gz", hash = "sha256:0c6d2c09117550c089995247f605ae4cf77dd1533041d366351f6f298822ea74"}, +] + +[package.dependencies] +attrs = ">=19.2.0" + [[package]] name = "jsonpath-ng" version = "1.6.1" @@ -4469,6 +4483,24 @@ files = [ {file = "llvmlite-0.43.0.tar.gz", hash = "sha256:ae2b5b5c3ef67354824fb75517c8db5fbe93bc02cd9671f3c62271626bc041d5"}, ] +[[package]] +name = "loguru" +version = "0.7.2" +description = "Python logging made (stupidly) simple" +optional = false +python-versions = ">=3.5" +files = [ + {file = "loguru-0.7.2-py3-none-any.whl", hash = "sha256:003d71e3d3ed35f0f8984898359d65b79e5b21943f78af86aa5491210429b8eb"}, + {file = "loguru-0.7.2.tar.gz", hash = "sha256:e671a53522515f34fd406340ee968cb9ecafbc4b36c679da03c18fd8d0bd51ac"}, +] + +[package.dependencies] +colorama = {version = ">=0.3.4", markers = "sys_platform == \"win32\""} +win32-setctime = {version = ">=1.0.0", markers = "sys_platform == \"win32\""} + +[package.extras] +dev = ["Sphinx (==7.2.5)", "colorama (==0.4.5)", "colorama (==0.4.6)", "exceptiongroup (==1.1.3)", "freezegun (==1.1.0)", "freezegun (==1.2.2)", "mypy (==v0.910)", "mypy (==v0.971)", "mypy (==v1.4.1)", "mypy (==v1.5.1)", "pre-commit (==3.4.0)", "pytest (==6.1.2)", "pytest (==7.4.0)", "pytest-cov (==2.12.1)", "pytest-cov (==4.1.0)", "pytest-mypy-plugins (==1.9.3)", "pytest-mypy-plugins (==3.0.0)", "sphinx-autobuild (==2021.3.14)", "sphinx-rtd-theme (==1.3.0)", "tox (==3.27.1)", "tox (==4.11.0)"] + [[package]] name = "lxml" version = "5.3.0" @@ -5320,6 +5352,36 @@ plot = ["matplotlib"] tgrep = ["pyparsing"] twitter = ["twython"] +[[package]] +name = "nomic" +version = "3.1.2" +description = "The official Nomic python client." +optional = false +python-versions = "*" +files = [ + {file = "nomic-3.1.2.tar.gz", hash = "sha256:2de1ab1dcf2429011c92987bb2f1eafe1a3a4901c3185b18f994bf89616f606d"}, +] + +[package.dependencies] +click = "*" +jsonlines = "*" +loguru = "*" +numpy = "*" +pandas = "*" +pillow = "*" +pyarrow = "*" +pydantic = "*" +pyjwt = "*" +requests = "*" +rich = "*" +tqdm = "*" + +[package.extras] +all = ["nomic[aws,local]"] +aws = ["boto3", "sagemaker"] +dev = ["black (==24.3.0)", "cairosvg", "coverage", "isort", "mkautodoc", "mkdocs-jupyter", "mkdocs-material", "mkdocstrings[python]", "myst-parser", "nomic[all]", "pandas", "pillow", "pylint", "pyright", "pytest", "pytorch-lightning", "twine"] +local = ["gpt4all (>=2.5.0,<3)"] + [[package]] name = "novita-client" version = "0.5.7" @@ -9919,6 +9981,20 @@ files = [ beautifulsoup4 = "*" requests = ">=2.0.0,<3.0.0" +[[package]] +name = "win32-setctime" +version = "1.1.0" +description = "A small Python utility to set file creation time on Windows" +optional = false +python-versions = ">=3.5" +files = [ + {file = "win32_setctime-1.1.0-py3-none-any.whl", hash = "sha256:231db239e959c2fe7eb1d7dc129f11172354f98361c4fa2d6d2d7e278baa8aad"}, + {file = "win32_setctime-1.1.0.tar.gz", hash = "sha256:15cf5750465118d6929ae4de4eb46e8edae9a5634350c01ba582df868e932cb2"}, +] + +[package.extras] +dev = ["black (>=19.3b0)", "pytest (>=4.6.2)"] + [[package]] name = "wrapt" version = "1.16.0" @@ -10422,4 +10498,4 @@ cffi = ["cffi (>=1.11)"] [metadata] lock-version = "2.0" python-versions = ">=3.10,<3.13" -content-hash = "eb7ef7be5c7790e214f37f17f92b69407ad557cb80055ef7e49e36eb51b3fca6" +content-hash = "17c4108d92c415d987f8b437ea3e0484c5601a05bfe175339a8546c93c159bc5" diff --git a/api/pyproject.toml b/api/pyproject.toml index 066b4772a99184..41244f516c95c4 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -100,6 +100,7 @@ exclude = [ OPENAI_API_KEY = "sk-IamNotARealKeyJustForMockTestKawaiiiiiiiiii" UPSTAGE_API_KEY = "up-aaaaaaaaaaaaaaaaaaaa" FIREWORKS_API_KEY = "fw_aaaaaaaaaaaaaaaaaaaa" +NOMIC_API_KEY = "nk-aaaaaaaaaaaaaaaaaaaa" AZURE_OPENAI_API_BASE = "https://difyai-openai.openai.azure.com" AZURE_OPENAI_API_KEY = "xxxxb1707exxxxxxxxxxaaxxxxxf94" ANTHROPIC_API_KEY = "sk-ant-api11-IamNotARealKeyJustForMockTestKawaiiiiiiiiii-NotBaka-ASkksz" @@ -217,6 +218,7 @@ azure-ai-inference = "^1.0.0b3" volcengine-python-sdk = {extras = ["ark"], version = "^1.0.98"} oci = "^2.133.0" tos = "^2.7.1" +nomic = "^3.1.2" [tool.poetry.group.indriect.dependencies] kaleido = "0.2.1" rank-bm25 = "~0.2.2" diff --git a/api/tests/integration_tests/model_runtime/__mock/nomic_embeddings.py b/api/tests/integration_tests/model_runtime/__mock/nomic_embeddings.py new file mode 100644 index 00000000000000..281e866e45c2e9 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/__mock/nomic_embeddings.py @@ -0,0 +1,59 @@ +import os +from collections.abc import Callable +from typing import Any, Literal, Union + +import pytest + +# import monkeypatch +from _pytest.monkeypatch import MonkeyPatch +from nomic import embed + + +def create_embedding(texts: list[str], model: str, **kwargs: Any) -> dict: + texts_len = len(texts) + + foo_embedding_sample = 0.123456 + + combined = { + "embeddings": [[foo_embedding_sample for _ in range(768)] for _ in range(texts_len)], + "usage": {"prompt_tokens": texts_len, "total_tokens": texts_len}, + "model": model, + "inference_mode": "remote", + } + + return combined + + +def mock_nomic( + monkeypatch: MonkeyPatch, + methods: list[Literal["text_embedding"]], +) -> Callable[[], None]: + """ + mock nomic module + + :param monkeypatch: pytest monkeypatch fixture + :return: unpatch function + """ + + def unpatch() -> None: + monkeypatch.undo() + + if "text_embedding" in methods: + monkeypatch.setattr(embed, "text", create_embedding) + + return unpatch + + +MOCK = os.getenv("MOCK_SWITCH", "false").lower() == "true" + + +@pytest.fixture +def setup_nomic_mock(request, monkeypatch): + methods = request.param if hasattr(request, "param") else [] + if MOCK: + unpatch = mock_nomic(monkeypatch, methods=methods) + + yield + + if MOCK: + unpatch() diff --git a/api/tests/integration_tests/model_runtime/nomic/__init__.py b/api/tests/integration_tests/model_runtime/nomic/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/tests/integration_tests/model_runtime/nomic/test_embeddings.py b/api/tests/integration_tests/model_runtime/nomic/test_embeddings.py new file mode 100644 index 00000000000000..52dc96ee95c1bc --- /dev/null +++ b/api/tests/integration_tests/model_runtime/nomic/test_embeddings.py @@ -0,0 +1,62 @@ +import os + +import pytest + +from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.nomic.text_embedding.text_embedding import NomicTextEmbeddingModel +from tests.integration_tests.model_runtime.__mock.nomic_embeddings import setup_nomic_mock + + +@pytest.mark.parametrize("setup_nomic_mock", [["text_embedding"]], indirect=True) +def test_validate_credentials(setup_nomic_mock): + model = NomicTextEmbeddingModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials( + model="nomic-embed-text-v1.5", + credentials={ + "nomic_api_key": "invalid_key", + }, + ) + + model.validate_credentials( + model="nomic-embed-text-v1.5", + credentials={ + "nomic_api_key": os.environ.get("NOMIC_API_KEY"), + }, + ) + + +@pytest.mark.parametrize("setup_nomic_mock", [["text_embedding"]], indirect=True) +def test_invoke_model(setup_nomic_mock): + model = NomicTextEmbeddingModel() + + result = model.invoke( + model="nomic-embed-text-v1.5", + credentials={ + "nomic_api_key": os.environ.get("NOMIC_API_KEY"), + }, + texts=["hello", "world"], + user="foo", + ) + + assert isinstance(result, TextEmbeddingResult) + assert result.model == "nomic-embed-text-v1.5" + assert len(result.embeddings) == 2 + assert result.usage.total_tokens == 2 + + +@pytest.mark.parametrize("setup_nomic_mock", [["text_embedding"]], indirect=True) +def test_get_num_tokens(setup_nomic_mock): + model = NomicTextEmbeddingModel() + + num_tokens = model.get_num_tokens( + model="nomic-embed-text-v1.5", + credentials={ + "nomic_api_key": os.environ.get("NOMIC_API_KEY"), + }, + texts=["hello", "world"], + ) + + assert num_tokens == 2 diff --git a/api/tests/integration_tests/model_runtime/nomic/test_provider.py b/api/tests/integration_tests/model_runtime/nomic/test_provider.py new file mode 100644 index 00000000000000..6cad400c069555 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/nomic/test_provider.py @@ -0,0 +1,22 @@ +import os + +import pytest + +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.nomic.nomic import NomicAtlasProvider +from core.model_runtime.model_providers.nomic.text_embedding.text_embedding import NomicTextEmbeddingModel +from tests.integration_tests.model_runtime.__mock.nomic_embeddings import setup_nomic_mock + + +@pytest.mark.parametrize("setup_nomic_mock", [["text_embedding"]], indirect=True) +def test_validate_provider_credentials(setup_nomic_mock): + provider = NomicAtlasProvider() + + with pytest.raises(CredentialsValidateFailedError): + provider.validate_provider_credentials(credentials={}) + + provider.validate_provider_credentials( + credentials={ + "nomic_api_key": os.environ.get("NOMIC_API_KEY"), + }, + ) diff --git a/dev/pytest/pytest_model_runtime.sh b/dev/pytest/pytest_model_runtime.sh index 4c1c6bf4f3ab19..4c0083a2de3435 100755 --- a/dev/pytest/pytest_model_runtime.sh +++ b/dev/pytest/pytest_model_runtime.sh @@ -7,4 +7,5 @@ pytest api/tests/integration_tests/model_runtime/anthropic \ api/tests/integration_tests/model_runtime/google api/tests/integration_tests/model_runtime/xinference \ api/tests/integration_tests/model_runtime/huggingface_hub/test_llm.py \ api/tests/integration_tests/model_runtime/upstage \ - api/tests/integration_tests/model_runtime/fireworks + api/tests/integration_tests/model_runtime/fireworks \ + api/tests/integration_tests/model_runtime/nomic From 8cc9e683631e155e9c7c1c282e91f9933f2858ba Mon Sep 17 00:00:00 2001 From: crazywoola <100913391+crazywoola@users.noreply.github.com> Date: Mon, 23 Sep 2024 21:00:34 +0900 Subject: [PATCH 13/79] fix: prompt for the follow-up suggestions (#8685) --- api/core/llm_generator/prompts.py | 1 - 1 file changed, 1 deletion(-) diff --git a/api/core/llm_generator/prompts.py b/api/core/llm_generator/prompts.py index c40b6d180804cf..e5b678451600a3 100644 --- a/api/core/llm_generator/prompts.py +++ b/api/core/llm_generator/prompts.py @@ -65,7 +65,6 @@ "Please help me predict the three most likely questions that human would ask, " "and keeping each question under 20 characters.\n" "MAKE SURE your output is the SAME language as the Assistant's latest response" - "(if the main response is written in Chinese, then the language of your output must be using Chinese.)!\n" "The output must be an array in JSON format following the specified schema:\n" '["question1","question2","question3"]\n' ) From bef83a4d2ecae95b5b07be68b10fe5181cfa0ac1 Mon Sep 17 00:00:00 2001 From: Nam Vu Date: Mon, 23 Sep 2024 20:32:58 +0700 Subject: [PATCH 14/79] fix: typos and improve naming conventions: (#8687) --- api/commands.py | 2 +- .../api_resource/chat/async_completions.py | 2 +- .../api_resource/chat/completions.py | 2 +- .../zhipuai/zhipuai_sdk/core/_base_models.py | 3 +-- .../zhipuai/zhipuai_sdk/core/_http_client.py | 12 +++++----- .../zhipuai_sdk/core/_legacy_response.py | 2 +- .../zhipuai/zhipuai_sdk/core/_response.py | 4 ++-- .../types/knowledge/document/__init__.py | 4 ++-- .../types/knowledge/document/document.py | 6 ++--- .../nodes/end/end_stream_processor.py | 22 +++++++++---------- web/app/components/base/chat/chat/hooks.ts | 4 ++-- .../base/image-uploader/image-preview.tsx | 4 ++-- .../develop/template/template_workflow.en.mdx | 2 +- .../workflow/hooks/use-workflow-run.ts | 6 ++--- 14 files changed, 37 insertions(+), 38 deletions(-) diff --git a/api/commands.py b/api/commands.py index b8fc81af673afe..7ef4aed7f77664 100644 --- a/api/commands.py +++ b/api/commands.py @@ -652,7 +652,7 @@ def fix_app_site_missing(): app_was_created.send(app, account=account) except Exception as e: failed_app_ids.append(app_id) - click.echo(click.style("FFailed to fix missing site for app {}".format(app_id), fg="red")) + click.echo(click.style("Failed to fix missing site for app {}".format(app_id), fg="red")) logging.exception(f"Fix app related site missing issue failed, error: {e}") continue diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/async_completions.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/async_completions.py index d8ecc310644d17..05510a3ec421d0 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/async_completions.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/async_completions.py @@ -57,7 +57,7 @@ def create( if temperature <= 0: do_sample = False temperature = 0.01 - # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间，do_sample重写为:false（参数top_p temperture不生效）") # noqa: E501 + # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间，do_sample重写为:false（参数top_p temperature不生效）") # noqa: E501 if temperature >= 1: temperature = 0.99 # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间") diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/completions.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/completions.py index 1c23473a03ae32..8e5bb454e6ce7e 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/completions.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/api_resource/chat/completions.py @@ -60,7 +60,7 @@ def create( if temperature <= 0: do_sample = False temperature = 0.01 - # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间，do_sample重写为:false（参数top_p temperture不生效）") # noqa: E501 + # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间，do_sample重写为:false（参数top_p temperature不生效）") # noqa: E501 if temperature >= 1: temperature = 0.99 # logger.warning("temperature:取值范围是：(0.0, 1.0) 开区间") diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py index 5e9a7e0a987e28..6d8ba700b7b1dc 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py @@ -630,8 +630,7 @@ def validate_type(*, type_: type[_T], value: object) -> _T: return cast(_T, _validate_non_model_type(type_=type_, value=value)) -# our use of subclasssing here causes weirdness for type checkers, -# so we just pretend that we don't subclass +# Subclassing here confuses type checkers, so we treat this class as non-inheriting. if TYPE_CHECKING: GenericModel = BaseModel else: diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_http_client.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_http_client.py index d0f933d8141389..ffdafb85d581fe 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_http_client.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_http_client.py @@ -169,7 +169,7 @@ def _set_private_attributes( # Pydantic uses a custom `__iter__` method to support casting BaseModels # to dictionaries. e.g. dict(model). # As we want to support `for item in page`, this is inherently incompatible - # with the default pydantic behaviour. It is not possible to support both + # with the default pydantic behavior. It is not possible to support both # use cases at once. Fortunately, this is not a big deal as all other pydantic # methods should continue to work as expected as there is an alternative method # to cast a model to a dictionary, model.dict(), which is used internally @@ -356,16 +356,16 @@ def _build_request(self, options: FinalRequestOptions) -> httpx.Request: **kwargs, ) - def _object_to_formfata(self, key: str, value: Data | Mapping[object, object]) -> list[tuple[str, str]]: + def _object_to_formdata(self, key: str, value: Data | Mapping[object, object]) -> list[tuple[str, str]]: items = [] if isinstance(value, Mapping): for k, v in value.items(): - items.extend(self._object_to_formfata(f"{key}[{k}]", v)) + items.extend(self._object_to_formdata(f"{key}[{k}]", v)) return items if isinstance(value, list | tuple): for v in value: - items.extend(self._object_to_formfata(key + "[]", v)) + items.extend(self._object_to_formdata(key + "[]", v)) return items def _primitive_value_to_str(val) -> str: @@ -385,7 +385,7 @@ def _primitive_value_to_str(val) -> str: return [(key, str_data)] def _make_multipartform(self, data: Mapping[object, object]) -> dict[str, object]: - items = flatten(list(starmap(self._object_to_formfata, data.items()))) + items = flatten(list(starmap(self._object_to_formdata, data.items()))) serialized: dict[str, object] = {} for key, value in items: @@ -620,7 +620,7 @@ def _process_response( stream: bool, stream_cls: type[StreamResponse] | None, ) -> ResponseT: - # _legacy_response with raw_response_header to paser method + # _legacy_response with raw_response_header to parser method if response.request.headers.get(RAW_RESPONSE_HEADER) == "true": return cast( ResponseT, diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_legacy_response.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_legacy_response.py index 47183b9eee9c0d..51bf21bcdc17a8 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_legacy_response.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_legacy_response.py @@ -87,7 +87,7 @@ def parse(self, *, to: type[_T] | None = None) -> R | _T: For lower-level control, see `.read()`, `.json()`, `.iter_bytes()`. - You can customise the type that the response is parsed into through + You can customize the type that the response is parsed into through the `to` argument, e.g. ```py diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_response.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_response.py index 45443da662d57e..92e601805569f3 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_response.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_response.py @@ -252,7 +252,7 @@ def parse(self, *, to: type[_T] | None = None) -> R | _T: For lower-level control, see `.read()`, `.json()`, `.iter_bytes()`. - You can customise the type that the response is parsed into through + You can customize the type that the response is parsed into through the `to` argument, e.g. ```py @@ -363,7 +363,7 @@ class StreamAlreadyConsumed(ZhipuAIError): # noqa: N818 # ^ error ``` - If you want this behaviour you'll need to either manually accumulate the response + If you want this behavior you'll need to either manually accumulate the response content or call `await response.read()` before iterating over the stream. """ diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/__init__.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/__init__.py index 32e23e6dab3076..59cb41d7124a7f 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/__init__.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/__init__.py @@ -1,8 +1,8 @@ -from .document import DocumentData, DocumentFailedInfo, DocumentObject, DocumentSuccessinfo +from .document import DocumentData, DocumentFailedInfo, DocumentObject, DocumentSuccessInfo __all__ = [ "DocumentData", "DocumentObject", - "DocumentSuccessinfo", + "DocumentSuccessInfo", "DocumentFailedInfo", ] diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/document.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/document.py index b9a1646391ece8..980bc6f4a7c40d 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/document.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/types/knowledge/document/document.py @@ -2,10 +2,10 @@ from ....core import BaseModel -__all__ = ["DocumentData", "DocumentObject", "DocumentSuccessinfo", "DocumentFailedInfo"] +__all__ = ["DocumentData", "DocumentObject", "DocumentSuccessInfo", "DocumentFailedInfo"] -class DocumentSuccessinfo(BaseModel): +class DocumentSuccessInfo(BaseModel): documentId: Optional[str] = None """文件id""" filename: Optional[str] = None @@ -24,7 +24,7 @@ class DocumentFailedInfo(BaseModel): class DocumentObject(BaseModel): """文档信息""" - successInfos: Optional[list[DocumentSuccessinfo]] = None + successInfos: Optional[list[DocumentSuccessInfo]] = None """上传成功的文件信息""" failedInfos: Optional[list[DocumentFailedInfo]] = None """上传失败的文件信息""" diff --git a/api/core/workflow/nodes/end/end_stream_processor.py b/api/core/workflow/nodes/end/end_stream_processor.py index 0366d7965d7c17..1aecf863ac5fb9 100644 --- a/api/core/workflow/nodes/end/end_stream_processor.py +++ b/api/core/workflow/nodes/end/end_stream_processor.py @@ -22,8 +22,8 @@ def __init__(self, graph: Graph, variable_pool: VariablePool) -> None: for end_node_id, _ in self.end_stream_param.end_stream_variable_selector_mapping.items(): self.route_position[end_node_id] = 0 self.current_stream_chunk_generating_node_ids: dict[str, list[str]] = {} - self.has_outputed = False - self.outputed_node_ids = set() + self.has_output = False + self.output_node_ids = set() def process(self, generator: Generator[GraphEngineEvent, None, None]) -> Generator[GraphEngineEvent, None, None]: for event in generator: @@ -34,11 +34,11 @@ def process(self, generator: Generator[GraphEngineEvent, None, None]) -> Generat yield event elif isinstance(event, NodeRunStreamChunkEvent): if event.in_iteration_id: - if self.has_outputed and event.node_id not in self.outputed_node_ids: + if self.has_output and event.node_id not in self.output_node_ids: event.chunk_content = "\n" + event.chunk_content - self.outputed_node_ids.add(event.node_id) - self.has_outputed = True + self.output_node_ids.add(event.node_id) + self.has_output = True yield event continue @@ -53,11 +53,11 @@ def process(self, generator: Generator[GraphEngineEvent, None, None]) -> Generat ) if stream_out_end_node_ids: - if self.has_outputed and event.node_id not in self.outputed_node_ids: + if self.has_output and event.node_id not in self.output_node_ids: event.chunk_content = "\n" + event.chunk_content - self.outputed_node_ids.add(event.node_id) - self.has_outputed = True + self.output_node_ids.add(event.node_id) + self.has_output = True yield event elif isinstance(event, NodeRunSucceededEvent): yield event @@ -124,11 +124,11 @@ def _generate_stream_outputs_when_node_finished( if text: current_node_id = value_selector[0] - if self.has_outputed and current_node_id not in self.outputed_node_ids: + if self.has_output and current_node_id not in self.output_node_ids: text = "\n" + text - self.outputed_node_ids.add(current_node_id) - self.has_outputed = True + self.output_node_ids.add(current_node_id) + self.has_output = True yield NodeRunStreamChunkEvent( id=event.id, node_id=event.node_id, diff --git a/web/app/components/base/chat/chat/hooks.ts b/web/app/components/base/chat/chat/hooks.ts index dfb5a1b6855d5f..64c238f9d1e56a 100644 --- a/web/app/components/base/chat/chat/hooks.ts +++ b/web/app/components/base/chat/chat/hooks.ts @@ -334,9 +334,9 @@ export const useChat = ( const newChatList = produce(chatListRef.current, (draft) => { const index = draft.findIndex(item => item.id === responseItem.id) if (index !== -1) { - const requestion = draft[index - 1] + const question = draft[index - 1] draft[index - 1] = { - ...requestion, + ...question, } draft[index] = { ...draft[index], diff --git a/web/app/components/base/image-uploader/image-preview.tsx b/web/app/components/base/image-uploader/image-preview.tsx index e5bd4c1bbc0dc5..096facabfd2a85 100644 --- a/web/app/components/base/image-uploader/image-preview.tsx +++ b/web/app/components/base/image-uploader/image-preview.tsx @@ -88,7 +88,7 @@ const ImagePreview: FC = ({ }) } - const imageTobase64ToBlob = (base64: string, type = 'image/png'): Blob => { + const imageBase64ToBlob = (base64: string, type = 'image/png'): Blob => { const byteCharacters = atob(base64) const byteArrays = [] @@ -109,7 +109,7 @@ const ImagePreview: FC = ({ const shareImage = async () => { try { const base64Data = url.split(',')[1] - const blob = imageTobase64ToBlob(base64Data, 'image/png') + const blob = imageBase64ToBlob(base64Data, 'image/png') await navigator.clipboard.write([ new ClipboardItem({ diff --git a/web/app/components/develop/template/template_workflow.en.mdx b/web/app/components/develop/template/template_workflow.en.mdx index 2bd0fe9daf436d..5c712c2c2979e5 100644 --- a/web/app/components/develop/template/template_workflow.en.mdx +++ b/web/app/components/develop/template/template_workflow.en.mdx @@ -424,7 +424,7 @@ Workflow applications offers non-session support and is ideal for translation, a /> - Returns worklfow logs, with the first page returning the latest `{limit}` messages, i.e., in reverse order. + Returns workflow logs, with the first page returning the latest `{limit}` messages, i.e., in reverse order. ### Query diff --git a/web/app/components/workflow/hooks/use-workflow-run.ts b/web/app/components/workflow/hooks/use-workflow-run.ts index e1da503f3825f1..68c3ff0a4b458d 100644 --- a/web/app/components/workflow/hooks/use-workflow-run.ts +++ b/web/app/components/workflow/hooks/use-workflow-run.ts @@ -185,7 +185,7 @@ export const useWorkflowRun = () => { draft.forEach((edge) => { edge.data = { ...edge.data, - _runned: false, + _run: false, } }) }) @@ -292,7 +292,7 @@ export const useWorkflowRun = () => { const newEdges = produce(edges, (draft) => { draft.forEach((edge) => { if (edge.target === data.node_id && incomeNodesId.includes(edge.source)) - edge.data = { ...edge.data, _runned: true } as any + edge.data = { ...edge.data, _run: true } as any }) }) setEdges(newEdges) @@ -416,7 +416,7 @@ export const useWorkflowRun = () => { const edge = draft.find(edge => edge.target === data.node_id && edge.source === prevNodeId) if (edge) - edge.data = { ...edge.data, _runned: true } as any + edge.data = { ...edge.data, _run: true } as any }) setEdges(newEdges) From 7f1b0288409626fcb87bee2fe12643ebfa0f3059 Mon Sep 17 00:00:00 2001 From: Sa Zhang <55871322+Nick17t@users.noreply.github.com> Date: Mon, 23 Sep 2024 21:39:26 +0800 Subject: [PATCH 15/79] fix: change the brand name to Jina AI (#8691) Co-authored-by: sa zhang --- .../model_providers/jina/jina.yaml | 6 +++--- .../tools/provider/builtin/jina/jina.yaml | 20 +++++++++---------- 2 files changed, 13 insertions(+), 13 deletions(-) diff --git a/api/core/model_runtime/model_providers/jina/jina.yaml b/api/core/model_runtime/model_providers/jina/jina.yaml index 9c70d6ff33ba41..2f2d6e6daad73c 100644 --- a/api/core/model_runtime/model_providers/jina/jina.yaml +++ b/api/core/model_runtime/model_providers/jina/jina.yaml @@ -1,6 +1,6 @@ -provider: jina +provider: Jina AI label: - en_US: Jina + en_US: Jina AI description: en_US: Embedding and Rerank Model Supported icon_small: @@ -11,7 +11,7 @@ background: "#EFFDFD" help: title: en_US: Get your API key from Jina AI - zh_Hans: 从 Jina 获取 API Key + zh_Hans: 从 Jina AI 获取 API Key url: en_US: https://jina.ai/ supported_model_types: diff --git a/api/core/tools/provider/builtin/jina/jina.yaml b/api/core/tools/provider/builtin/jina/jina.yaml index 06f23382d92a3a..9ce5cbd6d1afa3 100644 --- a/api/core/tools/provider/builtin/jina/jina.yaml +++ b/api/core/tools/provider/builtin/jina/jina.yaml @@ -1,10 +1,10 @@ identity: author: Dify - name: jina + name: Jina AI label: - en_US: Jina - zh_Hans: Jina - pt_BR: Jina + en_US: Jina AI + zh_Hans: Jina AI + pt_BR: Jina AI description: en_US: Convert any URL to an LLM-friendly input or perform searches on the web for grounding information. Experience improved output for your agent and RAG systems at no cost. zh_Hans: 将任何URL转换为LLM易读的输入或在网页上搜索引擎上搜索引擎。 @@ -22,11 +22,11 @@ credentials_for_provider: zh_Hans: API 密钥（可留空） pt_BR: Chave API (deixe vazio se você não tiver uma) placeholder: - en_US: Please enter your Jina API key - zh_Hans: 请输入你的 Jina API 密钥 - pt_BR: Por favor, insira sua chave de API do Jina + en_US: Please enter your Jina AI API key + zh_Hans: 请输入你的 Jina AI API 密钥 + pt_BR: Por favor, insira sua chave de API do Jina AI help: - en_US: Get your Jina API key from Jina (optional, but you can get a higher rate) - zh_Hans: 从 Jina 获取您的 Jina API 密钥（非必须，能得到更高的速率） - pt_BR: Obtenha sua chave de API do Jina na Jina (opcional, mas você pode obter uma taxa mais alta) + en_US: Get your Jina AI API key from Jina AI (optional, but you can get a higher rate) + zh_Hans: 从 Jina AI 获取您的 Jina AI API 密钥（非必须，能得到更高的速率） + pt_BR: Obtenha sua chave de API do Jina AI na Jina AI (opcional, mas você pode obter uma taxa mais alta) url: https://jina.ai From 21e9608b239036c98367193e32ade356c886de99 Mon Sep 17 00:00:00 2001 From: themanforfree Date: Tue, 24 Sep 2024 10:20:06 +0800 Subject: [PATCH 16/79] feat: add xinference sd web ui api tool (#8385) Signed-off-by: themanforfree --- .../builtin/xinference/_assets/icon.png | Bin 0 -> 57667 bytes .../xinference/tools/stable_diffusion.py | 412 ++++++++++++++++++ .../xinference/tools/stable_diffusion.yaml | 87 ++++ .../provider/builtin/xinference/xinference.py | 18 + .../builtin/xinference/xinference.yaml | 40 ++ 5 files changed, 557 insertions(+) create mode 100644 api/core/tools/provider/builtin/xinference/_assets/icon.png create mode 100644 api/core/tools/provider/builtin/xinference/tools/stable_diffusion.py create mode 100644 api/core/tools/provider/builtin/xinference/tools/stable_diffusion.yaml create mode 100644 api/core/tools/provider/builtin/xinference/xinference.py create mode 100644 api/core/tools/provider/builtin/xinference/xinference.yaml diff --git a/api/core/tools/provider/builtin/xinference/_assets/icon.png b/api/core/tools/provider/builtin/xinference/_assets/icon.png new file mode 100644 index 0000000000000000000000000000000000000000..e58cacbd123b5887b34fc8414d8b57aa801bb690 GIT binary patch literal 57667 zcmeFZg;!N=&_2BB=Fp+Qp;5X7Bo84eDIwh=NOyB60Z~d2kZ$Qlx=Xr2Iz+m=f7?gj z^}YYXx7N#A2R7$y_8l|#%r)1{5T>FegM;}96951lIaw(+06?Pt^Fl`j-qfUpS;Gt#o@*O}lzgVthoC1P{6Ynx zQlXK`k%$=049MFe{_XPgaD5}vh8aSyfgzu>m(X#y>7V*qni_A`RXr>y_N8wUWz1ek zOV=jVC3#7k`HG!o($WpI57KgvEi|83FAWE&8F# z1N{HRY^c)zXS_!Ed)IMd+`hUh@Sbch%CE=*B0D(kN1J{th zRuC=rKjJChof~Tn1zgXJr;xK9wUK^CFZP{_{k=7$lD?)UuWr z9!qT-_zHh-YPiCk!FWpg!)qVgBa&{9ea7(LJbnd{P!}fsdf3v{aYu#@zoQw$;8MwK z4~Nj5BkczZ`icCq4J4;O1l)+Gx8g#okadwpt2mjy@z#cms@^C{cXOJ?Eu&HT(t)IP! zazplSzhgI9wXh0}2YR~Zh#Gwp0}{QYlHyW`=S+v@D56(68f90}!Bzf`?5P=&P2wx{ zwulGkG08amt&&AiKS zj&0W0z^=t14%(k-$%c|pc$&67-}BB!X-EwW1pYz@IK%O_{&9<&L66r&%N3RC{S;P;Lq0lMM8{7!sg zocf9MnzK?OZN4TY_ESRK#u(VMFU5k95OHz;!4A~29r3t&rSB{UHUbW!Ut19n&@ldO zkFEO7uNE&wv;?h(d+67u+(8UW536ayF29pX*#<0J3aWnq2&M%7)t^T79y5SkO8%eK=#WdDb}ur4*nEHgv}?hc34w{wFgyQrP^r zC)M~-jeMlGyfZ(}HiN>@f7)gjZ}yU@*2!jmq#+ZiU7mx0i-RM{*_&$$K_~q#OV54> zELdvbuimG~4an#=tttyEt!6ulqH-c5p^KF)eUvdteJDSR_KQYR zS+r`zWN9QAWvT)Ebpix?mB03>)t}b=jP?g;zWU>E1->)-qg$$6hB19RT2a*H!)ZhH}C^6X= z#?&x> z7X5JELvP0()#w@@5FB@z5ZSw{Wpjr8B2>UDwlSY;*Ao|p^;d>7T5pezeiIk?-H61H z+Pz|QCZ;JS4#k8}(jAOiW{%Pz0zrQj!pN{}&-=;I%EKk*XuHnF2R<&S>KSdNVQ@sK zE6l*13>O_>`=2{Um_-Bl#WMA_RzuzU5S})=M*zUZ-;IVDb|Cz|hp?;Iuj5`p7gRxa zbJzZaw1e?CT|uLy$U(W~aV}H)L<0b){QI=@($T~cgC;!ow)!TiCLkO_LSuqm4h5?I zEx1|2R#SF}{d00+-ZlT?BqDt+pNwPH?A?o8p;W@>4ic}uiJC8YDewTnzZM*R-wru_ zAJ8_4Qt)uJv*P5CNo~=2noU5HhYplI{wtDGv`N9qmTM9er;FZ)ZlOL5O9$bXtIsQX zeJ3@Qg`kb6+qZfdGl&?+tJysbvL`SELh@ z#SW#=n{ASyix|5C`x^(m#jj#m&5MP)4bB65>#nbjx^1O?l^s#eBje9#Z5KBdSVeb(O)HGjAF6mENvK;m%7l_s7^ z1?*Cjd>#JL?U(5ipgJ~`C`m%4n3v_S@8H`S<5hINFPQN=EGUKtVt1XjkcOSKn^Tn*iWE|BOlOx36`7YLdy z98^o+i#a{?Rp0D?d|S)xSr<X46Pbz`S*1 z4}DCZSNWzsDyN(Xy-K%NNh4EODZg8+aa}Csc<*!chy=TV`r2jjPIp{&RH5$CMe(`; z-F{pUc&0gpy05EDwSGMdNAK9Q_610Di<2n{NvQ5M-G>ohVnv4? zCJ?WmmK}c6>U70xf}aEF98nZTCG(Iy?^D8Agtn3mAb~nN_VoY&L*+$Vg!-~oF|hkE zjT}94V@IC7TJ|ELN@BS%G;I>@8%f!`<6Po&Qd{<3OgcjNzfk_$8?IJz0 zOp__egp5N;;Cz`aL-$H5OQ!Q22Z-XR_B%@9?4Ps?zR>5Qtls^`guPAwWP?sjP_c_c zSTR^RC955415_4ByTCt(S%-jz43A}9ME_PV?!5O2#*tEgYz3=k%PbVbJaA)uX2n2j z<1E?_n6;1@E;Bop$Vm%}Du$|fifH5;a zEMK_HhdkW7G@V-ADDy;*8|$#|M&PTKUTPU;`#;Pr(F5@3wcdzEp+lk?knu&ni=Jg4 z9)WQpEWC*EWmffM#_!F<(Hk}WtX+{HP{K@cu+2z4IcW6ZLj#`h){I$wp zfFIt-F`L|l?Tgzwda~P%RATvT0oOS$?l3XAn|oT!ijOv1uHJh=$tR~u`#n%nUf*O> zIZR&PmkVN#_EpF93Q$^o^X^kFs^(N5Y~5SJ6GlU2@$>Kvdqf$fdVFswJAHvDdY%UwRmM9ErNGF1?6ZzwbJWqc3<45-OJzb$(Q9zQ&!(_*Ouh4)e zgm{jXtfsYO%b{v-q5AWO{ukY@T1ds&0T27{jyOVzhMldM$RDZ%pG}bgSulPOI6MT7 zrfv|57vr3_gnT!>8>v*9XkDyj*cHquaEJ8yZ49{(L{7;Yb}=~?bQ`H`(VaGg{o>ox z!UI`FC&+$(lzFw?EmW*q^! zHtPBM`vcye)R1(4+|*y#$&hqB(5%H@T~4U~6jV;v=#e>8n-dfsvXBHzYxV!_ja4!n zJcMS^WOXMbOBO(mk+nOq;$!$CKGG3|Qj`K-}~9gWv5tcLWv6TJQqnbL1!T&eeT zJmazB_RoNNrmvo|MmMiV_$U0_Q@8!D>q(%Ew|xxVwAo?hbhJW^_8VNU-9i2TyL>!W zJsb{ZH+ED(ugN$ls;9?*$h!OC8-IN{p32k^o+!R^xf#Jb9 zL!px~aCln{vnZ{;#)`w%S9H$h+YHk z99FKN>2{Ecvj`nN!L+872$!JLVpd)(A-605cZR{?K0i{h{0(%@w(TZPbN!*d)ok5# z+d4Cl%3}`-L0kJ&1ttQG?}FbxUD_|9Sx!!jeLD1NymYv{yq%Un;J$Mf&pY-xWZ-D@ ztNYf~v%F?64LdIbM!mfCa4#-rx@yoZeQ2|q%{GI|4vo-|fha0;yfThuqWZ=vt8evOG!qo>x_cPAf+Ou0dA`h$;2)7bF;!Z#{T@B}R^V zFnK-Isozox_@jhAOc7(ghJW)iefhCDnnIyY%biRX!n z6o&wkc8t8Yz>ZX(W5`S>L^t5k=(V@N*Lr>@KlQH16vtF_pDltH!&3BP@+%EY-CWV} z@cap=8|USFyl=POePVb`>B`6NO3CU9J&>kSl6_7@4FGPS<~5r*X>U{Z+ciG5ytjgB z8Q{FVUxV{$nVVQpe;0T7pnWzYT5FFwhH5TEF~j8g30u^3?qa&w*GcAh1tWtWje;uj zU3%Q%%_HsVb$g28UuZE_kCU%HoEq?{bvIYhRkO(zx?Ls2kk}#f9)3x0Q(ga%eE#-r zz~*69Ui{mmH=zvAR)mDNwgz;TjXJhm)@MmH{7&A@B4F4KSc;DkeYllNHci}Qvztlv#d5DL&|9Hp$K z+v46*i9FI*cJ{33?>Nun*P_3ACO#Ig1E2W*L=7&rGCgVY>|-M496E&8Rd{5-_msz< zghoI3*5i}$Qj%Qg|&MM0@kQ7%^g;KyCmz^b@#H_o`~?zOYB zWPRMlgZDfgM*nyrC850(F2P_`eYMza^L2&u%&2gVQ27ZN_jMOziTq2EpT5oE+ZV9t4{3fn zcMkFBz%q%RG|aFpLoc$fUoAZHXZZPyYB?o; zo0oOaMZs;*NProKwjs8R@(xx8PkTV}fk!?qRzz`}3f#I_mQA!~D!93{B*Yecj~YGR zG}vUlnMZ=OT|{~+zJ1s~-m-i~&{K&SAwaA#5HatkId05J`6mMET{r;D(`CQ3`h^rl zIcB$O=Br`V=CIcT`fa{R;GL=5_RnP{ZHkmPX7cX4dTYvKjV(2%34+v3hvWy$5&1n0 zA@L!|l%jj1t>~kpAtFuAAG%(6WYciRj&r4>{$!PK&rR!kpZ6kc>0OK3P3^yzyHdepDunJO@4Z)VSdd0(y|!2zt1NBMc>rd z#)Hf!)4w*KEBmoW-kOPj_t~#Zy%&~dEZnG7&PdePNiku{+zb(K=8rB$d0A<1*@Np_=$Wr{o*9hQR!a+Vro0ekA>E68 zqN?@^abw@y5R2GgzYFnKuVluu(MBf=-i!sZmaQd11)DG~+3GKP&>;5Wfn(<_FV9Y1 zadxlmiDAAy@`Ujmpn8(H7P@?1XiAQYe9D;;7da3R8^U9><^7s+?$;&x>B=cnyZBML zOuP1V79^xUTS=6B5fTx6gthBgmZi#CK8Jrla0WSseeHTN(p%OQfzcCPMyS1jN%`Y< zck1NX2>wq!48tDVu!vyQebrY9Z{xQpD58C76B4JnsLl?W9_RkN>(&L&S%iqFj>C%c zYp5jIR3(RVHV5#5A;iR zgwP*a31m(4-)yOdU=ofFqSiMi`ecTpXXylFk^xknT=RtiD5yQ|FU2T_>Q7(^Pk9d% z$nY9WcHdr-M^a&kjXQmM9hFic_O{J${hcIJ5m-E#~#<*KQ^O?fnquUQ4rpR$DNk z$^4g(?p%py`3fmxr6A)>4FLf`Y6n6|GOn|W()$}Pih1*%bR0@WY6%){P6FF|es1#p zA6!8>A|g24re}N=Ti!+dt0HhzkSx3c>9})b5{l%XR#!FIewPZw1zH{!C!wUL5>tf& zW3_}O6$V|#(H{LCb>fpNlG)>HsorDb-+4IDzY6Y*_Nv$7Q-RR5u_ffebJN#;H{|_8 z)hf5*&rYp`358{)U1mGYn1)ndAteRMh&$MhA`HKBwL6TnnJOF^N_Z#;=Dq5_*(l_aMQmkljdJKFaJ38)ntl zThj8Vu}K0#T{|EPVm@E&_-xnuMdBAa|{6fTco4K(fbL+pi(DqC()7W{A9`H zy1mwUt1As@Q|Ba6^#0)}=(Zw&=Fabkx%Qkg1JuGU{GhwA5dO_Y+ zW1n>Ywd3UbK{e*jQo)R8dh7>dh|JqT)hA?sk`KTYNQXLZhYkyDN>f16}WQa&c zNa8E%l#w?38S*jeV1{FYPv;HV)Mz-Z2yx(4(;%P`mRwh<+F{KVE9CXpku4 z40VwB7-6hEj4}uE+F|=de%B2DvHq}T5RJ^&@>iZ<#^&%unQ4mRTh52^9zE4_F zlqsC9QGKK`6yhi#->7?TR{QR`9aXPTEFlF1NCv6nBAe)6*jl%>k4qN9YK;!6Y!a#7 z$)wc_&<;A$g+q|jkfMa}sHyqy@3q?1BgN?!8vE~P$D+b6Wxc)$cduD?hA<>cU-&D# zJ7318E8kaV4mrF0{@m-20Do6?a&&j`q?SADfRduz!OmWS6CbrC8IR>6=cb})b+}n- z5#vvQ%@01_>9-i}w|XUYyM60aRlWB=6h7_ru(fdGDQjDzm){xcNjrCctVPI{Tl6q}olZ8cobQr5bIx_AXsZA%n zXL{lDQ0E89dkpAxtOhRNrfh8&mVVaL$ymOg-~W)mfcA3*u5Z8P+!?JBCL%DDwXAhy^zIU~FDhPERO=qk4A(mhCWEg^UEr(Txdbhv9Z?ebk(RC(&g?2F@8GyUj`E0%@U*LiGF8*vR2+co3d}U#2VR^W7j7T;GKIF;3*aOZU#OHv8HEZxwSG#7y&6J^cx}O5DEfy3FY}NS{mp$5N zNM0qMwv8TFD4pe{oqZFQI&D4K_EpzkcF;tiCcK>B6I${v zt`3*#C`42BCnCRdIdFbG`!!UcKa7yit@hkif_9FU1%Fk;9hdtCO;Fmz_vlx_6M`0^N zA4M_TH}UmaeA>3eOcKe|y0d*NjYdx-8Y_E#2=OcO3WAY&6B;6N_U}cu{PDdZ{TSy9 zdD!IOZl%GN36=%*k@-l5kn!DpFU8zUB=36KWfvl4Dv7f^MkGifd%OsR0D7xwEd{|c zcMWzSk8y_BzBX}!SwFVfqrlIU!;iW1&?*eWczRC9SOlp5TMG3vRv1YM3U48D@BF$M zmRHILh}N#ZP;^Je;NvZ|`&$l@tu|7R&*IrXQK+#3Af;Q1DDMpYB$sqk@L6BTGJlLP zFK6vX(Z5zVLOA{*gAR;5Ho64Vc)V!RSxT?fwQX#<9-km$O}AFeG&X~2v$6;6P-F|A zUh10Qlt?r~){#*s3A7}vNC3cHc?}_Ik4VAWrDl5e`a> zeq*JJ7*!MT&=uP8hhDkUEbaOwJ!>6*nb7NOd54%uUiN~|MI;u z)^z*EtLI2xDvvdL_`1p)OxC>3$mdM-7B$ZgErbOx5(5?w);6Cn?onV80;Kky`uWKb zvu;a_4}pm%N7}i10;mp3vS11-l5lb=LcJ-)z-$KOi70Oz_G5>9 zQGiVmY8d>P+WYTk9{D^*N4bGU4_tx_o0q0kpQ!M81yGL=;4Hu1Xb4fZp^65KL{VkGVk3VuFmz}f&4oGp)9={MW^iruX5V?O5 zNp&L;tt7^@IZ;VPEdS?B!Y?ikK_qhh?%NJ9YZN@0q$otBR(p+^z3-nh->s8|_JJwNQ!4O`K5`WnShsbNywl6Emi}!t`~(N% zQ6jc}%+V5~z=F)3zHB)yiH&14E)IBtH5^7S^%bWkC-!EOVtg-$ zeRRG?3o$Od{JB%4#mG@vkVJb$Ec-3w(hZxBMaZWTnMR zd!b}ge@G~XxmluSl5oq7XGxS~{WF6CnIH|mjYr?zVXClXs-pw|G+`m5b4vI4BFhI^ z%gyS%*zNJnQeoWlJY)>~qY*tdC!@9g*rT|za!W{nCn-3&fogtlkT!&~d(Y)HXl189 zkEF=W{ec-LGgVki`X=G5FMX!S_7Pf+DxahJNO=f~$ayfL&ym;@;}9o6t-Zf6Aj#>M zA1~p)mFV|x{Nxk#y|R;WwKI65IzMc5^DE7A93%uO!Yd7k{f^0*y;muSQ~iR#Y2v(p z1pM~}7N{a2nnZiD1)dbCfTBFzd#s##1W6yUd}jwGio?xcB%WJu&T-a?RXU!T^8u^ z`j*x&A4P-?L;(>;={hz-4H_zzKUFy8jjE<-(9USl&SG2#tj%UMbw%QBs;61vWn-}f z*W$>?vDd@^Km>w}9%{bM4)XFbw#|=6+Rv!xe&k}KZj>}cSwfH9{XX$^M_^o^E-{aV zF(E681x_Xa^e+~5GiD8MLx0WFPBx#;#{WKGPxuNWIGrjTmrB%fU zwIddeXgAxhAJTlj0KgprI9it$uB^=~s|&iqbSM2}zRfx0p8cTEC*dBJtTp|mCjF1r zC7x5FxUfKiwM##&Q>a6L8zNWXs*^VIlKPx!^N(LBx?-8~r>i4)32BsOX)g}M?8MOk zd(1xzaZwrzwtIT9X60&eJ6~a-T~*HH6Dm{_D6JsxP9em!8S5xwa6f5sYXJc^L-llW z1|PQ5-X7t5^P1kb3!{xV&~~mT7UdL|l+)Op#TY+H1vdy()XfONIyCP2D{mJAHFLU5 z+Yj0vOhd4*=_g-~S-V$JBqE@SRrMhDQythV9swe$lw_Uy2{21w`c*k1&oIdoX|zF%XiM*q1%$Xwj*q*;@VX zQ()(kpRfk-DQuW3_pD%a%BXe69=hL?%Im7=3_1UZyUTtK%Olm>t$1(*uwV@w;k@wdg`6%+Mf6?YuFDkZqy8La z#`W%!K+9^Pa%g_PqJasSA&;ZfLauChWSb;1z?oeaShs{neeF;^Wf0DdPvfv3DChk` ztj1gog3Av0V1PS>Y0*-DZz|`;p-@mXx){~Ayl{=zUeP?>z0kJhI_#4R>w%tZuB5@H zSO9CIjllfdH9z84JpSkoVRD?cEKHr5Y(F!eJbNvH30neF1@&tBOgfE2M^OAYsUB#9 z@Tg`Z6mdBCNa3>^!wc`ha{<|fcrY2^t~ZkLxz+^9NZ#HlBHK}L+b;D$FJq0mQ8B1W zbYYK7bzN@zOkSFK+QwnG2dSrbZ6;DCSMDP=SZwhChT8htDUi+W9Z@#D14oTsM~bA7 zOW8TJ9);XJOxDO!?8sH>(KSQnu~u{h06V|s%9!lrNXT!rNi7*YtSBBoY)*bzgZ{Pc z)XF$l0RJP_`_9T=DG4JN0Q?>q5qVbv)oN;(X;ue=GOD_PvV`^NoApX()Q_fOLE6>y z0y}|ap7rna7msc?pNrGI3V6@d4gxP=^b+bTXb?)z8VE5)XvAMx zfW<^`jB2rac9aXv-WGdBk#C7jZw;T*D6*PJIy!@f_ztYOn0evydGkC3VWGj?ce0`$ zP9})N_n$&M!+!-9x5b`XJ%0Xb#^5`{dEmvJz*$jV;0OYcGTM-6r%^M$d(Bg6oXb)Z zs9WxUZMFl!kdn=;sh@BnHzb(OL#EBXnkW$*#9XH(P8C)#aSp{y6Ob`63k3%T4;DDE z{@C70%$I`%47|KbEpnX2@zXYWZ!@V?#uuCLGT`WN8W*Nj1*k^I+=U#dwI*RNCiA$7 z)|Z=dvd>}m789BfU@yK?3Pqu~QFY&seh+wnM*9xuJ-rcWXOAy^V%&iN z&2h%#v2$^<7G4jnvCy)b3OWeZBO%7Omt`aVn_!(q*~o}OvFL4Rz}yLup3UybsJ*Dk zgA!qvu~`BGi&D_3y{>-8(uL6&8W6>XM{PtJ_e}2l;^!*vIoIdCYSh1IM90pJA@ig= z$9xgO^lwjPpMapLVI@biakt<+%kBHc^+;`UTO=V3yIdJ^Ujzvg{YlMi1OhMv22F7N ziNb6iL!07fC1g+0y7;C)b&CId?=!Ia=yx{6;b6Fi11C3)8RiX)Aoxi?hw@wGmjs0@ zE_mBISOj;c4W8T21)G)?8q8HNCM7^Vds^y!oT=55(E&Wj@PVOYg;u9IYIEmJ(zz=d9xzo_yGWJ zkiZ1iCv{KBYpH6?(e6z8r31ZTPtgVAmmMW}I6`!+n8oo1Ic)qMVdmb7PNk8bl|9uT z(F3@`f+cH<%qWe&_k&%MZ$>}a^*EvdjjU#A!NQrGej%ns7@cL64 zbWRM4J;@%;nYKkfeV9BEy>x#t=QTj_dgjWgEn^<{!HRAQNx|zH@!;}Tgcs0ykAR45 zy{6SMh&D}Xo>`0CVN!X0BeJ6rkI0mkvO17wk z67joIa32Zy3%P5IkILO8$7XMbEWi20pA{6h`}TTHW(#NEzxv%nCx(HBO!4xy^Ms+7 z%+hr86Mnt=x&4PPO-XXp%vyAcE`f%vdZ|qdUZ*!;Zv>h_A)||q5Tvp1_F%yW%^ACZtbJq1 zFKev}&DfoCgbi>%JSo7b;Pk**^)zsMO}yBzH>c0krV3d%LyEK%l|)6<_(n9aNcE3jeTIL;8F-&$MC@lW6GEtr{wOg3 z3GtjiZt42-;r;gqXAAMnpB_K!MxqOistjcdY6N%d#8bArD4E15KYQ{ih5CMM#RG5= zLCHT}#7HrPUHT^RBAMb}jgaSGFWY-04vw+pt+Npcpk_<`skrlExi+vR>vor7v$Oib z?||=Z!|D1)Jcb$8^e&CwMR?_`?;soYbL%9y@XILEy3 zLrq^=+10%E^`Y|)sMMddx?q_RK*d1qZs?<^F{hV!`Gw!*av__xRp#KVEzwxKw0j>UE75rPX@jK2g+Wq#bdG}>IHO0<}L6B~E z_JL_;*b_`VX!^rwIn|`#6~8TEaY=CGPaKTfP14_#Cq|CR38p{3i1yUyhs|0!Na+dC z6rZX2PMMgzOlkb!c}2YREePNPa53?yVTO-yzp78WN+QIxkxl1c$|`UOK)={fc)Wg# zul58F@KL@ihR7}IXU~53mGZA&>RunCTGPHJQVMm^#SPW{vR=XQRBUH8Tsg|kI_87l z)w&8CBoMH{2WJgN!UR&{hOLqpS+q+*ws2J6Qj-n00;?St9bx3Vy(>#yO@3Dh3DqaE zC>y=Op`?kqz{?n${LgaI1$_8p*Un7Gw?zm58ZMZ~?uJhUIl?k!ntMItN_@Ydr3f&* z-Hf7!=!OS~=H4%Yy$gf8uA(tfiG*XI<|I^#01kh<8Bqg>-t<{+Akm4r)-nPU&Z(;E;MTQy9fscV$DUoRw@|N2Tkdr@d# z>Rn3mtd*_7Tkss{QAj7Fb9orilx^;sFP`m|s zgl5@GRdV)=k(Az?FavkLr#C88*p)_Y0Ky%ttR~k~*lOvTX0rJ%>rhVQ_28D-sqi$r97>S1&$}+=Le^0y!@)H1 z1P(6xruoeJq21s%%;{lUaou}ho?{|#l0EQB*i_Vdc9vcFsf5XFj7>>ksGw%*!BPd= zQ7hLmZlvnUPnq`1$Ug<*6feQOOF-#g+Jk4O#JP!coYIWvQiMWsdDAJPXk1~-{_w?M zexg&c+Lvb#_UMO|^#$TgsnDnFs4HP$DZXCJ<(*roWx1%db5r|M2AIJ4THSwI$_{EN z;U??r3+SH|7OIbhe`Tvv;18X(X8Qb*Cu{m69vpa^kdbhZ#$*-!!Z?HE5@?RC<6zp-UCs~~c_g1{F}4FiIw6Mk~rf!;e9-?NhvUCdzS(^BPd#)V)YLUOCTFBhYH?u>6cmNrq&PRY z%}8DK4(vC*foEKeu=wQOC-ib;tX&3Q*!}JT=kK-SXH5;yu^`&TZhndr1J~zJxOE}j zK&0_2I~ie_c2nmf(v@H6^1qmO-35m*wr29!%^PN|3R$Ik^0H=+yJfzmQ{b63Ap*Rh zOT9*vNc7E#O~vXw?hL*9Mr5B(5v)irW8GvnM+`Fv9h6)hHvyX1{L=Pv*VTDAsw zQcfIEub5MGw#IE#Xsy&|w zyX)t%xS7bLutx-Bm@d=*rc>4Bx>LF$*WM!~a7f-191IMPQ2xFu0@nPZDXXpoGo2l? zWCVeoZJb~xM#|{CG3}`F5VbY6bqKA2GI_7S#Yw!~W1X5#`n@bg+5{K%Ef4VDay&0~ z!A!$@rh1J~A=tn;8pDo;o==E&>1@6&fyNXA&PR;(H~2uE>diSs&Q{~wbuo1F#N|1)Og36+i`EH^gB<VOPpho9ACeskw8#wl0K?Dj_4gDpD6{W zvZFN>R-+mA9v7e6*kAA4{TBb3&-!q4U1fUz$1CE>FtCN0swGE1YO$epil|2PS#4bN7O=5UD&^{8>m1h z`x$uqBKE^d1*)t#%x9Rg?j#7R zt1Bzqt_9-!j9ssV~c;H&R_9_NdmTViop=$#0shgwoG$-kHAB;9GU0XYh0aI?Ilh?jJgM8yY)2 zid|c%RR=ouw}tKvk~@ui-I?eMnt~1r-_>d8Vx-`8coz3t2X`cpF79Mq_Eier5Cede z|K$HB_G(thPoJ*Ci5_pcSALZ_h%3_4-??tAjCnp!U;udADg^0_;rh={+yF z0~CMo+fYy!7j+|tY-X!<2$GX-J>-tT0=4~^b<^s>sFRhGk7iKGLU1EbMUIH zQNqzD{F|m$dzSfU@+1ZshqfUPR=Zi-ej|^}9ntlx`&TvQtWttiO!5qwj$Dlh031O; z7jDxSlP<}GGFhn4+==t?h7xVnwDej<|8rH2uOGS;nkB~mz|*60c@Tdh9hbVz>Xj*Y zOE9gXHdFTTgoFG7O|i1s%YX~0VyDGHBO4ZD_qWsaczm!ci4*LRv0n4Ipp4Sr_2)6F zzGbGn==yEmUIEcO;6M#yeND}Vvmt3~ZHbWC5~&)fnPd)Qv^3O>s8$Fv8`C7sHy!_S zB08G-onda40M#bUG2u*}dXF7w#rqGT`@VdMRZvu;pLlGK%1x2X#4yeHicj-rl#Z}O z5Ih^X5@%yAvRVGjO^nw)SduAQvZ>P`SVK4Qij8uXHYD3#!g`0CLXv_i@0i2eA8fe+ zyCUqN>)u3Y9Y$~B&GxWg*AY#jpss&eBb{4`2pd^5cdBN?KtL>te*Ad$?9gsljk6{D z)!u$9FD}jAZljz?VFHBPnxz}VVfKr7N^MPMpUv!tLMB(f$)|i^nKuO$oiYaae+bI{ z1bMGmF&>Hi9tH2(Ru-?`_P~O8%ae14CosyxEaeuyNUloB_IQH5z4@Sh0Nb$sd#t`a1v*F9 ztQuj~q~ZE1OlP?tYJqDaoN)Vb-7ACo=GaQZ1WfeYv zHqE}qIVMQM*du>>Pxt}vbIlBb$0m^no#ezxuf2JSA(sB#7~rzFg9${jfKnjM;T|oQ zL1sH5Gv_PaYCy&9k})}N_4(=Ywx927!?Yc1{?byqI%0OnF8K%{ZwiEj`V88la24GI zB5=n%S~D3Idx;vu0*a^@N7yxjf%E`kc0G+t1+E>po(VQE^M_Mmm+wPJbcd<9?A$x? zRUT_ccam@VMj`h^hq{6_^nEKn=OPS94j)wFfR*H+3XAS zDG%BdO4*%5uX=4Z(zOns4bPQt$lo}u_)DwgUSBqhe$WYuc; zjD$uoSGA71@%BS78j~)lfJO(cY0#M$4)D#xr{X7C!IR5+uQkg)-SxKnlt&DgZ3dp* zH~~V`UQda{4@SVlbA?i1s4^s@J`e)xUVlPJt-1Eo6u1;vGt)(4<5D!u^5@(6R#h5D z=C!&eZZA*dTbeOD>T|JzB-s?`>z7!aK|mSWA4abXiQ8c|EBZ%w+>=NI zo4&*c;mw>Pt%aG_5{#^PnJEwkVS)3ktdmh3(#lhIDOvfV7on5UNoczgw8l|*q_mw? zPhDYPW1q}x_@T|Ycgobh{39XeY92KTqJ&DYLk*;_8n_XlZNWUz{+)y1;x&nX;LBp^YB#ZTBjac*F<%*xk|ho&u1Z@TQ!0%`R%=w7 z-cZvj_3xJN;?+5~K0A27$O>rQIgJ15l;+~q#4b_LU;cqZOu@wD<8_Lhi_PQIDCeC! zc)-_ershnk$*6P_Rr*mg01oz*!u1b*sp&oon-Z0zn7awe!{Lh=6-tSNPXg~@%xx=6 z38O)8- zE0O^bV%wv$nb}@E6{bzKYuTF;bC6MDn-s{}T9Ufl<@3Gdze>r9=r&y2cxfViK&^7OSg7MZ(flCE zAZ1lWX|`+fW?^!-QB~yU7Y2EH0cr21z2|-#`%*!KCXS7IV(+YOip+c3$Knlg?|P#{ zu!trY2KCY3zKh!|UMkt@NxOrff=NP=|0V$!6=3|v30!0uezOg_7)%*$YlE+eRGo0O zz<}CL(D~d%`>Wm?ZdRSVaA{6{=*>%)2H^jXrE8AMvwiz<)w0{dU9FbgmR+lsmhIZT zY}>}NZQHhO+pXog(|hgtz5TyF*UnBH$5%(%KhER>(LMo&Q0X-f_!`0wRMH^UG1vOs z4j+Iecv*L+vKK~6r^Z<7RO);y>FB(0ubZi9*v!Wk#&p<;)Yp___v183i}v?8NunGD z=+947eMJnmH|mXxNM1{oXcDn@nM^0?GhTmBl~*(pZpzP`PpWRt1u5o)piUv%$mJdMO)#~)eOljNfxQlk0rGhL7PBNOOx?UukkGf>5D~Jwx(I@QZ40vwOVorGEea)ay&Fwi#>IVFyNN)U;nS-6nG9M{ZIl{<2g1gT_td8#p2! zGHfuP>*J=d3Lcc!7Uut;Tc|ytCc+ra1~j5?3O*`_l-eHU`K0*_>G^DW_U4KZ(o^L| zSgpqnIgubZqG5dNgFf#xWpA(3}DasXG%b*!U)Rs*wjDPW!DoYrZu`<&awQiePDQ`gEt)o6zbV zyJHzXzqFnRGmm)p2(87-A5gT{^=D(@G(?+6cqcNH@L$1&C@V3Xm4wXW~4H=&nk9w8z7haYZ$Z$LpJ=$EZ324~q`m?klqsR}YdxR2jalD1ul3 z%EQ9Ds`$}#?4b_cXJu@pP{$HSQ}EJbi23iD^P_6dzHzlN!`ys0@#M-p-S$N(Clt%f z?0=VkKHSjvd)Elv+;BT#d&*FLS&~sV>DACu@LYl+ror{=(?cyk^^@n4!bdb>jBaj` zKrcI@w2pdmlfOTlE%(_EN$e4!RDh1%uwdf7>+>~y%T-C`J)B9*e)QMrtu>aH=hp+q zcXiDaM+d9=pA%;@yOe7Zkq$mi3}!xa&2PR%r|H$e6|D+-w)%cj;2|m)N@zg&DT;Md zyjvLiQ2Uj;g*f4&^kK$M$sK|TbH9z$DBh`-6(J^x-|_jz2(6XW%n26KU8#C}a}A&G z0FNq%be;qSH3qzi199`i&PKvNT$q=|Sfs#R4@(6EcP#RNZA|xObnK)-Sw|KBM-!-B zuBxphQA?6i*!<^MD>nc~*JPwOnuHld8qcbn*?b%wzB_%ATTb9$sra@chDC{jz}-FZ zZqu2qXXPigV9H!WFkZ#Gla1VJ!&eY#+=F|^I+9d+c%IlsI$Qyx`=dg>~G*3yR0Z<60wGvF{2 z(NI7w*qa#9*c2uF{mVSpldX>mn36%e)9WRb@8Wbd86TOQRXqnYGLiQ^6&w_xTlHqi zCsUwO&#C@&;u!wREP1iBK}mwgT8?u>b{D0OSh_C`EF8*j@9HI#egvooy;ME<%KUCe z+QqEz8R6}moCZCX${wl2WrOuKi@}F7C+ZzoP34!}C^lSyhfd!P zSL1~qtR{6Q-lLE9y!CuMh{+?gg4jQYz{Ceig2U+bI)$OgyFVZh^s*hf^IS!X4^Ft) z-@ga<+Om5j`#n_gsda`#S^jO=*Svsd_tg5~Ar&a~KJdz!tt$?)Pq*@j_-9PvfWna& z%gFRZoQmS{_m_|~jNRNulBQloB>vMLLv)2pnm$ia`5&M16(@0vd9nVK_U-s3*0Vm_ zFuD^><=B49cs7yzBpy(~rZm8X=S{tZ3u6{I;KHTS#4+@hg$}Mv>;0+k!#C$K{n*Em zhE9V)pj>KI866_qQMb_vtgrtJ7~M*cseDL;Xa|9SnWdMl%wm-=PoHfk>8}GztR*@` zffgU;*}5GvTgdIq!(8ub560iMBgm26^+KJ0q*FE2>zkED9Q0dbGJj4le&ghOIlp(; z&>wW6TVq3pn)(1}wZ@NFgbFigiejyIk%F-KLi&OXZOup8rg74DjA2vn=fc-vc*tT7 zD)w0CQBeL~tYxogS}DGa;ND)-!;i}#XisS+;3{UMNrc3fF(i7q1NE_QgVAbv`x+`{ zX1l&7J^yzmKB*lbGRXjQ+kbV+Kg(p_pU4An%eQ?C9Ig4Ib9MZjIe$D6)Um-sQM99l zEXcup3thG?ttLK+bnY)o=J-ziNEYndLZLznA61OvopYEF1gzBpex<)iV}|&FvRZm5!3nIZ+__ASEE^7 zc|CJGt36nX0@6S9F3^!j{6xh;mqpANQ(Ml4QTgRh`)!(#z?M$8wxinA)Jx10^*G1|8IXWo&YjF!}oM0T$)Q=Of_3lVBO(;Y)9kTQWV=y7Jt z-}7;9n$<3jk|rL}#4B6cI-Gtbl$OK9?1_`*DWJ(#*xvA(m(Ow*AZsJ3jJQKTxd zvJ8dhicGBknXCbA_jvxg4~a4*zsgT85{e(>0Y?wmBoUci&3N7_|Guw{OkW+&$iKnG?@} ze$&fdGeX#;>29W`ykXOPWU{4RtKkbd_k!W2y!$a%Iy^FkU!tyzfF|iNUE;1scB1}e zOzA>YnwM5<>F2|_$Gd~5fRsFBd)iItAgM&?BQQ{{8QidzDGDZCybeej-s6z~;5w!R+68hCwso4NlOBaqbbS5#Z9{^He3Ns8HkOb$Ndka~umiiH)2{gMMNKPl(?qCM`yti`>&FXZ%ienJgsjay)YB z=1EuGbPw}Q4hI4gR+PS>Ehisq+?eonRfGe>ypiOpLC(nNC8xY#r?|oaCa}3 z#G=GzejZe;p3xtSZR2}g?gl2^(brjooxU7ol&izlysK8cb&WT4e7MJ->bHV$EAQRR zN{pu>)qhmpbQa+uZM*rtVCGo6bdJ~WsJ?3THKKG%LGBwDS@FvEEq@BjQJZ#04i?6* z6hfM-nUS%RE$uphmHdI)V;k3~(-zTc!0muCsx2mu2j3r$JYE(_{bOfkl-E|N+2>6- znKV3h{`2Eu9;kn*UZgBKyQ^&O5>rQUT_JUHdWj_s#f)*M!iBpGQ5Zvr>XoNm*Lo%< z;4LmFuHK2DpnX-@ZKNgur0HhKPE@M@gESLzE4C37g6$(wV)-w~3#bW0V8hUCl&jToFaz630{Y8^jRy}DkHWUKZzyDb= zn?Co(puolS1Ivu;&`+~63m9G=($-gNER#I z6;D<|Nr7@sxI=lRnNO+Em?WMG7S1D)a)C3Sd%pk)Q>61zjaCaD%xL^dO_ZrNZIPCdRZ7knZ$;8tZLc`^5i6kt5Q z3sO)YWBu*++iamUQqD*)bW9V-rC{HVYrA*}kwb6anNF7LhL#*1sC;uhgrWM7yh3s@ z60<*rgqfa-@pX{7ViAup20SL-=sI9ATxI*tqF%LrnaThDg_{4K!es;+db!6J(0ZqE zemWWpgHLJ2nOc`>olVu=fgbCT@5n0~L_ zmA@llFRSsEk=mZ$d-D}4%}oSRZ?k4{p{}3NP~~w_H?WBg$Zx(6kifNZ-h%RF!?qqs zmJ8^(-oJQsb-Ibtxh-xmt_oah2rqi8)k%&^)%$&Gx~LcdS}NB_y}t!>KYEJ~l3?1O z$>$kW8=rSR*8fGow?Q0tU8p^=1!qeQW>Jc$oJu(oe^S;Q_iJ2b(zbq77mpuY_}yAQ zjtGO@SZ5<&_Jm&jckM%DIz}o#mxM$Ph0;?`I0fYdUOpskkx|KnU2P^=YErh8Rm3dPP09f ztj(K7)-toUoYpVC$9VYJZ>7n#H~~xYKKyC(tZ?+C8x=?o?fV-8Avs}+1Z$AWXG{+IA(1u?+Rkrq-cK?a(8v(uB@m-w| z=tZN6hx=;QDe(Sb>i&EpC~RM5+?5eL=A?%6ffLY|=dnLLEydnNK)5w2TgHQdP^KVwLmz=ceXPgP zCh$;~YUdLZ3V0@ln5QkZogiAbymYcnVh{uf<7An6N(01i-8N-1Y&k5q%vhOa=TtNk z%Cz?_k4PQJAx3U{2hb?XwUaHx%L;$)=zD^006oZ~T_x#VE&?a}`SgQiPIDN4`8YzI z4U;SQ=FY13dEfhcukXf7J2z_hd)HU1Enbv=vP4q0z|U~;@vuNLTs6Gte1kIW0IU~2 zr}aDO5(noH7mr^4r8r6(U(R>_Oi8C+fD}!yo?YRTjeTtWO(pA`HCx|8@#KKyj|TiA9ar3N&W5fq_?Q@t(pvI zMJRy39wUqCsJwYF76^Xd^>6v}qblMV-?Q2L6K(pf1cdidr7$8N?Up>U+joB#G-qP1 zHFZ-9|B)Bx=OT<=Cs0GOim<;DW*{zheTA1unD7p!9z_Ave2-T(qE^E@Z00rf-|F0X zvc_@cz(YIQer>R2A~&|Bt}+oQrjHXzsYInO_uzQS5e0v>hW$Py78AOXwLFGd!+c1Y z$$eOW#G;CjKTifa8paU=Cz_E25@Ls3?}@pf=X$PdJTVj6cPc0diF<_5Pj*ENe#kPp zMY%-F^LiJApg4i@7P>h&${!)Um0)TmEiYu2M1~)OvG?j7lbpM9e@hCnT|nY+Hk4m_ zbmvSInx_Uc|K7Zk_QqQ03H_5Ivx0ZxvQBU;1~JVMcJm$H<)op{CKhOW#*;GkrWElT z5v=wU$2KldFsw(jJ4wy}zPU;h{$Eu`Tj8p&V>0jY@PzU6a8o~Am$9)XlZ=Y`FV*2H z%hlnGrZa_HU{jjX)kKyyNJs6rlz-!sr~S?0A&rcF5spRrNd(R{Y(~#{lHBMf2>@R{ zv<%}de*N&bLt2zj3YtL+>6Zkt8G#1&87#1I4tW0#`|^jNf!-rnjuKwZM)7UC1BT5B zNVNHK%5t^S+Idcir<^h+9(V)U9?R#-+UHxrmGf+SfY)^a9;n`FW5IwTBlrOh-%nXl zD;Q(3kBhu`(@Rs&L(XfVT+|)hB2TMu9B^0%wVggGbRkcI3vpEAB1;{GZ)x8@orn7B zrJuIoYL={ujtF^}=O*SKmqj!X`cKHsTBwF5morF$B-c9XWUKZToy~@Ejdcw%5mi}x zVmtQ}3UF)e$+sJwbhH1ZaZ<*(0p0RieWV8s5Kyts@c3*8z)_Q49T`G0IaSOTdpP za=*LYIGh1!8>OUISLEdya5+EA`sWP@``6vl;1;j4Mwwuqd3QY?ArC|bun=;1oy`v9 zJf3QJJ+)CdzQ)MSD@Xm@xH8#cN5(+e7JT!yVj4n~Jsb%yPR{D0($GC}&u?&eCn~D` zGdNPf(IuyRsU|99BW52=pgMt$Un$x0S^$)9p|E6vzRNt^*Yg$Q?<{sOc21hZT_jdj zwA#a0LUG87H#!Y5XFg?q8T7=NNT>99;fDlnOGaj1*SWP-Qx_Gip(Il1v_TI&Wpm&@NY)=lh6y;3s>`&8~`7V~KVUXn1T2GLw+W;)2*|N zDlLkAr&k?x)*s}69T#5Ts<1A+Jw#{_L5vE%z9W>ml<#zBLE(4m{ui#5)1G$K-!cv- z8*HMyk0o~O@lzFeC+hyP8upes|ZhL)Ur=RR97G6)|}bK zdRSkKj=bXn?-$2X1vsSUoi+>fOni(^oqEL%a?V|e-(7!Zc^9(sU9Hm0YX$OWiRIUe zv;R5J-VT3jUb>81XCYc-dt9gc&cqwG6vh8@FIe2CQfm+WV<>xnpN56@%*eT*9E`R{ zy(~5uncW2t+1#K5i{w;5g0fgfl7Z*ym(PQbbb~qS1hdja_3I@g?=sdmRHUvF4(L{S z>;StY;)QeMQq*fcFIEdFJVNOqXQZuEy5HgfjK;%IZZ`|%uL(e0{7+1R?Q9h--uWY| z+*mTP6HF38WiY}`n1;VBfz+*7Wg=xCwjNbAhZM zpFi7gO32IBIfYHXcym|KV(_{_%2E@RSN4GQS|vTyPt@=G8cHc)7!i~!aoTWK#t&*_ z?$;!Mcz}bTIF_558r4~^s|vW??8z0gO!Ftj#nX-IFVH!l3o>#KQA z2svRJYUK*{zj;`~;ZvkQhf!NqCY{*M?e{OvUzGkpJG!|hiI9Hpy5&;IdE0O}N<7GGt_mcR*pui^uW04$iY`=#quOlyCy~%yy)<%>@ja*~3Ov2{ zc@5g#f4Jv!UCS+3mMp!7hObA%Wgt5J9JUA(;~6p0rZ7hcvbApci|CyJsL{`mfI`|X zM=GCdAj+_z0Hi-XR23Y@*U{{OV{2!}bIZ61b@Zj}Q!E@L^CvB22dZ_zveIqdp_mQ# z%A-sVAqxRG`fl$vrzc;ue*a0kKOVXFXcysTx(qQ!PlG;NFVt4aG*!qj zR#<2-+yIrQ%*1Ubd%0mV%!-oQ7-2lKuRiQQg(44xEaD<-j zBS&am(YxA<=hG+kYIE>@`qt0~iK)0&e2p@)22U@GC>Ig>gL*IdMd*aihfLHT0vKcK z*(;Ce#h$w#=PR|9RWFjv%`LFwYwE)__!G%szGE zjOUr=ZOX=;xnpf@pg-~M0N*4{b?@jrA<~aLe&s!!)I~blxD^Bm6yckk`Db<8jGg&x z0G5a|E_bYhwxcvM|9b@Jmw=nz>6wM^SwFWRhr(25x6E5@3i1MrfX?;G+po7h{G0BX zueBZz&&;zOK_;-NU4h9DQ`mo5%Mt-C>1n%nMaLgSW0{$gw8zu*KF>TG+@T$u&ES`+ zW53w@!%#+q&O7c-CdtY-m8Q7ke%<5oS>&)tJ;wx7{K?sc#?I_Sz2!bK`_%NWk*L+l zEQzg=9j&r+EYhdVY3Ub7TXi_x4`r*T`!%`}Y1jt#^sKH{`IGHAi8}F}SCr6=tV-}cT4o#v-ihc&+4wM#^p?|q_ zDiQfqq|#Eb2Cv7YKRxXV`-Ci~y=lavsoGN+m|Xp=agiC--3>7Q0A1UsMMB1XRh2dQ z(n{HKr;%m&zCmb#j!OHej$-qszo42y_syzjU%D`p?#N-NFyoZr%P|wf%t7WG%W!=B zpZJ`g#fCrUBhsTqE%x{PljbC2O-9(i_RVx#x9Ca?p6TpuARYezZ_(*eTnyp(ybNz; z|4kI&ESM0-6`GHjC?1QDZPun`n1nm|=Bi+|z3WZ#s`dL@l-^JD8?=Y7F#l`d|Ml-j zG=NXq_1FsU_^)JU!bQ=CGsQd`?su8{$N7=)}zXD(KZsfR1tQcRDgXv7ZtPsCcEu zUYP!=&f`XU!|UBs0Q*Ut=T-c-lXZQVr_2z9_apGzW@&Psx3Bo3XOL`I`NNw8Dyug2 z!sJGOogLoCtUQrI-osQ9?|A5#VkpgvU_;+uFOpR9OkB_VZjAOBLvLPKXyj_;8Y z$*nEIV)01JTynMZ-S}qi`PaHr%!f;{5=Y(?oWm(;JS%?*BiS#|XU5bJyL4dweZr?$;Z4pkh4#4~qf| zETrF#mX9m0_m7`uXTseUv&GY*l`M>JmIr}{Tk?-Tj=*i2Ex#<;^*prUUQr;gMxde4XIVFPM=@F|e?>f6A%*Ol z;GrN{gu@WIOOv0;vmFAgf_-xqbZTuAJo#?iB>szhi9xlU57kG}si_ctHSKC5f&moe zR525U70X7k%NG_4x{3@sAhENMcQq%8o54ct^{1l8f+M+TE6JH>IBJl9n_8T8G<5Kc1B3xi!vN43vAM zNb#Qer}_WcL1L` zu>?H)Vn}Jlpym*rX`aNteGkh*devM2J@u|YQ&4V}C=C?|EBI@Yy(=WWVB(ZM7m|8z zvyVb?d}&Dt}awLdjYp-~b`#5B-FFWT{TdaE`v zbKrBWZlI**=S{iO2BSNRWTL^NT3LWf;q%)S{W)wqlW_L%>d-(($>}^C&ynvhj%4S9 z@LO1&nnM@~N{%3N%3|_~r-JTez_(E6cU^JX5uir!iE_rM;@B!fa3#c)CC_kASsZ!q zdhK*-f5>a>hkThWiTSi`5S@^KvZD9N$B)znys_w+D?1pEZF@x%_5C|G&Ooh;B0cvA zljk?4_L7eKNJr6UO%5I<3cnB6XX()kwACIpmUKC9j4Q1DQ|V1|fwB^s?Mj{7Lk*5h zqo&-Td^vu=p^1?6yH_|Dxq ze@FQS!;f~_qvu7BmRhVI6@6$w6*xg3<{P;kzad!hmFZg1ca|W!-bbC>Y)O&G2*jTA z1``lubu+!;meAezW~Kt7Cx_Fe=^pBWjknUX8f!_ zkK=2`29Pjq^GCLih|Yo!A7C^zKfnB`pKOt@G?PZGum7ODQGPT<;Bw|}5BcK*D`hj3 zJ7bOSrnUA~ZRmo(({Uvvb7EX=<^-dWG3>&M4h1-3A>oGT$lSIKf_90pD{AC?-uvDy zN}M<#^;;j-Ge01W86F{b^R|eS8(eMV}2V|0^}Cjt?yJoY??zykMPs`{Y44wr7 zzb{^;p@6Hb!?!lT`DDcN3{!40N0kFcLpN?+u89WW?prS*S?+0QQX%S7QZd^1*&rNZ zO-+iHOT$cG?&-c{?}13fb6aC1|&c#?|)LZZ0*l&P0~VU*9&BnnKR;=8!iQ%4I|D9q+Y{OX{H_UnMb$_}Nf znk#%uH63z7Hy4|-3vvzS4ZGL!VieT`IXl4~(ApkJeRC;}Y^jFYUhhk=nvaJfYXl%V zE(fytYBuhsd3`Z8FJqM%h;QXAog5$aQI#vB8T5_}aXz$z88=z?j<6H7foK#LfbRyI zvDy*eEg^+)y(?S~xQ1@o#xXl5;K@>uQ96n1@@z4U%k{9r zz%d!}Ec(<|VyEythnMI)2Of39WHsAMO?ifdGf}AAycuSYo|4=jPFKgXuFf7?I&_6HCQ&+16Ij&RVOE-9!QUFHPCWOP+SbLOyw8NfUKe|TrSyMu%JmH z^p&8KAMkzEIbBrqj-D`gJcOmBM|ZP#=FSYu;cyK8j?UM^<5$e|yT;JhyGZ4HWwB|} z9o)=ZmRMH4bWtU#S?%DjoH-3JrgOYKqDic{3&2Gbs zg{jTR+Kv=aw&|#^9Sx`bDhh|%4in`Sk}0l5S&G+aGJAF#@CfLfqA@JcHnabz|%7KmEWL{qnjwG&#SMgt60vK`Gid#XLRo9$L?jgV#T1(Z8#ERn)RQeb(>2Z~)KXPIP31ypuk_Cab-B!cXavTd8er4`fO6t_Ucih+br7s z%sV`SC+3(G67KRVL~4jsN-5_>I0u+EBMBHI#qOAP4ePxYTa53&U#d4Io=Rk1+R*a7 z;NX&ojA-i#acFXmJTmh)cyJAFue(x(!*g9KMU|n#J^?QHs!9R@+MCc18zR1jNP=F9 zt9x(xcrltI!9@oI3ERY6A2)#T0Cb^b(vLcp`!X^oXR1&)bVAxQ2Yk)#KpdbtFd~N_ z-^WAbdi8m_lz+Z7=JnxCM^QF7ue^TZc6?#rPH>^L4zP9cCr$Kl9f>p+-pF>t3l(PG z-@J>{eORST;Y^PrjgEgmY$|qvjkTtkccN((t@TAM!PxcZ{H?!G`rdxEZ1DJ{(PC~Y z=#!16`W*H7B+Ji1E+|uhnQV7k9(5elvt8B+~If!)!)a@w_+(? zgD%`;2707z5#iBA=D(>E6s*!v??7j+)^~JGpf0=&&eQs0JsqL0(L^r1${)2_0?6Lb zfV)Z8_C~9SoByekUcos1NcT;^G@b6-#7L?v9U)(OnPtL+IADL{U)f$FaTM;jGgJP^ zb26RymWvf3y>DkjFYrb#UtW-C2p@K>T?K+>cThaJP(K|bwCHgJji;Sc@{ zOG57zLG|ql#0sL`*p`;4jryDq+|rU;;St;_)gk*_Y(nu5EhSk65c^y4jtzplzaPf- zpd0^fP04Da43!`>7bR-_4%g1E?%!?v$!vx$Ivt&=j{FkqH}o8X{!mjeY3R{$I`)7( zwlNQj+*;7v4RBR;93*F=vN@0Gyr2BD`FJ5rC91{=N(QKUI=NH#XlU{R%aqsDxKyPJ z+i=Yb2;_D0O=@2n1sM`!D;$V^X!o1Q`H~HiAqH4b;x9aV$B<005-CRmuxIiafJm$r zEFy6C1^8G~no0W( zAgV-_Maol_~+kjJAR9l7n!RQ`#YsP2#dW{Yy(JwqKADZ`Xl{&|f#NWP^F z9XH!1$apmPW`iTWD7D~89fM^0RMFuYY!gL`U&``y2 zHmE#$6*I++O=U~(ucM`mTJ`A>;FC9gq-mLfb8TD*;Z_-hw(n`!IiWO_QphuKx_#Al zU$D7NaWy*fY@nlWhhf}mFU=dC>9tqJ0(Ge_4yUwOLgtx2=`)?Lp+SIrLF=rColV9Q z522qMs5c?@4h;?(v2S5b@Ap^=WXSkyD5MU1`>Qu!{g@l5emz4vt}^3jm{M5CxY0zb z&;64yc$gE2b)ut#8yv?TQ_&V8R=?8-Yw54698P*}jq0Kh|1;M|7L{Wvqcq^11lC^L zB|_@ING^IE&F+baf^}#WOnz{6amikJzW`}!oIh~pI68j32LMqLQea}-H%xv<{pY&D+-#0lp=;CfQ-1QtZsJiMhhOa)W zdaO*V$jVA6rN>!(snrVE*5?Fu(eT=k`{8`&9ldTqe8X{`G~EabNyn&!yM&(cOP}iY z$!(Fezf5(|(2c%9cr&p`yw>O8Y}8IMVm)eb;5Uavi@FCVT`er8a+5af^cJ8`zM9L2 zMP))Z4-ft8u&6JqUngIy@j$QjQN2{V6Lp`#J9k?V?1Q02ROBPlQJC6teQ~KopTyeX zEKpz9Ccpl9qlMmz5qJ%gdwwM89$cNyJXHcf zvxf#CAlr2leJ0r)J{s+-`+b`fClu!fdKn3I9>IhUtzA`7A)}gVx^b>5Hh*lhKp;1Y zH~7(j<5|(~-b3-Y45D?P3|b~ygA7P&jujEaDJJhTvw)?~ zjsAv0U*lV~)Xn9~@h|{s?z$bCm;XW1>Uos(5k$HyjHev`PnwDznpZ@HdzMNx+qr=Q;tp5Fi4YU?sbVdMx%*yowRq*?&o05={XXdUcXa@K6_MJzL3od=O8i^a^j45$ zyLk|JqH(zQQ35zQecwoz9H*Nfj$kf!c&;N;%IU6S%Df{5wx{LZGOvgwZ1uRo!v+Zc zmsaa}D^FIb=;4MeLoxdb{!KYiY0=%f&o1q|Tu>Fo6#>fxG*xceC3cpfEo-Wd@v30h zl-8};2PC21*eXFmR^UGN0ZSnLR-AOLvsYXm1r;F+YXdn{t)vaHn0SX4r;-eQvFw8~wo~pd6ulv^ zRsVAO5J+T)pVk2Ehc&%a;L5~IZ2ZccKns&7!>ybNE1BI?{ZCVK3mYAE&;K$xb^?9Qjj6scM(xT)1S|C zb|Xm5m{TKKhnS&Dw<7o7lhN`~n1m(M(7`Ww9}+cp=)h)X5x=%I0KKESHWP^RCB&U6 zLx##99_fu}Y%;dfS~WtHmV!>x(tspqMDw&C$&b%u)CW7}>p||eI}Qma@t=G7jZ~!x zy*EX5suJbA4J2(tYgrtY{p8#gz3IAQt^M(R!AwP|6S+KZ0c^d0m*3wis1FC)+VkFf z&6#vF`xn5AUB4b?8__2&u=VeZ5fFu-r}A45e9-w?RZo9h`lcM80r}421l{9vv>)Z(Hgx+T2<*(+il9r{ z6GJ?o6J=nVEm8e#k@{LW4wjr;7fT)7D&5yahZ(!{8GmHyJ}C?R{JfS_y|ol>xFjzG zfwvaW4nAxyIUfb%gi)M2oCgxLF~2(1T?Zw((yU@dOimQZy*qbs`~9o+195K|2@fbB zc2>#t6XSwpb`M&j` zIPrWkA(0?keiz>k0D^)-{q85K2#}z4TG~ugw;%-j!(Q(i-ty2Ts+CZ-lkoVioU=#^ zwf+?1Qg>hG!`Jis-pz|y0mZx^ubiUvsy3$VHg&R1Q&q+o0-~AjWB$Tpz({M}inPp0 z7E)zN_j}V|Q6*Prd&#E#tmNS+cgY@O560?b@JW6+iCzsr4;W7zial;)ZJqWY3*{HQ?Hykx6QQ@Fu zpo6x32roS3zH*Ef@&Xh}f`q)OY`pTMl>}1|VIO{#$@SJ_FwcYI88<~w^@9Yv8Tf&m zDl-TL)t8G>@Zqjl_cLH+%hgwozCY0Tw`?~P$07`}!E8ohL&0i(ha*VU_MjcfZU4Ty zei}!Y%&47YHYN8n78p_{+S0s4J@|wt?=#Zu2MK#mOTTm@N5K@r2Vs9GLLjc_sHd+k zi^{V^u>YH5lp9-2f~8}VLSHi)4YT%0=G>#T4JIoc1y!xJI^$c*ovZfXbDBqYE?Xw; z)Y9ei;uriv&=1QYOb-w>i$bTZWLOMD9ADDXg4-6w$cb#HD}8C(Uv!R*TgMIG7@sUA zr+h|H?l{#6!VVpcK-5S4(DqC_)mZT-&Cl+qMh;|4RRkf3FqPJrpvXfxcPm=4VDVrXx?O=2W6V||aw8m3m z85;{HnDJK+_$Uyl121r<-ndg6B8o5 zpNo}(yQ68F`?C=)2c)~c7Zb`AkB7=`*I&hKCbkS!eLIRt`C5)C5k&x6{Da^-A+({i z{6#I>z8b^_hmgx5h6_~y*bnuFdNUcPF2(s=&ohx*u7D(QWb+EqG;77N5yDCDkR%9R zt=hu=aY-tCIu%NlefKQ~Ov|l~;I%3rH{^#8$xDp5q4@m}lFKtqjz;?T!yeM(6E~~V zC5(+eW~@*6xs(;cpEE8PX}fmf!D>q>N0YpEou<7KYX}Z5$VIwv4L1l46UwC~ z=X0&fA2=MPRwZU$EAWo{lJczi-@wpQ{)qIOqhQ+zDG^g^uITOI4`;~0Hf8n)_W}Gj z%`ae99R%yA+Z`J)I__p~ThTW(-}@xKeKl@^5E)@uMm*2+v;d1#1nx|qN)k+*n$qV9 z(v#C0%&KYdssItq_7Ku6%ssv5xTJ(Kz*P7io�oZP%!}7y``yDY3s??a+BUanitz z?SdZuayqXw;EBkZs=qqQ&;CA)QYe6R8a9)Dst7AH^C3THBD!BVj)Q3pdFe&(>)UhG zjmJLECbN6Cdi~H6Mya7a)dEsA=fcF*#LIzm)J&{g;KZO|REfNT`UO8auqhjQWDq)^ z-YTYc#Gx(LkR8p1f2H2iA|p`9gFYV4C}z^m6wBcqFJW(F^pE?zLW6ao==JXn_0-m* zuAr3a7!0AGB3xxos*pFTlb9yjJX32s&nOLkIS=v$i{nT^uCaPW9@c;$tJ!vX&LsxD zMZL|bAu#5g3#)8M88_76>n?{ z%0|pK7w@a9$y6%*{y|A=#n?5uI?j6q*Ycb%Q1KaVXcGuaY=7Lh6dfXf5P{;UcEbC0 z^xW%?eJ&eS`H!l6f0j@=Js!6a-ro!*#v@-nkSzCRS#_xscvA#VAzj9eYLTeCA#GAS z;@vRuwhpo-;!fz~68%m|>84j{GP3FmvTW{0`t3&;G3d6KsbjSINC1pHyTfyoO})gj z%{1yV4WQbtgRMc|B(z2c)JS2vzYvScqE9U_aFy7 zGCit7jwkr#*OSd+Y=6xZTVmz&s#v?pK60l*j*c$}b6MLftViYurp>;8-<;jKf7Q|! zDzQs!cXuY_J#^J#+jV709WVCOxsyk`pELIFgSO0tN9JP@t!K$ypF&xmw3g?9&M$&s zT^ia;IoB98Z`rL<=>N6YL#V)H3C<TNA4fokEa(v>Y@Ip ztB9|o_Bab`|1#MtD{&To7QOosif>SSvvj&A-n-(Z zilk+iapH+49evg8`(R<~99O|w)Fn0v-}5p#u?yUCl^}QvSec|6*}6}_@n?R;i3fQ`f2$=@yRQ=JuGNAbj{U1$V9Ta8%y}fjIODrkf zARW?;k|HGy0@A%KDIn5~5+dCV4~=w}ba%tju;lOd`_9WB%rLXV-f=$XT&J$<=;QIu zIrKNnD`9%BjJR1}7Y%d|$&_Lu`lLJB9Voy@(S~djvdLFdGM5{T7lK9H|P%?$!r$FBP>z@ zC4=j~@}dHq`v4q1q@IB{nb5n-aIxucJD$WiSWpkgPhos3B0C5LR^Oc<8DTFrFSx*9 z#TmS#UH)p;N&lda(gOhAG#R&$*;3qoUJ@`iELP_g@zrBx(m@az?>LTVh!ZM^g4dVP zD6%_5(;BWP?92BmtublIB-8=;xF~3`&Aq`pdHSUG16kyOd{k|W#Jhx5mM-jt7%<`A_7rlbO+yl-r}r7FLcnlY|=J-;7) z>~$Gayc%Af9=C%y{8j5{lcMo9g2!m)=LA}>o*&Fx$n=LgXAD&L`&u}K4>6S{3+fA&IJv)YGE-B4j<2+L+@A15+9I>dkFM51kCtKoH z^yhn%iih!2;}~qIu)ZZ&A=X!i>NpyMY(fkO%5l6zh26#3=ZPB_-q#u<=p(&8qDa7+ zSQ9UkI+)-2%~T&$F8T7<#v>6}KA3n+mc}589m++(iSZp`V(`kg!~CJ46yG~Afsazk zF8W5UqCB*k&P78}j-$WAmvu|^hu9DMJqzqHw0%zRLbsTJq6$_tOJ1M{nmuO$5NJ7% zMI9lD(PltDXb^x}q_NRKBIh;%l_~%kxzqEo@2v{`MUGEf8df~O#K@18qI(>!3s&PQ zz__p4Iv`fvkIA^)Uj`!82v$}zNmu$Ljp6Tij>o@Ed`@&RQ%FzAKJ>aqq`4+q92Z;- zSlmOeq#k220!>0Jr0#A`VY+s^oY{xE#YmGD--K;|ab}nP7#3i(*YmJDJ@vZ){8P{v zv{|}4@k0oTCorfIn2`pZ)h1&SynC#!vgvzi z7jM0@3Rf9nmP{K}MW9iR6NAN;Qq0UkC03Cc=lUeS6ULD!B{n?x#G3XKhu?+*5b5IZ z6>?Eqw2)L#(Cx?6fo!rO{?gwV@lq8+EGfEdGu;#Tz{{=VvigP)%%wEZKY<|ghU4E_ zI>>c%A)ECO>+j>v?^R~Y0|hT;#`wDd(B3(PKZm4NdFB@jE4>oK%v~u$_br@Q4PX=S zAB32e4OZ~8+*|KT5C4{{qoOHQ3D$4F^q|gYzs_{Br8AeyKwo$tyZPXLMgGk^9yQ>Q z9a*}v#ub<_5IQ<^5(n7jt1H|>{=V@Ad;U+_Li1}nk4_UXs6F;k0kUaSF&kgL2m0_(Wb4gu2K+lQb zi}ip(4QBqT!H*qki5xPIRH<`BzF~x&;#c!zL=z%7AzsL3j z1)J&$#ttARu>itTXR$i|;Z>?tsZ`<+`j7*f0K~dMB3iE)eNXqkiwbTn=hOd_3Ew#K z5ISCph@vVN@8QE036N5zazV{0QehC<7WM^psc#oW&hGz<$w6+dicjIGvs< zOP2HKQtPkpS9fe|&ibfsH$042S}Vu!k?wTZ{S2Cav*zA7=vy|0gL9QKxqHy%^M zB+wG}OA%jfYYSB+-IwqV%YL!ew@JPehOyJpuSCSwRfA*3(;NKH{GSt37cq(*RoJl6 zFYv$PzFK#{?WF0{7_a~|79RT%R-UV*OE0HX0xeGKQB+mA$P${-k z&?;|cB1shZ*4AC+Y8m%X9ii)%ic$-HGuJpQY_Hd!Z%{y=*H+%Av8QtwkM0lph>lqN zq6b_`tv}tb@?=`Sxj%nG6P!kI9 z6ntTj`A%?1hae-*Bu!&SK&t{l=0pr#iE~7#iFPpfG6Vn1E&)$k&0i6NYT0xk{vuld zRF^VulR-@zM?nn-7o}<$u{ZY=)Ac0GR9Ib!_PL5DE1FdSe<-K}Y1Su{V@)1+=z(?1 zckfaG%F%<>^0B|&O(<4{oQ2jP`MF71(5K68#Wq;me*zUnys;GdVL5v*Sw+i7KxaI|zt}7U`<+t`ReV?RU$Jb_ zTNXIKP;o($LRHh-x?uvl;!>00F-5ln5>OjDcl0(>ZD}LY$1E3O?{>Umec`6CLgu+M zjG)*5z78KCtk};t<(wQ5``o}%^mnmUCuHnhgs>ax1?JsY4PMGzm0F64)p5JKZbqwu7z+G?*o2f7Uobdy-IeoE*FOU{%oDD zt)Pc7#bPSKVzE0D(A zQfp#h7)4^D^}Jz%-@)DsBQP!uWYD}yoWDWx6hT;#Pj_t;i-qsWk6Ab!E#s93h11U@ z8~#R`O;u{4Xt;rvFMA+XmJt2VR96iu6G>Py?9JJ3nOU6{r zt<-#mW4dzat8c7vZYzpp)r=~vYC}6&jdOUg$?@>%)l!1larlAg>BYOd?s*r2zR*^! z30`_U{{kf@L1?z{%2IIX-`k5Ux<5aD(3j5){1jC&|0KPFib07QX!2FNpyOPV{bOL+ za!Ox7<+qI30kYFF2j5lL8P9BWgNWyYwgt0xh1T@tQt< zWg|7k3j)>s`RT{|T|XaJ3oHgiY`(Txuzk75XJe1c*M;bT(42}4yVdjMe%v*MyL#s? z)W0jgNPx)ao933y3P`n!H=hY+j|H@8u-BS}m3`^K6_M}aOLT4doOvYXDp@wnF|WPH zfM~MXSuG357_<>iR1sF$n6lU+_#KLv{ARj(rUM09ma#J5Z_bt0@p2dcmXpLQioa%S zm+T+*F8QSw5A?%+{uCym_WoC{GV|^nXcPmZ3J7z#zY_&zenP16;eLdk-G+7uStz1J zmn=;xN?hQN<4Ci8#@P$$4Ue(GCs+K5&w|GIi^ns}83PFiAHd1rcPs0(lD`|FjS50D z9Jklt@RkChL3h0@Dq!CIB;??$*U7U;_G!pO2LFc$b7KkII?3n@>Q@C^g< z5-^|rE4JyvID0k;>4NOVqu$N4jKgN<=ZtSB#s0+vNB6MF&7U*=v8fF+Axm4MiF7hd z0iOn+Sfi))PnYGUOIn*5`@6A*v)Be%Sk29c2T5jp_G$H{UgqeMFZ1Y=Uhja2(lVO1P3M_2 z^^xJtTJv)6Fu%Lx59c@akU$EO|^Rz^uBL3$I&!r;jX3P7MOi$qmA2pycXpkPzgfIv`_qye_zfl|#UK z|A}9?vF!707Y&?h>1Vv8O-X!R{p;&hIu2VA1vzi673AdbE`8ch)W^&}>_$CON*Nic zGWt{Udz3zWNMSyym|nQ8Is}K(q9O_g5Wa(ji;I0aWfh;V7Mwzdzs#Wo=mrjP%Le ziI3abWB1Az``U*PW8~(5fi6?7OKdaNj#q- zOmGX_=B__UkatY=KJ6#ES#N9mt@JCTuc86wZVLhczyCm2y>tAD)8hk4=8dyQL-qcf z(K1NwkDXqg?*?!#ye%6z_Th2y<)mj+4_6}6ZT_$tr*87EBsuo4m^UHa-?%AnosT~OIROM2ci`>6-p&I#E{*l@!#qP4DeZTmu7wWk zKC(W~PY$}f2^kqdJx=`_4%XJ-;`}=vuG|UZOZ#BbN)(kVOlpTAje*Xhdt5e7Et+$s zB3{KDHPAL{Htg}o^l(`T<;E0yR?@~LT|Wb3R**W2r1`LxIX&=uamcY{sT|tzrcn7kzp+>g$mj03x zL|k*Gw|2_kw!X4U>dsm=C#JJEPPbA#+{ZkRWZW2p(OcRJP9ze^647ipI6DJf*H0Y` zT?Vgd9y8%dio;$JDKb5JAkW-DI4?Euv8Xu?`7SBytWVg{feercUhYOVR}zjI zW7;~tQ48gP4CKoG9ipZnxqleA4Pl5jco!uXDEkI>-sgodYI;+TZjgnMn(L~zob@8n zAO8z)ECWwl3!>33$~`TjEk14l*kTosyEubvxxT`uxl^wQ)60_s^^))YyW(8izi#7W z&&Hq-MN&Q#WBy~4adS&OowWX6_om*|rtWI*yF=ea44*P3A{%62N^!bPk;!8VRu~>f zupYN#C^^cee|itB#zFRl@o3KaAWgTA5INQtlfSJ}+FFryhp%4RR9{KFW^KeH`L}%q zqT=zkXwn|{sq4}h;4D6N){;WYWVaE~W^w)RN_gL5SB`UU|L?;5%Fbij!mz1GkRNS|5AhEla{M7V{u*R>kz|^!)Rp~d%qt|zaH3m=; z!`Hc|kDW*GT6mG9GhQjjXcY>WiPtP<-VpMVrRsFvM2cbG97RsGfh>ObG{-RF+`AdVte-)6d)urS;x>8xR;L#ksAymYA$|n* zF4qJ3gf*@v*ii&8srIkA8dtIQD+p=J%?|uL+Xe7QnHpiW4CnEOk30LL zr%n7g|Ac=)RLs?^EfEMnrrZda3=5`*x z>9+WvqNvrCBU9w}5V_kgsEY{cqDKe*j`IX&eH!T0?0k{xqc;*6ZIP*wu8;U=$!;0_ z$3CWFwDy;OD%^Crg^~Jz#gybb=X08=I)ohr(SP|Tf{nLKkzr6@ta6)TcQOnvDxEfW zbje{sWiUP@#=P~XDX}qb#}_8PYC424x!ZQ;*XZeGlhUvi8;-brb!irQ&-#9Xk}`&d zu4>$DqgtGZtNean%;V&TORM7IA0O_hXnqe#fEg<70!DF<^iS6U^lKxZtJT7{nqz;z zn`Y&i_Lm_nJU!CeNA^Lt=M9(a2a~UaIFZ!j?$#najRr%2c$rar1726z$JA8$8i--S z6qU|O$x(kT84_Dj>3k1TJM?IQ`R zFQu0R+zW;+lyS#Z+2P4&&MuNF0K>Lp)A!W=r#6L5yfYgVE#Ps0CY-Ba^Spx2l#GC8 zmUIxeoudh``tw-`Ij}9vt8Vu8wyVFjR1gb$b*Z2zBaTRu#P*X;(Vt?)o%hNi;W}pN z(%A$o)da|9Kb&`GTiRD!Psh+WW5`iEKv#mm%MPSI8{u`rzFi-d|KU6FSG7i17Y0-y z%13^{GQRw??-Jhcxq=g8#>unyDY$VbTDZBAA4m%_>vwT>y$4CfX!&8G!In3z_qeIZ z(Hd|8o{H0CB0e5ME}`uY%&S3};M9JZ@SaPaA1heF&3V#K&w5|J^4bgw3j1iAsl+(6QdHh^r~V4;NQ@8vjnDIgBrTzQKf z9~8A_oOW!pQ$B<|Q5mY}ho{}8cEUNC z{%B&#Vt<3Uvl`BrpjYHOI0W5$cjEl8)E1U*{>;LiFG#hXCA#4t7r^&}=qV&#X%JK{W7u%0 zh6FVSh9RMhXa;A=hYi;XtJ^!43VxHhE1yxV#57}D0~pt~vb~+CD%aope3%*_$T7v6 z>gGbTsV;F0HxwVi2M&lO~Kd&s>^3RpZ5Uw z+MVxfH$bKK-I8&@IKXh7ymyRk*_0wqB9U=D^q*_+!bb@fz1(Gwon1!`mY{Ne2b9I9 zv!vbK%y|lv=3W`A3pLviDoLU*VZsHWB5Sc!?|bG9LEqz7+6&Sed?Jo z1L(x_KNxB!RCFAa2BDb7^XOyT<1NfX+6siK5|>;UPuBM02X{}8Hug(}w+|R`^4}K2 z(E8=KKA}0gzy5g|HeKwtu=0tn>-Olgxw!?a(dS-s%cHCw5Gkk%;QRx1Mq0%Bh-#0{ zpV!dXw%$+*wlzHLF1@wjvFU`4`_5=(l+|oP3~$6ws6mdqxFmctsP^5t1b3>rcTEuZvR(SIeHNg+ z(9fW6WKJOrZV!4<_$2`Xo)0ZSH>%734q))64f^Qc722pd_#HMYrwYJJo1&zY9*{e$xlh36;W#OGvNZZu=|v( zAj6E$T7~!KpBt3hML&Ybqm>vAuNGC@qPitQmrR~8 zIVRC+wj5Tp%7?~ux=0j6sSfHoPD3LsiUv-@9)$^F+PB(9%Vf0mqY9sw{UDD#fkyY4 zqLk-6v46zoy#!E1gy;Vq8OhESM@Y2$$PUTV%tb{O_7qWDwjF^jHA2gJw<=?kp&8>8vyL0rh&zJA*=(i>a|8 zSj89JxO<4D*6HSsOF$+4?g93gvTV5cD=pmS)P6Yb701Lc(Eon^H@^*X9%6>FikGX2 z)7sk0PkFs`0Q?9XMmS^zmIv5*Ki;_s{7ld1kGsYlQt+=T1ETJeT!&MC?W`UwmLp(o z7R7QKG9Q<(oxcCnL?6xnX%gV*hlOgG`ZXoOj8cd6s==~mBZ(T8TC~oc|2&H&qD?;= zcASRl5Td0ASL>d~p_nu8yb@2*h`DptA$bfcrE)b1>Dch#pC#wIRt|Rahq&JXW1&vC z9edMx1T{Wa+>YQW0F&07$`-m+#5@fgkpT=J|J?U$HL0xw>G^{~@ixi}ycgx@*xZC4cn6M#q^rT+}m-ueoh#OQ4Ge@FeoAY4!ljY9Bu^KTag>F1XF@Ad<+sY#buRXx_|^`b0!Esh~oDKbX$ z*ecrb1$eCYHeyvOI;7gC$6M-?;q$Kg*)BgiFY92j*lP+39edV3DTyd=WM-R3nBEUp zQaxrRPK`XHf`DLXOB65tkP`y`s!R5>tSo3&Xb5G^^;18SpHI{z2ZzSPks->&)!T73k&Au;S zydm2qC0mhDMK+@%b3G9B&#$3@3>MzWdS2!Adf0CCRrVcrUA@vCG>l93p!*Mrd(AyY|v)h<-^_J6ku?bpX0SjOR$X;{5`MG_++ z&0u=}lMywn$om+3XK;ep$tJK~}gx}6d8C0s3SclEK*gVvz9g2BiTslhzmf#Nz&O$N!UEd@*ybo@?q*8D#+G@yxChsW5HmBB%99O1GyB3(kJkBWDj-<@zxA@$Ulri}L0#0dj& z3RBM3!hh&Qkd=aw-xAsj1)6*woPKCEm?<(T;j87lEH62QG5?O+A1W&=JIuehm4vs$ zmR`IHFRD{J>gccSozmC_P>S2meyB~&Nn!2eb{r0j2J8=wcHj|7VGn2FG0f~)Xnijc z+B^-GI&LFgigLqd5E{O`FW4ZT3X9FVZ4R^`Ke|4F@cy9~=YOnFJ0`z4_H({wjHMg> znQ?WQaaF6cenjb-YV_C}DWDH&x}ek*qbH4d!$#XlK;r1Ms1k9Js&~R<@aLjLEjVRI zvu$0qk;Ihw@y>)%^HgDK5s+#?)u`gj(1|G>8IoR1yj;e_1m|D$8KqkLqZ(*FXO?T` zuAjyG8k1Tkoc*T<ms-Sbw^5i&Wl(?PRtR`K{9r(Mkj#q<)0bqTZ*)QQWO|LD6?-zbpJyhRFY$$Zd zeB4o#DhrJsUUIZsgC%IF)Y8aS=)t#F$m%B3e)?&jJ#mRZEG5FK*}%AhDH7?Xf(IcW zM7lswxK^S5!u~4U!3H|Ae~MTDx)nYoH=xbv5#WyW`^|3jx=+5;0KO?OnRZ}RTBJh=xp z%5z23lNRQqKiA30h#!2Wlk6Jq{lPPP;eYr2?z8{ZSjEm*c;LxJ{X=ZPf1a8jcX$1( zIg+_0Q-;2#<^f9wa;ofv!nm}fL|JXr9qYrFP7#o^7qN>Ov@FOy9(B_T<6gMKTV^UN zebm>pY*MQQ+B0m~hi|PKhQ#LE_o-8i;b}u>@uy~AolN(Kv%=<5=^!kbp`IH8{5cE? zAcUBVC%T?P2h}-bf_OYfmdOp@6DtGeF<~&%Fy4ripf*O+=I~>rjpB&!nS*Wk$%-$r zvdwU7r1Lt0RpSsD-8e6?v$KIF6+i3~8ynKWT-K<02Q|_gr&jV%ch!P?i=1D|@>Tn|ast$p`iU!A;7t6w_dkNc zO3W(2@TlltNy`wCirOJY)X-V(eK#)cy7s%7#iLCu3ccdYJU>>K8dx;kofoAxXrUg= zt)#zvDP*>?YRE{ffhDM`mqL`89f7~bmydpi>A|C5vI`yBiBNw760N~YLa<)Dgy!sr zE34E03N11~V(7G;X0tD}5mOGe0YxVH43-N#7na6s!=+k&a2z7J6idP%{r(3(BKwCs z466O#zADfyucsKiZWLSmYSj_^(NOi_*SPo8K2w=y8`p4wNNmnX;$~fc+)j16Nw%fb zDune#m0v`e}^lQ-aA_TFXkKW8Yf7K;<|tm+MsH zlXRdYlnNA;w;_9FU4pyu#GUVYu=5Qi?sNC@NX@yHc+}97Cc^@hYLGbqmJZVDzBf`;egqagQ>@gmy;Ix=z<9W(W}Yc0IZ;mwJ$e-PP`tV}-Y z&EE5|BYmwK+S9SN`Zc0DjA@$fSR2x)d2cW@&z5cHuVSTxs&yh(-r7;=V`7R}9>|e# z(S(fEf)Ch|G&|AaZMTNF_d5_pUN)5lG1mQqW5CqUAIlk652~eFP0m*42-|n^{=5bE zlT)%}LK}SS)Mt7%KZtxdibST6k&uv(MN9{ZESmu@8 zFvHR{lOnyN3r>ES_-*K(kG?Q(^z_~(l$k!q3bs5>!eh##9*@UcZmb-fdMcUht$v;9 z!GEj$;)|(t*@~o5r6wR5ahQsN`j*6B~y!R)o6GooaA+Umf7b z9&TfuU(AbWHh}xxzRBf)KvaK~(dyT-^d$EKW7_WGLDnCRwz8G}tLDmF0{0{2{f*H- zc8q_ed!kg9a>Y&;g6CWEG>Nx0Q-GM8YAzVCvo@4x44;XtYor%}{%Op$XnQH7rc`Qw zt`d8E>%$JV+dg@SZ>)*4uFdFww_CqmoV3EmypC3%)+DJ8h2tCEclyhPd@>4FaU5-$8AyERlCQsMgbq;Kw=d zo6GVL?v6z=rK!gTMY}oJs-G0Zd(^w8tFN)c{vlkTqKz(bDx1lFI5aNWFCOyPXl(Gp zmq6*{;qcYc&GFL~pP>%Fo@Cf^@&mcb(rprzYzgTWFFUP<$=Vn-(eXok10T3dwT1FVf8O^lyWTuIB_@49_EmNhTp{`FSG0Sc3FY<+syjk8> zv|gpHQ98T4#8Yfi=3u*z9xHmNIVIllciV#GTmbK2l($=v%#7?w}az3sC}Gw_tMIyY_q9q?2ow z>u`Sx&OVc_`!>py_Z0I6#$88C27-TXg=gCFydjIEThkj1FkWIEih)Ho(|tejUEbr5 z$+uosBQUr0{y4j}Cr8zLgID5VZfg|-K5k;?8QhnpMjNq{MS%5mj#WL%Om#>&$NPLb#;x08 z+;>nu_S|?XbRZ_YLqM>Wez?5N_di0iGk(cAK}xXw!A2*xj| z56B%Ewa=pmBkaXMk;Xs2wnz~z48LLy4Bv?mL(@~Q(UsisG>A2uE1_3t8}i5Idu~@O zW7)DPP67ZMIqoCh=%7#6eUSX7rxi_i95+$-_lhij909j}h|j4QiQ7+yZUSpG5t1~7 zk96pBFy;T+Sgn41V9L8Siudu~STsK1s#N)uVqQWA&1L;7ZKJJUIu7S+q#6Y3KsR^` z5OO;gHh8&;C@+^xUt;md5kK2wPxYCQL)6~D5}Ro#5OL+}hf;wcv>DCO61hv_oF=cs ztWY*w+VTCfEK-C`s`STyF3J6LIjKCmog~0ab5{G!*6T5AF0`?%EY?s)iMcV)$$co@ zZbe1$>HNs4UO!^2E)o1*TE+VRu;&%uRoD*Y-V=Gp1@^g<+{bdC6puU+Hnof7|{*@jjW`H z5&EJZb<&e#k!1QR-q%cp#*K{Ev+o|)YGc6f`M)*1>EWmAsT}_%)FFP}P6zGf4-jTB z5I<0!tGh2Zd6OK+sX5ybS=qn%U-qN5m4weIB@L#J?Lu-;!pHjBEa8}XuHi3*{?nz5 z0EQ={_4-=a34rH*o(-7Wi$p`N$!Kpb>S4JS8Lc4T-?YotZe?vG=60ZMn&0Sm&y@U3 zboSkRiW^GX&q^}+3oVn%+WmxNsC7R;?*cXv5LX{^$DX5GkzRf}?`5e>vzSDV zx@NUDlgl#?3K&&NZ4c*o72tyD|4RJ>vr3ebd^Ft^EyECh1{ep7obSZ%)0UCcIBQIf zPgkM^8B)`kuRPZy(PC{pjudYvF_zF=zS}K-3&%=?O8>sEeAb@IyqlBQ$;*`Bi=#h4 z5b(X9QP`zVPv)_ycW{()O53z9#AmQce{;(l!W}nZ86R#M>S^BH6RDInyI6M-8XLvj zYe$o?Kn@($He%_M^J}x7DA6tD zK?nsOZl4|PR=jQGbj`-My(f#TOIPqHnvJ#4&rwgw8)Dn=sNhC}baLFiHI0h5J(WTy z9EZ;rk=`~xH`Rp!U>BABva=SG`rQU%=O}lv3iiya<#fG-dfkO!iP8;DO*O$P@`l1K zwfjW$DlAxpA(oB=v2$d8Ie$LD6c^bN)mEq5+k<637I$cyAOO!xXm+wvBOsT4G}-R1 z_)i&-ermHUf2iG%)%>rV;16wWo_{_Do%|ra%rT2W>|B0JJ!^UFD#18%d{I58Nj+Dx zO;)z|T3ox;U;24RKi8cW5<90}l7!i?fhSGbOAp2+dD2=cASY41YtU%!{SU~u%)(8r zt!Vq-T4VYGhE%9Kb=%a@{>b+6;=X!qi}=2!h|`TslAj4TU7UovL z)kG3Cg*XH&$^+y`TpozBf}DxusJq<9LrkWmn%^?t8#D7f9(KQLn575Or-jG}lY;ux zTD>E2$`k;E=#32zFYBPH93yRSmO)>!A?12WRm0p0j7QFHefA`7dd6HO8;^&-?V*DQ z!hy@kfud|n7ncyB=*mnv=<6_9+O!Z{em#cQ_&tOfdrCg=v~NXw^Kv4{NtNkmeV57` zZ$BOm`I#f2>Av$ks`Qc4Wi4*(=2WQ1pwuJXI#0J9+n_tVRz>;kP#GlZ`iblA!DtAU zlxnv0g!i-P&+YtW{rg^igdYBYImF^I^=<-ly?^gy%HLhT+3SnS?WzqcchezfVz5#` zw{4U^_QDs~?n4;9`Y|MHw~6qO(=Vp@dJ^xX2EN;oU?enm*Uxp5j(8LW}?F+t2{6XF8fexPnqRVO%)AxPiVLFxE zK0QS<9c{|(tzfr(5%+$pVi1+zY-(6>uy+v1f&nmO#>LgD?o&^+2Pl<-oy-8a zdtRAcL8hRl(UsajgyoLbCqqK^dLV6n-Zr626L;JOWi28zs#Z%vkC5J)y(p#rx)x(!<7Da zee`z$VwCE*0%E6>gR{W4t&3c zOTyX0h5oaX%{KeqZu|Y7P^U=!?8Hw#OwdzbZR>AiX%&s-_3xS;OXYkF47=?XSxonc56G!DzrOkVaHP-{7aS!o2CBENFU)PZ$ILnKrAMNP7sb-Z! zLGtv#KasYPT=RW#{N_G%Ebxa`^(8WANj~=>G|#_+ZVBgvi0`v0Yj)}V^6=(3x9Zm7 z&)41~qwaUY5jR+N9DWzkOw}au-=)9wVs{}2l%PNL?N9Phr>U3;Qeu&Z}`oaEDkf?av zPwWWDo35G0^(60>f@=WS`MT(-_o3!F{AIM^tne_m^aKZcXwZUnCIpA+}YQS84RX+8C;<05y z`2`cWTlj&-RX@@q0#ht(r2K-lVeo&roMVf+yS;@dhHkjMzhpZa@+;}#9<_lrcc|`` z9Qu5F=vh~KxfZij*z|~LwV{ZW##6{(I!exccJfH$+mzf<#};uAVNkMK8nJIg|4RH2 zt6+vN!d*k*8=|q|17YA*-37rkIt>dLNQw~!+#R0pG>=F23}JX95AT&sW1^Or^mw%D zT}b&&m7I)s9!ZK#{-duSu2kNcMtSSING2D4Hj_Rr^#s+=TAs5~{8;d@2LA4%afrE0 zX^BnBJ~_zWi&;k!_YO93QUHK$vzF)Py0w10Vr+SQ)hn?I2eiHocLX!BghJkLw z9LV{XISz)FWeEAgl4PPoGPs7%@q2aRQ2!Y9=>t@)+po2@c|;E`uw(^^}mLdkeiJS4FqL^3Ho zFkOk5i;3pbv=q(1k!zI^`^)r~%;E=i^yf&L&6)qPO)W%jg1`QTk`|Ql#kI|dHeI|Q z;NA9-&iYq#^RuGu_Did4ue%j?`fAg&#UAaQuP5nV$NotRoiQiEn-%Um` zlQCjqK``%UDFe3Gvf1p5ZIvumZfl*-t)FGt3J%}5xU27)yr>6>eeAqYn-KOfaMDCt zTcOOM7CAb?YrfboRcgO;VfkqvkMnvlmT@gH#SqS7`5?|5!xSFNXSjp4Q^dEvna?QU zGJE@970fHf7U&AY~vb#5nk$&BiIC zp^aa(nae0?F)^OBPwCj)rxq}dlKtj(gc`hB^X}GQo4?zo3_fr=7rGrQ4)1J(cQdjrFAH`rxhG^WHZ$C4;!4iMgrwqqHaOiY9S zazF5OY!2rOvGH{bT&rLAQEO@>=zjr+MnZfTRVvy@|9O>niS^D7`xs)4fvQFMZE8S*X5T6 zrz0xE&U<4Ww_+fe19S8LQ=9scGi_qh=)4&6=`+4GNU;vloJo&aD>RPai5+*D>vmtC zwPEQX@94x&qoaxA^5xO1w{$$6*7ZHl(<@I)Dow7e2t!rV$Namk&F^5ZlUDkP!Yp0Lj22be& zRq`{;dwI%SUB8TeQlASpyE&3-$!MT)U`JyZ$+xzJAi7XAc3xct^vjerQ?f8NCo9s9 zkhCK~h)HGFgJK77sC+l2@MTb8uNq?C9}#8Sk3;*sD;5DKS2OrXqbOq`f}1MZ71Mn6 zo!&6kT{8aN8?eGfZ;P?dRXqkBrLh>2pO7L**+cRTPWgpJ)(288xKc|SkUNwOm5D}u zI5@t}*FEEnT5NCMUy9i+Y$q77bF;7TZa@;{DF<#}S=7`xRVf^6h@Lf+n3Xlua=F5F z*XKOAL}%QXfXXqsnR$4;hreuYcZ;A_qiuii#v;D|mVV)5tb>xBM9i`O*vFmk-nIbr zkasbnAU~38UWd7DIN8T!*Z0ingHDbJbR@(+gx;KgkR(epGKRqQsMaF`d_;aBP7AxI-)`%1 zl%eC8Y0fw@)|srU?C)Y-yMaxVDE{T}_!P10a=&2K0Ca>B>go|M_D`S369k^v@0XaO z1d=~sP%O_Cow!nIKV2OCgStCcm)N!6kq+>GU4Qk~s_WI(gspkwFF@&UC2<|`oBZ0? zV_}ihU5!i7`w%_PQjga5`5r0t@(U{#eu$dmOCo(&`1S+TenvKYc4oMI3|eZAq--&G zE(B%)+&asLquT7W+QJm2`G#z&_fqUsbuKBNoYy@qkmEDVdQCr8zs+^IRu`L7+__>h zgr7l%UIq%jl+Z(GYP-WcKgMS?AF6s8Zam+u`c%yqGjqsDRX0X@RJRLfrea;I_XyLERFT!=_5y$^)=gJ?U{J#GfBO*%^C2LtC zgtrpeOIb@PnNOlgNY9hB4XKv3#!Y_iy<8 z_MTs!AI_a~U+3I&?>YCJ^E`j4f0KVTpF`*!d&6WnO!xSFs)y-bET8U~Wo4!~q>~R) zYX>_1#5cx3prYqFVKI0wxS$Xj9g$D(JzWSdE&C*;RsbI9DZa+?q#g<*X~O*yW8ymc zdy+YW`@sFp!L@zf^8p6bdd93>&d~n#>AZOb>?5OTs&%DI=)vYh8DAXr##K7*c;P-* zpCk=C;f6=mM$QQ+ID^=vWo3?yMePcW?(?BR_a=9)C)9ugk$>=@<$IaP> zPAoWpFAGUZk6G!z^XiXo<2_bjT=pQm%1`UBXi5II%kl>$*vD+V}B%mwj9@p4?EPzE8(!9y0;pVCn(Lob%Ft`StIa|W#v z{2x+~56gg~rwTVddGBdnwQvaGoq-#r9@k5AQSl)Dq*T#75FP^e#pBg{gGnmzH+^%1XHtLCKpF$h?Q54NQDfq^mzHwHeI^vG5rm@N$@>QA6kQh-&H5pSV3{ z-z7+wdX3DC222!PTaCUpN5R_jvbP&FEVi1EA{s_z2V$M-)p}Tr$H58?Ypw~pRT0a$)Xr)yNOc@xjcnN>KHUwePUJT@Cl25{RE@GJYD?G*saCEU-4bpTIlb_x?X|Jf@Uv04>w23?uXa;H-mL>OXJiE5j()*=!KdLqS%r3I{iNccZFHG8e_TrpVkwugA=h~yp$qVqGvY$f| zT1FL(w?>T??1DTO<&-WyH06k&csHeCsPFxElYd~2%VQ1?1sBP>!=RM*=I72rZTkuq zB{+OCd&>I^AermY0F_#oV`zn&;F--N}Xl&DF1tJpG^zx!x z)6V#(>!03KxNeHQg?@js3}qbe$_`YJ9>OWyGFmezcb=^O<5V4Rly$tKZ0?n7hfBA! zUq$2EW+BS7bqWPL?Uz<{P$|LXkH>qiHgLVv>gBhf(lp!QZAhK=Z#hmQZhf>WRN}o8MKwAIe(eTvXm^M@+c*I$J$2Kk`DIuTUMfQ5Cd|63L z8#xmyZM8A>rHt6F%u>95<0pCxFB{Hg#kpe-mHNB6#px~zH>KAT#G`sRi%9ZP>Xa<^8+%jBQqXfx~ zekaN*W{wwpPQoyHnLk%r$ge8OXDz~sa6xfeW%f|K>w`|UA#zT^d*)<3`?nSQ<7 z>e<$y7WF~bL8HDJH0I_h1LkDdR_ZL&k_23^DP)sa^KT=;kgaFA!Sl!I@s&#jmyu>kyk`cCXfAf3vZ%q+0GKT>4aZ9 zjrSgb4~gxxQFgXIJNy%rG0WbjvC=bV*u|M+^{W$yn=6p9H5y4~27?)~g6ZrUtv^FK zu_9zO;%Rr1PQ5ApWE8JvFihk~&R;UVcY9%LBE$LeX`0I(jAXgF|n|8S;*Z z_D=IG&22fA0KHA>lwL$zfB#q}jZOGz<~n3k-euWhZ{E=xBP#vey77tq6>sdCy_9l5 zn&4LT2OMs-kMl*0t;Wd;o6?#wyrm&sHkN&WS3yuv0;19%v?oMRa9)dO!Q9xs${n*u zSa81tJfKD}r+*(h{A3N1a5(Z-VPFG2byK*`du-sKg{8axtkD;*Wc>KWHmkV2OPE|s z%{Mnm=@i348OZtZ8^h79?@ADo67Y+V{3a!~gv`cuG5rO5QLI?MY$RbQ9`=XJF=pR;ANbdnNvVCaYB1-PwoDmWOqkPIf-t9-rn4S_@|!^q?JoVddnwePV3gNcad6sCp!qIm!Myka=dat!1Fqh-P9_enU) z7Cx81RK2NP{9K4i3uPJx1)0U${!Mnq`5~Y)V-RG7;Q>E=ty#bIU6$6xOK+|CrWnT9 z*)%2LHgX2W+lkUEr-i_gwiS(oQw+IOm*hNMK!g?O+1$MF^IvDdsC;&(GF8Und}5CpdgPt+|i~7%+E-9O#2NFJ)q+8{|gpDHfvoBKf9@ z3Ofheu(7yiYx;g5#0_xmTIseX2K*pUqVY|R_qEFx*XJr zSYq2WPeD5RLe%5y1!*->xB#GEzc}FJ4MAUyWp|ye{!Tpmp0AQ53}iT6fwRR0>Y+wL z5Dv&1L9fzf* z>)(yWO5aIYHw1W(f;cRs99oLmD*8NY-g7|c?+8UW`~GGctld3ly@*~ahBM)yudD;; zSSc>+%q10(!(-|lvA`1(NlE18a1aG`Gaf`>L+fvRDC zU@O8v3Uc7RsN4Gr%l^W7N;Uu)!%e154P~R4V*Zz+J-*4hK8AIl8Rgb>#+}7C@!S_TPxqhM` z?Xz-ZmF+G@@#jDh7qG)4+B1L8!d2w-%8540`HLaAK2C=rvduo z4qUdk+6Fs#O)VH}7xeMLTMd;y8N%SduI7;XSKU?N4+aKIa}OH4bXSp%tny%Xi!@jO z1AL>jZiRVa(9Lj42$>45Os=2`|*Zy;xy@+ z4X+XXQlEL3yLGgrTHdU-7xE$L#q{^~-p`xTyW-9B9nt-9 zDgxZtK|HFEXi$9KtRcFB6jA_6joM{4ifUQ!i2QqhO-S8CX!5U=3eb~eZ=CF)j+x1!~z!ULZE@Jcw^~<_WVDkg%b*mGh9QsgdCn1lh zRSVXoPV{KRs2`K`sFRZk(zj|u(^zm0f&*9rAmC50xY?^+nG?Rj8P~Z?oy+8t1=I9B zhD0?U-KVJ4FI-G}|Nl%DU2fb1>!>4NwLJpLvA jtAu|o;eWmzI*g&gu62;ASE;z&&1ZSR<~;G-?Wq3(I(xn5 literal 0 HcmV?d00001 diff --git a/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.py b/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.py new file mode 100644 index 00000000000000..847f2730f240c4 --- /dev/null +++ b/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.py @@ -0,0 +1,412 @@ +import io +import json +from base64 import b64decode, b64encode +from copy import deepcopy +from typing import Any, Union + +from httpx import get, post +from PIL import Image +from yarl import URL + +from core.tools.entities.common_entities import I18nObject +from core.tools.entities.tool_entities import ( + ToolInvokeMessage, + ToolParameter, + ToolParameterOption, +) +from core.tools.errors import ToolProviderCredentialValidationError +from core.tools.tool.builtin_tool import BuiltinTool + +# All commented out parameters default to null +DRAW_TEXT_OPTIONS = { + # Prompts + "prompt": "", + "negative_prompt": "", + # "styles": [], + # Seeds + "seed": -1, + "subseed": -1, + "subseed_strength": 0, + "seed_resize_from_h": -1, + "seed_resize_from_w": -1, + # Samplers + "sampler_name": "DPM++ 2M", + # "scheduler": "", + # "sampler_index": "Automatic", + # Latent Space Options + "batch_size": 1, + "n_iter": 1, + "steps": 10, + "cfg_scale": 7, + "width": 512, + "height": 512, + # "restore_faces": True, + # "tiling": True, + "do_not_save_samples": False, + "do_not_save_grid": False, + # "eta": 0, + # "denoising_strength": 0.75, + # "s_min_uncond": 0, + # "s_churn": 0, + # "s_tmax": 0, + # "s_tmin": 0, + # "s_noise": 0, + "override_settings": {}, + "override_settings_restore_afterwards": True, + # Refinement Options + "refiner_checkpoint": "", + "refiner_switch_at": 0, + "disable_extra_networks": False, + # "firstpass_image": "", + # "comments": "", + # High-Resolution Options + "enable_hr": False, + "firstphase_width": 0, + "firstphase_height": 0, + "hr_scale": 2, + # "hr_upscaler": "", + "hr_second_pass_steps": 0, + "hr_resize_x": 0, + "hr_resize_y": 0, + # "hr_checkpoint_name": "", + # "hr_sampler_name": "", + # "hr_scheduler": "", + "hr_prompt": "", + "hr_negative_prompt": "", + # Task Options + # "force_task_id": "", + # Script Options + # "script_name": "", + "script_args": [], + # Output Options + "send_images": True, + "save_images": False, + "alwayson_scripts": {}, + # "infotext": "", +} + + +class StableDiffusionTool(BuiltinTool): + def _invoke( + self, user_id: str, tool_parameters: dict[str, Any] + ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]: + """ + invoke tools + """ + # base url + base_url = self.runtime.credentials.get("base_url", None) + if not base_url: + return self.create_text_message("Please input base_url") + + if tool_parameters.get("model"): + self.runtime.credentials["model"] = tool_parameters["model"] + + model = self.runtime.credentials.get("model", None) + if not model: + return self.create_text_message("Please input model") + + # set model + try: + url = str(URL(base_url) / "sdapi" / "v1" / "options") + response = post( + url, + json={"sd_model_checkpoint": model}, + headers={"Authorization": f"Bearer {self.runtime.credentials['api_key']}"}, + ) + if response.status_code != 200: + raise ToolProviderCredentialValidationError("Failed to set model, please tell user to set model") + except Exception as e: + raise ToolProviderCredentialValidationError("Failed to set model, please tell user to set model") + + # get image id and image variable + image_id = tool_parameters.get("image_id", "") + image_variable = self.get_default_image_variable() + # Return text2img if there's no image ID or no image variable + if not image_id or not image_variable: + return self.text2img(base_url=base_url, tool_parameters=tool_parameters) + + # Proceed with image-to-image generation + return self.img2img(base_url=base_url, tool_parameters=tool_parameters) + + def validate_models(self): + """ + validate models + """ + try: + base_url = self.runtime.credentials.get("base_url", None) + if not base_url: + raise ToolProviderCredentialValidationError("Please input base_url") + model = self.runtime.credentials.get("model", None) + if not model: + raise ToolProviderCredentialValidationError("Please input model") + + api_url = str(URL(base_url) / "sdapi" / "v1" / "sd-models") + response = get(url=api_url, timeout=10) + if response.status_code == 404: + # try draw a picture + self._invoke( + user_id="test", + tool_parameters={ + "prompt": "a cat", + "width": 1024, + "height": 1024, + "steps": 1, + "lora": "", + }, + ) + elif response.status_code != 200: + raise ToolProviderCredentialValidationError("Failed to get models") + else: + models = [d["model_name"] for d in response.json()] + if len([d for d in models if d == model]) > 0: + return self.create_text_message(json.dumps(models)) + else: + raise ToolProviderCredentialValidationError(f"model {model} does not exist") + except Exception as e: + raise ToolProviderCredentialValidationError(f"Failed to get models, {e}") + + def get_sd_models(self) -> list[str]: + """ + get sd models + """ + try: + base_url = self.runtime.credentials.get("base_url", None) + if not base_url: + return [] + api_url = str(URL(base_url) / "sdapi" / "v1" / "sd-models") + response = get(url=api_url, timeout=120) + if response.status_code != 200: + return [] + else: + return [d["model_name"] for d in response.json()] + except Exception as e: + return [] + + def get_sample_methods(self) -> list[str]: + """ + get sample method + """ + try: + base_url = self.runtime.credentials.get("base_url", None) + if not base_url: + return [] + api_url = str(URL(base_url) / "sdapi" / "v1" / "samplers") + response = get(url=api_url, timeout=120) + if response.status_code != 200: + return [] + else: + return [d["name"] for d in response.json()] + except Exception as e: + return [] + + def img2img( + self, base_url: str, tool_parameters: dict[str, Any] + ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]: + """ + generate image + """ + + # Fetch the binary data of the image + image_variable = self.get_default_image_variable() + image_binary = self.get_variable_file(image_variable.name) + if not image_binary: + return self.create_text_message("Image not found, please request user to generate image firstly.") + + # Convert image to RGB and save as PNG + try: + with Image.open(io.BytesIO(image_binary)) as image, io.BytesIO() as buffer: + image.convert("RGB").save(buffer, format="PNG") + image_binary = buffer.getvalue() + except Exception as e: + return self.create_text_message(f"Failed to process the image: {str(e)}") + + # copy draw options + draw_options = deepcopy(DRAW_TEXT_OPTIONS) + # set image options + model = tool_parameters.get("model", "") + draw_options_image = { + "init_images": [b64encode(image_binary).decode("utf-8")], + "denoising_strength": 0.9, + "restore_faces": False, + "script_args": [], + "override_settings": {"sd_model_checkpoint": model}, + "resize_mode": 0, + "image_cfg_scale": 0, + # "mask": None, + "mask_blur_x": 4, + "mask_blur_y": 4, + "mask_blur": 0, + "mask_round": True, + "inpainting_fill": 0, + "inpaint_full_res": True, + "inpaint_full_res_padding": 0, + "inpainting_mask_invert": 0, + "initial_noise_multiplier": 0, + # "latent_mask": None, + "include_init_images": True, + } + # update key and values + draw_options.update(draw_options_image) + draw_options.update(tool_parameters) + + # get prompt lora model + prompt = tool_parameters.get("prompt", "") + lora = tool_parameters.get("lora", "") + model = tool_parameters.get("model", "") + if lora: + draw_options["prompt"] = f"{lora},{prompt}" + else: + draw_options["prompt"] = prompt + + try: + url = str(URL(base_url) / "sdapi" / "v1" / "img2img") + response = post( + url, + json=draw_options, + timeout=120, + headers={"Authorization": f"Bearer {self.runtime.credentials['api_key']}"}, + ) + if response.status_code != 200: + return self.create_text_message("Failed to generate image") + + image = response.json()["images"][0] + + return self.create_blob_message( + blob=b64decode(image), + meta={"mime_type": "image/png"}, + save_as=self.VariableKey.IMAGE.value, + ) + + except Exception as e: + return self.create_text_message("Failed to generate image") + + def text2img( + self, base_url: str, tool_parameters: dict[str, Any] + ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]: + """ + generate image + """ + # copy draw options + draw_options = deepcopy(DRAW_TEXT_OPTIONS) + draw_options.update(tool_parameters) + # get prompt lora model + prompt = tool_parameters.get("prompt", "") + lora = tool_parameters.get("lora", "") + model = tool_parameters.get("model", "") + if lora: + draw_options["prompt"] = f"{lora},{prompt}" + else: + draw_options["prompt"] = prompt + draw_options["override_settings"]["sd_model_checkpoint"] = model + + try: + url = str(URL(base_url) / "sdapi" / "v1" / "txt2img") + response = post( + url, + json=draw_options, + timeout=120, + headers={"Authorization": f"Bearer {self.runtime.credentials['api_key']}"}, + ) + if response.status_code != 200: + return self.create_text_message("Failed to generate image") + + image = response.json()["images"][0] + + return self.create_blob_message( + blob=b64decode(image), + meta={"mime_type": "image/png"}, + save_as=self.VariableKey.IMAGE.value, + ) + + except Exception as e: + return self.create_text_message("Failed to generate image") + + def get_runtime_parameters(self) -> list[ToolParameter]: + parameters = [ + ToolParameter( + name="prompt", + label=I18nObject(en_US="Prompt", zh_Hans="Prompt"), + human_description=I18nObject( + en_US="Image prompt, you can check the official documentation of Stable Diffusion", + zh_Hans="图像提示词，您可以查看 Stable Diffusion 的官方文档", + ), + type=ToolParameter.ToolParameterType.STRING, + form=ToolParameter.ToolParameterForm.LLM, + llm_description="Image prompt of Stable Diffusion, you should describe the image you want to generate" + " as a list of words as possible as detailed, the prompt must be written in English.", + required=True, + ), + ] + if len(self.list_default_image_variables()) != 0: + parameters.append( + ToolParameter( + name="image_id", + label=I18nObject(en_US="image_id", zh_Hans="image_id"), + human_description=I18nObject( + en_US="Image id of the image you want to generate based on, if you want to generate image based" + " on the default image, you can leave this field empty.", + zh_Hans="您想要生成的图像的图像 ID，如果您想要基于默认图像生成图像，则可以将此字段留空。", + ), + type=ToolParameter.ToolParameterType.STRING, + form=ToolParameter.ToolParameterForm.LLM, + llm_description="Image id of the original image, you can leave this field empty if you want to" + " generate a new image.", + required=True, + options=[ + ToolParameterOption(value=i.name, label=I18nObject(en_US=i.name, zh_Hans=i.name)) + for i in self.list_default_image_variables() + ], + ) + ) + + if self.runtime.credentials: + try: + models = self.get_sd_models() + if len(models) != 0: + parameters.append( + ToolParameter( + name="model", + label=I18nObject(en_US="Model", zh_Hans="Model"), + human_description=I18nObject( + en_US="Model of Stable Diffusion, you can check the official documentation" + " of Stable Diffusion", + zh_Hans="Stable Diffusion 的模型，您可以查看 Stable Diffusion 的官方文档", + ), + type=ToolParameter.ToolParameterType.SELECT, + form=ToolParameter.ToolParameterForm.FORM, + llm_description="Model of Stable Diffusion, you can check the official documentation" + " of Stable Diffusion", + required=True, + default=models[0], + options=[ + ToolParameterOption(value=i, label=I18nObject(en_US=i, zh_Hans=i)) for i in models + ], + ) + ) + + except: + pass + + sample_methods = self.get_sample_methods() + if len(sample_methods) != 0: + parameters.append( + ToolParameter( + name="sampler_name", + label=I18nObject(en_US="Sampling method", zh_Hans="Sampling method"), + human_description=I18nObject( + en_US="Sampling method of Stable Diffusion, you can check the official documentation" + " of Stable Diffusion", + zh_Hans="Stable Diffusion 的Sampling method，您可以查看 Stable Diffusion 的官方文档", + ), + type=ToolParameter.ToolParameterType.SELECT, + form=ToolParameter.ToolParameterForm.FORM, + llm_description="Sampling method of Stable Diffusion, you can check the official documentation" + " of Stable Diffusion", + required=True, + default=sample_methods[0], + options=[ + ToolParameterOption(value=i, label=I18nObject(en_US=i, zh_Hans=i)) for i in sample_methods + ], + ) + ) + return parameters diff --git a/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.yaml b/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.yaml new file mode 100644 index 00000000000000..4f1d17f175c567 --- /dev/null +++ b/api/core/tools/provider/builtin/xinference/tools/stable_diffusion.yaml @@ -0,0 +1,87 @@ +identity: + name: stable_diffusion + author: xinference + label: + en_US: Stable Diffusion + zh_Hans: Stable Diffusion +description: + human: + en_US: Generate images using Stable Diffusion models. + zh_Hans: 使用 Stable Diffusion 模型生成图片。 + llm: draw the image you want based on your prompt. +parameters: + - name: prompt + type: string + required: true + label: + en_US: Prompt + zh_Hans: 提示词 + human_description: + en_US: Image prompt + zh_Hans: 图像提示词 + llm_description: Image prompt of Stable Diffusion, you should describe the image you want to generate as a list of words as possible as detailed, the prompt must be written in English. + form: llm + - name: model + type: string + required: false + label: + en_US: Model Name + zh_Hans: 模型名称 + human_description: + en_US: Model Name + zh_Hans: 模型名称 + form: form + - name: lora + type: string + required: false + label: + en_US: Lora + zh_Hans: Lora + human_description: + en_US: Lora + zh_Hans: Lora + form: form + - name: steps + type: number + required: false + label: + en_US: Steps + zh_Hans: Steps + human_description: + en_US: Steps + zh_Hans: Steps + form: form + default: 10 + - name: width + type: number + required: false + label: + en_US: Width + zh_Hans: Width + human_description: + en_US: Width + zh_Hans: Width + form: form + default: 1024 + - name: height + type: number + required: false + label: + en_US: Height + zh_Hans: Height + human_description: + en_US: Height + zh_Hans: Height + form: form + default: 1024 + - name: negative_prompt + type: string + required: false + label: + en_US: Negative prompt + zh_Hans: Negative prompt + human_description: + en_US: Negative prompt + zh_Hans: Negative prompt + form: form + default: bad art, ugly, deformed, watermark, duplicated, discontinuous lines diff --git a/api/core/tools/provider/builtin/xinference/xinference.py b/api/core/tools/provider/builtin/xinference/xinference.py new file mode 100644 index 00000000000000..7c2428cc00534e --- /dev/null +++ b/api/core/tools/provider/builtin/xinference/xinference.py @@ -0,0 +1,18 @@ +import requests + +from core.tools.errors import ToolProviderCredentialValidationError +from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController + + +class XinferenceProvider(BuiltinToolProviderController): + def _validate_credentials(self, credentials: dict) -> None: + base_url = credentials.get("base_url") + api_key = credentials.get("api_key") + model = credentials.get("model") + res = requests.post( + f"{base_url}/sdapi/v1/options", + headers={"Authorization": f"Bearer {api_key}"}, + json={"sd_model_checkpoint": model}, + ) + if res.status_code != 200: + raise ToolProviderCredentialValidationError("Xinference API key is invalid") diff --git a/api/core/tools/provider/builtin/xinference/xinference.yaml b/api/core/tools/provider/builtin/xinference/xinference.yaml new file mode 100644 index 00000000000000..19aaf5cbd1398d --- /dev/null +++ b/api/core/tools/provider/builtin/xinference/xinference.yaml @@ -0,0 +1,40 @@ +identity: + author: xinference + name: xinference + label: + en_US: Xinference + zh_Hans: Xinference + description: + zh_Hans: Xinference 提供的兼容 Stable Diffusion web ui 的图片生成 API。 + en_US: Stable Diffusion web ui compatible API provided by Xinference. + icon: icon.png + tags: + - image +credentials_for_provider: + base_url: + type: secret-input + required: true + label: + en_US: Base URL + zh_Hans: Xinference 服务器的 Base URL + placeholder: + en_US: Please input Xinference server's Base URL + zh_Hans: 请输入 Xinference 服务器的 Base URL + model: + type: text-input + required: true + label: + en_US: Model + zh_Hans: 模型 + placeholder: + en_US: Please input your model name + zh_Hans: 请输入你的模型名称 + api_key: + type: secret-input + required: true + label: + en_US: API Key + zh_Hans: Xinference 服务器的 API Key + placeholder: + en_US: Please input Xinference server's API Key + zh_Hans: 请输入 Xinference 服务器的 API Key From 7c485f8bb80d170b5756575c150e7cdb41613145 Mon Sep 17 00:00:00 2001 From: ybalbert001 <120714773+ybalbert001@users.noreply.github.com> Date: Tue, 24 Sep 2024 10:33:30 +0800 Subject: [PATCH 17/79] fix llm integration problem: It doesn't work on docker env (#8701) Co-authored-by: Yuanbo Li --- .../model_providers/sagemaker/llm/llm.py | 20 +++++++++---------- 1 file changed, 9 insertions(+), 11 deletions(-) diff --git a/api/core/model_runtime/model_providers/sagemaker/llm/llm.py b/api/core/model_runtime/model_providers/sagemaker/llm/llm.py index 2edd13d56d4d87..04789197eec2a1 100644 --- a/api/core/model_runtime/model_providers/sagemaker/llm/llm.py +++ b/api/core/model_runtime/model_providers/sagemaker/llm/llm.py @@ -85,7 +85,6 @@ class SageMakerLargeLanguageModel(LargeLanguageModel): """ sagemaker_client: Any = None - sagemaker_sess: Any = None predictor: Any = None def _handle_chat_generate_response( @@ -213,23 +212,22 @@ def _invoke( :return: full response or stream response chunk generator result """ if not self.sagemaker_client: - access_key = credentials.get("access_key") - secret_key = credentials.get("secret_key") + access_key = credentials.get("aws_access_key_id") + secret_key = credentials.get("aws_secret_access_key") aws_region = credentials.get("aws_region") + boto_session = None if aws_region: if access_key and secret_key: - self.sagemaker_client = boto3.client( - "sagemaker-runtime", - aws_access_key_id=access_key, - aws_secret_access_key=secret_key, - region_name=aws_region, + boto_session = boto3.Session( + aws_access_key_id=access_key, aws_secret_access_key=secret_key, region_name=aws_region ) else: - self.sagemaker_client = boto3.client("sagemaker-runtime", region_name=aws_region) + boto_session = boto3.Session(region_name=aws_region) else: - self.sagemaker_client = boto3.client("sagemaker-runtime") + boto_session = boto3.Session() - sagemaker_session = Session(sagemaker_runtime_client=self.sagemaker_client) + self.sagemaker_client = boto_session.client("sagemaker") + sagemaker_session = Session(boto_session=boto_session, sagemaker_client=self.sagemaker_client) self.predictor = Predictor( endpoint_name=credentials.get("sagemaker_endpoint"), sagemaker_session=sagemaker_session, From 1ecf70dca0b313966d8955bdaf43fdac5515cf99 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Tue, 24 Sep 2024 11:20:15 +0800 Subject: [PATCH 18/79] feat: add mixedbread as a new model provider (#8523) --- .../model_providers/_position.yaml | 1 + .../model_providers/mixedbread/__init__.py | 0 .../mixedbread/_assets/icon_l_en.png | Bin 0 -> 123637 bytes .../mixedbread/_assets/icon_s_en.png | Bin 0 -> 37303 bytes .../model_providers/mixedbread/mixedbread.py | 27 +++ .../mixedbread/mixedbread.yaml | 31 ++++ .../mixedbread/rerank/__init__.py | 0 .../rerank/mxbai-rerank-large-v1-en.yaml | 4 + .../mixedbread/rerank/rerank.py | 125 ++++++++++++++ .../mixedbread/text_embedding/__init__.py | 0 .../mxbai-embed-2d-large-v1-en.yaml | 8 + .../mxbai-embed-large-v1-en.yaml | 8 + .../text_embedding/text_embedding.py | 163 ++++++++++++++++++ api/pyproject.toml | 1 + .../model_runtime/mixedbread/__init__.py | 0 .../model_runtime/mixedbread/test_provider.py | 28 +++ .../model_runtime/mixedbread/test_rerank.py | 100 +++++++++++ .../mixedbread/test_text_embedding.py | 78 +++++++++ dev/pytest/pytest_model_runtime.sh | 3 +- 19 files changed, 576 insertions(+), 1 deletion(-) create mode 100644 api/core/model_runtime/model_providers/mixedbread/__init__.py create mode 100644 api/core/model_runtime/model_providers/mixedbread/_assets/icon_l_en.png create mode 100644 api/core/model_runtime/model_providers/mixedbread/_assets/icon_s_en.png create mode 100644 api/core/model_runtime/model_providers/mixedbread/mixedbread.py create mode 100644 api/core/model_runtime/model_providers/mixedbread/mixedbread.yaml create mode 100644 api/core/model_runtime/model_providers/mixedbread/rerank/__init__.py create mode 100644 api/core/model_runtime/model_providers/mixedbread/rerank/mxbai-rerank-large-v1-en.yaml create mode 100644 api/core/model_runtime/model_providers/mixedbread/rerank/rerank.py create mode 100644 api/core/model_runtime/model_providers/mixedbread/text_embedding/__init__.py create mode 100644 api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-2d-large-v1-en.yaml create mode 100644 api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-large-v1-en.yaml create mode 100644 api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py create mode 100644 api/tests/integration_tests/model_runtime/mixedbread/__init__.py create mode 100644 api/tests/integration_tests/model_runtime/mixedbread/test_provider.py create mode 100644 api/tests/integration_tests/model_runtime/mixedbread/test_rerank.py create mode 100644 api/tests/integration_tests/model_runtime/mixedbread/test_text_embedding.py diff --git a/api/core/model_runtime/model_providers/_position.yaml b/api/core/model_runtime/model_providers/_position.yaml index 1f5f64019a1663..79ebd007647c85 100644 --- a/api/core/model_runtime/model_providers/_position.yaml +++ b/api/core/model_runtime/model_providers/_position.yaml @@ -38,3 +38,4 @@ - perfxcloud - zhinao - fireworks +- mixedbread diff --git a/api/core/model_runtime/model_providers/mixedbread/__init__.py b/api/core/model_runtime/model_providers/mixedbread/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/mixedbread/_assets/icon_l_en.png b/api/core/model_runtime/model_providers/mixedbread/_assets/icon_l_en.png new file mode 100644 index 0000000000000000000000000000000000000000..2027611bd5e8b4c7d06a5d00e515a0db70e67a17 GIT binary patch literal 123637 zcmb5W1yo#Hwlxd{hv4o`;T8z)?pj!I3YXv#+}+*XJxFl(Ai-S{2o{{+@)x(eU%%Jy z8@cZv;}lgKYMisrUNYyJYj47p6{X%H5+FiAK)jWa7FUIUKrn=WfW(7`1)o`(ohSys zK{}~Qi9*y&5*>hl5eI3>m?WJgjm4%hTF$%H=ngTLNTxEE& zpdhNmuL3Wd9BYKMqhC!EmQTJuDhkrxtT|0xWSt5)r8?z25LNF>LqLChQ<2x&&AZNJ z)iz_5#cl;%Fo?<~n`}`WfzBL&+Ortip#m2TTt)Ypm^tJGRVqlNmTA^0NkQ?*7;a*N z3Y%20C6@~utEH@pL-B2IR3vYDPfL)yu}OUxxCrdeRxPHRJ;yCBX|7vzSE|^UsE=To zD$2)YxNWcy;voL7#||CjKH~H)ga1(J{HQWbE4(ZvuAc}@2SA(p&{YI|7vH4PJOZww z_?~e-!NUKm5>z_H*@-#qO`Q&b5R^+rBw7E5+kAvEC}OE5AmLNgvAu`ch{;1)AfwN`SnzDQ5az-vFwv-#E~x?@vjZh!e>o`f|#N&%s5I= zGL&a0#hp{rDQ%F55xK(o#WD!`U*yVzHLBn;Y=3WdGOKVp>=}z2pdzM))SP?s@N^^<-Kbru2dn!rr(}$m5dY$yBSw&ROucI~6&-RJd z@5Zqo9t~d3+c!#v#X$3mZuMYymaaBGvnA#I1ZfrN6wV8@74LV$F;)1h;t>D;dfvs? z=RGnn%e7WvJBT=61$ZM|J>hm%-gGw!VHUBA!X#H4{;_W>g;RCoeZ+#hwl`NpN^ z?z~7zlfZ(uE1Vjw``!NeiutSt?GZU72H8Ije{lr)OfK7~1y7s{^{HiZBAn0~_uTLL?u-BIs~h6l0Ef!hK*lj6Hqv^o%%pJIh) zXAd3u%AVm8eqZJ|lZXGr%hlODYi#S$J0@N4nAOuI_wd0q=}8V?@cTwbEgzpx zbosFMf%)jk-9dSM1ccD(8UjH5*FF8vdd8jbX|Se-kQ(K^a^QS1u6LD3$n{5kZh`CcW;s?Kk6-R~&$R!Znyd0ru>D84U{LwF#UTLtE>CUQL$Z{N-?1Bvv7(j*lsq4nrcHez=UJyL-ic^Z z>9+e&GzmD&x0@pcD$rF&2PgW8ZLLospBQnz|6VEfaH$^}uz8NHAjs*ox$7+?cpHAh z^+`{EObUuh_^%_56uLw)n-`O%Xj8;mD9pCtGzn8>NW3G=7iJeGFYfnyt)ctcw}f0( z6Q~(B)c=}~e;M$9BphL8w6m2N=hmKgl(yX0pInUdYIP8;1&uxVGS@i$d7GB_ceSnc zmIo3VTt&CC$M2&WG6-4I&G~D+uosdN8xd-jd_9(_jU12@{_FC>D+o!cskx9a<$H$7 zhA>Gbg3yqC$+akkzC7PkjUO;N0I{0=-Ef=825(+?A0+|0)#7c`@U;Q8!ESowP^4ma z*2;^YRdYZ2>Hf=6Au;xY{P-i?M~SZ!ShpLtw2@6rHXP$*a^4>NhZ!6*{i1O98j(m@ zrw&4XI{qNY1=5>F4l z*2!WEhw#ZTJBQYq`BcDW3$y9QnR^T#gUmmRN?BW*r=z&ISQF=c4GE6T#)$D9ohKrK zaJV^g6%lrET|^Sj+vQlBl}1Y?S%!1vZO5)W*55*DczoU2TDs`I5a(iOLMT2sea8un zCn)05UPlQ*lksHDEdAp}<@}}DP%M_>y+>kYx+ydr9#8kbSE+Jh{w5`NNP`oRK0kwA zT={{e3;j>-=V->|GIflNnbBv_Cajc>NxG*ZPrgTpeyTVCI0fFT z_{%B%G!d{w;=k92W73W#7&}v;QLsAw{E7iE5C|gf3U%*R(o*(p!yHp87&xIfu++)>%HjJcRdn}_U5T1!`+H_85Iw8WqmM$!t$_^SN8GCU7q zQr=Bp2x~4Kq^`kNPM4R%u2hb6MJ<~`nNR6&hY8QbS9ZKv=vSrp9bo*Usm*P(7_?}?Qo^Sn6eHCN_pi1*%G zy1f5xXN5bZhY^<9lYi$e)wsX)c_fY@rM7Q120{Arc7x$wp78p1`+XlOt2`7{+-K;D zhAPmh@J9cccOI+5c%$X)bBEn}8*W3@Tden^|HyAo+h%k)vizsyjHDRyvW}c|3@8^Q z3fkKw^zJi7GTwo`TbCB62NXDy8pM4p{pO-~@{GkAy^)+%iCYi7Q+|$1LU}6-y0O zXT@;EzQ|);DLc&rpUox?VJcmWjW~+Df^9jmYOU%7DEpakt+k@R)O36wVIz^%QX>0$ zvy7aJ2UPmw*gs}Lz*T*&qUDyAF_>hpZ>OyRcK4|UT_p1nBRGeHpGI89!7G&=Wd7Y! zAj8v_k58g>4m=mFiV(7d9N&g^75S8|9-dw@k8I5z3NX1nul9Yn_vsWP=J!d>C6y*9 z(<%I@DR4zc%lpc^33ThJ=u=8NwnFkZk%P%`#$wkp)$J7_656BX)^P=As;-=b+zGd6ZGhVIc(e6_Z zsu*^5sobY-B+&29VA1J5G43|iLA|2YJ_%%fjuXU2 z$}i=UFQ{V}lvMJmq|l}ZB3ev{2A&@obOgDDbaDl%2Tm{4r9YKYI$$;zZ0*cnz9aCq z;@iQLV07s2!6+>hC4Z*;pZV~=&E#LG=#Pw^2L%l2acf#BDV;anY0a;^Yk23Kh)~wO ztYDSU%kHRC`7rUO;+@>8X#K2okG&|)_YPj|G+B(je#C`PHEqao8t@(bq2aS)I^?n+ zLCSsiOnqLk3J?}tpuw5I&Crg5wZgo2p9wOxF-k6p{2$0w@}+|}_vMz{J?YFx0OM55 zm>%1*jN=KrrT>Y-{{hOX<}~_NI$a?KEcow`H99TRu)jLQbL8os)u&W@6>3)o80EHY?SjjuxU%TG?pi zIVxa4$z%F`+#zZ7b?r^+bIgH=Gs~%hK+X$0SuA2jFo8~9x)=z zI_oK?$V)!piz{K`y=CCECPF~6UG=G|sF1}1>^sH@=L&k89`sJLyDA^XUrrXo3ApFv z!+X$7PRUC(+TKP-Xq4MH8do2=pyM5L!vX6`K#6Bb$$w-kWNz>~HL6N6S(KMahvrH5 z>M`$1#W;T9G<~&GelpswbYj5M(Q~?bZ)F%NiWh$a-B^6NdO!aluvO4f`)mH?NiXEp zr+Rdu^T^n65lO6x%W3B3=br|2F;p{5(aOro(qf#vE7cazvo6s`0jVIWTnH?3q2rW@ zOZCt?*PG9#45y#r`b|6O8NZVKkq?7eP&7CP%!IUk!sgUI#Mr?)QsUXI`?}SxW%n`L z3h2Z-F-ti)lzfjg$LUZ4YTzeVcCysw$?){adDrsU>>laO;g4+U?+&k)8=-;({6v;| z3KfS9>PC*_+ykkSEM$LxYPcesw7#7QWGM{k9jy*7rVfN)oW4_;98|@b)F$@YBNeAA z{JxV)?zkvduT0Ty`uuoz(vUh@8&3UilKroY^>42QpAp7_Qtyd?8e;fXhwM_UX)2RJ z#AMZK)u$RL>Ea!ZR>&-uLBt54gfBCLm5BnY_rc+l(Ye<~>;6-$zAejewd{fN>+@nJgJxFfC@ zUNo#A_lwB<{QOBm#IVV2?jQZZ4|+xz%w8@1*4F6{b5zCZ9t|hu$$AdwA+fWLw3f{@ApN7l$nwuyRXPOhzOOtY!YQOdsv9F&7V;Ma z+xAzPt)mPv)PId_41NTo!Cv;DF3DL%LSk@F4Ty=P$m0*@qBuG!O=#GyhRISebc{96}hB3u5NQP+lpE z#T1i*k|d-zbQ$Qw1ILg(2KQUzRyp)*WZ1yiuem~-O?f*C;dB~sP`kUkvuOBTS6c|g z7;Jr!(4vU3bst=pBKus#O5P~uV1hYUM00*-H4G{cI9_xE!L<@9{JyBTp8^GC{E83`3_ zHpN4w6*eiE#L!W-4(wD48K9J z^7=Br-EfW^8S2B03nw(Y4@D znFYQe7%>aL!)t#>{J@i6HD|>$7Ck;=TM31;V{dDa(j093Hvx?Q8o*aUMmSVQSao?` zO>;|DW*KoPvT6cVLxs?+{jLpC5|=?*TWUWML;!zDrOy5kSHeR*i`?2ZbSy00bx6zU z@1mi>1i35nkcLO2rr*_-pwe6;9FrRd7{tPu)g+=dX{|vO$Zl>SGh1bg4D)ZL>;GlW z5nk{dUF_2ZUWhBjM&PJ0&Q_(Q`i9s*2|cjjbQ2yrqtSqgBe0FSf=FlWNAs##VE%>)H-;!qV00 zx)^B<*F*;JhA@}=oE+rTk`X9jdztGM-2Mp5uo)LKp7*}N$!~+eP^X#() zKV7-w`rw^E&>I%V*>hpNr5?o0_i|xQs-uspJyWRmj;wdy%eRRWAZ8~3Ma-(89=)w%qWv+1s&Gb7}!a;z^z#@vI|z+V%= zT=j4NG(>Q@CFX~DwIQIH)8ZElf=e|My`y&?iz;v-J-uFcmcc1UB~zU8$4F7qe#V`p z4x>F{YuiY4Q~|uQov>IxtsQijcr)5NB`|#uWbUexPo{@TPEJ-+zV?;AQ+AD~-UyLS zW41<6=SzqLY2UP@R7ay*>(|8)IPzsPz5nybsQwkV%$ih*F*z;;Uf=-^eizYUG6;sG zLqy_8Gl%}#Q~{x2r)NeZdig+qoLODf*CctTAM=s_rqJ2KnEz&V6bB6YJM8}F7Hf{B z{>4m8OtC->jn?svUB`BTWiD;c>E`2LG{uu%#Bp)Eo)9@y#He2`jwplSeLohI0iRwv zp}swY-zuYg$n2cR+xhc=mu0~OheW_>+G^!gbloKEdC4&*d%~TS30TF65 zL};D7w;0YTPfl26RO2I7aWSR#MiGIg&v--17bCmB!w@|%;3YH9=v56bSQ+4cE~0K- zFGvB|EcZ4u6~ReO{MKAond;OPGQEpP2SjC zLtwS}DBk>ZTfguI`Wc*e!@w4g&O?ahxaQk#ar{mJ6tvDLkOp1lNZ&<}9GgfBQw+a$ z`oJ~z@rZ0M-MiO9z?q7bXz@rvu@QBAOMu;ap-P`Pk&OoXwPXkEF^Vs5%c17PMYL&| zc_P>>)fut)=(5OED-RKbM$-yGga?l4OZn6%wKz^>aH@;Oz9C5bbK0Ux1?PZ-aSqGk z?SYQW#=ZoVIF)t{;p~6{_EIm*f@r%$%70{aDX;kNn4% z(X%Tr1W64kc*Ur$GmRT*`I**LedlAEU*N{H#*?r65ra0?Wx5!7s&0@yObtkwTGFUX zgN?x+PbF%I!ybwyQAfPKwB&9;5zJ=)2%Wd&$fv60;AqnWOgcYs41$-{y;;mx# z2qPSF+Fd;084m^_-t^ZBXFbAxo?EWQjMrX)h){TSzj*j8s3WeL;G8L-y1Ey;D#Z|W ztgwg2ohj=}idq5S8YO5di{0lUV<=&R_6#pz*c5k64tt3x;Ui9~_kLk=Oo!AP{~k(@Hhc=$JmN+oPkxFx zktGGa(c*_){0){j$}>cD!5Lt3JiVe59Fz?(Zo=V&&*w10wkE==Pq7Bs+6p94L%w^L zvy1n+h!ZREH5_yhu7U$bNPY#^?7<89?Vg7;lYq#|p~1xVQQ1a2Ow>VXP*3B?Who0_ z)V4-m2NqA?aJQh!MAjXLZ_cMMT4(cLCPjRWd&1Riw((#Y-zoj_=Y+*03QkxXg7KLg zOgWY+<3^#@K;GTDXkG%<4l88^*T@Rdb#x}Z-bW~KdF)}#ubz6Hd>h65SP8tJm>lS1 zeBv&y##bKjoTv|**T%W|qIH`F@`AxM!argHm@;v$v|u=VT8G|PKg8{1Af$ekxBP?P zlf-09%hg1EZE*W$DH7jMrDqmO#g5Gcs%5k1N7?_R`s|gmOAH0h(?b656e!Sr43rIT zmQ?pCv`lY!3)XLv+5!RsawQdqv|bVD4~Mh>m|_*@I8bnoTv(<%Np@Ohh>brLOTd23 z31QUlz1CzcatEn*Yhha}Fd*08?p(0hKG0ZboS_%Wg#z>t88l#C29k^vs<_0^kJRhK zikV)x!1)@J!FJfhliQ4HkK$BbYmfl88y8V~JQ-9i(rtT^CRoVks+`#kQilTml_u$^ zASB3C(9j2w9QOJU7ZipAA9ER@40#bi-s5yiDI{2+_w;Q%!j86|9exb0Tg>|M9n*(k zEYvVs6!QDZY&vh6cehy0)4PSwFrwvB{JDDFA%Jx2;ayxwCOl?4jK#N+{%oo4Qr2y* zV_G-w+OB`>gjIV#puLCSP-4aY@QHNn<1Im4}|6YPBL-0 zp8{Ci;KkQ;A+tisHY3UR4jEDVv$AIKIre=azf9|l z5l0>Fz@Q9SIztL>Rl11s!i?;;fwO9(JvskL{g-xPupIg&=)f_f2X(q_eMS_(GdTb| z#g*gQ&~bGT6GRBs7`$ssi;F-0jngWg1hC^jhwOu4y z+`F*dk+)8roc2vS6l3Ykq4b!WgjNC?W=OD8e$?*!Hd7v~YW?U1nf=OdLHu*4LzpS)zYBS3}X92sxc zoTw;>*Xs4HKtjYkm+(*i3%=jK#JubxkXID>VHDs?ISx=Ep`kM<-cuMfH&8&;E<2s#q5`cW=U zl;)!CS0j6HgM9uu_PvoFQiN8zcE+^ULT)=FKgTPg3FBb+fl7GL2p4bTGF3l(@G8ap zS&|7&PfxEXwqVb<+k86%)c@et3HGjQX7?S5%iPPeT0++4_AvaqK3?5E!==-X<5A!`0SjNcYn&X`uK5>7 z;k!e^xEo8i+`nsgf~7*RlAGW|ezfC%?rsjljgk7P*|A1xgvU1hP#i6&$%Tb+gxI!ygWRDh#k{fMHZp0bp!#)vOq^Q?rZI z39(FGs2y7+lG{-1c~>@3R!V^{abLZYl!&aYCFu#j)Y~JYs6Z`wbeTj)@@5>!(nGN@ zelBeo6~Y-?aG;g*@$JR0$ar%KM%(BH#!cg*a5GK@N!OdWoInslOJR+eoxD}E*=63a zb*_TKGUHzYc8454=k9v}ws_k1&ySp^j zPG}q^9V{CwtBhEk_05D14q(1PRqsL*cPev!g~~U zay|5zZK&d@^ZG4}cm(?P=S)sD%>?FGhd9m1rtF7Yb1PX58E>p`e-T|tYu^Eu?35Ym zWil?3N37mq*9D0Q@~COOo1Ez~Pv)v^v~CG;VF2u4rx0p1h*%p`fne9Gyp!m%K+XrU z#V+cRC;+#lW-m46pOMM~yuZv&R)DCBKqI!K!E0wG1OD=f5=f>drNP7m%fH-T3gJq! z{knfu@xceBQc;~E;)m(!TI(gDva^6 z{iuq&%7AQ-tc7?OT$bY}SpjL4??v~aQ+o^sZ*bXgDgUl8>wisq+n8wfzXwyx7UQn> z&Pz^Zd5P7J43o$7777`(e<-b1=?^6BuibJu=ga_HrtQ>!|17P>jA1s2n zWZG048*k5Cxf<=3;7Z#SS;BSjB3r|mG6{wsbrq9^E47b-n+aiqSFjJqh)P_A*jgla zGt}w{bIT*mm`=oF<(6EwquwBc7P-cIu@Ka1{my%`zQ2?1wq&5Pk}h^E!x`1%4c7fB zPsIgSj+aslg?}Hpo^Z2U>2^l!zSSE?blBPF54Ti#0sc;dK&()jTRHqb>BK6*jK+Ny zd~u$JEjVh4^mFe@9bh!^84-~%AFD|V|F^z7U%l6sI|#e)Q0H61*{!ru5+5?6ZYlY( z@mB#c8TtqgtGtW)F#&zwMvN&G7DmSfjTwfht=)(6zgNI!<(Xy z@B^cv1RPs(+k?k}FIGlI+ZJMXxOEPe&~{;ihO#$wRkZ(asrZwYOr07BGa0 zkKi(l*#R$_+A60p7%W?2M9PcNV154@Zv4Z03$tKE9B<_QradgxS+(Ks9` z`RL4|aEQ;l9MMR7QZ=c9rZ+Yam%)W&m6$Z<_9N|>P60Pvgr<;Qt0QRSJJ1~}xb-C?ylVpo*kdF1%yVdX&jV;;PUnGgoJ~@|&4Q>VEKSllZxwV3sC^k4L7bG?G?q_k`5qgWdqF)x4B8%w=P(AtBAp2E0;XvOP*tT}tS% z1nN<^Z1NRH8=0BsS9V1S?@$#*@2dt1QjrZAWr9z17vH#zqj`~$K7yObDC+W};AwJ$ z9{NsFw_@o+&UnoambIwIwwzJ?FD@?5;h8_Wg77!l-PbbqrO5=Llw&-l7>$h>fU`wH z>rgL@UYfb#rrJM;&oiGGoBNeFOmEw_RvB5MP5w3@EunmrtE$MoVAcjUwh)Dv@F06mbC#i{;~0!Ilb!PPY^inLDGP zmEFx$sr|}3mEG+hdp9%(m~E8BCyY=vlMel9=}nMA!(S3GAr}v3Q#0St>t*@7v5XTGt5{LfqP0D`;e(mv|yfhKQ9LH4V>4K3{^X6Valn=5iv;3|1H>!lmQ?SXr7_KGhVg%- zCR?;_Wg!YUh4VE5H#rUJoL7X}?F!H!1_ z`{TTT;RY7!LjWt`q$)K2?@|q>D!J@fv3Yb9Hv$mc+^RHIgV(g>_Y*1$)dC_%N4JJH zk0=$mV-qX$4-v+a*;=b=*?T{Cxmw`UU^ILH271f&8f{v-5>Ck)ywKxF%)y@OVxjL= z=oh>CFaXT976K*S9F5enTIW24zvYV~QU~Oj7Q`6ks;ZsmlMV6vF-&ZkQKtljm$!8OPky2r@YG0m@eFrcxk~o5)p*s)GwH+8%Q|#zmv7W z32s^sD(m9`GY{$xLzn!qjU})lY2_h;`g{aiUDt>f5s|vwVFjprjFVi(864;u$ctE< zKKngcRrm1FeU@F%6`lLbdv{2xl<+Hy$yz3@_6FG-GJi(U?1JU|@C9y8Nu&Nx64s&j zZ)9~@S|76Qn(RqZkk^PPiIMBIX{~HKaVZs})TU`HefGa}hdzu_<-y4s1CP(d{M zOCgODmVz?wJd>dH)FMN3&ofEr2#?=;4&90@9#odrjh+C2@#}x}Xa2Vg`p;$a{m{Ttaky7b2S2bM*I zb#r5tlm&oyvia~}xjTfRCn3nGu~?T$+j5srtTMno^+IgnyI@^|3u6dlV13eTC0ooe zcM(Ircr&*n)LRDX`wM?1B5&3UhoOH6Cj(@xFvA;;F$N#>^?fv=8hJj!A3xI*&>~uX z|8lkEfb#fvU2U)f$h;EvBEzpXiTsi{F4J?|=$SIWZI zzmklZ5zOn1aJMpmmYO|z+_A>#xynghs7Yr%b9;I#S<);61jZWcH^P*31m=PaeLwsqbUUyCPGUJ%Ay@E;3Slp@vaF07=J;8uv=fXU z28#TvGJot{y`u0x+op5CA{G_@gvfc&P#H@eO}c5+egI_hjFTPQoL?CBFurBRoUuuz zwgXM7smOeuzVdIY4j}npt@TE`SoVDIqx_)*R5O=QNZ%aM>earv{5cZZ`vrDyMQdkY zlVlXNF)NKmMC+2`N0?n@t+gwkEKXf*iSdoiSoE1Lu(IDWBu$?9pvE@X%?HdV8*cQi zQSI6tj?ENmwBYdQ`O(xMkM)n86~Z?N@}j^S-c1HPWI=Y{XD@@B$nlD-i(kXVaGwv8 zRQ^scB=fhSj8g?=*#yQTU8H;{@F8xTgVu_>HwnFlp63P7Wp9uRYNJcqO5M@AwQy@q zd5%JBXsq2u^O`P}SG&qdZ&$crVB>rtS!)+g9@0a>^h^<+;3T~RFltZWKz?iSv{d_; zQvXTVL3YuZx$&ghi<%=@;|pR+B}$W=f$1Hh6#2>T1$UDPe~g8Mg{OVUaRB4vftG4` zLRBc}?t5IliCwjP{>9xxVRY)-`}J>1D4)LX`^S{yiT!28>9?Q#=s`ua(ETIZxIe-+yw;7H|V;=y`cygNF|<$ z?eLY=5x;CF*<{Y;oTMO)>)EFZ3EGFU_* zWDOeT_8ve|(wO8G%NE`(Leq}0sYFr2p(=k7Lp(^KhSVNXr--`~a3R8g!7#JhT0i%l zTY6s*lanG!s%TC}>^B#>?B6Z^1z)WEHJ0Z-{uVDE#H|#<#djeHYJEF?&otO)YhI+x zY63UEWS2t9`nCcE(WLYHR|D&q@|!p=D{p0=*4+^(PJzjbw6B5CDA1v6$n}cRj2ZY< z3g=Xy?+sePHK6F`ssK*n@tJ-`1Al-_XKhy5lX>6D!Yc7>bM&_gL+&c|rp9R7&*u3X zX>!=jGSzdNjH@~Px8I&xu63_A;$4To$o*y4I3lP*<*&S@Wl-Y{B*DP|m7xg<^QOi| z{-6+SxgUlw*4SLMGHHLeZQ!pWw-tH-0|M-q)2TsNawiquT1;2gF4R>J-Uuszb$uFE z0(wT`Tf`WJC|7=9how4$15rvSubXm|B{GRdd%U9~ZF#+X)_s$9UdaKB(HLj8>31W0 zH13bU8R^%Qzf_nqJQ;CZG0cxXfd=h4>pBiD>X^Qb$DNuX6?^rzY(FKcJ~X}7dfCvNXbK+-_#|e zE)~T_rqr=$&KC-Y{2x^1DbgdKY79B`0tpQ%P&qBM?ehug#_GbB?-am=U6|O zbi*|t=bH52JV`SEzl-gmAFT3^Q`A0qeZm@c?H~78Zv}TB-p7|F%AgkIQTp$o)qO-2 zguK~O0oakSH=YU91T*;SD?%ZNm95;JE~B~^mEL`jFM*k9_>0Fadxt8 zQNQ1i`JmPqeL6fkx&b$*G-|+m4C3UWaDaPwF*j9e(~Cp-Q_-T^o!h#bi)f4YD75Ei z5@%A|v0F>%JJy#QA-CCIeC-ELK^$3E=55I6afFO|Bxw6j+SDwLZs1)NpAJb?_P!t7 z68y?V+GL=cut(5FK5T%(sWPkXE!6oirWrRo)DYP4YYpFF%!{CgzlHKRfP0oB9S;6i z+hK)qd)I;f_+|%6gvRGQ)}>LGG0oy%Gs!6~Rcj$KRd@<#x?SH*XqEshpV$pI4jZS0 zmK&i55L5V*=++t)dl`(-L_7eIhlVgnmd5zKsy~{5nP=^Zz3Crm*%|$@gaW?LJ8i?^ z{X_Hw_ntoNDTy+{ZGj-pd6CthK;S)fz6#zG2jxa%NpuA^kKUzR_~LOQs`y03|18`6 zJ1AK_E_O&HBqTW4(0CxDNhAGb(~}6)^z@jf-mv45I)=Fg<}=NN!HjY_81WTE%^y-_ z7sc-Eq`C9kk3j4zE>e0+vsIRt%fL$ANR5qiV}wZxV^zBzuW-H{rp>pHY?$W`QWEJ- z311X#5o+F)^-#HQ;%7mL|8bn1MDD>#PAV?65pzZ}U_0PyA|o>zu{)kRP6YvWV`NB= z+`b@d>igS+{Uf#MI=b8Jw~oF_4)4^H)OPeZcGkQu_%LX7!^P6`y&r2($m{g@V>&~^ zzZ{pn6l+uhJBej`e5v9@q4s^>OrK6lb@QRFjtvb-w&sfalpO#<*Z5#vqfnx2Lko&? z3u7dcI$@rkJ?1XGz{Q#VhYt7pD2-_se_;)cC!#&IpDM2X-6*Gms}T+B49bA@U~XQFeGJIGHvwwj z_GtR$d`W_zxe&Ki;A4ASih7QHRqL)47K4Y|9oU-lHK6)prN&Q(#Ns8e!D}O3}$prBR{+#O&`L{+A^^*(c#xBrixnZ}{73A<<0{W4TK6TJD$uBBhxUoC(y;6DK zkR&G|I_c1J`kApt#RM6}^$*87zj@l4AX`k%Od1QumgvYfbJDi zmyokgI(6=#?+DvIT|S<<>OUHgedISIiCq+nYF@B388?*`jzh^FSrlDtyxfZN9 zckZ!8l>W`oXC1tdfd_kexn2+vKn?B{UDbi*QtI;Pj^Q}{B*=A^7Z3(#nESyvy9a{- zaC|tUVRK`22~P0c?2r0R`NKhRZ_N8sq4<{@WoIu~oCcJpsR8fT0=2G&To#xcRE66+ z?ry-}&E)(PH;AVq>g=oOnEUY*oxPmq2KGGvhQo%Er+&>L#Qjb$Up{zuXLsX_;Q9ro zH7C>;5@>!_%|?tsr3i78=>H}6L-_M=nX4B3C7qem0v;kvBs;jRH}K=$VqMZH$OR+J zH<1(rq4#_G2aqQ$xP?2M~fhx z?X@G(nt0Q~fsS)jr8hVm#P9gxcKXyn+SB#bI&8+3{5o=wl5vr~)od{B6khL#xQDP; zC8h*+)?y2KJv4~v{NYJ1FRrPR6Y>oe1Qdt)Bx7%T27p^BIQ7%1I{%_#Uq08E%D$N1 zHuaM(V^ixQZL6frOG>o`pPUxr6=gK3XOs z>Ynji<+HCaa`fXh!^DY?A|a!dG5g7w%5^p7>{R%oj$5EUsTXtBUdf)igP9(sGsgms2<=#GSF zzyhZ+z8d`8^SuiR)hTbFcmS+u375Y1WOI6(cKQ@C*+p6q{7%_FeBP6Wgcl0L-HCN+ zXN@%#@4iD^xlohJs$R-LB(>i9N||6DjK^M2ftkRLf4RTXWG$44t*R#dm*(#PWilMY z1YR2c0FmK=Q!{n>ku)pfW7VfzZn z*M5L*3QSx$1pe34=rhBS_wd`3V=1E*(>vHSG^x{wh!#8f{SbG`?;bu}WlnDi^fi!9 zj20za9eGRps^4?cEP@R6hxSa_*NzwAhjzucm2==dWrO23HyF30Sl=pKgbdG5F?o3ys&VN_9klm?19V#$u#K(vNdKL=rvbw zPlV*g9uGui8TalCbwtVMZxzBaT*ZB$rC?07_?9LAr8>};P{kEWZhrb}rlllqa)7Gva#gx=J?`S)m{{0s~-MGDYj=y3r zFsEgOPv!9b4aMnj+Si6_xJw%Y6Zt*)Gj6!!l%s9=TP-Kd3uJ~x8XkMcsg#G0=-+oX zDEB5iZ-zaYB^$yNbKdNH7;~~Ee*6SzM}LBwWys%cEjAc&0=nv?NOjdj`;WI7GMSfg495|>d)(lZ_H$3&Z7ZeJsmzI!3x0f0`+ip;FUqmR8)1)T zB3RNhfwFJ_u$RsFYqot>*yr}s}@rm@jD8-M5d1%Lytlk7QxwiR2WRT zcb92?+7O?_Y0LcGle7GWf5`|tWVk!?E?=|1wGzN;8{-oTdSvNSUHCOQBf*Oz@LcSg zeJT&=>j>GKx_sEg)uCroRUHDQ2~Q;F-W!CS4O$p$0Y9dnFHZDed4#6+b9p#l3MJb; z8fJcR<>OttdhaE7O)lR37PV5{3cRWqs8wUiC@<#lIak)(7l+nCWS_3mOO)g%K^kkWoEMQMxlT=p#$-!^T zApg4LI-M4FVn^c!RT?6}+?Y01ZMZaBsLy+9+EQFe(U){M`dwi@F;e%xTrB#zkP)Al zKIz7N+4*uBUL41uG6OpEE=&T*ZG^#~6GBCRyMwZ6f)iUA!}l zYldERSaQSWq|j)9iASY8HaSqX3a5hjF(cNat-mjNwaAh;bh%wf=hhM9^LaQO?J(B4 ze|s07mH32=-s4Rdjt~4}`Pp$cAoCHD{TXd)vEW?-h7?RyjZraFit24Nc7Mdzc``S` zM;UbZF}{KiP7hiRwCG5*yc-_S5zXV{mEpA%e;$6uf1C>u!!&&(SclxS=XB=wAWBxxckpJ-M<2iG-G zDmH;a$SmDVt!|CJB%kmdACWH@#k*1H1j!U8`G!@;%%_#l3k%+HhusoyQAR8{Dqw^Z zz$esv5~?`d+To>}YUGTQt7;4@NR1K+HwM0=y=05+i!{Ed{D#3M*pBcujm?CW@z==G zjK@B%hD9_WS=?Eg66d;Ndv~KWUi#!y!4~bQZNn)i?Gyq8Zb<0S2d(Z{ zyC-gi(=Q$S>K5M>3BEjT{KKy)JKIQdd z(XOm~JTMb@(L+O!tLJ!2V_B(pl!zNTEuPL^C%ZGd_x1}d8S1?=3Ate&q;Fm0jpjMz zU@R~0=!Plw0?P2ZFo{(`t65>JZLsQG5{w&v&9b5w~r7YXX9i3e)+jr}o=^En+6rUG}=0Oq2`m{gj(Hs!9gA~_QOLIC~M z2ZSyjYe*!Pffw-y!8m@KaykU|4S#=sm~y&KlkzV#LVlZbO~kRL8uG-j-io0Yk}g1* zY*V9e!>3{|t>mMFCGFhlTWKMiJ=O;`+_S%p0<5r7R-WGNe8G6$fxj5>`|=Te5Y5A$ zA4X6QbFZvoRS{OeZL4EUvo07uHaHb>=30?v^Eu`n>f*w}^TF>$1AD7nzlUWRh(#Ke zq8F&5ZbYAtLBKHMC+4hANC!o}BSlG?Sz3^(;ed4IH26&)<(J&B2S}PMed*>N@cIUV zZ(hJ2(N&53=kc?ao|cyQLrY%I{R;*G9@@Gs3`@$HR zuvpH?4Q5?cH;H9q%$>Nq%)~x9$w!RFPI`Xk#lA!ug~Q0*JZ6c9TJcZjk`}puywTGa zvW@M5gm6F=aqeD3+>(iky-3b`qI5}PK7O!U8s4&UKgP)6Ns%W>d~?i|5bA{YzIP1s z=IQ|KG`U?cQ%1NIeEfjSLOLYv3uj~OU(aA(^m$?)kVwV&fpHgEw}_2 z+}%lV_u%gC?(R;4ySqzZ2oN9y8Qk6d&3E6uwX5dm)TuLj_U>M*yBBrqtKnW}M<{Gz z`Xk?Yrgqnh-0e#Q0GV*h8>YeI3-x7REWNikH9P%BFF8Ki5 z4szP<5e?YTXVu_eOT>K<`o}7~weF`ekdleQk1fjt9bbN`R*Xg}MDae6$YFh~Q8-9; z&$U7hf6-;BJ&An=u46KS|0AH-xb3fhqHV`^;F1CNRy!QI zs`3iMBMbkCyVJ5Df599&V|z4#(XIR!EL?vq{-jEGJZ%zOF3&L#(DI2)sU-?sF^73j zJt*3nF^19W^w#JjH5iyCSB!RqkA?U+Gw85K11_OU&;Wd~@ z242}h<_LoFYMo0%S59gEPyO_2J=n`|{aF2tx+D1s4Gig%Qd`U;`Qa}^l&`az8i zh(~D7f!*rdFSeTXV`F0izaQ#ljF8D~jWFo0nVtEcqZLuC=I?^-TrP0hp$(rsX?rlP zIO|hW0TqR@3n6M@+YqH%^jahhN1qE<|}-l z^nn3PQ`&x=LAP+i4dhV2cEAa6{Kl8~&l!ZbN2gO~@GM@2Vbe1=P@Lq*u#2k$7?{2( z@q648ww)ck0(Q{FhTLMP@E?^)If;>0o=`y)fe?pvDt)2o-{n7mv}XfBnSlCphcC-3 zl-1e<|IlrK5oWU&Ht`IVvvqgVnF`lQs~dM(tAc@tbepdMBB*o*q0RmlYl+H5`%|`K z@U;~pJEV~aStiBz)Kj0|DjO5OQEG|Z<{f{R9G>2-~5P_KMK9Bh9F1BVHjMfbG zY2R{3(<0O}IgZ(n<;*!D0{_o@3(=K^R5&US$v(}(8==Jh9BfS z;;qc#hcJS01m}_#7p_%L>Hw=ms#h|>n_fo5Ln<6m5p5Vl7m>TLP-Yp5g*C~r!A=o^ zDXr^YzT(h~v57R-ji%@yuFoqvAR zbNZx`Kv4ooFc|Sp1Foi)mPIGASe5MhFoCYNM<;LIn81`~avBgAs+p$hIB<_}>4&`K zo9_wqj;Of7gw{_LR)h4FV58jRJGA*KoBS#mX6|LYV1};kvXzu^>&=xV^BAlZ`qSST zm2q1OMY}IVU;4uy+5*WF_dmQKNoGuj>WY~zFeSedO3Hs3eln{mcH_&SdiPzU@7`Zk zM<~OlyAPXwG=Ie=b@wr$z&gjraSmQa{^|!+D8`lpeT{*C^=AG1J1)qc5C4A^KBkyS z(B#=+nf2(SKHF!iEsy^q&`C6@PTeCP1Xx0=IUgyN&FfMi6;B%30qFowPWr);lnkL# z6q63Ro{-?i+_@YhAWO<1-qTsDe}y~1=X>ahgv;-CjNH@6=n(VkBy+QG@Z3J~SD*^qg<=R^FZOFw=& z6A{}4!7^J7NR#>%A6Hyua4G*6apL7phx9stf^+MA+z&d{{%i!uPr&(swDD%3fK%q^ z7?Rm1gzhiERHu^kB_lPOR;>e@-rhF2kWRoU6_WJQkada5+Nb)^?ABo?GPryYzO-)x zx0HKcUR_R>YJ(3)uVnd=!Nk-Vt_}mh9>879ykkcyE!XC#vv(k>jw7l_)`5vyo6E~v zt}*3>Ye@sw#Fh)fJqT*E-t0w*q5D+`M1nCm-1c4W`i~A;1jx{YXgn zS6M${LhEJNO#Ez7RKaDHtNWqwxWtulovs(5LgIcVDy)a|dgZSJC(_St17 z+9gzxwLd1Hd+d8&_~o5y!g|EP-^2sm(7y%0FtdQJmLj*6k>({x*d0?^G7W4ykx~#S zbw(eDm44@d8=zA1RTF_tvvI>km&X46a|W;qA@c*O5SzXrmy{ZcwlJ7)+$He|&?9g_ z|3Dz}+%Ir+@uXQwZYC?jOPV2)ANlf*-QzjwjjYv|;#|pMU9D&{eW6(+ybJjEH*NV~}N9t>AQ-tbm75dvde zg;gZ8Sowpo_`bqx4yiI5)s8OtXXeJd9n2jRj=KvLc%L7(Udpy3gP(=44!|Hj0@E++ z-}+KqX)f05ho=P|?7W_$1S&~anWkR3z!ra3(oN#KQ9!BnQ(j8iq@f&09T;XDDVLbY z<_V(p|Gp^|!XNv%Ey{Itz0qdS4C-#)c*N$AFe!zZ9T4C?^OE=c1qDwcIi=u*yH;Wt zxGuQ&?ZY(cIw9X$z7S1|=;~13#F14EhWmH5`G8>@$xfZvnTJ;1U6mRhRoDyK23VPi zxWI!}feCo2xx*>!vj1D2H=v|oQ>hCXBJ@iRbkyRk6kst#urPf_RPKdd6G6qgFV_Ps z$-q2G4;p?=CvSPva*)6Zz}f0M z?~80Nv8Kge+FfNmw^fD0waCqLBb*Q*d1$SD6cUO-NN^=48d=Jh&G5aIh5W=jaOR3C zeP4xO4G}`ttq2KodGL%GLm#HiVKB!JIVGVEjN#~N0wQl35f1i9wa~((u;k8ZpBCFy z;3MT9)2&5>OHey;E{!s1Ut#HTBpb66;JTzl>B7uSk+SPuP?P#1oN%6V^$}eiQhwQ6 zt+YXH>0CCgWeOH{?NpUy-N9b|g;v(0RLqosb2`g|q9yl6;G6!gDfIk`61Mj~KOQ|D zgZ|7?fz|X6BD3+&W6(EyS{UWS*mC_w^1Er~Qe%tW!jnmZs#lO2sb2TsnU=gTiUj(D zsi)h)mF|7-DU;P7j`{Wg6iY?=+{|FuNscm+;#0DwA9~|X*S51juRgoDE2A)Soz9?( zY@{~gvi`TDu72$1O5^aMPQ24M?jJ7JfbR|F?ZYPcnPvGHb*Kt!RL6TP?cv-HNNzSq zZ#6V6aAN4?4147pi2HU8Tsu+b|1|pX8|l&#FCJ}2NUG=o=FVml{#cw@!;2Q~#F&Ut z`{1S9<~R9i#a!sO0Z-SNKSBkyon4hLCG#^SS+zpJZ_Dd^cvD(z+h<1vs=FBk?>ow7 z@n+S=aleHgKf7F|pD7{|j8aI+^`S75TD>fHZI#~@rt zSsMDK472B}emudt3XQID%yq2&yHTlKb{+veY|Ya(HlYpsN3v)IQjKnF6@I8mX7PK{ zwn#wiwfoAw;^acO&)%C+UsD=w-N<@^t^O>V=hnl^g@m3(h|yYXN1N6s44)T~=XA8^ zcw9!xo5EthJ1$3sHKC(jTN^a+8;t!cfjr+I#?diN-1?2M%;;UUOSeYkiJMsc*tTGc zNAR0nHWb`WziKBf&%jsGQKOS{m;&k`3uN+O1H{gi=p~DZY=o?IfScZvJds|%1MicC>VpK~!xJxVrOALxk? zI)#MPfEQ`eFwp-aba}&NdXR0S=Z_V$U%7pPBx6Op#gnvIr&wpSyPI@kvsPuzAf*9y0dTF_d!U*DwwxyO`VNsHf5GLqC$+p`D zL4febL`Y$Z1#W4exWth^eR5Y5x<{aMPtwhOlsi%oxYKIB9=B;b~lPF>PFjdc`Gk`k_le6Z7t^jjx+$v5gAly`0~<$+HJ zoQM$rxf?-G2PgXrb=~@=*h(*57~ql|S(1t(!%|mQS97c2N$tQ3bBh*fS~6LYaIgCW z8S2~R_<$K>8>JM&>n!VtkO^v?1|e$Yv?dJX$%&doP|3SM;GP(mHU_!42#|ba05Ngo zB+u5KZrFjY8h#gQ-$-&GQya-|I`5B(3{OlMHx;`GF%)m*Bu~C4_ThX;tWyQko{kA5 zl1g}Ffi@dzC1daddpV6jwdJ3u#ZR%$#0h`RF{+!vD)W@dv2S^o8J_UoR!N?S&Mr!w z;Y$}qAbhUj)L z$dmti`szx}GUni5(qZiN*#d&hI;vX45J zwh1zLy6N6E`N3d-?J5z6)s9pN|Kv_s^B6&40d~B(^gzdP4K+1pTI!HpXs>Ee}V(JO}$5IR(Jr zgsCXzaQ}2P9p&=mhEcW7f)7tpGCunpxQY}jKH=oms4l&n5v0hC?n*d`MHKp?;^e1W zhledD1M$>HU1Rq38O57b@{ey0c?dku3)%rWPDP!0UL_wm6cB|L84bh%eu%7Kg_?}BfeUaev9N4D~xcs_EM%}m==!Px&E6s~Fd zb<^R|{c_tJ?Np??x>_shT8|;(BSjqQAq-VDi2=iK2S4hanY^$!TA1|#>&ZHu^$u<~ z--PDX^wHAGnE)3T*KX_v6q$H&9asW4pnjg~V$eCug8JsiI0Tz#+trgf*Q;U-Q-{c@ z*-W_MQPLk{d(}*8iiKy4KIB@>;n!fkp|Y#OaGjwr4X;PnM5hHaD(;ZBTJ z`Z*eddP#=4`~i%3-`034w5nZ?7Vs+TPS&w-&VjZdyt-a(+)!+cw@ROL zOGdgH3$1X1@kQ(2$H}PK5fVVskKwg!>f&zf|lA0I^T`offGR=*hUL;@e1WFKHcci1|ld7(sM&3zI5tr6$X8eT1BT#GL6#2(mwOo#od z`JY4q*`gD;%aSi-ogPEey=Q_wY3mITMbO8QkDMH+n1krUHtksXbB$=n=cCpJQHZkN z3km#Z!ViBrR!NpQD&fhRI1!(BI^nTR%oKH z042DP##e-f-&XaNi@0R(SP9;!%6QD)r)0U}wTX(LYO z)*e|x;4;inGcA%49IE%u<<{W1u^T6~S)#MttRxX-kY{s!Y2pMTBVi00P>qKSnGMA2 z_8WO3>x830FJg0zR4(ESy~86vo))=&fGbjPVct6r>iGjEZ@8Y{H0BF2`-5vzWMkao?5tF{}Pk{lDa2aKJrpb{Jo?863|Bvq+7ZKF^Ese zbh;VP8wYRM8$bD*v$#R@8SkuN-Xifc8n;qa12*W-WpP`#AtD4jsJ3|zJjbehuMi_r z&_w3UhHYf}^YMw51n(6oL-g0;bsW1tr0S%>B}fITLtm{F{~h1~VQ5p6JRss+@rio_ zH4miVFS9Wxy>EleS=6}jPk zA^w$@uV#6qs#jE}IP>LA_V!cPys*W;oy!Wgiev-c;cJ2(6GXlHEu6#Dby=&Ik)YqD zj(TG$>az`Rr=3O|zPRSvtOE;-Q**kG8@IE85V_9WQN^uryT28j~_Kl9InOWW5rR!_B}P3FNqhC$bY;c#qU zrYXu2zgFS^G{3+{WbG8EFXzb1I{}vz_eX;AP#0v&rTu{*ta{`F7q0ttE15)jvVLuA zg+n5achYRRXixOUfd`t$>5*nWHIK|a)xCCizinpRnLtJginz_7`iMd}Bb57xt>-(n z{>{L zmcUcLa?`=6Nnh6P;CG$|CO2c*iyJ!yqvGRZkqEj6q~&IL%wPtgd_;@98)iL!-;ai- zhkJwK!tkM~!ei9L*Qc`ZDvrTxftihqjSK>e4Q)P|?=R%C%&^h<-9e5z4S%)%$cf|F zU6YMze5@?F5$};$uaLZbVb=I@Z?xn~knr}mwWzl=(RBPA&~aScr9LNRzesd9!Gj3m0g@rC}9fo35MB*rvuaV2pIQJ)Jg zSI|>npw)f$T6U-BG0VYP_wqqAh}9~yb?GnUV|k-{ICpNU+sN zr^msL`6%AZVn0Y*Y805XF*ET?K)f8mH zEU$^za5NlgxvrmksO_^Zudom>k*^;_5ye@5DevX5m4wiL-+8X-! z%_FO}uSXb9_AT%(5kcU{DFWkhd2FFhP7G|`KoSCVp?q2&hOuyOG9vya3McZA^L4eS zSVR#)nEuKo~NP&*^%%R^|QoHL8_Z@8Ccs*Dt7N2LswE)~p{?f~28(J4;4f z6+%f*=bo#D`AW&QfGl(b&s-O7Y5jqATh-SY_ot_c!e=D8ze$4!0KHN;PXbyRMMi~& z?$B=gSeNnp3Lv?3(0lMv))ir^5&0I1wE^f}XABjkeGUOIP+xxuT9;UKk*fUEA8EvC zSUatQgn_7@UiszVaM<&iy>j$@;nxmN1md4d(;>nGsD>A#itSa%FGIYa$0wZ|cOgVcE^xwaq#oGs`=QMlz}QO*^95Jr8kV2z zP(Nbwz@z%8(|aAbK`l4l`w=wm*^@K9{j#BPbG&ccl@~HJ%RB|_C~sO@w^%Tn)J(58 zHO%M28O_>t~H0Ut_hcye%!gF^m?qWyd z`b_#!lFjhmO;G1SbrUi0=_L(_?qEeBJT3>Wvs+J7hWNLg7M6$Z0-N%cNPuhc{Bb z=)~>rjo{w(q)%m@7JQ&a$kO+dr9TZm{3BBQ??(jMj7PDA-V@WITH#1>*E^m-@du#j zSy);g5oSh`b=$2K!_R1h9BI?Xg>x=7=?Y6CCJj;qhk>WSv8xqy7NswWec-pZXaA5F zS(CZh6a>NiB4XBNnZWrZrG{p|kK|R!qcP;iqT3t&2hS|Mek3Vj_?_kqS;R_RrIjLO zARvvOxLIDC%*rNB!(~Xqvu^cwhz=y165prnht*5UtiN{rhq>i|-i)ui3#eAN$b9P* zWli0w0TTu=x9OL5af|fhkObM%>~@BEHxbtPJ-$$N?T=XUEAMIp!K;8+L+i~j^c(7* zEZ!$T&I4I=H}Cz&b^jq;#>bc=B`ji#nB~7Vdy#M>tqMtRK>uz5Vr7_)RR^y*(r`hpgw;}Q(RgP%lSp%t?~EK-q61nICFU1 zk=fxh0~E5Nkba*Xpl^FOs$)NVYZhYSHPCU6An+_KjjQ6`TkP!qTxfNyoxg8TO2PKC z8WHEnV|!`NhcuTsJJmUJXSswKd{I9j4(=rTP}c*uQV0E^=g|p;G$HL$ zh+gL&NZ)u=Sf_Q*@g>=Yc0rW44qcx#b{rOci(74#U}OwGf5Lpt zcUe?BHPw`ff4uGRQn9cq!cXQYshP7T`u&){UI+(={noIP!imDFFHhNMoJH9E9xrz? zn01x3~`qsebtAvXuakfQVwD*2w#FdQSBt zxZlVTgw(TAmB&I;!CM+5=ik*27pn(5s#h~Y-<+ba&c>(-oKU6vYwW;Ajh8g5SF-blJ*^#EFCUFlV)lCpoS=r{cy z>R!TTmDp+wgt&h|`NICs17>y~g|%#AcK#r0i#NIJggv6%#aV zb*}B4af1G#_%Xngi-CD%MMN%HqPi5GGtXCEb(-Gla5vhQn5O{8wEA}v;e5-iq_9X( zc3+YC6E@`#VldNm4@a#6c>*xvgxAUyd+8|8x@Lv~KgS6Q_#1iAF&+}vOVAUMtd}dy z>{bB#gXHn%OY3mToop|dpk=sGU2~`jMLWMa3IwWG z`&92&*L9%K#~$#!9YL^<%p&(xqR`hi!!9d^DjCN&Vl7k-S=lPLyZRNv8FZH|_bkQQ zdDM_+XS`%rsBkBh?ZP@wti^cX^DIVg(%S`>MF4si&q2d9yjYvT{O?40Zeev;$E~D+ zxGE$5&F9VIF6kPlWc6(>4ubo)H`@lhrAZnxzh?9`+OvP;LZl9Kbqy!-x`?z4^qNZ+ z^xhBD)pgDvx}xVj3Bc2~#ZLA0?v#(gv_Ax+C4D1*$YMGoriMmTPx1$wE}i+G_X}7}Y1!8IeHYC> zSsOq$dN%5^NY0zJ{FCqKPt?<4x3if0PJcAwcpUifuyJfx!PC13;aGd3b-)2pJye=5 zu3l%6?30k2)Q@1?(#BY2t|-{VI1*lmksj}xoH%Hg&^Voe_u|Q3O+Mq8@1!4e?gwRS zPN01oV3G)+sHGuhv`XOzD3H4xFlrcor6N5^PItyCCoc~&p-M`9&YG&O#c|QC}!RFbast%8+ zl!JsG0-Ip;zR?l?#+vE@(jP^B{)tCfHfRbqjfEVBd`RVIj$ocx;|_V!vMdDD3bO?6 znw6N|UvbTfNv%`^!NK4kO0EhV>arq=^d?vSn|cEm~z$ww6)Q_B}08PDHTNd+=ql`Bcl`1r}! z*eJ*Afdw~0OS*vrb5|U%7-|!tBi>N_jwX*@7$^ds1OL7g;>Ps+EbAqeYF{YcaeJl5 z6pm_m2G9~|9IPR!j$Lm$*WGUQ4A?;Tm``8m$*xW@wV0ccCye$!&An(}z`g`jr$3VU zoEi9h2QI(8Ax>#IrVhGiu`?QdDV}^1oFCN!Vx9E4XR6rVujfd%8UBA%BkmgX)K!Z7 z>m^*8MrKb%bl|>cQV&Oukzp4$Pm<~i3LYM3_-_ZSn4Itw$el<%X^tueZW^B%kqQJ2 zhZ-0X<%>qC?w}p{o!f~0V#rd?nZmjsBrSXW)IhmHEX-i^vleaN7W-Lp%RDY~kZXme zB|l)lrd%P@2GO$&8onS=e85c+f+`5xpBo1WQMa~(c4hd{Z7am8Ngqo6hEImWExj6o zd5sWqKL+8Ncw2eR)L;2n%u`_(CntA#*LR5+O!CubIPDC#cA#RT)*)kIKEyWQtYO`tFx-0k1&U`e=+$U%!2O%QF3g$H`^1^-;=A(10iiOKCieB?8pT?LMh3>AWYKlaI~SFZ!o zo>denH1wl=ap>qzT6S$I+er_V?nBvA7?HRi0`_-PRWeAYU`UNOJm{vhRCaEckB%e` zz!`ZANsT)#NfrRbFpy28ngW=Q2-J=7w5`>oFHO_5P&O(a25dO5F)Ur~|Le+Xw%|)i zjmaYuT#{Hidb7Ri>MPV`BmZqlAnffzSr`qc!Z~1ZZ8st2{PlRSsDyUoRPLNbFWN<; ztK(Dq;H*+62E8(^I+y!e({S9T#Ta|bZ-H@SxVfK$j1Kg20sg9G3j>u`VZ+g9G>WKc z=s_V%=Ljlhk@F6=q~?AJyhzIli=KHr<%5~J!@@Lvymtl~ytVu9O2bVjiR55+_j;!a zrP3I40v`RhS{&YQ_rB5kj-!$$ygh`f6*!I9(W@2=3I3jcT9>8a9k(0?T?Sl=-1v&f zG(rr~YW5Odw;NTIQW0^Gl;$;=2>C%~R1e{FYE9=-s%}Kt0Tbr02p)~d`hVCkIBvhY z{0E`0Nmn4LV|SR1bC@OYILPEf0BqUX@yCziFq%AuFBGS;h>t|Uvqxk7Xuf`ztRwdb zkACcrPcB5yS@pQX4&OYo5+XB1M74_Ys7AL+*>wsZQY*txohXzI(n5yaQppPEIzu@o zK$pB33Ry_yHouqh&>dSsd6Qt3#)UXW6Boo2_*9kp7L+#nur-Cpwkrf}_p)JIZq&`T^ zZ?`R;%0NI?@0}OI%yBD(s}0DR`@<((j78W*05}tc2{pB~Z-M%`x>!0fO<{aDZCr;M z0SX?%Y(H+eICfwHGh!P5mTf`#6~JB4B7Q48XENb%?B-TlECGBzhSuWw|nDu(YVyL7B_CftCx|s&&&Y*=Ze^u%1^AGt{dM8cVd+Ogm=HL@cL-0K@lTLZr znRxSqLyXa^1iQP-9A8~}KC(4SBCRGhe4V1R0A@kh{mt`+s|xY6KvJ$8Re5>5*2XLK zxO@4QV9yQ^kGmW}qus6gU`dz{!iQ&pID4-29? zumR%E?9IccfS)n`#XK4zN|8bQ#qM0K#g+?Q{e`-tTf`2jVpwLmWj^504Jvq5e#Ul) z)#Qu)Oe7pdsSpJGH4hptmPaDd6ryl&nVDi?p z@x3r^+HL(_!byCG<9_Dpu6xymuAUULzcLpa$1haISw@S_2B}fg6>wRPhg$@}Czto5O?7!LL5}Rg3_*dWz;o!}&DGAmAmt>E zsn4NGa{T^^=f=VdqwqUaRV+J)~1AWjZmS|LeIgeGu;y#!zkhQ;eoa|rZ^J4X)BM4>u;B^(;zA&?q#yBdf0TFWV` zw9@DAJK1W&fa9?aF~=E!2l<90+kY0wbj-f;^T(1CtMB7p8Ofy{#EE2j!(?YJ_YT4U^@RA*@%prv;Ow);AF0LeczoV zS#{&&Hpp&a;0HL_&vzrigF@Neg5D+z@Fg@S4D6ut=}k7c-Tb2_*WLeCB#+M4vm-n; zTz|3nF4m$gx|7*Dx0jgSZ~++!WfdFFrx3X(IskoqN8MpT14R;~!&>=DiNL~qI`nvh z^$seO3&PX%8&rx>_=C{D+^E8F&aS7hWjj}4yE8*@C^337AwGr20*j7y)1WfO-rIsH z29O1+?vQ@aFt65;r(J(@WFEPXBmf!15G969{2@dQ-x2*HC8E`_Z9nO~J@9FjoP{RB zgEWO+7`KU#Wj^zYY>9+swoIg>ISDD<_^%_ogWylNBUYrzuC~~PjWQ=PhU%K258};p zEzB^x;{3e<6Z4|#wonP23iiNPH7l-ob)Piki_mvD1S`hU?|uD&Y1R!LZP~Dqn66)> zIS~fiRRpg_uI^^%y{-dK_EBd3xeetSv)X&9ew;|0azT#BkME(DyRCCcxcC^zpYY>( zoQxqgoBl)ojm>D?9`GqgM*v?5x4#N#S@U`o<)JLE~5dLu6YmL_KL zJA*~f0S!?XA9XtT)HSFYP)DI01L-&hbu&#sb)q+=+@0LV;}6`gE}Yi~JE%IZV3TY?>sScFO-6vF)m0oP&UY zZ+G)6U-b@5@Ym__kupQ@rn74FAi?{EV#W<(+q8C!1iB1@h+`q*W&c1!&Fr^?hk*h- zx7zq3A(3;4c45Ersqh|J?eM$W9w!0t;VG%Mf_Qm-geYY$(QvNiAv3@Vrv*t1i#^Am zJC#Jlo8wEZtLl<1m<9mSwABc)v1RMwx;R##{WvK<8|8gCom7MH>hr46zt{$j;ql-d zQMny@u{HPg%|qN4A4dpnZ%@Ce|M>Ko7KGgslB{21{1%hc08S*MpYQM2hL|#jlEaf+ zMW<-=L7w!rBY`0{-V?uP5PTq`jL-F49Y7bc_p_EO2k$d_le-aPPkeg4y%yX9%KU@b z89=?Seb;LvcFG;{fAVVMJEP5rr!Gw)ST{`IO7FvimEaS~ns)Qc4>Hh=XW{e34E4s3 zuti?n^J4J~Xs?CRBc}1I7THg(`kw46f$Hnr$}u8I>pu7&Mf3Ss>9Dl-sxx6AoTp_e~blFj8$3 z3s%Stq}Efx5SDgcv{5M4VFYrveZSIO@S?m`fm&WneXoc>8|Uj@AV;7#%CcTHy8{*@ zv<8EH8j?odg{HLn(Y_4TQfL_%tcM%?QNdxsbg6PU5iqY7v4e+w;U6c>5`AA#B1&J} z6lqG*;5JL)xm4y{fW$K&;^fJGexBZKS9Jn>4H|?#SD;S*zbt?(43vauDT)))qTXz= z3Y$>ouuxc;8>PkZWn{83cF&SDUrRn!CPQX$vpIAz`KIL!sw#-H&dR(w&&|8{?fzeX z(W`Ol4Z^lH%h>1YGBeHFp@gkKAH0@Sak2-4xbGr(LY%7!9QLnCtwCTS;vP z#%9u4u*4$p*T>XZ^?@QoEiH(5Mj8hqcGx|k{uC8|`F(vkTb!H6g8AM~l6TvvY=P1g z)YUW1vN(YI4<*Q6fx@^=W7LtZ28!!s^uDU4}b0YdD38;?BcCYpy$yJPH@1h}@{Q z*rNPKgyae%-Mrr6?N6;?;CaCu-O(jv3Wc(Y4OWW2d;qyx!7ke#;#Ix2F=DC-W5yF| zW=FmE(_SX+M7uFib&(Y8Cdyc1=oq!QW+|dbqCEbE5ydS1Cefx(!z||GeWlTimK_uk8RDJ> zgILznJ&Eaqr-fyyf5D1idC9g3MMglZE7_jr{X*E^Iy4A9Q(hk0KudbK`zcm*i|T>) zFZSvTbUnXj+j(t4BIR-CkM-zS+F$~KZX+E~e_f7?6T9{yrbP{@Q%|ZZ{8G$r!H9HP zz1yv^#tH<3qioA}eirhYovzcXSnsu9#Ofew(iboExfI05&Zi7sZqHx&gdh}39RB6X z$@EWE76EyUv2yqEzg>4Uct&0&!ovzg(4D%o9rok`>);rAe!EIB+Pj@YN>TRmUC)$d zFc}h<^w&>7>>ZixD#K)zZT*?9j?Rl!9U1fSpd@t(iLW)N*T}x(IHMt7d5%!wr1?L@ zgADkU%j?L3I&CFpH`k|d^mIo>WB7I%jO6edklOd2^Se|AvDVnPDQ+i9O0cjENw*fAsr$^~HWFq4oYyRemxlsfD#nRT zg7{ASr4yl-|2Vjj#8*Ud@X98fZF&H>0`*mf-1~kg9F+t#t~*I#SFc9c7a@G&w*_y1 z(%90yhyol<+JNUr++Sn!P}APC}LR%mROH&xbhWNa;q(hZ46~ zX;VB3VL=Tj*a#*=8}%}U3uL9DG0qE0>XYH*Z1?X2b2*-|97`m)RWl0sF2A77`BD!L z0&QY`N4VO5R^2o$p5PnsTxu^cnK+Jk0oid9>zkX8-JkxwDV*DNEC?ARaLzpT(P{Kw zy`O9DKaxp{P)J;hjJf+P&Zw>q>bJ^jMtt%AMX}Sa*}+$IQ6kKv&4xnYp)(c%?~7`( z8x(l`Cs-&j_ED5xHjTUCqN}UK@1;~$J#UuX^gldI>zOGeHQsztC7MdvC8MC~yQu`N zUMG*rN^j)t(7C~j#mx^2)*UWdiyfu)bAT(*y60&s!VgCTi;|L406B`(5WEDJe^c3+ z$*Sepyo$@~*BZVwG^r7DXTpfctJSun8`>dV2`sDnG;ZF#-wAPrxCq$zYE)`^Kp48r zeBh5#Z(KQkfX)5+2KQYBI4qosayrH2zZR1efLXgil-}QO0bD2b_l;e*=Q9r>Vd6OK z9w*|%@Ngu?#kv?(pc4NX5*B-36xmsK3|P#->l0_lLeED1j~|bwSVd=+wPnmmCt2q0*OqNj zr3Rq;8q*qX^Woo4j*s6JTz~mE-WD>H9um)Rq0^F!)-PR#K<+N&7Fu={g6SW&r91>K@-GMIN>Hsud z4e==5V%6uhu4do&qJ{8Gpl52N2Pa%ZZ0?teQl}Bt<6oW>zJ{iCI~m_E&rqL3N8$`J z1RAGs?w^21Q?ZG%S1+?(#N{;=&Vszfqu&RmxobZk=f4zOdDUdYWTcjGx3XFGPRG*b zf>-$vV)*qmBuIbS5mDH4!8d>nVuS`<*E2O<&)mBlsCa67qjSa{SbJOiQnl_7hg2M& z>%P7n?yWZ-P36uqetTwF&w$N<-P(epCw8vJcB)EK7b>emxQ6ZQ8i^N0>VGvewL^dH|<)cD^aN@zJU zHP*JzFry1}*an@nv{Wqeez~cMIWwNOtmF;;FZa0&^jU!{FokkY4DG2FJ(ADEN}mD` zxLXAHvyH@36N$grU?P>VewF*7*TAo)YMG|Fwf?=oiv2{2=f0fz(yxk8XF&W^V$MCN z_zbv86o11URtTFa_WSdB%X|J%4^nmDS=5qT0|#XPtd0(X8wSD7L}tS(_{5b|pN&3# zq=^l^*5ku>t92cP4BdT%X1zxg9T09xnmDN8zOqqZahoLUrz(WOp5t_C`sb{vA?7ji!2iAKqbVV05Mh|3ySzkdzNk z_*F_Q^fFbjLy*i+edqbwXKuB8TZt%L6Ezo8-Wp*0L8d_H$bd|HU;K|tj_WxnbGpwm zBs&Fq@d7*jkZTCSwCCKh_OcL+@mpf55xY|t0>mif_1VmP;NJ)C^CY>#tHAkznuG&%4v7^<}$s!g)trxohfVYA`6Z`-BnaO@YhXDw(ROaNEGDVeo8Gy6}zzy zJ?87GBifSbmLNvZ9 z1Bch?NCrD_F;?U9*xM=3+R1SGkNv@pa=v%A*6#+l8C9y}Iq`2sp;ye*g0TqD_|vm4 zqLl%CIOap`A0SBVVynfCW)jUUkNfNAq}MRa@V@l_N7Fm7$JMs|+p(Q=;>K!h+g2Mp zX&T#_*tTukww;EJZ5!`g|ND7Az^rZ7+7`~^hyBzjIA}(LM=2GQ5`$RiC&fJ!D6Oom ztSf2>fP7n!fN}JqVd7J=`Vi)Ii@PZ0`6X9O?&p!zQpYdD`7~exG=@Fs4%a>t1jqzH z^0W%ctgnkRSg%BlUXub(38)Cn8&q-^nedjtSb-PIbcByGHn zSyUd+A zR@vUgZ!+&MokqhjrWyE==k=+RLDQQ;R2b<|Briy<9Jk)sn3LmHV@9-a*c;8Cf~=KI zBWl&o?79;pQWc z0wbGScB*VMmnvh6&c@5u^^>fvjN(l%4r={04j}$0|KT(dhe8%d@fH8V{4&i1+5hY= zK0AZ4c#4cwbl#w`Sjw-Yl~Z+WCZH&8AEk^S7wc}`^l@#uD6y5KWf>41ED`$z8Jd~bhp4zK7}Cy;f_*TZSAZKr{c-! zVrWSk^sY$m@2vT6x&mXQackm{KvLn;5jM%)2uiR2n3LMEv8!7pcPWhl01XE zq1g!Hi($K~)FJn&faVgDuKm#^eENOluz`or8kK@k{gG{bMjI?>ftG)F^ z0NN=nop_c(^>+ezWWUUO7@mi&h|i2-FHV}tv-0#TuRk=i%R9v<&Z}F$80q^>YjUth zehjE(38Jxr8TUElu5jhLeCke`Mpm!M7k?ry0zXjF?Z!oe|rCYW)CxG3;yO2Sf)@Br`%ID zWw%CM4%8oa6W87%(Y|$EI{V&QgAa14yFm2Utx&=G+}CN>_cjvW28mAgR!DAflM@Tr zUEa_)fh|YgM$64&o*eT^tzq7#D@bO%60K28SQ#UcSSmf2_S7ZS_9xM)BXlQ=MQ{zN z89$cx{;g_$&z?OPzPbj{{=wPn1`Y%d(v@mZ4sgTTiB`rfi7C_EDU})B7b}eDH&rPdI9Y zzj}A!D*F~Wk!9Y+kq$ZjslYp*c0%IiyfiUhNmQHsOV;T8?ZOd$<=Ls}ebw>D=I**V z#Hw>wd!RG5$Kc_-M9@N1=B3fAbtjkX7(<56hxaU$Fha=w!Pl^3_i0 zk{Itb>~9NbHB;dwRE;XzO^lU{TRsH%mDZrX0m`n*IV+S51>q2BzMY9%wj>0*SwX+XYBkGP7NBbBf7tv1!@ zb6p>FO)%&N)vo$qjZIvK8eztb3C)|1>k{s@_D}uea#fv4oX4X$)4{x{{9h0vO>q7Y zgY{&-engn5sKzu^lkA2sZ0#i2X+_jIKKmq`Z2|r{!ZvoHMDa55EGa!`-_p|yJ&;%W z8ML`?jWQ})5OkJbU18X=?(A;ar>dyv1n7X)U4Fk8_32HV`@Z{Dap0$E@t01Qd?Nbs z)&{t(q@Q(0#EFb31SLEg&^4uq!_NMlB?7g#acOn>R<4i4B}*wljCJ4sro=IzNoc(P zfyG82UjdHUFWy4?^x;CS#AS`2cP&0RW7{6aT98$^RRd-^odInXSPMnwj-~_2qQ57_&fh=LidKPn?%TXj71fraQ+2ILB|!vs z#MuFS*2#qXzN*TPXyfV_7TRWMkZ>NDPb~yR2Q6&;J@VL)-o-_F^8QlpE#%<~ zB`%RKGnC)J^I}}bK4dpHOOTBJ{O5he+R6tA*STs!vR#1Xz1b=R&Na_IaBj?!0bGHK zR6>gay9sN2K8CUjSUJ?+no%Q(cGu(CUFQ)%dP0cP#z?n^#IA zrCRZ&z+?J8e41}(lm#&C7&zR|217#3_Zl79N__KAH4tKBm5BRPkB=rBFJ=aa;2WSE zJ@1cisyzO%<+T66_P_&_=Nj=#c)KT(z~4sXS;9PO{q-uN6SP98UEUvHiA_=5$t6&Q3Beq8xp|uJ_=|E9G{!v)A zPLoq%lM{3QXT4+hIG)n>ZY(lr%%v*}RmG57i#^F*ACcoEQpeeb8 zW0S#UGlJ{l4iwz=a`?`Xx@05{Nje!kXj2dvWGcMwcjE_WttKQRqVGX`3;^zWYgdVy-@~9`&v@R4_Sh68Had+%41NR8 zO$m9&7aWyWWR_>i7C{xLr^(Xxn*sk6o5>Rq+$rMA!5@ofoWUmI#R*mBaur5n`@L|{tRB`XoZL6g1^|b*3WnZ{&H1VpWRP-_|I<8KmoaZI`bChp5l)avI`T+aOST_WOKlC<=NQYijO2qkuEUKDM48qqZ~=Tp{^o|H#`{-$sw}9)V&0EF2^SI z+?LLDg&u-g-(^j=axtv-)jdKUPb!PO>ol~9>Pft#RXGnKOqR!SMSe~YbmRA|6$yk? zWSq2r$cyv%j*QHjYsjKu*?GtHb=FG#*LlrwjIW#KQZ&H@nkbSrOG83AAByL|f)EaM z~gO`Hd7Nh3O}Y)D*Mksaj;L>P+!{eCHQ!`8BrGUy6bw-ApHyeqvSpM)IH5M$JSGreQTt5@_*#%p6sWotwU$Gz>X zK3r)ogjZ9coY#OKhpjEfK<4;Fjlzg*$i=0MvWiperK6tEZ;`77QXRvk6M5#;qeO;s zJUoll=6u)L%o%k{fJ@4Y@3*sIf}d{5wFWxtM6OK5ODEiA#O^7t-nf3l@D}QOIAbsU zd@0afrnz;`>o`$+OOl`As`I9n16Y@(zJk&0qoI|rBcdMQM6HX{CBeNA#JlM8-L+?RUU&|HzBH++k$$aOZeH7$#l(WyHx4VDR7I&HXE$ z35TBpOwRNQrC(HMqI>}EGI`hcOgB&ieD5^8mr{!5F?L>qawSI$53OS7BAEf>H&|BXFXoz^^hg}!mZQOi*d;brRKM)A>8PPSDt7Os$*aN7xjVlc3*tqR6~2dzgny~7 z!XX$bW~ANrJ!|P}9VB0w7{C1})(89Qcs0h`W=+L8Zeod#l~d=iu2nwqP>hlCP~ zK-nuvf@ezWK9jWw=6xW+BhriGPqpTH{>(_W>?_D>_1Wq#bnpkkh9q|jflL`ke4nqM zcy4zP2SHjd2s+FDM+R2Fbm(q zE&GK2 zg&$VA)6z83>5<%|zx&;xwXFM=a*1e`I-y{|`wa1yeJh1rJl!}ECdMP_)$8^a)mJ}g zU?Nr|m}S>#C=1b=wB||HkX_Mt|DwwX`W()z8ep051?a{!7!8FB?{jCfEESI;A#WO^ za@vepEqxttaSaEGPy0)Zbr)3<4^2j-oqhUOz%6BK)e^uV!t!9jU_@LoE-fz7X}3F>%O|*5Kax(IhTonj(-><-y*0vM zRdPx8DhG$kBw}$NPAGEJsX`(?w(WR*aUFNcY`#*vGFH z9p~YnvOk%y+TA*rlNbD(A2kW1&d!JR_U8Kb2cJumz2uyVCqK5wt4K;>fyzi&z9d5Z z9bAs|8*hNi^8tlyveh(9-sCzP3i6?~Rt)ukREhzFL@n9^^nw#ZnJZvz6r6cUB#(QW z+y@N7Jjn(jN^5lDN?m(r6DgY032vtSwrAdG>z{?zjOOv+kELg4Ma(DXnJA*2i?dLC zaZo4SLQFfb6(5+QS0fB|;C2JKaM85?DheNg4{qoeTdUl9BQ@n5vR9X|dE;8j8L`)T z=~uUyfy{4<{BMOsSgd54ztAug9m=$}tG(S%K|Ba*oxX)O`^hAjP7C?*KSffN?lVG- zYz^*&%kD9_tc9=ji+}4)vOR@k67%>FyZ80N(-x|;C7$p2%9xVrn_S@5TRzH__A^=h~ zeNg_>q6a9-W* z99PiZ5Dlz|s=~o#9AA%W?-&RZ&q7(3V=LPOZVE^GUr;W!W|TwLV^E2vJQ4_LC;O;S zAqZ^I3;h9r2e#-m3GHi*=5>hMDcoy$1&TYGAHKE?@mIMOIQejRiqn0^_k<5*WuTjE z$(R@ka6H@BGs-!Z`sf?!9UR;zu63W zTpT=3FbPehqWA+U2n#6xG zSjLa!IESryK8zrpND{eSWI@F3M-Tv(%-vH+ebKD$89)NdRn&kaegjJ~MKq$KkG-9Z z!>m%2n=kJ@5&-;hnsM`vdx!nH*<F|01{I3#%4Z@+&Q>-_c&Y5r&;HnX^_v5gICf$DVqpZkDQP`>p}g01?=2r?APzjBrg#s_itgiZ5)EJ73HR} z22d3frEBL%R&?gBv@pJ}=M8e8lL=picx^H-wk+0IaU1`?!cwd&N-h4QJ*5K1sL!{8 zA#C~QSHt|~QpE8zKj}r{O`E~{-?vWC3F{?k8v+qZ=RW-&MoLj$F;7d((?!knm^R2XrN;9S&9ko;Y>14k{^*MmdC`X%ptxC8OcNKVan z=9k7Aswk632pH`H4OYl&C(59!&GNqsS z4XX*c;Kd6pcISi9={@u7QWJvW=On(;p&G(l0JHtDNQvJB#r8dJ{|{z7{GY}g=o|sA zS~y6b;F)=WqDa-{_#bA7HvQeqiK!gI<8;F^(!H>NS`MKJ;=QQ&q6CITeKT-sUp%up zqjOu3VgJdwr7r4ldciZ^ou%sQ)9Ur&K(AA#XI58&)JbV*+Xofi$pt#+{Xj_GQrn0! zVA&q*2fknlu!DDt>L-I%Lt9iLs!FK8M~#X=n&t+ zXD_P5MJ6&ni^Wpm?Kk&;J%nrQoC1jUR|gI*OP!TQy{+&KjlC6r2z)0O-`+izH67ph z@uh7q?oYkc@2x={`&jc0aIX%TdBJO7yw_hCmzEH~QR}g>%e?MqN8^o&UiJFYQa`M1 zCuir=mBwhMF>0R%&!QpA*PAg`gp?t7?}UUgQAyzV~wlt zxoy3_ac!Y-SPa!o0JVrN1yfAJR#3A_J^acs6p9?vYNpraJ9w?h7Br9 z{xVG2b4sYjN{*Lbz!N+&9sYandBiahgK4e0c@>d1eeicFTFlQGPa>r3NhHs+heFx;bf1wn2%VGqbuo3UcgkNRKTFBTTX>R7YWfVYm}P$N)7c>nWX~A~G-smO!@3jN+-9tc2_;n-)u+uEu`G}V7+_!Q zU@N(;s9r@KCM^8u0Xxh=2YKM}wz+NgKMeq(D1(tQ!q3vxhTI5-pSW~zg7U2yx7BD0 z*nFSEJwP?mI4-#FTQC^w$fWI?jx09x%~ec>c}=~)yioSRmhHSp4D-Y}pqT;2O}@K1 zD_~#xy<+BbGKp4wlS9W_1peq!zF(GDJHblJf-x_$-AOfGbd#bbV15HcrDs(ATM_Um zCVE|VIbx}wOaCa_o9JNc$NZMZtc(+GK%x$BGLO41=r`#YkNt>@mH8h%QNTT^&w!a* zy^Ahlzd1LN65{D$G30h%yV;)4*f!giQgnkGdI1zA)R^FMTl{es!>1uO{VQg45mWkH z!UalB`e0bDv+;82Mj~UgneXE*x4(EkXS3$U@eb)JdvRVt@S_OH~_XB14+ zDh|8cw0d#;`4wu--QIRbk0{a#)IT&GD$|GkPYlO%EyhUKpQiH{_k2fJ(dupfXE0Rx z%Z0WH@XbS~yV~YrRhl&;rz-+ z6^zFdC8G>8BG}9o*Un?4;inyI-T%C0@e%NgRO5JN=zj9hxPHGt7A>(K9JBc=bfSsF zFoG1dGl_~5KHR8Zb0rWbe0ed zp)2lt+B(m=ouxASdG5=-S0tb&Kzyt1e3U4L_(l$}v4ube{-4PqmDLtxKWA|LMnahi zl;xHh*Hkt`B!|O3zV{_sb&;x`SIc+$Yzs=GIFheygrRp^Fh10j6S-^qASOeuZi78oP`*gf+%P3<{do0T zc;@^hBuo2U>I>Zasu{dT+-HKel9{IrSp^ z!O`f=0s9sR{zP&iqnkP^{mdPZFX`8?FXf-q+9Sm4!g3)&Y+Me+j4#>u+IA%h*AVRk z{xGXzEhP-3AA;N=jC7o;xYoLUL&-|TlR#>F>_F&bK)2bZ5r|3RJqN3H>{TxLL}6{U z$*Vpd9K=-mbo+7s3{u-YdEFEOP^`&6L;70qFK=)@8Vo}mu8|Uv-tIxy$6*Eikmy{5 zaE?GT*QaR1Rs1MVz@Gmw5i9Y_-xJIE#0!)|%z&L6dFP~RCqmw`g`O$A9;S!pI>G29 z=?VJFjb_Uq9u-H6!j%5uQGLr$=hczC#I1jUGZj{nsj+L0Rnt6ZoJ5p|%blO(_2@<| z?_Sp$D3R9Z_<1mt8Ne!ez13=?yE?nbm6==mFMO(RifL6sve|t<5p9GOlhD>!#*B;# zz0ms%@`$TA&DX69ZW*dVetW1BvujOD-t;|~n;rePImTa`>tgu|1M-(97pg}1dHd~A z5Y%HXKN&8E3X9F*b+2Yf`iy6d7P(oT%* zbFu#4>dR6Lvx{L;g|rET58(1RKTtt zGW{-~j@sUU=fM&KP94sCL(%^CEcL4&FgOj)Q7f~i5N^7~kCyaj4?p3|7yZF*(XAvD z+ep%G^E2?_zbK#>)Rl_F9pP!(@SRHi{II|1i;FQ+0^Bl=3B)0JB$inr8XJr!!PVl^ z*-xYo$0d+}2}F|ElX}pMf05b|`8q^gsbF@8tbXH*>K8CU&_Q6pyi;=dQPfOiT^2RPNkcJ!CAxs!PTEC;R4!#0R7+GP%*e>u91eG;};o{IjIlrwsBzTPrtl ztEs5oeDES4G1ay!n12>+kOY&rPC=+=u}~f`@9@Jyo73d+kpL5Yf%dqEPG+s&I+y;2 zWskkLu=Xp!x;>)UL2boAU)HW=YI!lYLf?I#3X4IZt(L!GdDyY*hfQ-mg%lk3I!;KX zprlJoR&YAn%7k%PN);Nex$(oYz&4_5?Dun{y}CFfu-^_q^dOSycn>1@t?#hvP#=%C z;}Wo!Tovhlkdy2AvtSO7oS(Bh^|JBzPQwUTRrN$!omNZ4*(>2?G!kyX;-ZE$;!rii z$=heP_wtg8&0=u`VrVTndBX7XPs#uGS33h1U7O&?JVY#$q(WmY%X@3 zArElVtl4ml-c5V;vBSGF5BGCdx=F*mtfM!oL~Y-a&-5K|S*JoKu$LyB{43fX8P z++5HR<|voEqM-#b^J7#w8%ZACNqvg=^AbVdmMPt&s|#G}@Dh6A8V!nqXoDnSG-_XI z$D7H7pTCgiy=zeiI<4OS))RpI44uz2P4aRM__J$$S=~3bWk+iZThpNMnu<3w)-e5jS{6wBGK2g>VFz>Q0MJ5dHa(qKX5=Veo4@68W z>REa^o`+LX8s-r6on$+ew+WmBuWrsWTjae#!{;WzgO;p0;tNW4#lMGDVgKx) zK`PzVa%`HNkdoGc|0Y&H=fv*#_N2TA9d3JmOx#WvHLUVEteIxWb;F zEwyAh;Y7~@B~wVVw||0+T}R^cANXGAO%8=tUq8UJDeprn1#6CO zF2t$linO(1UoySE<;)Oz_ib}}E2Hq*oc)N|rQm!UYUg&|G#PntqO*0BQe*r2%5>{F zM2ELG(u%Z!dC*dw@_&LJJPUHjK9a#n|IT*>C$=;L%lI$#O>Y~;3T1wm`rJh^bhx#N z{ear-Bf^WV)S+vQ0aVSz-(^1XccH3>XL0O}@bgs5#a%t zNT+_>Uv-mxqJ(SC#(s0rf>3oHw=q^J!oVmXc zfB9H5!rXvaA#VVaDJaE1h2^ek@jOkYmNvz29KZ8(KqL-uPLO9{E%At=jOffnYr zUUx9a zaajx?J!jh3#aC{}GVLrwVbkPG?WlyO`d2mLmn=kY48rp|KM2qHsPyFc56=jTN97gb z@s+%PHWlnd6VlDm5V)u0|nVpDpVcnj!N%N_n_J@eSSE9 zCZr;R4&#CMTkE0h33sxom>K3S^G2H%8MdAyS302udAt`KHSM0~vWfi<%J0zM0qYZf zCM=Kc^Np4*I$I+t*4j@m0Ti8bHP7Vad@cm;VT3Dv#KL(edDR3#)i5NTL?PM`p4?N$ zjMwoun0U~h-+AkIh5yVSf1~m$n^S;XECztl*CGf8pIt213bN>WadMLb6r}=MGuirt z^6cd}U-anW%u+b`J#imh&ny%V9v`6<^52tuP&ezk?2D`3ICMPkg!Odkjw`hk4U8Lu zAb%fZ-!wmYb4Ynho_AM8^M(rUBurY z0-K9z`KAdUC;fJo_^eeK7b{Sin~>+ko>?Oo@exI;^c-d{KOqFQs~&bTTe1MZ)-^^m zLuAE%;M{l5d>|0O58Zdjuou%Q-jzP!${lxgi!~V8k2Fj-V8&@D#5p}H6p$Imhoq!M zA9TXK4E3pD#Ssm+!M*}MgA5`Y3BO971&c$#Tc)>b?RNU<3v>F%@&)~hE$z>~W>=){ z6?y}4A7PIS&OMYs!etxrn^Gz|1Na8VwY*-D;9MuEGq^U@!aLz>jsfeEAU3_$y^pqA zo+WEo%K#3=+={OEFV^k-F#&2pSz*~_?D8nV?_@(Y24f~*HfdHOnUmcq@rEr}t~gnz zXIn6JxmTW9bz37iu-%vvCcOTb;mZ}rSI)l=$QGe#D(-s;Lfs)9u_%A$z9NWdO<-iF zAQST3yXaQoz>u4(eh~^JBVt=6r)L@W`>Mkbkdu{iTs``q{V82uU>%IIs!c>%^>RRI z6o5y!SdSj6>?iPtCo@C3gsjhq*?CU72_E;)*R~;4v6RUy9nm%JJ85Jh{K zc0FlT@74mEN4+)*WI0=?Z~VHzD+I}OosQ?m}apKGdUK=p?mhtU?$qTpw^7=p7#?`^Enhb(e4$Gd4N*Qp%HAB!p3?b=D8JMNK zq5d(yh(c29dxM+6;0GQWhUmuD3U;+$O!i)bo9=SLB{?U>4UM5KlUQ z8WS$u+a{bH{fEOnc|re#USvuvyh=gY6qUB|6V7Gfqbf?d2;DtqiyjeVFYKl!X%hI) z<%#%pkxlMg9}__Fzu&RmL?-RV z35B|1>-z&tC=RMBA?{)w$Cz*ej>*%F>$QYjSQC)syuCrZ6Qtj*V?%{r-)bXldm=b4 zD_GeW~lJ>})6nYF~QdU!LhLS1v6#Fe?gnl<8*3V%lGW&xh0 zo7qRtch_{LsRGWDZq&55YcQhj2&`HWSn6EFnp|Bka97CbJW&@S5yxjLb_GnU6jFL? zJ!WeCCDp+Mmv_P1Zc_-1*$T_wDI8+8(J&|)MZqv?`wKfcd>@fIf96tij4Am&4rZj8S%%H24m zuCl%c7hCQ%;uw9fI2Rl!tcBTpBwmbzX7;r6O#eyBaQZz~p&AiygRz+RFwf8*vz_hj z?KDFUQQ|z^83{@XN4mrteCf7@s{SIdjoQe@oPpEoAxM3;I9~TN!MZOprPIAqZNF+^ zhDOKkkr|gfBhKJ;Rlq))^2({24?07J4`mMuP$ema`$V+zsgqwoZO0d+2`?ZHdAns9o4c~8^p45NzH81 zX~>w1qFFW9L$k>3Qqr#eRLfl$R4k}b)tK}93s$=Biqqe|llKhFc3DaMD+W>+=B!HgpeO&CSA>pnff%-J1D3sC_-ODZ|BO=&t5|1Du& z(gCmJ$!NW64MCndJplV^i?3E7kQQ#imC43R(C9g!x6wQ+m~i6wmwQcdlrcnCAyp9* zywP|a`J@eup6&%21Us>fAlGL0Bjx{-$O~~7Hk~+pGPq`&>hp{LeaU7xG{^nljo1es zXxGAO2mEWo45tkt3c?okdUY!^J!;1IUW?4qrVnM>)vXz+o)YBBzoYX#XY2$F?=psD zQl+a(2}{pr4b5&D{7K}M@{Ke&O$kRTC7d(zM0FQIpzWhoLfdP0Khg%X zBtlar5i#3iCpxEe#Q}dWOl0}HUHtOilhpsMwcPT)twi|BGe7Hieng;`sV_AR`EC?eLioSnUt4=jG;()s>b z54jN&B}z=oZ}hNwCCGf>r-hM)kJk}&4j&C>WkF2zE`}eyv#mtUujll`L^2UBm97Xz zeh^kNvCfCFkawfnGU@~JxE8X&-m(TsrIbry^R9mm9tnyHl`nTkJv5z4v(IqQ_r@sp z)>H_uq~@?0_bNVcj=&nsKael)3~Ti<3(GZ3t=LsX@|d9dfdD?l*YD89;&5mZ@@F)= zc{e#G4Dor3tzW&c&SGCuX-c_K$QOKAd|S}sK)sl^Kkr+v@JB|l4J5}7$M?%tc^^z0 zc^0;+-F+5&m;4_0W;?ie+xqy3SN`s?Yh$^Lh{)0MXquB>ZuRgLkWVwD6B%3N3d-&$ zHMW7dR7lF{SIY)k+)%@V4Cu54#EoLdGrVbM(-nAY$!F)c(He1B$E`NCI4h>%TuUdv zOwSNotFNx~SnALYD)yB7d5Yz6-KH}q$JLSiEp*|Jp6CfY;Q?})rX37snVj7Az!J$d zqb~38pC_AFGT6rZ^GrD~&pSQ2P)Wm5k({qM*&WWw_6 zGD)n{6oTRza@-}+q@XzqlB$WnJj(okZYL*-8y!TBqIr5HqC$SoqDV!Mot@A)I!5$M zZ%Z2#_)dDSt@VG3bfu$^^)B>&IZ)9K;p^Ys!}P3xB6!0xk_)DCy@fgWfxzyAx=!HF zhFXaULX{OnJ&bINIn`_aK#K)$^+6()*gU-~GL~nQD??D8&kgwksKH>cy^T zb4Bp0PWKYl?So>7FZP~&Px^%ePg+hjsoTWh|9L=TEL&F4$-<4<9G2*ryEf;$|0{WAk0;QrqYt2H3-2a)o(ebIBZxhr+il% zQ}XGRa0cd@Q9G-(-PV+Vkfe^c%ceFD#86ui)30$?5`o*zmeB@CP!wif+n&r9vERS^ ziF40wtZ^ouf*pX-g$R6ZsfOPOs_!FFT2am7u2phtRJqi-Ca~PC{K3fBQf944*@i{>Us zWMNo+_WNpMY2VhA>WNE-ipz0JbTX_7YK(T$e} z?~in4wX9`&%fHo;u3xQo;Y*&(EnZ$col4yvQP+`QWageFvL#G2=PC+co3tg{Sw4K6 z?k3*%_sri;$-7>me~wJhMc!MWQd+P$GT?4bYA?uSF( z9ZqS>^12DUoDu!ZDCL&QEiBKi_2W=7Q~o(F~7WNxRoH@+Hm8&8Q2+7D?z0h z2en%pf+zFOoh_>!Ycz!nbbiJc0AoUlLpIr_uB?f+*1L^1g**eZP>TzBX} zTr!cS%Vr*X9>sqx%i$h0!taK;Vd8BAap74y0@0pztIvE9!Uzs59oKqr36c$4C?{yh z+v8hGKkl)ocMcR^4r9JycY0CEfQ!n5Catrs_ z_Czl(>MC83?>LG*rcbzrazm2{3i{b@0fE|T$!AI8N}#lLgYjWg&iA2EJ;)REWg>kU zj;uJH>B5a}6%0@IoEx+)x{pHBIsU~vyZxvQI444)$>tW7Rj0qrB7Z|1uG7lU&S-$( zO|Bbmv`sC{0YOhNL!9U`DvMMIv90=$9&~p+T5}f!p^NYfa-Y-uZPhXt&ukNv=S=|b zN19{2+G(?&_XxN7K%0?vx7Z(?kF$BAw0-i9;C&WuxE9>#To&@)KoUegIiROvSHp9T zN};&>6&@&#H!8}A)aagc8xA$7~1^mnTKA_eOefbZ7inak%>nR^M)n#3>&;|i;bYd#w&J7 zV%2t2&vqhZI`SzOY$N`-|Nq8uYVO`Zam0Vf5fQCL_YE|Oqsee8kz}SK7Ir*S7*K7T ziEt;$aua_3j$nT8Tf)1rFmc~kc$vH7%i$3-*aJC4FIscshwdfID=Lb}cOm^Pr{pi$ z_2A9fsSp*CS^DzoK{%9j6Q;GnGLxg@!ek}G6?+pl?2mBiZfqXWid2!TG~{eaIQ5Mg z7PZ#(ZC!_C^m74gH~YdJ`PApxh<2=VUw$~(Tq{j90{ItRxt*(TP+=T^@fNS>(h+JN zw;8_o>wy|uB?zs# z0|3YfLcV$;!E2@JU>Z5F0V}sM&$(2-H=HTAy=cUL$x09EhxBer0B^a65%oFMjSh+z zV-M2jUCk<2*iDX$@*8kwpfIl$QW$jo1sK^Cq+8RDEWEySMXgY~^Zxqv!<-?D*spC> zAOvlIlw-QK(i#J)k7QDtVfG=B9XLHl8qZJ~;CJg;%%J|F*aMbiU2)ydgg)WBWpR+? zD!Li~kIL~9BmrIlno}^hm$bPC1~}{b`D|u8tsCv20iX;4du+vGNMf%lmD61oga4wz zhu+=FldSK(nlzH*koC70GYPv&2sIl)ouSv;^4tJhRcUHzXk^i|8-4}vQd)f+Sv68G zSEmiy#sE*|+-qJ5q^1K(vjgwQK9z={Z1VsoqAg`2`w z>RDauHew6Q;r`zxW;46S;{{XSnI!9ID^iD|ee6z?e`R{Q|6(=dr~YbKU+iks+Z69n z{_!uN{BiLs+maDs=zGudD{YB%toi@F022wAA|e#=_EFgUQ=6aG^9p~|OP!3D+`zo? zQ>$i15p$xwN+wfBv>pg0a{12h0k4UsnYoAp>{#tW;*so@E?b0V-TI>md2Qg)#O*#T zH!z4T?M(n`sYhc(z6A=HkKpYLhd=9;Xzid{n};NDF9!w_lr)HgaKAd9yY2YNSMZIM z)&ej#zU_!Gey!L!tR^M4%TnNPdc^|iSW-rTj!d>6_V||*tRs^j_-mPW@-+P=JHTo0 zu?q&_2zILUpB+2v6yk;Fs|=>K#BWfYZlmOQzS@XP0GSNPt7}Cle0a+;%s&j8^(Kxq z^?_%rdO?@#;=Eq7hucXGAz0+7$w=yUOjKCE-Wc9SDPhU4^rPma`604!So{_+zm%@% zqo_)BFHu1NlcvURv~?l{O2nBeLM8m>x?}3uJKDM+rrh}j&6Vu*7`vLFuUSvd(Rvn= zA_T=}?X`$;12Rln#Hz+(5nAX4h@CIv>C+`lc{Br#;3p~j`}X1$`U)4QAJJj*yLIuv z6H>1?rOJ<2teN8x+97JTG3ENS(^vBP+P9ZbMEmS<1h@0_nTk=f*=ltYZ8yy^F|~zX zmPC*p`h+R1iJOrP0$KyQ@*P;ngqVCc6Gf1C?2(p~HvetJTp`S9M}D9cKzDq&FYMs|2Jn9SY>gn?K<;D)_?{yanhJ@Se+lmu7$@#{84 zg5-{w`|REl$>O7i$y8$@$-F>RCaT{Y;eqqWUq z)0iCgM*|T2a1iT5Zc$;WR;1y#dquYAhFO-;6^$91ZWdM|(VA`7Bv|S(tBQ+Fe{jFl z6Vi|F>Lq4nS~Do`gE<@UyTP(d+s+fs#E=rM_NS==E{SV|m~Y587=P1V5$vme+aH5C zaQ9nZ)7^u}68jfT8#{xOFt||_mMXS$%8iWYB=3|=8%WIrsbKuBD{Qw>;^+?yfX1lL z{6Y1#rygSvd(S6mHIJIP?5#{o=0r5#iEu`8l?x=yKbmqu>~gy?T>-<`SsL3s-42uL zVs4q<2Hv;?*gL%dA9GFWc^}MhY@G6#k!*D>0yBlGU%eL737J8a{GNg|3HLLa;2g#| zkkmg3|9w~EA`o1X@v=FLn7Ou~9dXkO*zQUq?{w(`himxq^I1k}Y1Tm;#wD)f9zeuA zKOcSHjEij&w=+^$4u;ba4W0jsj|>9bp&IJz-U?NdAT_=&4EcAjsp7D7#(w_3Q^J~J zQ?h1F{52R5CN*JBporHLz%-z_ggZhQhRMxsS)U+>0@VX>TXtwJy5&IMYRTyNY1ZuY zU!zAX0@9Rf`>4;OrB4UZtj||KYfubZ`{s=)kC6mHB=9X?HhgHqfq0ETD70*Ws4(@1 zK}>)vmrJT^3L%6KBcf(563LP+vtY&3aD3}hvn%r_lU9R}MWC>p?Wnlu4YO*&=F~b28nrg!%iz#VhGzR>QW(Qa?O?6Z@>byUcC40eso zI-7?2r<&U7Ryi8R^?86l>4L*fO@Kaej)t1Dvc2WRT*OxumO1x>!rr{loKi+Mn*xny zErN#o`@J3?Ggr(|l9h7i#$t7!Wc>kVt%Z@z3*SgpCXy(mpYFQ5cFHvf^JdPf;J>FZ zL5L5DjXcTfyPSnhPpGUt@+Og>HDf*7J{?;^#MgVXk$n3LRRbVmt&~m|6~z_QE)7uWTv{djojRu3c&q9rJa(H z&F>boXMqp`Mdmi2-d1v@_fZ>R4$l9^Q4Ysm9jaTzgk01|P>yZyh66kHBNWqIR*!Af zqWbSb>mPIhy0@X<^paB)lxZ6-wQYxp;%aGzNd-*Zf4tX z4wp7m5DEHqi18n#4YXATmxNr37xN~uiOj>J2t;vO#g`orprro@XBx=9Gt!%F()owW zuJA3n@{|fm3}?y4zX3vv=57Z)!{?rS9^1R(LCP=xrsyF%COaHU@Qu*#^ZKRuNAQ76 zA@uuf8`*w~^{q8pT1A7w@C8u0dwksFs#|;bd$@O@#HUSm$-u8iGlge*rzcEf4mq=* zKfLy#^F35DsH%KXDJ7e<*%;aN*`YpM4sdKj)_2MVyXfXX<}IBR82$Yr!g^bdhVWN6 zhs*s)uZUmuTp5&mcLaBrye_^kc#?lRHNJt$an9tlRtlxiM>mRYR_Y4hk+pA{DoK6VJ+;gHi26%4G_?g*s;d5wIcrLI z5yH`$XqF&2kyqT12siSm3ELNfll}MkjCzu(H!+%V`O%{IrM;ld-+wzAhIDKL!&Rp zPf}C!Pcg|@X)BfxviIK<6JiJtQ0L!AYg0uNODd!%V0r^fo99(D%&rE+Vj2V?76gcX zC+&AEx!f_%37(sEYhzyc27p<%cYVsnvP~%cvIIUd_13^7p7|TcDzN|I75g#3uv6;i z;q_I2dni{sJg8uuXEm~HP=I&HF_&Z;;8^$`T2FIqrPqax&v)3PlD&GVFHm@2AQhJ@x#4*wl#0I^j-q4$bVUw3qJJywIpzkgibj zf@b7;(%w0;qf;yp?YPdU-+vdV$z1mE_oC@BFXJk67p#v$r+P4BN_TCZ&;XHR}$;= z@Ds8Aa^|Eyc=v87KKH?lR?AGlwR2y-vlDLf`&9@fn7uhPj^rB*ye<(CAsA4I6IN2? z-}Nx1u=F6}-=XI({mbp)_IBYG@|=L(&$5n_g~~w9C7%=Qb5ZYVjo1Q0unTO*3F;?z z)7qqQAt#+h*lIH;Xf%QCZe+Yr6BxF2hmSL{CCUT!rS`sQ z&G)c}Cte+w-V8p*`(scqc#4&9E(9j0d#|(bLeAg*5UzFrE0t)n>o>dudQa0d(KM?| zn^7s~!y9g4p~I(R8%6y_7`pRa>}lmpA_!y>Jo|CD&5nFdGhAki{yDCpB^)bhH>H*I zP+c+QTRy?i@F&OGD(_XdFY|XVnV?Oh5{KZeCH_@|Bi8I7ctFqHBTH)f;_hRhP-Ed4 zybKL{U%6tol=&xWWC?RG!d*LS6{V6NF{oZMyqEs5TlY32C9ueaxgVWo(02wc^=mq4 z&f{q!sCcc30Nlw!#vfcKcxMj^!MQNLaq~uQ+fQmb{2T>` zhO+j6AxsNa`u)pd|C}DQzp=E*G$URZZ6WJg?+53;vj^U3MOaK19CB>c=0z-_^5q_m zP7PyVtN6p+Xuc3F(3rMp%0BGY#@X=dDdF;EQ#v6L&QZ3|nX+$cS^ybU)EBI%Z)HEB z!`_!dNW0RxcO}VyuzdhS9+6YrghUlYy#VnI>jYd3Nb&sI7;QFYF6HCq>Jaa=i<uywLsKJFZ zn+%OZh(c^-rth5end0w+X%-Kt;h+DGXmM{xef;Y0yZy)L$z`NF_)VP^&3VBSw@dj3|7__ggT3 z(n>^Qm+DlUG=z=U3$qy}NASb_%A=TRVc2(R<(KuvY&iSIqm8rk2b7l@@CR4Ml2t&A z^H_%|_Nf)lTtKe6mRlT*BiFG_v+v21xlUui7ipT)Nvzt2C%j9StVhu{rBJM>Vc4P( zDPxezA{UzK>G4)b;_m?ewwh1Yy~jv-3%-H@HJ#hv@&{(f9_wbK+lEaKG);I3{{5WX zde|xG<%}B$xXnSuXEUL%ScO9fn#?;={~kfXwS#=C@|-m?ZG;B2Ike5UD!#E{Sr5%y zy1Snv*7{SgQ%64N`HmHf!HGUdh(KEV#n2e>N(Z{GYRl>~}lfY9Xf( z{olx6Uggoh*EW=TUz)WxZmvSf!*${IPX*<|3lOJIm@bom8>b(om&Z+h@3C3ENakk* zl7KY)yRkps(q#>{Rs9=K#sst8bR^jdVs2#t<2)4!yJ^Op8DTL80)mcxp1=DeBoX1^ zjy^~d(D>>qNLn6Tqd|(d#coC4%5@LGtDF<01op$)z?H&zm+m;3mYg)9){69vR0W#gyu9fI?zth6Kt@(n zc!~22`ZksCuPNWp(@_UBaZhTQVZzwz*DqS_LzF$AOhGR>-lqCma z8b`$sG*~ixX}QoWUf!=+A<=gf23Q@ukO&1ZYjk0y4>z2e*cug=immxm;}exofAd!m ziU}9t$5AMQ6PLSK?)3iU6jS2AuK5(xpaRtCB6D|7+q1**hlXFfvAcYEgZQgr);UCz zi7xpSFzZBVw+zhveOCv&ofyJNAiYHA4{hIwA6hsldQ0##%LMTxjCKO5N&&bhJcdt# z;hACbF@GRm4hqGWRw>)+t4e?>S9SRms=51zg8fkyh(20a{U?G20q0F0c&(Abl(Hx)3+9@DUs$wrLLvev>6efn0-H#)TrQy6_`RpcsLQP{C-&sR`jtVSgbo^Ig9&%?-FaA<2=k4%rRt}o6 z>l&9H(~qh%wA{rO-j`XHbLsZOMG?DfC-f8DOUYAe3si}5{hVws|4N=2){}>xqP#$O zrg#{fNWBf==n5BS6~^QP|S(9a-VA_A$#Z=S_MPuhFr&>y&u8Mc|9Fb8Q(Pmmvd zh1PqP5~Xu>^-vBLf-h(jNCfT?a(~1nwGh!UFT51z33CtcG19_x--n3!CcwQ5Fpk)q zqd)n*kIdZOY8K4Eb~f@mAOv*wn~e7rFPifjY$*{+Jn-WxNm7<6=L;jRy1yVm22}Fn=6W58|r9OFaQ>tZO42KO6_+UC;1ZCH+iML~52QJ=2 z>g-Rxqpj^zrN_GMAbM@;E9~mEaxis!#3h3toMMZ!1eO*LJGqH?mSh7+i`WJZ^Hx^O zFa^i3tpLNec+yzEqz!TxWtxeXft82ZPZeeJrvi^Ua+pz@h1) z;I?i^EgL@t&j#d{3gX%lrMn|}w8|G%Yb=zElt22y)D-97M-kMVz{6PQ9wGuT-TDmr z;`bHLk&P*$3j^D+5N7WQ`nYIBEAvO^>Yn%8P8PN+ZQ0zvLKQ)g5yaTQnb=f z?-g=*{SO7&1GhbxNI*hi&j+;^B^2!{(HgWbeKl$XNLIGMmU!$=d1p?#a-R&#FP#1B zcsjE*M$$i|0CvG8mP0PstoFBXwaC10RTn*!*8TXrxzoCHh1rUH=&3>F2CLcSYk!gf4>~hWVvwpbJ z7}qrhebw)pJ8&a_f|Q^NSSW`-~wzP&*E(J3AqH?9zT(VU7Wqr1X3hM~;*yN@~aFC;6aYhbFBv?Dq3cQa77z zh;}o&P#G-bdU>9=!TKJN)o~s?d3zx(!EvP_LmA;X!?Tx`zcENIYM-Q@Yw(6=zgJ_> zjjr-fT>k+*Jj|DmxaCs~vA`m_+>10}hx1_N^UEF#f)fcxNgHCrvm`Bjf4@|9s@rp` zT?)r_H=b_ax9Yj=%pyrOw4bo_PQVaulOYto@)1bJ&yokt)3AoFPJ-n!eQ;k{B{Rg; zRae8;r#!yqjD$T{@w+c;z`Ixx1j9-L2(xWmcbAlSXnOs}u70VLLY?p)@QYQvb^PYL z(-+nX?at+}M4U|9+6F6oe=0ZEQbDB>{SwhA_*yhQ-Xn~z;s}7z6h9Fz`Xda#$cy^9 z650E>G{a^>i}OFZaXX@#!2s|*SD zQONdcbWvf}%}#cX!=cKdX8uncOge!J(iwj}_HuU4-ZNg4X`|E7{9$)~0#9R-v&@S( z(Oiz+&cR5)$;I)b8d4f2{Uez~d9;}(fhKb7ElJeHYtnL-HN$av0D-Cw{S zrDe!9BOKI(vuAI|>XQIgtsa^SDfODGK(vLmddm7Njt6}8`g%kY*>MU*s7V_?q#M@}ab502H>A}b@swExc7 zj!)Bcy%F)d(M`Um?swd?3sFy2EinZV8z`y0{jsN7N zI;8%f7;|D0f=Qqvma~^}cqPB2q|HP?Zh4AHj#cRnCr?h=cs>9c@6A4coe`{TA@ue% zty8Nljon5RUuh7-KPXL`q`zYY_+jOXy0d(>4XEtTw#dlUb5>FlnxAKFZv|pl`J5eNQsG%J644NK^lB`!&Sq zwR#dM%75+}7^Hgirlb*#GP!;Pg)sJA{z;)#e)kJHB2I;rXD;y|+@IvjE}S|I#LZR% z8Wy0OVgya0j^MSO`qL1C7Fn>hapJUwM`+so)=#i4-;$e%%M`e&e0`+C1>#uqYTNA%So zb^8hJiA?s;o}KGQQW&F~9I}_0o()4`uba-Z<73^R^;KV#UYU`0rEoEGJ!n#%i=0ORzoNrX>I%Wqe^@G||D}roBB8bZx~GQ#M%XPwXwEg!O}`IX z3ZEm(k|zx3TKy`^u(!S1P$~bp5CqfLJ{e(y(C7tk%8cpr(*}LYDRT^F5GRZaz|ADD z^aV?ctCLGkR%^T-y?@mL-n+W2Qz&ijv?D~#A z({%IJ8MJXAo)x@d?ubC-M{?L%_-!%ly?joKWW(It&{gRT5I`)V_%i4JGP$cz42xF8gO6Tvp`87Uk zl7KWM!{>F?s~+lG;lRXkv2;SZph`K)03kcD#!G-bK=Uq#$I-8*-ZmA+suUSWr(d@? z@F@dftPZkq3X3ARmF>_f`mPdOmU69RqF3b9l7#xOe~BD-P=qWN@dLP;HIt`lO((5l zLRocmy>(UKrYW3%d#XP@Z8?8FTgS|Ys3T!AGai)}&kNQ_I0?dCQ7 zT^rgIsJwM<6Pw~_LjaPf_o1VCSMTv(IH{irzaQ@J-;MCP34_t3v)fKSWROm3rM3kS zNt<4li$_yyxxfvBKpl1*C3)cseSmx!b4w^fpHU0>S|piKLAV6{T;rwAtWJMtff7fxx} zDPxaCF=9Cnh4fUOkiP9!Y;+rl;iDb@y8T?!`+}P+*R|-HrX=j`Od{FXlFUklO2$h(zh{+qP=Rai!88_fCz@?|q{KQrw0@h|mZcmKRK zCSuHE$7-%eM{cl!Ee zAM?j}E&oSbKJDMk)>9Vk)ebde(=5ZZ{)K0u+lX0Sq{n4Qss7&@1_ZY9+i^d>5=Aa) z-Oh&YJTSkm!F$JMj+_iu}zEDee>Si8KNg&)o#SSK*jvNLj)I z9VZ-)vWO=$jzD+lk$rVVZ|bWQZu@+g-BwdM-t6#|>ZR?7IcI=`pU%$`R0Pf(6q}O_ zOfKohNZ&sZc?QG{dDvx8@XkZ*P45aAzq73hV>T~L1=Zz7(?9CIqyZ^|Vtn2OaES4> zzJb+}OtsNKHtR)UE}!;K+e7dWQ5O=2skqc;H_zSt8hzAS!)@}xA$8>qgJ6sWa z^x|xB{5^<>g75D=Jl~6Ll?)2m;7xMZ?BkGKC%xu;ZuI%Po3N-)8xGmfkh55`=>QzfmKy_8!{*7(>Uu7=f2Z2RzxK+G@ zDola`5n%7#C^mBy=DzXp@Nn50Y@1w}d3rubN9NvGwMmJ$Xc@vLmnz ziS$wgfjNFmq{$2zVTlWg9Lp*bu|D#bW(JeZ!tPryi2%_#3akG7n{;Tz!Fn6nOXR2A zp4b5I*=0++9cX^r{O1AVh=RbBGn=S=l0C)mR~e^`vQo_||46znjO;(ub+<27x8S-@ zM53p?dhM3G)d^2}W?!X}zMdCeCvI04bN}0o;S|Bf{Opafm78EbA~M#T1Ne?1`tcWb z>hPw1EOsJaBp*Z&$%nk$zstgz<#p-b;aAAKRe_+7G6cIwdvv+=4RyYe#k+QnTSF;? z{nYq~0^bnT5ENV}A{_8caT-Ob-=859hPV*g|>jbuqMUEeb208VrID>Sk0b} z0LT&4|MGVMIoLCpLou4hyy%O#R4$kV?C^}f94}jp**#{~@HPcOWohcBcjv4i-5NS8 zHy{cnIuIKBQ)p>;O&8cLx!YwpNLta92a>i(d$cZkPc%7F(>#(>NwSR(H%Wm49WIfl zna;Os=l_TWlflH+|=xU7X$PSDL2z$!f}L}9M@Pp6)T!WH)HQ@Ps5OTO+`*B7)fDPm z3{TNVJ890~AfEgKpv5&f#UVdHgo^D8O#;ly{?T;hv>6sU)PnOIa^b}IaLj4Ilw2=Tc|!zG&3oRR$nngkdz z<7sAL$>^g5TDJC+YxMJ3q|&2V08YZD_ZtIKM=<(jz?KAXD3PeFuUdD7KJ37FoL^2@3 z{XEfeg$|$NVjzH83GE4spMeTCxJZ=t+6BhQGAe-lc^^m6O0HVv6Hug-hRU|&j@SfD z>k;mk@hGBod+Wrhrw_N|FKNnkiEXn(m11`~#UEOXdA#(>&JR@uO6OJiSWj#={8_~*}n3!NKT`k+U?|4>~jAQ6wlG8cpEU`Wa?MO9 za@}Uc4V9hZH)%-441L+^JsyXgg+=|hF>as-xky}8>;Bpdh8d4D3BjeVL|Fz25H;%} z*EA-hBO(+gJ@GV#zBNv-Y4l{`EnCqBlz$G$%^_tg1`SkPyi>5+NvjD;Qxsy7VDP_6 zN^Q6IR=O19g0zvI8rL}TFy|)R3XRs_R<_kTCTOUDCFHV>CE-+kv$x_wjdzM}IzGrw zl&ru1;OIK=)Rik*b397vv0YR#iVR`GDcb;FqVWaQIuy;Zcn;A&S9kWFSdUHD#sp#Z z7O}ruJXA`Nrmo4L34GGL+qjAC>cg0+F?LBKDDbNj_aH^cqmG9Hfx@bXF(MwF%J|Or zy_5)*j#XqC?`%(q#PJ0imbe_Xg<3uaA0!;n=$*+>{Pq|`Hi9IlN2%rYa<8|47$DUK zkZ@#HXF$<#3MSH#zxvdxGZ|*J{H(WK__k;t6>S69I%iw*Pju3ZnXK7QDUAf!XR$vvVGNRSy!dzy3@*Gh)c|ooT^vj3T^_t>dNeH>G717dXBmy*#<3UREtxd{ zv|8o*>mj?Ns67bW_U;I8Y8V;(R%4R-d{J&v*;B5#QO23MTnmRQpBwvWxc}Aha(*uo z1rl3S#E`>lUyQRI<06zt@&>uk_ZSmfb8Cd7$Amcb>&vt1bo8_q3WOH4&Ik%L@-w=e<68bdUJ7d-OnYlPMRPNqpotQ zfeKPw9}vjN2EXBdqW)zwTjXJ|+YC5OLQ=`oE@aJw%>WaDmdXkYKJC%LlWg-T;$BUK z7@-;Xq=#?e`h?`yzBVrjUnzVBj#0tHj@-NN@A3Xd%eaTOoy=V4dQU;;e6oEPe*wGY zl0E%SMwgWo!yJ@m`zIuZ;&3ZciH9y*VkwJ#cIBg z`eT+oQwk;oEt*Z0fj$}~P>CM_vttfDkF{RRx!4gDRw3@Jq9gST zFK5KPm^Ln1_8?v&YuKa0CHP2=upR)^!Ix7JZXUobp{93`rDV3xk@-QrmElJo7tWD! zhJBnKpKN+bk)5oZ;cNWJlHoL6UdL1lToEGr7bbIJ7^6+f@#ce>v4;t!(kQ}KANsc`Z=`aIh$H-OicfEUA}BxFg(`|4$ND}t|vxc?^07p z-z_)Ln6%&GpF~KmsU4BQ;+M$+m4u^d7-+`klgj+)xeE>($*B-@4oLDb$5pzZ9Bq>Uzw?@Fi;{~zTI zfXXTOaFLfErvUrFtoA{C&1Yy>>;a1d4Khg0jfvKE{$xELScK}$5!8#@SDio|H#2Q% z2}{Xz#RiOmb{TUlen-+{D;yqg8~;J0QJ6DMn=Q95$+eWK5qQ@)m^+600Hj%45*NIG zARQi*?|^P?q^+7szQ$c{dt(yI-(4WHrGC`lQj{+n&)S%%>cg8snt{trb&v;DqAwL+ zldg^udargzSR(+xFq=!eFi23&?i81j*bnH{Kr-r>ar-QxHA96VGC$%%ms9E0^&nlu3!F}C55sZ8Ntk` zV5QY~IvM0XS=AlFdlhu9>UOc&^nhkkob&9PGkqE)7NQc=tq^X zz#Y?AQzh5aIvufx{zg?JI(IpTj)wD_*=bo`$PE8X2OYaOy7S><)B&aQ1TMHPI_59D zx?eJRjjRdFR)QEgL%>D)gsN~?Oz@jb>Wle)!%dB!lVUyF_6_e)UWRXa_qLbG#Mms- z(>bPu-V%Z>Ggf-j#OUGSZjW&J<=hi9dN*N|RXCw(iA&UBf$Q*73 zX%qr?*rI7RD9HEAX5YhP{}zv45Yv9R-7Ojm1b9FKrQkPu0oJVz^j5sXysRO=i3_ts zGB01g1W%e)c-jom18={B0A=h)m{F?it*k_vtVffi3E*UxagbzPhPR6C;L7Q*HMeB1 z;ptBJtBl-ehm=GX<}w!@@R3wT<+}o|c^V!B>e7b+QU#U9#(y)$%%T$Wo|7sXwu|1f z@-)5!fERbISFyNo8Xqdff5vXufs~Xs&GH&Om$$Z{Mw8%TD?EMb94q0bMZXkDl zho4ft$W&};j!8XI$+f-7iCh4iH3rF3`^)unP=5o6zQ}PVi1cv~+Uqh8#E>Oug?~T% z+nco%KXS#s2a5xptVfOT`qJA}2(2LK8>Wr$jyLvTHpUZiE#<$nFI0pIZLqc<880K{ z);7*Wii1s@bjm`wuJemlvzwd*{o052qDbGf8rLOjgV^~#|F0Lsg#yCB=R*{|yx+|g z{@Dn6U^i#H)2=&61mN*`mmr)ePI}#G^^a%%8i5Ci1C|nmCO3uQdBeCQWpD}4dV$*sj1?7xFSz*na0pS%PzdRuE2H+8 zHh5kA?Soipwn2+)_dvm5wm-A46Pd#K(j^BC>2>PEWPlEA^WAA2;PNIz1z^(^Qz#Or z4>!Dt$lyz{sbG&5NSoXt(PGG6l5s@ER_1!>XaimS#TxgAIaKwdeCw@vIr@#luNsto zovK}HMqPUGnTn#zWZsdCSHCq=x-NYb>PJ?@;wJUz`?g0xnd^2RN`)pcewY@TMC;iwpJdGAlbC#t)38c zfN`c$=|@|2>3*eyMa4Omz-;oCh4>?4f3_V!1!1b;g4_#F#-yH!J>d6$$aByz{O#?n z9}xQW4mD@x}-mfbf zg6+rG!~NjBE3cF__9zQO@3kg-WZ6;`)8HW<*9YXPf1u>hNbZLdMC%UMmPrGzf6HTR zMQ8)6_MTmHzts*~;f0{(cw6q0ZPyAF8?Q}P+)4u+!wKi-cG|_>wys5_$c#WLa2uf@ zd|x9R=aZ{)*x)L@84q`B2oxv0_on1>Negz$hWSvs6_MLM0lpf4cNyQFGpgbxes8d~ z0;wpkdlQ{}Pj6t@w$D_Ts)L+=Hnsd1XN>MFj`M?Zvk{7Zc3!QMX`F3Pqa}15pku|S z8dNcmmRSWaO0=3e_t#fg-jD4XFzx%DdLqvT)N`qkc#*7>OpB|lF|~c24ofNuZ;eKh zE-+}Am4@R{k|}mECkU}zjM<-SM&={#FIP@OQoqXO8s3$I#ASl0C zv)Ze%I<3cF84s-M&wa8-OX$LnKEzR92_Z1V3_kqr{e3rH&N;c$EN*gH=g4S2Kq7>M zE29C=EwM$p_)R}M$WJpH#`hPZ7cr3Bv$#O7h&UeuJ`^xad+tMZVquZ3S+H+7n&2=+ zOp{RhH%NLlmSTYL^2Md%1W%8oQc>OB+}*IQl7W)=$CzqOe)kwe?9(eO33C>|cJR|W ztMvpnd;)?UrMmp2ss59OQYv<^(C&UZwrQoR&2pO2s$F&Z{kgu6o)dU4=u-k0$td(b z=tb_v4G`D92p+C-A6<~TrT6S5*|SItqi^o_=syq2{qEY0*{b%&Px{Lg^h_D#<5uqj z?5RD)lKXZgHpJMmgVEbgOagyIfk^j_dl5ld&1!<2L{>HOb%7$PsUBfqb&VHg#NP(Kqi@O~R?9#ZXk&K5}yko(i5KS`*hP1$Dg$|5IV9+H17K=LHx@-GCpKK#fuZ`hZC$==NLyh$q zz5EmNs$pP4GSP%F29J94ksSqrQ6hmErE`vk;w4LP@--_G<5yxB_*XLr85rf9Hl2t& z_2{XHf|ux4W4S@;<*Z|oDN;mXV2``v;90ev56O2j(cwhCx}PT$g?>U~hZgTp#^gxQ zvi6-5o)>sec}fyiAAlNzlC~;}i*O5y_k0-WyCZz1AEE8}xb$sQe_wBtunv_FmcoRw zAbMfQB*Y1F`TKXty; zo+<|lpjIg-ttSEoPiFI`^p^Wq5PjQhhR_V8Fn(>m>Yv&6a0+U?R|#pi5pj0(my6l+ z?=0d#MmEfmT**fa20F@qpU=E7Y=&Cw5{^}9v0O(<#I+>~uk2vBtS&&QA4pmVt@@*k zXX+_5OzpuH(2D~1!^S+ZLC&B~_i|F-+BE!9SB+tW!JB3|npaFwcuVVUy&Muj38)E4{0sMf^ny%1 z>7I!K!s!fs7E(vCW1s?UNaDnTVgbY(>OJ&;pwbr8Pnam_HY`gyPns6H#0j-IJvR?EnbLRa<))!~az)!_3u z0LW<~;>NBZe{B$kjRuHDG3(ZIyK8s~QyvHFR|y#paK zRBCR^`oC)jFw@ak?OX`AiZm>{)=jNy)l*e`_3&>j5@dPRkVlr*e7g8fp;}3iF`~ai zS3BkGlz_tL;#8`W5`n|$AK@~|**uP`WCom%Eo9AD-cL*h?XywWl24~${y1lIeyvC* zXC8WJgUB)Uhp5ndU5I{dD?PueD&re=i z$2O>H)jqgs`oIl$ognsVmL^=p;hxOkqFG86#m~!!}7dQjfTz7_Za}7G?aqx>}$@OSx7JWG?+Z%{ZY7lbs}eSIc`P z9`NO|X#J14u&Rc??w5rh3dHaQ`CngX9{;>D#h0YyMQluM6V1AsrwrSlw>Ou+Df#@+ z7|Ykjs%e?7z^?nTS>?(c_p!nTbe9IaEy_kl&AMCD5JLkA>o&I*7n`9Sukb`eK&!uJ z^2})SlFHgpbrvIssPZL>M4q;5lK#Z+Xi|BYajDGcsr!S{-KktTn(|^U6K**w{jR#&Ho(WW5$>aNxh9E#qmm0%%nyTEY_NE^Vq2 zZyhy8l87C6P=KD+BShF?T!1mH(yX-35(MjD6x^dDjkJ)f=6HjZ|2wnsAQh!D zhwG5IGahyCPbHd7#$AD0F-i8a3eRgO;bo6?Ui8VOP6R;e;5q^+K=M6?!tb*2qh zf>q|UKt)|gy5`REa$S?c!syxrT|YFP=6{@!9j*rjMkyyXAix?cjd4nWlCoepm%!j{Q^l7F61YHZ)LLS%%T-iJQ(`6Ib4UwHc2U1AbOcekMm66!W4 zV7i)kWAcomP_1!H&>Oz2PEj&IS=9<+;#JaFHMEpmq$nGWr&zuxVMDub>ji_Q{}BnL80*5%1T1=hlppq&!ul|D)_@@nd@m(S}%!L1Q>(&-Oq!n z_IMpDEP%>%nq(>_%zzx@SB*6e=;{DRxbDw~ftErIj$_&Q|JH{oc+sOEF2aD}og|vT z#!Tt*PQuBY!992U%;xZ4J0HstC`5$1nbotG)&T^;5&sE>-*FvBv~yySgXJ9y?}>;q z>#8F)znvA24k+g4=Kh&FBDW!>$B<@~WjUn{O&0W2F7WX0_&DmH zi5T-Q)WRxUcw>J;$BTIxX_CE(zalkBY9i}U^7gTi|1FDb`}m>j%}8+$Q3Ct5%^&$y zkOu|YEa`LaTrRZl+7yKt(R&hMHJ6{1+jQ?vcH}(p{@?e<&UH3Un*_5g0u<; zQX-Fg9yZ490tyDa@ocKXX3;MqL2!gW|HY|`sFHc@DQPf9WO%$4PbtWlI@aNDB9K`+ zL3O^oMR@jGhEKft_I`G{EZ(?odfb&q!mlMKwjKSYQ<22l0s4ChKd+fU>izY29kzH# zkC1PJ12ju8Nb^rK0wrMRe|63feYw?jw}))Um@mp0@u97qO--^Wv)3WGYmnUaQ&?TM zJ`7*{804dh(Eu*#H=%}riQmGueR|~r*gB1_cx^i$46C<7?Ki}{fxSbE*rLbcY(pD+ z@!4of2@ru+kWVE~I-DuiC=TXr1j0_sDgW>n$Vra(5{BD8JLunCT*&pkQ*>)~!QFne zI$8)|Y-XbvqJ^0KO35M<;D%tYn{H=aaj~TOdDa)aOj30eVx0S^Vkz~zDxg8B`Y6M} z*SAB!NVH?eDgJbX{@a2AWRG)M_rzx6NZDODJ)o)G;}{;u`&b^TNNEBo zpiNSucQ1PJFPF?YFJP2`HQpt{j4$&eKB)rE@;)NxN=}%DMT8bC&Q|F`C|e=iX*xWU zT(!+uBaGqp*AImtF#N9^HW4^lhhm>q^dIB>W~||SQ_cqDC7wlk!dYCf65xOIh*Yti zYoloi7Cnxk=!>YIcGk%<|9rUYjiuGKq5*xF2K>Gy(k0RnS?hGoXp`nM$e1qLBwOpf zOhC2Cjs|v~{jy9fAo;(Ope!7ym9;frB95MMC=+rTs47DE*gu*0MyHn@c;0Qr!Cux{ z|G{pQAf=qoSG*=8!pG4b`!J79cyn<1=bg}w57Kxn21ke%wjjqY-76=pA%W=%&nZ5C zM~IL#jS)F0e7SD!A1fx?HsVJ6QJ#A@&3F|kr;*7L;Exuptb) z)s!yy+^f?;W)s-C`}ze*ts{l#Ope?XSxj`#J;_Rza3Fle4r%_29Ix&m&6%OFg;yq% z;jR43!fngIAPz+nDTPuYtX>&;ytvn~+SaigK&si(V8vL z*tXfRZM$RJwr!gob!@X^?<5^N-LY-kZ=dts@4k26@0~IB9$BkuRn3}n78+#gV#~4f zyMuUsO4+*{$tz@Va6cksBJEe&tqF{=fG&>_J84PsxVYGu*ih=o5Fl?4z^m5tpage` z8Iey%9Qx+~=U;XRI}Dg^s~u0wR3CnSoj~Qlz75N}E?&@wKqf|&Z3atXqeEdgJIL1L zKU!3)2oNmOjDm|dzDNe0sM2sVcQOq&hnYrjLF|6PLw?-&>e$O*VpTX!IMCGUE^Wvf zl~bOsIAH-);T|{amYai~$pdJntne{dkaH#`@261JJqGLa>7B z`S*&z^4F(Xnmge4zsXS>0w^2{00n!;(XbD42nkeKk;dy0=abIrG;q=K0msVtk|ck0 zGaA|edVp2Mxo^jMTRA|icSiZ(E?A@ikIDcZ?RFCW1$7~?2C@T;4_FPOR8N}p zX>&RFj#U&3^A+MDDOIQlJkRc`N4}zUwx8dCS#uYz6L1B`LB&#^Qyw8|6t#=>z*;2+ z8Dq9^nZx6HSf)}*C2~ANVDjzy7U$Io&(0w%Pk!AO?Pcp z%HwZ_%r=TnVoOkZ#r~ynU#}5~ddLe|8(|^SnAX`4MviSMz7!M{;8wUYMn5Z6Z1xO) zYOQy6SdI?#U?XQ5w^OqM?s=G19>kEKnY&b|I?{D?YK9oHrU`_fRHQX%h{fyP_xp|Y zkR@wO?KHo$-vnTurSM*m22WQ^v^3;7e#2T52A5S~vn;? zAwkIkLEz)y%(X_oUi0gW*UFYADBQ2pmgf3SG{8-LW(ckQl-D7Vu?@B)%taMU%R!rx zIlIvkHVe3b_;i`khojbKgtpOi2$mB8?6NQ9wSSbpBy5@!!!vj;ZF}Pwc%Y z)hZz#EI>>l!fepGMzz^2jGigX-HFbeqgSWOXXQ6UGt#MJLq|Q>#^SWzBKXDyloMP3 za0Xmzj^_PpGA4*-ZA@>yM^uM80dY_#{GHKXpt9XzoKNNr%`~c+ococ#-xz(5Ngcs; zLcX7jkf6Q4_KUixC~%ow(GQ+(T8ysm0mizvm2i0f2I_&X8}GYhjOlr#Hr(bGMk$rJ z1;nmw`WIqIZa^PtmHEjY`#FgNdtW06v@l^s#Ti2qzO@|sso=MYh!t{?y;3AYD(2vh zj;5xj2FJxDdy8n`U%H8bP??;}M&m(W4iUI^I`F^lAF+n%UWM8B^S^I~!((@rC;QJ! zn=u>hz;%*@GDJ^)v#i=d5!T(S%?Mr;#ba!ez+u$-%Y+s}{2~$5cps}K7AAWyDYs6c zbZvhe#gJ~NzxU9EyfO&m1%&q^`8)v6>1T1+1dmJTpq>P=9-7%NRKO|*PU#c52F*hs zUg(AkGabk&c7#yFuiJYrA?EZuuovR4di^nR(9CpI#d;h~pu;>Q7TRNu!rktNp$~g^ zWFDgS0k@~1vb4sY#z|sC-q~u{z-^Qh&|jNVf*4&FfE@cdri5FeP;3s z@5_LQs>jfl>c4~`V>O9#D=ZWe@HP4ld|vEDJ!W!Z$$EKR4irC~E(4vmc02?92=-$v zp%b;qB7Pgqk25shZb%`L8?uQ={A}YW0dA0;G$lpOz7Qx@z6_lRi@8cekwvb1F6bPl z?RDd!`1eTI#f;7tYs(uNo;VvyUwP2nRx41mr5Y>5Lb`0|zcPTItyQ81PQv{IPME@2_oE5GO*Iez#+iZw~*#FG%XkjOhMegmkm83_&#r-801w88SB-{3uZg$u`kzO_zGfdS0&hhUl1 zx2=L)8kF0Ig!1OSG~9V~&=%`rDo@z$@6D~<;43&v6+$nQPRZ)UBc8D_65*U1`|6ym$2N4FKcq_yg|xA$biIRd0zInL<=p3 zF7L4y(ZS^_7{aatb8;U zMs^q`%AiYiK;Tw!+oo#M9l~xl@Zqlss~`D9Z{=vMVG0zwLVPONYLZB-A=j5*8rPWr z%KAF!kJLl~bLNNP^^I{P^zNDUyP<|u9=Ek*0S?$BSEz^n$PMhx*))w)LHv>Z(0AtF zY|Rq#Y>SRraS;{JbKc`-Eeu1=-voK_TBI)Y|wy=fCqTfE16rdK@f%yySj zUqBaTXot)X_4T!l(H7|M?9gwwgi;6&1S@G^Br^~`qx4YL~=itxE8qODQEtDE>-Lr=h9SMI1A%#rtUhjElM zfn>ojOhtgyYVLGb06$IPV zv+JJ3Bvdu${ERr3U_;lsOJZ)QE6If2&YI_*_mu7Y7A(MPE}11DiYtqkPvr6S!6z#5 zD9+n`1bAa{)nUBrDUVb0!Cah7_coB)9cp7?K@)`(SAtn7_#-fvRV-hab*&08IOzjI z5=9igq+s0|kOSQh@lt)){o2@mueyqrmj7Ya&Q8J*05`=Ii$x;kA|o8ftasW*+{k^p zT=%&?oGwshcldE%{R6Zkfx>)akeYl_v7`5LD-`I!*enp!iK(mZH%2c;Zv|gmyyq|< zPYnmk#-9w_Lp7(2m4+`ZFkL3Ww5F8LU7jbKclES>SXqzsN3F<{fGj3=R zAP^2SMN^s`E^lfNLHd#f2{6lda`%?Nq5>udQ;*M!YxP4nH>!MR9#lnV!fX)cI4Wr4 z>W|hCo5Xws(CNp6)9CvMODd-pq%p+7#q`_GlLIA0Tr?p3L~q; znnQ^U7>6A{Oy+=BpPoT{71g<9O;D?s*gDLwuMH`k^R`I@ESRc3Q3brirUJ1LJ*Hgr}A-VZ8yo%7cFYBy>AsUo} zqYud14}I{3TPF)ul^AF!^8$_P&sF-#SDfR1OU=kDTAgQ`D*6pZUn9qT|TMC zv|}K1%X1N?!k90XQc^oeyg`vR<>d{SV8DnXYbn4shn9eitsN+Rvu!6Fmp50eH>pt3 zDsW^ewW3*gC_omGDQ5(s6ydP-cVpwC6)_mrIZPCy`ywvkqKGai+))#D7V9whiw(>G zLb^E7UdhtSQ8l>VVFOjzzj=-AD)XB{8AOYuSc5Udy*S8LOjer& z2XY*5@uqinf@wv?9X`B(BR*_k(VO#Pka{t#`?Y??)c8^D6E4|TLy|zf6l`PJbg?pRh z4rhteuruV{$`anE%p4q}L^XkgVQDl6JqsKWC2@ zl*AX^@s308NB z8FqcA;|ZRYYl68|OBV#(=8W>miNUPi`AoE$^O=v-e)??R4(-@a2OJk_Z%(1~w6#6i z>FRD72%N^XDcVPzyAp9of;=e0rb@DwA*QIq80sfp=X>rd>2~6CX8FrTQuD&`?)A1iMQ{Gz-<=h!3~iE}UHMzJcC!n!LIakaTp#8wYB2qsG6Bsn zsAK9ur6dpK7C{DI(_KE-kY1imfX4A*{x`>5+7c6y1Gj7dmiZk>dhSjb*PcclPuEpM z$CUzZWqz<4J2)4YT z+buC27O^?U$oRu0O69$C^H8!?9Gc`MnsrIBz88uG2`D|A{e@_N6vl+Beg2qO!nHaD zJ!I9eC166n?cXkp8^&nx6W2)=SS0_$1VxMm39Gj3W2s3m;Z$u zJDJ=HfJ~LyfAIwxZ{&0UjW;q*0?~0bb%oInel}GJfnADn_D)G-Z8jG8Vl^W)+b7Rf zgND^{;$8FBK;37jd>dc$brIYr(^eHNS1sFU(SgxAe*HKN%V09Nwod}##pQjjet!ff z445@I0u)cdz1xs$Krd!O%cHiRR#i`5E;kMyBDqJd`N~gJ;;xALPtN0D4D(S|@?a!* zoaK{{wqSRIeo}BKW}+KGR7nBhK6a{-#BxVrC4r3!>ULLklzneq@rvDoR9*t%Gglf8 zq`0VV4Kuak8cfxApoO0O523il0sYZo4*R9=EJjp1$h_XVuAu<|1Yzo&OZ-EPiM_bZBw^}3uyjM;5YN8f%NKHnhO^fVo2Mlf8YoC0`Me*AZt~eY zM{&jX91(R&b=sJYy<>}%?C-)oXMelx0F7uSn;Q;ubS-LdsjjwImlcNSM@F4`cXqtj zdLQ7nZT2J<)=3Mi$Kd12dtNs*DEw^r7lHprWB9MsaCRAHGEd;meYBq{S1;$cD71d? zB^)@r$K|+7ABea9K^RWn%u!&o-^*>;YJ*JqA7I1)1~zf2V{!V-Aa+Fn`q08Xab`{9 zO~J&lJr%~+hv?nAvy|D?C?ItP{VAZzvu1cVWI?QM8Im{l*OZ!W*K1o)%mS!{W5?8m z>vu7Du(Py6K)8zLpZS`M`|u)r*y|s)y%&|tYgYjjVLtpWs2Yh_td??>l5a(d0IaRw z#n(;OX??UJt|qJK6hm;4Mo;~h5gSNv1m{@G@_9HFx<9 zfe0QP19ic&&eWHx4dS-=^mCVC=HL)P*=<{z16sVsd$fN=v~Rl^ic&z+FS`)~7>>?S z%{kgvvn$XPXmbbK6n|RC?8%0|03q3hbesm=sUzSO~@Sz ztvJGDbXIA9Iwo}b3`-Ty`boeN`aBWBlQKU)evdX`XCD|nS}laNC46pHtk;ZDy%ZQ8 zbS2gE1wnkzblLIi`+-(U^(0bIyTm*c7IUMbaXeoq-Q@)`VRsgQ$8Dhc?mkyeY0;cN zoV^y)dl30A{P^#|{~M%0ie9~>vFRBRvmM$Wt@G88#s*)j3S(BZ6y^Pl;r^LH1JJMe zTt3fF!go5O>;W%qF6B$GuWYidoUsVuAWlIYWh#B+8rlwx)y*gaBu;&bo{Hiw_?J;O zrhTa}#s-25(xG-AfE5Ol!r+fX-faEUBWf?JyC3ooFw5C`fJe255*yLrRcrHZ{KivU z;5A@@sHkf$*sxN*kZcrwk3)^Z1*K5JTgYrBdt0%4%!XlKgMCJLogwiR!aU&;4l>=o zQP)B%D&mm<8Vc2VD!mE2f^??vgJ+C~J?@XGf2GOJ92@BNKnzh_EMivI|qp2FQ zF~sEX5|A|{C2V3*pxYB4XM%7J@UCmQ`${cU^jWRle2BeTno7HYg~MxNaTg3WvRATgVI<3ay3Z=@8}4!4)s3>%fDmLSxnFx$EYt5S8QGc!Gte)a=(LpZHCsovM;Usmq{QJSm(M)u_+R; zAumraIpUA=?dI79s#+8vND#^UeamVA%6?5m0wRESsxBM11P3he^28A6km3Y0E82_x zgHy;m^cz8(dwlp+zkoKJJ&}4%@*)R*CxyaC&xe552q-TLY;P(cB=`o_`oeM$=2-+l zv{bZFVWIe2k(VF>6(%GIjhdm8iy$m1zrpor+8=Rs)>BohUEitV>RVyaH@vJtp?JT# zkT-66>(rvPGHKJpv<7%1Ylp31W?5tGLjO!;SvacWk8jJdMpA(JUOhc-<=ldvnc{}= zYZ+>2*wi0wTjfCeH3W5pLdKhJ`C#Qzc>0CzcxnnfO`;MtZSC&4$Xw>$s7r^4C+$K* zJTR3}`(x_P49wox`#xsI$nT}(>7?lT!GlInB1JRjeeRQ^g!@AG! zOWF1QNJlNwAdb>IZ{NS1VZ}YQ5VbPpEvwF3B1IZbRFWB22y((nroRYUZY@34$0N6% zT!!H8W)rXjqrfWzEe;gB&9XutMVb~c1$WC!(hP0zG|Jg^kMTtx2k|;!?%uH*hQUFb zPZ7owp`4Nc#Ty8oQxzD>usz_|Yi)QLaHt1&avo{Y+QzH+)@Gec!tn1*$X{<$u~L5b zQdC7|vDW`0qyLlQ0WiNqL9i@^f>S^L%Ggusxqs&*g1VP6C(|l~T%%~9p?`nJS>l}3 zYqE$>i1kk{THM(1FR86{WCEd**KJtFrTM6cDwIPC;?fkRpzV~CvV}Mwl%al0fkzW* zAA$N5oc&c?z&O`dyS5|lNlN_QSfG?Nz<`cr?!dGU{%3I9kdnJ^9tAX2zxrs+9JB^z z+zQBiZDz42?uBk`ZLozfvy}wnzpTb}rp#f7Kq412HL6TFvsnyowC-db@mqiZMSCPg zAIoe_RJ++qNJzLJ`$5%Lw^{VD4*sS0AOw0Tvu{jCuYV?O`ZCtQw!jS5(CC`m_1JT}W|Iui3gUxET zz@df(g;3XcWFQ=VWZ$LZ<9Vh`2y8RJaHMAhVY!r-9mGB8$o^xZE-UamNNjI1;1|Zi zHv-S~EKewB%L;=N!glxZeBpv!)82XPN^>+(Em039iDy~}>G3j3Gun@u{_iu=A`YUHXGQu5uYlJK z5W#o4#MGL;vv5f_x;clXZ#IaPGn@x{dg-@2%J72dwmAquigcTf9x?9=8Q6D&0w5jT zL^C=60-XI-xgX4Icu43r2k|9`DZ|Rs9c0dgBO7>DKWnB$ZSG|#K$h_OJzDt z!MT;I?9DKk7F;x~o)sE?WfAxq!79OnCO$TQ$0{=0&$zaa#Rw@2aao9CsF-S_${s4v z$Od2VC>^#x=dN{Gpc7$*Dkpy{6G5OUw%f2i!(jmjB(GIRvrGu?w@?z46o?n4y)a#b zJBH%j8OF#)^v)n)%Fs^N$0QI1K(|Z^n(Nq3@}OU9kuIu!$2tZh{o5c~d2N*`Rh#Or z9K7`I#t-}Vr2sH3)YPfqHOgrK^;J2^fiERE1J6W>RUnwC7CZ^K+7B0$Uyw20($bmn zqQW|d#1eunIA3d>>50pxHT>_m4~%e6MTUY3^M^~sUUw;cN_B?8TU)!FKH0M;vTp`^ zvvV6!F9y~n`iPZMq~GqBOfUPcd@tXk`1RX1Ze@(_c7O_X=itfw3)Dk^CJBUxbDt;0 zXMwAM-TrMZS?(z{YRRrCDFaN7jGpqW;)U^$7X>&>Q&CU~O3JD96l8;(Dp4v!k0zNT zC{^KY)MSo7q+g0tHtSM8f!9gfHS|N)e@gxTUIt-ul7@e1@!Y}QOkPtceVHybLpWJT zpd2lrNPj>A-@kIX95*Am?3S}4-+i3P{M#Jr6}kSi6|vm7jY#(pd^SVH%H=B_J!qzgjR4!{dqd%E2 zvk}#-9^O?#=MiEcS=^LXt57~{B}kA2ds8k<^yl?KAn|)+%>bgH1!b?TE~y4s{!UZ9 z=ap~xppGk2d8$&`^Kg;O);6~-8CX2zb0{TKXYt3?(tHmD%jOcJmm_cPUTs8-D^y(e ztKz2CoQMe^VF!=Kmt`)>f374c&<4~+Mfn}!K@Fm*tEP$qk|7(u2mMYA1_VFp364!_ z_k}O*WH)>oo&Sd_-rlQh$KJO&sP=xwvxxY{3Sk z5Bm4nntL3b2B70hQ%YwoM(?{X>D>YNwmq+u%AXc4h90IO3wop!lt(TD4xO2KKrUzL zEAHQ>KJJ@R-8STxzXFD*Pa$9bTT}jf8?2&%a~@>aSw+&r2BQw7*fe=2zXJ>m!j!^z ze4Xpkv*%s}dE-REm;E2vxlc0g)H7^JhV%5I@HrejFd7tImt&e%mtG@4eBh-jf)>=i z_1Nlc{m0*SJn`MW1e96PPcJQ9_u#v!-Xe6v$=FAj*-VX}V%NviK?j$Rg?aS= z<3wo8K_Du)KoV4aZsjO{@$`xMnP79gK*Om2Ar(1T(boqRz=ZGiux0A?sJpziy%xQn z6u|}hXd4<-{eot0&$|a|pGhtWSb~kL-vPR(;EQJswU{*YXZAlVyQS?(o8Y$-LG~i1 zhZ>!C?uPdlIa`Tt1LV<1gr851y~Oe-gHdKVBIaJb;!y}8={_u4*Of6hu>Wupl>VR-SiF~LFTF5eV*8-9~9~!kwL8;q-D>@ zi?6!AhO=M{H$JCJ+a0cj0IFjem{(NNX9(`y-yrr!N^12S+VX=a2Df{UQm(*y0lbEl ztb$TB7vz&2vxRUgvaeJ+*&*1OD*={crpO zO$wmF+8r)E=&ja8*Nv1z72{_dSB6b`9|A5^WDm4n}Fl`>A;?J5{F#-?Gs{(@b85~G{f zB8c-D9FE8D8yL`p%Qfrl>FKE_vwU18G_Nl$;6B81%dN48Lcn1*cv(S=6b330F=;7R zL9ulo=&c#$9hSp{x38yEBeZDtxT1J@#P2_q#eMRaeG&_jZz>VivNyXog!?UlLXM~? znMD?QRO^-6_$52&O}Qp)N=^1>KY=|o**0G?R2P~YDeFZIkAEd44HBUl*e~&2FV4x1 ztOqtwQsptdz7Xe&WjXRZA4;gw8wX%y_TYa((+C#8UO*c0>wX}gf?1l?(%ykV{1N24 zJ?i_khG>pqbva(ttxuf~LD4x#z;HjYArJJ4PsZ~?j0!HjmBd+WsYvA`C17v#8%1+u z7$x{=o;|P!@{Gb8CYH%U2qned$Op+n+rEp@Q2a%8aYPn4n^Z|mhTJ@Bt5YToi^1zh z^OwzPZ6MG(N3tw$*w^oB7fg0^$4hccQ06N|c1uouppk5$3z=ULj4)7At7|g7>3%~) z=Urkkj+o%0v-=XHeUC=+u7H-r`4qGgHy?{ZPKLh+f*w?1xend)2 z!|(a(+;70}VPoB4zPPtvh0B+Sd_I!R&^ku$IAeD&h#~4_Nv%NPH4G0n^}kFYk#EEQ z!+7@;sgSX5vwgy(`4qk+Bebp%xov&{0KE$U*KiB1BHLyptDQ-y+&Vp|!Yl5D^i)BK zvW4a(;0)l=TbY#wV^4-pf!R7&r8r?n=B0PJypops2r^%+haWp zBi2hMvf1e08NqOl{BSy&sN5-MJ({k?{e#)Q+F$* za9)?R`w?pNU^YT}oiefPLNnbs!1f-z@15`+RsGF1y!ttNG zvlZ(%gS&NIRyUlSlxP~PUJrqkVJ1GtC43V(A*=}Uo@|kK-+-&;umhkcaCd7X_um_1CPI}UI7p8V4 zc&)^NEoH;HW_dy}-8~h0khzE}>gA_x+JQ*t-N*7(7d=ALOFe9Ffl932Dso(JJI^M# z_c`AFkK}Ry#E%4!TKn0BwC2r{I?FkRnM<7-a;tVpE-3tH;%3V_DEel6Q@m);!xka1 zav>uG&DDPizQKKYVmT{zUS;wLYjq<-LY)2-9F%98f&b`-LK=OwIet z+G1V7HU>@h@iyDV&QXxE2q}wXoN>bHFv3dVt2b)|K=-i6fhw3`?4N?O>J73+3pnlA z?}6x#bU^CU_ihkIhz2Oy_GHU>u(HozYg9q zG}h-S7mIbMhF`C+m~vd2HCiHDFUljm5KqyfL{GKtrhuSLP*p_h>G1b=y||UbLcH0h z6c)CXqa8-NIMrRD`^DUqh-VRq)~LT$R`;(DH}yT9W3!2Y845Lx|6~97Umyqs)J`h^ z)^5^SYX#wn7jb$1CnQ?;Dh!m2E^GJhG$cIMHppdXE-xEl1Q(3Y`A|p(A*KVCH=kzB?UD|B> zHd9ez;ShKFAa$&|Tc$>9?gdy8oyTT$09fG2Uo8lSU9Rnk46HuH5RiE5G~kK>W!t27 zmWUxx3sWxoi^y4^gTg=etsvJA`Lqjdzwl!swu%f#aUXsA#rt}5N9GdsStv?OQU1?} zn37vVRwOw5Nq^BvPX+N?(3l*Eh^STxa7X*57XpY+c$n9w2F3O>mh@eB(Qe&G1+3e}2OaD1?g%8ZJLNYbc# zK6-FZIBS80wY=3`7}iRXXlJSSV;I=q?{keFy#~GZ*HP(yHZ;8>NHCQ8iQ4U1+^bh~ zA24(qu)&PKAMA%s1zvA7fz{fda607rJHqOipkg9D`K6^~59(jT)1L0lh}E#R!h%;RlqN<^zDUM~A{hL`AxxO5$DU2^hF2r}vOsQ%%f_`K|2+Y-n{Hcx zxXty^HlL%Ojm!=FzG>#qHY6OHJD%r-fy{i2mFdn#x5=?pS}B4R@$YS zdCOrwm>S?+I%-MG^KziO#r*u^iGZFkZtMSB`2jJeU=W3&Uepu6Ntsx7;vY)N<1aG` z))0p#s8*O*pH_^582G$GC&$NH^D8SS2Iv~RHex+1XFFJ`VhmG|)Ulz|v2#w|P2VTT zY+%ZcmC%sdSHPgZc{vqY=ykstF&2XUg24T z=(X~4tJwvP&hJ1yDb#M3S=;myB~MwF5rp)Q)ySOpWWdWk)F_?xNfDjs5np#+DZlXb z=^G-zw}f8w`MWh0R}d)4oI5R%b?XOO@#mMKv5ZDI1iTE0A(fbYI|ELH%bvGx8b#=9 zJ^Rz#YAPR9T}n;soHodWN(emzrufCk@Hy8WMHPxrnPehL zvzfhin{(r2t^l^%Umy5GaJuT&W85+U&SS_5^AI;j2^^gcxV%z_2erpV4g*ar2e(N5 z`SKP9^t9Mzzq^!Ce0ywZ4&Qn~jN)w9ePxWUzbfcbsfXG#-Qkx>s z`OY)5xd|K&fo9^LG7)|~arxUGO9f)a&YGZn za?}5JNcGP%Iu!!Kxq1w7zLEDXh}@^Me$8*;$O6%j>A{fa7*0``A5Nu^^9;7$7c3tK zWV%jH%qB&U@*gHYw(R;?Ae!I6ag$F;WQd_K<_e2&3j`}-X_kDQ`=$zs=yQxi{?cJ_~~I2?{Lauz=9(Rw4aIW;GFKB z1*-})vMaXfil0wUPMp^+R!fG}(4#;uz>`20wc_tQ7I);gvr9g$Lx;OrZSlxKg2m93 zex5f-KUg&mpQG6>Pq4XI?NI?Jn!rV+0~5}m@-N@k^!fnC4W~bL$_K&irb)LizPhfp zxIpRgs=I@3y9m;c?2pD78XSQ}5gGd~Ji449T9=S@K!PyN5QOA$=f|lcu$44S;>ZHY zvz)aMT1ozk}AX{(M$9iGmj8i`^{L z11X~0-B0@5v@}_2eO((C*j@ke%%+v{Io^LzTo@C=1Ik;U{Hl8|km$_jg;?~^Q6zai$nsu4hnaQ5O|*iRRBz7ZkTm(U zhj44Xma5NrFt@|*)RF1A8d`b{>_19}|E>0)QGzHP?;@oWI7yF*dZZc)8rggI(UZ}p zXd(eA5xf0>>mR@s?LYwxB7Q(d)H*3T5RJST+Wm;VHb{7Np#f_6wZRRkf`o3;dE8I` zs%RS4y}VI$o8Tia>U&zx$@r`iVoVP7d-J~eqbMW~J-iGd z4w7{Uw4gXVP}?R3chMCpV@qomi>|Zw88$8L0w?BV#ZAIR58(!HbLE+wYqpSfT zO$CQRO-&L(LMBsx!EAeJ1WL zGv_5he2qWhfifvlyaqJIe5>{O_z0a*UDa5%G3RufqPwISn%BLax1}DoSoZ>dOyw>5?mgTQ=s4p4C1H^aKBD#3($pIU8+p{Q7S(b!PO|Msl&PnjsWHs!@CvKw?)LM;XbrGH z0GPV=#b0kv)-g^mFM?FZzqNkMJ7e=jx2M)0wW1z04Az8jOgNjuBk|3xf^Kz8#J*4= zbog&@ggOR%rY#mNn-a1y-{*m^MQi{E>ybI_mZ=%{OxNfvQ}2uCk$Ph&t?zFyXV`=v z3%SI2wYYN{I?Te=;~N2UU3Is6pG^+)9lx#ly)cwlIVtZyY=npcqquF7O39(4Uur+{ zJXZ>I18%;cw%*y^x=RLi=$xVSp7u6A`fl#F?k^0TA+cNx$zq@YtXK&3Sls6BO&tvf zTxL;MubvHfi^aZQPx_UJ6LEJ9*#G~7p9K5`8Z1wKY2c`?=~eh^V19DV>4t;dI->`} zqA!>2NjyiWtEsJ*4f1NjiN2^nR~xW4%Z-$6^U_h+g{B8L<7hQoB_nW$ET~o7Nz!`u zm!?&M_eW7@@YYbTIP4U6=c8h&RWp?I@^wnQ$o3>vf^(D_Eb^y1+GHML`PBwFX4!!$ z8Ier9F}uH!{&CXbk6@iGlE=4yu*6$U2AO}7FTk78T{p5q*YP=kjTK~0qk8Hs=#C(?OoBM-+-}A~aa1iM4;(Xj`GJOt^dxfUB2s%%oqF9lUhB!DVxG-=HH#0UpqNOwK;8(|0lS$pm@gsVX)}0O`=JwF3MeU!yF@? ztSf#7y-~j{!@#S~?tUs${n60n%2Wbdbfp4H{=*f@EGG}>-}0)eXBSZ@V^Zh);i#$k z$SHE8xfv)!akp3Ep@7CTO>^uO+kA|Ws_x&Hq5hZ!Ef7ES@Ug@ZA*`2pSH@d#f>(w* z&Q+aRkaHps(nZQFdAh+&{B4X>%XPl*<&2YoA2EYsKG%X3#AO?$E^X`AS_)rQ*zMvN zHOj%t0%}FJ0xcopAB)G_KOmi^5gHe7^rlZP=?o>F5QSI)tKZwb(Im%TE{PCULH}ME zqa;^5f9Sc;nid}Ws^(6Fb;;`NJTd811~MEV#a+wxyf1>#bq~@H?Ok@-oqInxys(8G zx$O?$fz@t@p9(CI#h08t3&}hSj39Kle+cR1GzGej0+QZ5HV-KX7=~WTr>XF(ZwM7N z4pTsJu>Gknxs0ORC)+xkNYeuk_HQq+bH5NZpmiaclm}@^WP}3di7&YpuSmQKUbc;c z_g`x478aYX(W2Fmdq3)siJT|&RMQ|M(<+vc?q6@G>hIrnFNj?}_UYqxdDXny9o~8B zP5ST%)U8unLbwh2l$4*&^YRV70`P9^G5$Aa{~O7{2slP-Ra+^wOWf=LLwC8>=_gMJjSFW!?L7gzzoS)u$xT#D;MC|K6_f}UZl9ZDv|}(k z5#a!wtuV$r4=;38ik1t^3-qBX#FW1@#y8|IBpD}VY6(77G9Ol+-NtIoMgfpCH}W5O zCBn`9?6^2G-BZTvu7kB1^WA?X(ZT(;wz38w1_=&fakBGQQlDD9g&D|9akYAUT80VO zqpEr9pCXdMT03EI-2EjnT6@h&fM3OR6ydC%p&)ZVWg zLH}1_;dm70OLOKiX=A!b=A=$c=f6<=-$aAItSoDe!h$_t9Zb_v1|qako}~LHc2#1VSe9Z}M?dYxzkbi0bsZ}aa-F?6 z6XH8hC~#h0xxaWIWM+PT;xi!RbsXvP&a5Y~Dd4`^P8n2qOXwPZn3xef58JyhfAYyi zXl&JYn{=8a(_^7YL z-jny~!3HGjk!|9}XSz`H=0QF`sz#Ap$>p9_8ROb`CT&sH#uBau{A9`KE}hMSD2i45 zRclgVz}MI0jg_Hlblax!=*_$gb^W-=_nr_`#~j^O)ss=Ziaf72gHh!JT$?L1sKMvP z>!g%=tq7(oisltC8`$F1I0`>dU*9RaL|HM8Fy zdJ!4e_V>sAOsimK@O~p{KNL5ehf1!Hxn~{OpuqY_b5)nk+-mmHCB84({XnWCpW2~E zyZd$a_T}g)WH~U)wtDCiKtR8&mkG!NG zW!U?F6ct~sd~O!c9QL|a+?&1SG+8>n0aiR0d=cv#;Fc2!+CIe+d@FX;s!J7y%Xy=R zA#3w&d6x&}jVcb#$j=tVHp?T{1D4*P)_tls@H+|TZ=Judp*_CM+vo2a=*hoMvCV~a zukTkkR{5_J1xgQlcTN9MIM}qQ!=H(<FKt$hpR+q-%TWcx7r>OCp*z++`B>-kTq z_df-IuqF#&k+nQ|33>2+mzw#n$6^h?s>DYyo2XOvH3y%D3*}`viP<*tc%L>Pg;m?d zPhVeOhB~E^NmiX?y!`Xh7A~98NGyN;Ti6!^obDED5+isGN=0onSRqmi2Mv9U3JS5F zx$-a$p^ek<_bh}_60FbrgmkOje5kG3fA7eqvNB9Qad>S$7C2E(F5&Qcl;bXKip(k* zRTxfwb!BCxY_%C@ig^Jf_}_;Dd3>UTxGde@w9i{VDz4_zbz0?huh*3IK*q<%?*%R8 zD37}&{H@U8=h(o)gD;L?Cb&1hqk|ubi(fM`-((Z?(`6zvO?_|*r(z>Wzu}gju5Hc- zsQGfH1)rYUJa}&vny#P8dyzYBh-UGtB5+4CE8}b%o3hqh$paU%_$%$w_*CwF& zs(u~mZN;||*)i4W2-~$B2FyO?AYZm0XJ#09zVp3nCF%kz2nYWzY&<`e2s1uGEdibr zFKnemro*pj2BI|9rD}ox#HZCT(#5F9m|ZG73z2!hCA$0#9e#~@-T@i*8hIbga^#*n zMtf_k(jbKUuX+1Euf2$gMIVLR03(9_p0xvwixmmNQ9pIat(7aa^E$HJLrHGqIJ^&6 z+#8-JfPp1(waEstYj<{acdhKk1ebQHEa+9w)at*j!GG?{9j+1V8_{G@H*`rd=^J`j zHKJrMqMb4y>zMFZq;jWKX-2lweKGT#XO3h}Weg`GN^i`MU-w)86chCa6B%3I+pQ=F zYF?$af-`}`8rQ`d|0o{&R64}7)~46;{B z9^CQ<0~vVLO8?%$G*t2)uriM;Y0;GtNWIy;$6(T(KTEkS&0ky3SYtL~wvV$HVY7ZO zYHW0NYZtIka8mNT7eEMmDy5vsq-Fx69ncoT_v?)&fsZ=57|(ud%YCES2uZ_2XfRo| z$h(UtAc9q_{3ym^fq#9%aH$_rcOjk`>T4~qilM&yczenRUk@PN&$yNf+evx)z6dhW zH+|#L!~cLG^z(14j;i;NHRsA$z$~S0pXsIdfu#w1QR3@wdfTA5dPYLGH@3xRT3Ikx z|CL!^gRRHcoc<%Hqp$C)7kEn@0p~0T4HlN6!S{(`AFj1)$2lLbzKXoCdnM)1Ul^FX zm0F=)Uf22gTN(15`p#3f#NNJ4ZG|_4_qx+tmwZQsPy~lioq$DGONyF81?>wsDrJiB zo;bVJbgx1BHvY$N!--i30Ts{F6TP|~C$oH4>rTh8Z}mzD2+zB<99@pjqiRdeSI%^v zviLdaEGDwu4%!#2BeKm;^bdtG`gu~;-Ftvq)BIx4rVy-?7zoD8)LH_km+=jT zKP_Jj|EFQqggQMt3rsDFQqOPM=#?+@77cEN>{bl}|B^&tz(Vm$QSaWI;CYzipz{+v zTD4_6y%ik;VO=DN)(D9YzOw`p9(-MHc87-R))!EfQ;l(Rq~_u_pg`lY`+3m{{Auav zI416iQLWBACqdTzL$%6~{<32Sz2z{Gv(z@+QuB8sbA=!n_7jG)$u!&iVdF;B>5*8Q zWVU+hlmU?+8W8xf3X6Y>Bbxpm&7#!rpJp8H7DdHM?THoY%F5j~4+@S`;>R2V7THtO zVBEXu8Vri{^lUuXpv^7u8{}e8T~2O}5eaYa+>zOhlAH^6_~uJ;6wgQ^tYqho@s0i} z#&kYP=a_oJ#oDiQKz7GLG?=-*o#hlI}!QNSX1k)9Yd#-<2F7RPh%Bq2L76+n_%yKK z^nuek%lQ6Wc`0to$V;A-#%3JeM_86%bcCGl?CTZ>CO%w??&ICJx2R*o{|{Af*%eom zt&Ktmg%uLC@Zb=f;7)M2;O_1a+%34fySux)yE_C64#BBgJ;r^<>HY=#%i3$s_2@9Y zHWP1GM7oj2L51nJNkm{;i#M3|iu)|9t~{j2PX=)_`NtAXMc$OpuL+E_x02?rJ?=V5 z?FE>F9;9o=Yl$^Rr+AP_mC?zbsHXhWM@08=l@t1ni*U&O#@`PQ4I|2DV?5WH%u^Wr z_-*PP*=XaY>I3z?h>!|_OsvxAJY0^=GQOuy4B8sokw4{0bQl;qj32Q%be>wYRLZDr zK1LUqJ)I*zRvy&7t2LUr>JA~du6qqzw`K8Qq#&4ayUyWpX2!Q9#u{I9q5 zY!#^2{M5TKsTvWf3qx=Cb9Ns&Z*+i(>F=wxtq#|TuOjSe4nkH^a{>z!_II2vjB%{fl8U=NZbD{@WfA;@vCcHX_K zPVMU;Jt;V_YHm`^|1Qb;#PrjzSB*@7_(xmOsIOjXvonf^86HKVFBK=Gr9)Bi#^)^u z7j{;V3FsCHrCQjcCF_>y&qQA+$I|OGM@A7w3oj6Ny93ZN6SV2)+M{X0Sn^?(Yi)UV zdqhzC^jPWuMW35*@Sp78CXhJ;>@_ZDRci<83Y=MbB&-6V9LDNS;48|s6dthlXrlpPEH7gKeN7p^-uJ+XJAs9B$JQ7YE?+ICdD;Wff~p>VWrGKa9}yetMxG!!Rdo1?`r zr|&Pe?;8?D!>)!)7dk@8SaM_`>_rUmM>LQRgtw|_FSNNxCfKf0FB;B3YuGpY}D`-hT44(H z$f*~_v`894sd=uip^C~9@*6lZOpvGlr?Rp#+#{HDyJt_X*Jg;BYlL&|48%f!C&l;j z6^1EiK`_ICdg}B21w3|0G$Za)wR-Vu_}#&cn`{u<3jg?asA9DE7=W?hB!-6HMGMAi z{s#V}=b#cAo4k?A#Ka^;v83R&t);8=pSI$6Cg^)D`vC_h)5+AEQwR?&fofcY3|L+N zYb}SoV6jbaqwyQcMq?E#T8Tp}Q}@fKT+#yWVYE!k0l{TIGm+n?bF(vHYLJzCxG-(b zz_1ukziu=L!{E)cPEKKb4B09U{@q6Q*GnwQ-J9LPbIN-F3%y;nJx>CmEB6`}kvmrS*}V zwriu7YV)f6V9_e}{I6g~Z=X>c2okOn59H+X+Fo9_!9jIL9e|4PccP1i_OBo5Bmn7^ zx%sdxgO+~2%PTvNeF23L`0rT5rr*DbyAf^YGs<7~8#+ zneYaCAzbfp{ePWwmf0rYq`}ZJGCupJY_$+E-q4UykeMrec%-9ps4?7lQg107*dwa8 zNycBe2U6*K+GxSr_&t;8*^G6mWUil+yeA2~u z@rR%QA=`H^?8&k1Xw-JK9QxCdH9uK_ZDN1Qf7`O8BBZ%E-tz!V9P6$|xkq;Ud;i-?ers(1h`JET186s2^6SJiF)+LlGSupBj0jSG=!6KG z?WOSw9Z)lMf2~CCZ>e1$RFnmWpQot!8*^gczqSgTRzK+7CHKj2<0x8B&RIF7DbG=mQIJNcucrIwf9cFbZ{iKbD3M4O@A|yO(?p~+8z6p+n!VR%pvY;jW zj@E><*zH?=aH2yTcqMZ-@N^16qjd;XdK-$L`a zj8|iZX2Qtusc;^Acet4F57%{99xkT=`DlRfIzXio;7l<0z%uzGo{X^NLxaIuKwziG zmjc!}_saL}sAY4_bKSp}TlrOuNB3atV$gzrNXq7%w}xS%Z{VUjBNA|z zzFk%fy%}-^HB|H!I*c$WJ<>zdV~E9ca-et`HMQV@vn)~ zuISTcx<$0%ZRT^N$WZhscihOnzG{aVqC^92l&P)4d|Vh5*|nM$fzVadRN2ny9_6@$ z!!|B``bu|qvQJkwm=tEAdwK4`SgmX64M9#BvzYN4W#oNNHwYM%_cJ!P6bX3utiGsh zsV>fAiIdoc^Z?_-x1VmepMlq;Fa^!~uEz&>fu|&&+~jloe`!%HZ`$j`z zXskB}Hlw{plz17W8N{+qZY6B(-DkODcCktQ(mOJ+vSg4*_+&|&M#+K}w%z1A!N$2h zHWk6!D{A~Z`P`a2iKsO{SoPQJv2T-dSkmkeK8I~*;(jI`wxqZ1hNs;CUTYSbE z;0)($(wz88gj$?HQJ$LW#gO3{Na8f>8yrTp)7)+bw^nINn?S6$%vdiw;OZ8{c9O7GwC!n@Y;a< zD%nu4pI%?MqBYgRx8?efXIl;L_Kn_4$Iiec5U)y8x+*>#A63~Jg9R$5p$uDTRN*@B zU+$<8dq+^r_LAJZd*?6PoFTZYi+BE=>hL8>vwuPVKWF@(weHzKumU&MvD`mq{l~}= zRZC`J`d2L6UU3+|9iUlJO9rRIr63muFcZyzJiKL;sdVcc(FNUW!G2Ck-OMRR2}?RS zAjzRSkt7|n;NQb{T&h_Y999txMfq;*;3{8gn8g9dr)OQa#p93SEjw=3qPGtiHP+Nd ztD$hLh}#VO36hi~$sLRnmEDoXB#iw13cWYyrxFg@#j)6aDI~zE?%fsv9tP2}tKrWG z>VI|?EcWC>@qJm5w2gRDa>Fc%x@DuO%trU!;GgQ*FrDX=AkF5LUKgp~7h#Z!za{+e zh5m0WfK;?EO%^78A0d5^ASbzNewgEwzvVs`+cZrUfz5PRA7%*n9Bwmwk&!h7>$Y4u zKLy)f3z;zQYm`4S&xo=|%1`jJjRyonFr(eIxq6HdfQ^P61At8V$T5)gRCa?@< zl`VEuOA1*I^B3bhX4x!m7j}?o5GQNiW#bQ{Zz3OAmyLfrE&ttT!8FrDzs=as(S`g- z49-?xV{QC)ew);GHEhaRSFfhARq-4c(6_cUVC&7Tlz2#Mm0?SS`#UAcJMvHqZ`mZN z;5^yP`g%@ROue0h<8VD<<)h8mboqE}P=sa5CrrG>r}=NA56Da1%E9uRNq5i3;kL9o7Of6U1GeO9&;Jardc$4pYVL0kD&$SV zKTgF^$A0eM@Cz`yz}X9Zpq#icgG%bW;y3=J{!CAfiZ3c2{QhEGdtOppd zBqu0na3&+DqD!=*vb2T@1`0!THQ#l-gw4%}bo7UzVJ+FxDx0Q5UTm!Ow$4B=^|pn4 zutwJ4<3AXk1e=>fB@K;dl35fC!&c3Z6?ub&&(?VI2*deq*~Co9IYY{wKqem+;(Je( zPiM<@fRw|N#K*O;%J{o0fDt{2l@cpc_76@l6npPl8Z6Ahui>c7@b&d`97f%@9~+7d zJib-TM=tX_4n25Ep|CTN7g>Q9pi^Fv2WHzMgY|sU3#)x-ymlB{+Wy_ z@e%X&Xy{>@%&j6&P|bdpiE&581^OXahBU5~odSo0VzVkh17H8S643uEaHlvxGLCr; zZJC!*RRb=$lg2@N1^*%`g;?C`l6VnPC9~E2!Y({+YhXOq8|TDflkd(@@gj$?fy#gJ zTBqNer;JbpxrEfSe7T;(+@#_yU=zl~bd+;c50^>ZAIv&z-@eEd@kV?whVWmnglb0E z#$chsbk#iHW$`j9C3A$WT$gIKJyxFBg?6ifF##SLlf!p7JktaC}#!Ivsg=0`+lsObDpxH8M_!qw_xi?5%_r)sEk`*<5U+Uo5lKpe>?mcZ-{bjr zwNBIot(W~vxRanpk#Z+9=5(G60QPSx$Dw*faN(zmNNUM&JD&p}AaF23K}JmMqw&!UpLRfr5H9Uj z5Oj7F31$-4O5Zzp8$xsCL+{x33vLWSh)xgGr{`ZFGD%MFE{%<`KS#g{OhR1B&ED`Y zcen9jvW;3SXtu6I$m{30tH$5H@D{Kn179S(X1Wd3`nd1qf@xpgdASkwco58rcVZZ( zYkEFe3+-b)~4pj%&74D|g{7Q{k_^R7=A18b209>(P8$g>S6@u%~B zGIqtep%%jDF$Cv=`nq{5=J+^}&4jM62{LEAy|hxFzdvb2!s2}PnEsx0`^{+WeXIfS z*Ue-(uo97UC?&(;lY3{c>$sdzQ90FZYOsfwif!c^ee{(ijcpMo$o||%hM9Xs;d-KA zEu70JD3rFs>>druhvmQ@YDF^p5BY&KiX_#50OV9)k5nw=4oay;Y32N#dny7Z)XpKB zd?mW$_g{AEQ2tdrT(#=@tDT9DgKOsj&$Jj{fGzucNu~JXiu!MRR2O&n!ai@-PHS?2 zt^9pgM}Bp)&rsj%i<}i>KA+itLlR%L03h@1-~Icqdk@G@>Bfl`8kDMU-!|YIgN%R`LAyU48CA27Iaf+Yil(w1e*Gi(1wl>Tp$Vvm^BI~G zj-nu0r07oyi@u!xDzyyXk@ZP#?oZ|pA`ThJu*XO*&lihxHF_FF(Z7&V5^doIJDnzo z=Oo_CEW%2Ne}R9bvFs5GZO@uHYcknv;GuTYrxRB~vx=`T3K%(Dn8^`{n9GUD^R2#x~1h#T9EZ zb&Dg5XI59dyjLQ)S*WGPgDb~w=|0_v{N&8vbnLpV!JkyW^u65w_KJsY-S|$a-q0Ja^@tAl`pJJd zBq%=*&urGOpEc%rPO}L7`ud(!k0P0hTWzZym3}{HVj{-D%-%IBTesTe%@s4|V4(}4 z#_n@Uvo;>%Sr0bM|7Uc~8wyr%=k{hwAC9IVXQx!l-MPpsG!JO+~HEMp`;AQuH{yC9U z3uJ+Eko7OA)_T4z{mt4s>Job9Mn`e=CD~t_S+MH`#^M5Dfd^TDsdHN8g-h<(d>#o< zT4a=jBEAxyE(fUKpt!Cv%Q-Nb{vzaf80Zeyl(~H5aJh@rU(h zlRWt)(!~>|62frlP4H!)D|Mt_Pj1d`b}2!)-m(Ml4?%UyF4`RSL}(kHC+pv`!>$6% zm}V*cXFEM$oKXb~s>LfG^cpSs%n1pDri=21VJf~y#mE@3a6#qZ@$HEy6-lm=7-fF) zFOOYUM`Q*Wrh6cX`$%y5WNF!1r-dxPg zzU;HeS)R;5Cu8=7?+#gvnMYQX9WRWcaH(+aVoy|b%hhMnVeUvTx6eBdDF680>@&pK zJ8a)u*=L#^`kl@Kg{Il%T7IRPE%ZF>@_8?o)*AlL9YY1IEf0)L`D8?ii7kDyF+DEX z|12#cRU*+=niigNP=6U}lvo$&=t~+Tt=5kl@LeOX;#c`gtV8CZ7%~wn>=an1-MOrJ zMjb?&71T5Kb8&a^n|j3$uabegrDDU2Z`ubVF#>qHjW0`qmTf7fI!H55 zgl+QO@0fe1(}|#|5Gb0FNcey|nWr_(fHYw~4%b^N9472;c>zMxCL~*y%X3#~nGDet zCZ<^H#Xqr$dp=uREtgW{=mG(u+oYf8r0^}jr%_wt0Bp40__{nFeP%^sv6UUZqv8G9 z!Amwo+d)cD`$Y)dZQgkV-!DLC`;RAvh^*1ktu-t0Tau<5IQq`A*aL#;wiF&RpVH$N zW@>li`yQTW$PhY}eLhzFBe6qCC<4A5^rt^J08Z|Xo?V~dLTa7BR|AI52{0fk-Jy>C zKm!N8;hBlSU4pqXPi$PJ&6*{8$HZ#cVHSIM7+#E?UBPQ_DgxXZ0`!qYf!&M%Bw+uI z04WTLaxE78h;Q80Ak$Tnz2Pl+p?TzhnaXzQOuEzh?$B!Yu%&4RLmmNtgje_Ccra0!|jwF=3S)<#@4 zua$c7pFW|g|B(;v%jxm8Q$0|mqF$=hiXmNwJtkBG`A=JCpuOhHMz#M6TiZ_+mG|gO zITR24~ z$G_9m*4`6>)wud5E648atRXaWx$j6qqmwzhy&m)dA6*lL78@l^kfSf8D8; z4Y%AMom-c`$i8wlc6WesN^NU4-4UIU@Seq-eU|ATcA(Ze*2)b%MHtvPe$+1Day(@v zL0cX5(-}B8@$G8=eSH`zswbfeW`)bvOQ{+5u1gH_VXU#6ulk(Pervel`_#lEo@#CX z@=coV3Jd?t8C#RntM>M7>$}~5kE2{NKw++s21kmUcYmwWuFT()_D?x_uz%7>LUgEg z1hS=hl zi1=snuqzS$faMm)nI1#x>CWE$jwcIrAK%ngV7&PL&zUOcY=6UJhtHUm2o|&BQ7v^Z zgQe38N(405*7pUIESV_V&p|k7L{>~5O%eC}3D(1crasEEADV#M0Vtcn3?-%QWOR0h z(YuO1OG*s5e20`*2=2vqkukUVy=f-E4r9*@o|ury@Q?i6D(LY_?Kw!|<#phmjeqqh z0cV~%1L*5YOQgPpda2s~%j$g%@kt2LI};HFPTKw~Lvsuo#=Z=JI&M^d$(RetULNV| z6;qRY`_d==lP>-rX##n-WIPV_qSPxnmkhe%pWT?{iQ_SN!p;7=ukZ~=sqkmedSxGo zeV1`eDwG{z1$c=E5?_qzHWO|hJlW!~(#3mkz{Fp2%}8d7Y?jU?Uz*VGKL>i32r(+q zoGiq(-@Y6&E`*sMU{b-rFx0HcKAvO`%)Lr8cbB_%4qAlfNMFHgz1rM&wmZ`oaq#GN zow=EYi*P)*o)0^%F?~Gyd%nm@)&#L9{}+&#>~9L_jtHQ!6y_aJq_LMthVe^KqL`)f zd$%+Nv(pwh!De0n!Qr21)m0*RjT0GdWs*sDw>}H%eeHY82*aQPAZomGXr#7P_-GZb z5*xFq)~fW{H1y|ySD4)H3+JbnFyavayUPhFL=?mw`)&H15Wms~YTM4~YGIa` z_!JpJNK~_5vHmG|euQ{(Fj}0Z=LIAFxJ|EQe-;u5F%E+0GkFmK~$Rt|d^+kCi18}QKr9-bg`eZBhHdRtL&C+iO4Bf`y3M*+owxq>*vMmb)8!-lDp250s#>RY+WJu9n(ATtrJR@Jrnmdm19kc@3;xFc z$p`%}$Dt4-oM>K;2X3Ve$cBnAIVmGvUKPc)Gf{G@U*!Xw(@FN$bB`jrNT_;W4pkgao=0LIwKJ?Csppbxj!ew)H|v z{Tky63tktsP8C=FDF1V&syc}eKvcJ*qlRLE8vqA^6f{`voHWhVZH^epB24xqyTmVt zI8U=|L>Yn)fU23YT^|@G2+q)(H6>31#hBx!6R}!=epfwK5J$BCtZzM*f*fVtA8ZR- zr2=}zp?{OUhwg4xsrl<->A%hE37|Fy(ROW{vEHM`^y!=D_4Og`@Ce(eJ&4i^5oVN3 zo$Rg7e7&o@bx?o7eJ$7cfP*Q7Mu&sp2T~H;i6|5dZ3NWf7K;_XCMCBvc>zToq`Kpw zz?gT%{-7w<-hj{*;Yh5bqa!EM&Z56YUM`pwzB46}7(eAMrzm-&_R8h7#^_;oV&V&` zo9d9{F$;D-=Dw=q=PL*(zl5P;5~vVC*M7SAc{w34ns1gGb+4arn`$ygu~iJc933;o z9qi)etK>{!2~$g4!JCDpl5KV1tQp$&@vo6|T|#uKcXy}f{+pAemh#^fZvKk0=epg4 zbpZX7TU>#n)+}HNm7?v&Ei9<-?GB?ra02D}-fGQ$w$V7tro5#rMi5t}(G_!)*OrC;5wHhB76*L-Sim^gZJyo3b zi129W9oLVwPtH{dmD=Y|COWo#r#w^1hxA`f{;jyglW9|RGc{DKhi7G;QXNEWHNy$) zC)J73OjkO?0X8)wF)$8zbbjI-$@~?>-S`wONkxCQbCA6PqRyQUBOY|Xt#z`hW1OBk za}51+zS64uE-@qNQ+(N7?}Wl<2G;m*%LHc-G>%a?4=0O=xJ8CU91cyM%c#g&AbQ!F zVB~dD1PZa7yGzxp;{ph{h?jLSeB?#VsAQ+$<;m1&MCgi!Y$p z$}=58LRSlOPGLW1%5PZ@x1Kqw zL}edmXp5G5wC`{mqlHU^IZ)^Zu1tPTUqPtvh}6lZYPCkMN8m;#6Ti46%ARoVz%kB; zXuzioK2#Yhhfm!EmxtE=Q`#ugSbmb z4|`_cBP`;EazVJ&FYsn7e)o|qfM67 zT5p6m$FEF`iba{0VVS+>$x>XeJEaAzt~Bh4VY>+Em6Y2y4X++UHXeTogsr3f_Aw5} zWQ%%(<}30g{lAT^a)#+S)}=ql1lIpYu?>M8!9#Ss_!t^vX1FG9q2ql{VD4?`P58`HQk|PPl2I)+2toB1A`j9ThQiS4 zmNlI#w^By9I;o#*n3V7)$OQ7DO9foFxS?(eK+pw+l}wZV5bfeemPeDiaT&y!J@QInArbF?wTVR5ov;8%PN~F~^`sVG?*CA1HQCKn;O z{*!&yeXw=`hor297*8waHo+5jV+%}4_SDTvH14!qh!-2F@ZEBGC$Si^jZQ3dh#h^{ zEuIP2iD|7Hb{eI`0^cBajdw3*HX;8Ha0HJl_?+Jz3=|>P91@%_82kQD4aVa84`;1w zfsG`Il}Zh=p~s`c)6nPa-*zv8;ZGi zZ(LbSU$@mh@7{lISA(&iHWF$X8s05)?2bzyvREanM`iLM^JNRemE$NDg&P!`>*WEL zH3ZJRbQTQbv%&ohn&nUpe99Prp<;sXEKOAt0m8N3BNJgITYR4x{zwGx8IbX9MT|a4 zg^&v{p%=lb%mZhOlky(GLBW5A?(VFaF+rIN4v@Ivm)~AwJ6dXLC68+~Lw-GMk$KP) zO`R!A(NWtL_|zFAU?a_r;aVf{q-4QOy7 zjrvub`)|C>&)UY3qO(%nAo$lY4O$PDv( z-aZhec8m;!zz2qvQJ$BWFP_#clkYG&dEK zgmrF0r7{(Hs%%_>NY!-2HdihJSa<41Cav(2}mUcs7>a<*OTQOt^S?wJk~ zxPM%>!eU~I)hVS0(oV3@tsJUX*H5j^xiyktRBODBoxX!@tD{|tCvBKIRvZ4viG`&1O<1w5_N#m z6GKtS1`#@Y#;;t}gIO~*?E?!+%v%&Wei64m2?YE9vH*~uWbK!zp>S?FQiElmvaH?j z31&8n?8KM_SRru`pz$gBo6Ih#bz1vAKcok9eiJ3)QD-CBxJm@>1Wc$$^yZe#l3K|h zTdfGZET*x))lMa3aBy{|hR3K6a(hD_ykM`ABV^!m!ejd29;QrIJssS%UxG$xHt4(( z;BbXom%qpQ@9CdKL;(Cf@4U=K2Z?bO&X}#aI0Ni@-+7VncEeS<6Pd+)GrPxzz9K>4 zHsKTJPN@SgnhbZB3@0<8ai0JOzP!pWm)7IFVbwjKDi?$+=T?tsI%ia``Cr*-OGU1N ztKpJGqxJcTVX&j{-*Bo+;&bU(%<+PFq`=52HoFP_{>B^?rMf$qli;iIBwCIH=`P27 zxszAZdIE%29g^M&^ORq>^3PvkjS{vK_rwdUu4*Ey$U90qERB91@Mb1480HS04dGSn z%!_KcvS(PL$9CJ@R^9<<4=JYpzsiM;>zzk_s99e^xWB5Qz^Lrzn|48n6 zWrN!yM(h6;wApu%pv4WX6dwy+o2l|N-6w3_&jpEx@p^O`^@n-I`a(WEQq|}EnN9Jj zfM*%(sn!(egh-@9UN8~N7k`8BvRn=3U5(l;f_)yuHU0CZyVnUfG)x?lH z+*u+s50X-RGn>^~h!qG7ti-cUkyz_Nw(Hn`3FUj$F zL58x>S0&AwxDzUc_bElf6IssvTa0#DxA^W}eykK5Iudyg$>Zb*^6;SzM3slSX7w-R zt=T)}Q3mb^3pE=uz( z9a(fg9H;5RZ<~!Ae0-XQbBGfvei3Ht#JSIANOoOE(vJd;bWeOM?|9l#i6<{8-SL!B zGf6k$Kl^UGyWU!TOynE&3bvA^8)NDEq4OueQw}1%Z%Ev0@cVtuNToeeM-%*MIqMEt zjWIg%<(?rjQaQpO8`u8w&+}}D2uq@eGk$xR=w47_aRJ~X8Ulkr%>*{@Xs0VZKA>pJ zx))~DLX1YC)OH?flni7awJ6mzMGJhZ(Y5Pg54^jr%o=~o6qnQg;yT{NM^$Cjj!30I z3+RUYm?l&@L2r*1ASDKQ_4r_0RNTZG(WIMvb<)1P%+_8PD@V3oc}&IN;S8)oGQmu0 zzRvSqi(1MVg;mNgqVG=7pQSFDQH)oq0ezO7lz{Ejo~dXWi3S%t+!5M9u%pIzB2)i|cm^0||an zV3V}$WRmW!^niKOrwH0r|MJF*FnqH0uf_g|TiL*uSU;{V`{mb6I$|T1wh-TOD=^I~ zOO0=YzKCHf0KdYhc=WU+XB$UGkD8HjWp|*cHCG;Wn}QSB_Eg6f63K!DU|m*>@Kcn$dE$h$K(a{UR6>EuW*dpA$@j~KyT zw;K}ti&7hYhxB9e_Fm7X+EUwi@yD#WVup#wZ+ok&?JnQ-CG)nGNQ=+;k7qAT)`f7( zG@Ml+tcL}G>B0Ed2xDotl;cxAUNzc2Ut(J~iCa0ZQ(;#Bv(oF-2gF`Mh|0j0bfdJU zILyr*v@)&L$ExuMl(~K6?-Rg2|C6$iZr7VD<*`^X#G4n6&z*>WXxu10*O_e^DQ)gf zG4#7fzx?kdct%pWE|-;uEJl!jd4*6bXw6NffqAeTNp;XMBSjh*+dH*z?|fb9Oa<7t zeWt`1kBA%Zm1&-sglaf`b+ME=bSrx|wb};eu9k|<_-*(UgaU>cMj)(TdlPlsnW>55 zwpdN&9&+ROnO8A=XGe}x*NZc-XkIqKPAHRtM#bKE{_NQAAv);!cw!uA)GG@)0^7nGK5E{t_^){ zlZkxA?{4uGtuoHt7H=pE1?{hbt z>FFFBf@sKdoXkF08>p?j>-~?-+pYjZ=K@lAd?B)qzj7zHF*sg=+ShWw8^>I>?Qqsn z#;DB{d6-pJ+QXoRt?OnZD=Vv09Olf<|M;{=joiOKZw`Hs)+*{(iR|C4rPZ%25JiBm z$OC14HKK=dxfAHF{afqZpUU1+O?;>6I`ILq4{Am)^KrsV=~gv|H2ZevUSIw!b^8e4 zVS=o-lTkphEm;B1AjJrw_Qu0FDIk>^K z=Q<5Dr@5O`UG?;vP5-`dbvaSKjRwBB`*TnIM0RL;nZi-(DZ>q=rg6bZCq2u0ahnI{ z6ry`=--xT$Hs7{qe#W^FtY|Hgf@Fdz{_BS?b@NjoeYPg7D)VyGRyjv5fC7vcM|EGv z_l82)BX2So)iIfx7#`hO4$F^alQeKfO54o{&+4!1sn#iyN2LHIcCG*NH|FBA+Zn(L&10v@{_D%pt?Rrn2{kUMQ@p@AN?jZ?zP0*h* zM-B=*@xsAh61Bi1<7lqUU%GZfHKGmx%{DL8g3-faT2+*efa=skkzUJPoz?i+pYl~-mPD%wzb1__YW1oc$~biB zIlKrh>>E6P4uTx$ZQ#2b@%SJ4Pjz@D^f#;$>eevyLR}$Xsxsvff4!MX)xr6$)al-p z7@C8eCotMZ2Hzxmea)#?_veB51GnaTpKi3rUzw2`q>6CwG_~e`UYO^EAc+{agMtEQ z>41)V+ocZg$lJYD%xoNQU%Op=rj+A-JBYKz%NmlWfy|wQx`#`5jL(#1c3@Ljqhg?0 zO*9M7V8g+K&A~$qYX3U5b;vCBtPoOji#+={8E%rh3%suKjWxU0gW?K3mbs{T7JbUJ z*eK6S6rw#C^V5DO_SO@>0eg7i3Rd;zjI@cLpP#23pJxr1CktbQ4U*rc$!s5giC*_H z-l1c3@*zB4Dn?$Xm@*6h3iXewm#8xWVWtFS3sOYu^_^&MVA4b*Vm+8#<=|;J0)iHYhE~&+Bcd6`J$Gw+b`}<(&%Y z!hMI9xauRUG5hWzha9l>C2`ImXlR%C)5n)RU)-e+M8Zi_>Vq6EoQI9(pyRv!C}M#R z^Yy;tn`(C)_iVV3pZL&CvLSODfYvN6xPexAfJ?}K$>8S|4 zc3gBV(}LgiUV!(d73wav^+J4~W*pwTtY<#lVG%;u!ZSd^rdu6R zl}*XJ2MbIu(El=?Y(y^U%P~Ypt&I~(02D8p<)IQxnnybjvw}X@FQWqMoly5g^R(pG zIPxPm2mi4#TEW6bx6n+gnk0RM`xKV+0oo26654dd)Dl3Mc(cPp$eDWMF5X;>Fu8_F zqKRa9ep;wFJ9Nnbey`uJ<-u&JiIhc8KH67?eZ+f(1ZUqXvddAVcX$v=%+y@2SjJ0_ zH6FZ>!Nimwh?tj*5Cdogt!m}!bX}wgM9-J>{_?P(|F5i&@u2_Ce(sz3)TzUV%k)IA zqDyD{a}Y9S0Q6GWC8d8HX|S%G&f=L@dBy&oM1Rgf+Y#NxCxTkOdXR$LQ`O1NDIeixAt z8O5A)jqyA24{(6p-vZpC)v?hNltlnv%gK1wBoF#tJk+lKIIlH zIu15$POy-WPJovC8l^m@{2Vs2%QNa5FjXHB+nqv{aP{|30XxJ~CD#N|Gag;yQN%5REr}wZv#8--o#mIjupd6an(U#I(^@eHJd2(~kEjI{3+_S( zMSMkdKZfVjgWF7F8p)NmAn6nn_qMLd|6&G7O`s@7I{bZH;Pt{I$sAbVnKhd8{SXo# z{UR~Cjknre9*Ut&?d_Pu7q0@Ztn?G=f5LnbPYY>)%|>{u=~Fp)$(>D~5WGkuL=Hl0 z`HDWx@=lZH3*b-)7kFY|hpGH|Sy;GqMpbiARm6LCE{#i!OFn9Uc9G`;3}f7 ziuH5=>4dDL=dHX))Bjv{@#2H|c~u<=Fd8*e3pQ0@|J;vD7M)Z(#2kV*>0di%d0Z%6~29t=NSS8g*q z1A1bdUJuXLdLOS%6qYjVub2qhnac#nL-7KT?n*f+uf=Gz2Qp|B-F$L9`Yhr zpd3e?0bgWF%{m|mqMcZ$_d2hEXn61yEO(LnPurF`sz9PU`dr=>I11H>+f;ZGE03B= zW4=v{O%$2y-6(;Vj$_z1noTzUA5-5LU02t39jmcz+g4-SPGj44W1Ed_G*;urNnuTZ{}hsa@bCtq z?-kW;kt;QQY=Q6YK`3U|>vy^y#X>U=3?DYM=e`n6xuiOekoH#Mw19VESID`u*Awzc zC-^=KzJDS-(G<)g5l)r;6uf2+{llXLV{(l4r=!zCXnW+gVvvK!a?W&c4zFl14PSxQx3Efs>YVZUIk$_c9{=!S}B5{;#@`>b0vFrXVG1gI5an|cN6 zs)(0~7kR$wJHvi(5510_Zq>Uh7wu@5x2`80^Y#&0y*W6brk(Dw8}4R|%?ys0N|)n! zWj?o#FTph}5aHeAGhzCP?ZpnsHjR62$>?TkYE6^^rTz_e6@c@! zH_#ZTwdiT~TI%s@&l;HtG1yovJ_FVsZ;-V55%B+HTW*mHw@@5>X)$;?n=bX&&X1k# z@fnExhoB#=2h3{09?BHq*&B#8IEGXhl3OCIt}G&aTR=m2*_^zcTwVmljQq!=uRM|c#$M}37vjzROZL8zxL8ByvJvyu2Qw!D-Zj%sOEwttF6<& zAb~-FOdZu{xuOKBpA0I#f-Oh!DnY{Fv$lT2V+G-)aS31=5FPjxUafIO=CxV$AJIHQ zZQIhbQilw#w!DD-v_|5}qU&laP8Zq*QXMuz)^97!kUa{MyQn#!7j!!AwR- zh3%hFA$9-)=F!t^VSn%=5w=P7sycHz_}Yjb}&R*HEr zpwz<7L~xL@{}CJWZ;_I4g*Zg7d%CF$)Q!7j3h8Z`UOHM0?}FI*Ulq_0_5c$`+5$#) z?QZp*jj_RoXyj@RbO*?1alV2nhxeeAMlE}N4K1zqw)*<3c+ky}o`;@PU|sZ>?X>v! zAV!+yZJ~_DzUzYtX@>Swzm1&A_^zCL_Q*efZ0A25nd*fssUdfu|CI4=d9{fY4CRA< z{=9@DAN)b_47C9t;DPe2a{IWdxT{VUI@LXimkH_1qD}7ELRcvoIJ-cNXUu_?;HWJ? zmthG*DH~r(tBVZMzcHI?<0Z5Oa8A#nC6bpBim!_IB~lx)hpM@Lm=2Gd*YUu0Oc5%k zH0?ka>Vx9^x&c?)M>QKtPJ@!_Mp z<02)eDyP0-i>A;al9XZ|+lAQ+^#Tv9iiT0QjEwNb1ICW%h-zJ|k7x-M7BjjHvebMp zab};SO@HG~X_lO|4|VMw0Ol|0(rp5+(!l@B6D|hVPVOXIvtDr3`#>9W)VHlCYE1}) z07rzu7F5?RUlvS7cMe?WMpW~+heV&gzE>*M2+jW?b(}2w+L*kf{qxbI?CUC3tzF`< z4MC>=nfpZXc?7#nqu@ezuiUCm*j>HV*k5aMNP9yMt{Ty*t>VY*xQ2qE^hq!Ij~I;u zr8ZULf!@KEs>iC=QlXn-Py40vWm{c6fX*#|IJLn`hiPoda?UneZ(7FTZ;A6D_jiL5 z^)kv?m2!?@2(AT`XKQKwBrnN5i=~*51--k4jP|p0Y_m6wosH~W91Hfql)=8Z&*j1t zYxLiKywo_gbCaHb^-iTP6#?0rmWIC%fhJ^lFX@{UJ>rrp9pqg`H6(T;WQxe&wSL)C zRSTK(Ln=J-$@E_yY$~5ByALn-sTE3GSE!-32^~GmXi*tQEl&FTCQg-0U;CzOek^G* z2#%>s&d%8J)}3?}T9LB<5(>F#UHZrP^m5c?(PF9G!(%x}+r|5;!}1N#U%HQ-QLZwZ{xUReh%U zT!Sh9sJ0q}6{8@{g`II>Y#l5{!OG7CFZ9my@Hg22J8dfiTECEgaO5P%z(%!qpjy-3 zQjS&Z59<}@(=3Gpd442CW3tvtoEgQaCZ*j-eZkc(3hpNESpjdwxsuF-(OgiKLaQSs zk62UZUH*bJw;dVkBRKg>`%7qfn8F(NH>ziiivFm@f%vqbY69K1wayvyc{vEZdyso{ZH_K?!_3pNzk}jus5d2fC24bUHP054cpcZx& z?C@(%OwBUW;X~fkgpB;Fnb4c7uhc`mP>fv;EWg^h$Z@R>OiXf!>C%De{F^|T()kDV zfx&6dVXJLq-xj;k_ku4%7_UKu6$!BmyFqkQQ)%juYv?49@$%0w_D-*MK}?XA>;s;z z-KRVTlWwo5Z~#I(k7S2%3otcX@`DTvP1?y-$F+4|E6J5MkG?I5*mHC9B}{-rU+~@{ zeOSzmK;t`@_C&9D%p3a#seQIsw3fo+)4osMdV%zT)V`jU23`yz9FdXUf+%C1U>{L^ zT0z0({7<~CB4YL5S7Z+xpV5K)(rJ(Pyo#i2af`h(-h#Py3fMxB)%~taRA(Jk4tlZb z{fYqNmC68m5S@ABa4Psw<_&Q`3`cH_psq);Fd4u0hq0uXxY~^Y@$)R^4DUZ_+Q|%kUyk%!EI3-%e5iI;T$JBiRrq!6AhKQZ; zXF4j8fb~zt*I1BVz%3%HGZ0QnN8~R7kS{MzpkP>XbY%3P1I%p(tbx^dBiT-pdO~o$ z6@<;_+-|I{dc$5WUWb+w=HX9i2Q|gQAl#3FhW#^QC3ez0coJ~SDqeuHgj1h!RISLcw+!Ez#Nht-UF;52=34_~)t z-e4Rcjp;ahdqUTS?u3Sdo6OXt#Yn-8n%Ek6WIsQJZ!Y23CS}T~5%nAV*owZAOzV&kW4|dQ zYG5r3l+F)S#6JA5>cH`L-tz3~htA}5uc047O39}A7TAhc zMtODv*MMAVrCMCZGIn_VeHTH^&VDC6S6&OcLhp|ACclBJxfVR3?>zaKgNhjwiQ2r5 zYM`w!db`a(lI@Wa~{;&FGk%KhrW(I;RK@0q`d zQQd{HWy6^FV`ED*;MzujYx%YR-|!Y|BLdFp^y4D)&#b3%Ey|{Ec(nL1f9mwLQC7b= zv%tGs2bc8areBQ&YdyqkDQX2>6oV*wl_)lOtP|%O!7;Dx_%v|@unQ2A;otx!PYc(O zct2i95S1)2r0b zCJ&f&L-(1ouXJWs`}Q&-zV+W;Km38alw*mr$`q1sF7jceYVkw+joI{NJY9~4VF#}? zR@2ef7x%qgE3AF|WapeiUvvlU>TqgS5QXVh@5%d)ZQCRNDFh;y6hfz7{ti3x+Ru3*ouyyky--!Ln00RHhy=iMja`Rb1ymI-sG9LZdyfISlUe;8P(Q{jB1a$ z<961;;Tkvj{-GWXY<;)4I>H)rAe4brtf4wob#0c@Cca%O{m0z;cw`8OIG<2KbQ^Y~ zR|iq#sd{a{aJFxU9sKr-byiwO@~oVBMx!4UPD7GjGDXbhp zD!mJ&MkF}>ariDYAWRaEI23Db+tihU`WMw9(1d|b%kkT=s1JnEstOUVIR7=~lFc$ge3fJt#g>YFr6%y{v zO`jelp+apy%SO+vQ^mzp{y4)g>INDIe9q6QO3>IGCTc%L$y^9Qa5^9)s4os8l|4F{C?v>Jsu7M20jQ~bS$!V~iw-^6wmJ(D|5o`KAEJ;J!FMUEV z?F&(O<3cFZ@_RcH{eS%~4NWHb&Z?=gKeW7Qhgln0IG}S!l8GTVJ6<+Bti|Mgan^K* z*atXzVDmRSB{s{)Z)!EWNX~D>2nN}cq+ta0us-gy0_i9n4h#CrMD5wq$hNKjs72T( zdrJK9x&Qdp*(B30wTh+1 zr4}=p3r8*3woIxZi0K9@@Nu?iu){ogqq4tAZy;X{jXi+)d=;+h@9|SE^Hk@`zOFWrE@V=)f)VG>FEUZ>1 zWzrt@4@~>t^YuSsCy$op{LJ9$xxwST~$>VrjV6~C|`PU;5m;3v)d_`_A#5DVUMSWDuUI^K%XD)s!yI_ll%Lx|VN44)6 zh0{(_r{!1FOV8S49dkoxK7i!};4hr^*%cFr<)R-f!5 z-Po)L8oWR}Qb;!kaoR)>SM&giu8gzdc%PBGg&%ajFh4uwX~Pl0!|K*wfE+Me{rV5k zQ$=7sBzCkyvxL9r2fuy@bdHf z_pi6*drmeR%jYvx8keR+s@Xspon*~WfIIV@h8--ZMx`yyViBe7M9=WqgIK) zh;K%|Ub}zD<$R_%J16xTn_;qUh}GVAWxL*-)2+*{rLFTHyz?YCm~zEjZgh*8hN18^ zo7F8jZ4be!>jNNl&~Xg5`d@uwz}qM9a;@ ztq!Lhl#TCJzxHx+a=PyB?z|wVbu8kn#KOQnTxRjW=R!{qA#=*=LP7`Nl5P|QwI$o+ zM6D1CeVLB>61zTLZgAZI@9J>+AJX))mf>8q*#oEfPG#P>ykUM?hzy$8C?%cZT}gf% zObS;==Ik%XaT=b7%s#m(76Xhhl{>(%`H8PynymrUcRR@2xgNRAUrTOV`!=V3?2T}N zqz8_xNKzHQMf62_o*EB0>wiC>^~ZbCVao&u#3qhr3pPy}JhDrS-rp)Szvy`5HMn>o z?4@wsO(YSg7EfgXkqjD%YGnspBP3yT=!-xKmnPYT(++vfQQhE>=?Nv-I#SW+G@b;| zv{2%Zk4qBlXw$4bXvHIGRxmoUwjJlomv`ShDy(zAuE@vHOzF_HIHa*Mss}b-a}f@QSQEta%pS2Fntja}J9Jq1M+9?6gG{ASW>%&|cblkp$~y9Ymo-ca zuxd_DF(Xy#0;^m1;DSVovS+A2C?l^mxCd?sV6X~QvtMg(T`5_?Ha9}eqoQMrW21`& zgV`9sIK#XTkdgC$M-H2c`mz(pe&4zG^EV>Da7hAWC}|ji%qkS73sg=x7<{BXCncpR zjVShbfsxbnb&xuifNaI$R%eIPmLKr?zUN6}xsD%DU<@Lfp5=ASyujP|@G2$tePfA0 z#x%?oSK%9z73Q#6D0S;#DZlr++Uz9AMm|7%-X_tn6fBqeJ965(s^t+fK=_y)`gc3i z`sXL{)1pN!O^K)nSK!*+Lc;oUitr&solX#nYc&x-I|6tYx@ID=wzgh1xZ%Kggfb@e4FIFg{%dvOiAwh#5Mx&&OxsH$ zzeH?C#|pmw^E21)qvqD;!E{vBAd2RTzn=7t2LwnDaO_t}cc9GaWmRPFa+#IqwXT9y zCsCs>ir3l9QkocSEv@x(W=2)sC?iH&Z&FvRb(n9wC~m73v}xL`Dhd9Lwl5n(?isEo zL!vR9L__#jw=WbXNS9o^x6$mXqD|k}z%f|=y=XlkIJfm2`;B{!P6av5n1-qNqL6_raX=i4AGxCYRSZAR7>agMZ%-6|Vmc^$m>} zsc?*eSm74j`n4#+nq5)vmZ=$fi+Z5fJ?FZ@(Tw32cr}MmeDw_FT4cero9G53wF(-P zw{%x%*R%HiuWA&4#-B?{Jzs6(gwpH%HCIvl*R73qhqt0_eK{670eyvhcSwEV(7!NA za~x?$G@jd2pq9SqohQLAm0&MQ#eqeHxmgj*&G(FHhpb6^4E_oQD9gB!%}vB7YO#Fi zo~6diz^-s)4qH3(9QmXF;v`Nv8_45t877#r43G~1V@-mi$Zxr`e$IL2&ZhcrJe0O0A)-`{S#*437XwMZHSXpzBhS)x&Np0Jdu z?rvq;1Cx}05^59kMRH5H4Tg%wu^Z*{l>xY*gPDUyGS~g)CGmqh>%QbIV zeT>N3EiKuj_DMDi+61pdv`*-j&46r$MIA!isAHxDM=8vRK=?Ci_G?E7YLkDhDstAq z_h(2Og*RxaqdW{9N0XOE4ZO7`HUjHBATP3*~o0SS8J{WHsYRxk}Z0%&X&`#c) zl`Qx%p7Ibl6y$}&4JK%dX*e#6q+jHZ0nM7dqQS=>L8>JE-_?)u>i3W12@MWwtz~IK zG_%_B&dd9tqhIIy2H$ogKm$6R=UE$i)FMmEL0|RyF=#48ro?pv`xH{NVV5a6auPz8s5+5nvR;r>(I(Ruu~xo2`K6g9HNM%G zRF|I`HDXN*$<(3B9K!TQZ9Rnq+F?vuHL2atOLm=itzesK&g#kzI zWRJgiNUDx-$B7gII~KVuBYr7VX5udFP+MRBXJs&_`fD)@$&p~D=V@3-z&uma4`WmM zAvI@C74hR8vwM+Zho;L4)an*ni=Hb|Dkxe4;=#0%jdV1mGzayckrVDIbWNOwmIKFy zdk-BtZ6)g1k**pi^8zkiYTD?3IKLeUX3-GUUP;hJPyMM%p5P)JFd`OeQrGYS-qU>( zEZuT5VWu$7|3zCobstDAZ4E&s1vZv%?CO*kwMJmJ$hGCbDKlHX@XcWl@za0~9K z^e{NDTit?|QFs5v=fSR&McawSnB7mQk>r)o_S6u<*3LV{uCc^%@^0j_sLYE+Oxmgh z@fmGpfB1a`6WCy5fzz1R+T!k>)p3*jfgxn23$)OVH0&Ky#AtQG(w3aNXakH9WGpZUy>Ncw)M0td@%QGxsOuXnlGc$ ziL}?pK>Zb7FQH|Q(X9Ep(#08C#AbuBVoek{X#Q%<;lKL>P04DrK&fcvCFLJ;3{Qh| zR~Exj|x?+WG1$}i23QH-Gz2CY1Y%T&>qvznu?r>9rGpxRh_@+UnQ+_?)p zkmT2r!GX{eO$(zZy%S>u#SORW6dR?_><-+U8-f@A*rq(x)nEQkY5#iLA@#Xe@2#(< zaC5*cx1n_LiK1uBelmX)3gc5eQ&6G|2190kLRS!5Hf6yZ>Mija^yPe64VIKohC9ix zCl*2`3YGi&Bh{PaMclCtR1uVG_k5@PWLEmCa(C>Uv4t}4w_<1!whdc8TCq9A#hdE| z9WY=e3l3gCa%q6krg$PmUNplOA$ng04v6(0B+{`9`xZ+1eQrg6o; z4;mBFmCma<@E%ibznE>t)xT3V!BVY0kU*wLS=kH1{|)>Zu-s3jKsxo|&i6RAC-g$! zm(U#DP9eaHx!ct@8PI~;Cy?i?`w2fAa<8I{X8^XF(A)~|Y+Sq9Xc{IZ(#Nk9z21C2 z2lJo(V!@b;qTe@;vQ(K^j*pM8aGf^V?$i71S7?w)A(*8q^P+1@JBCH=f%GWM55`o= z@y@X2+>1W>{>(Ys;=rGf+t0FTL*%9m=d5UXzwL}LHhC*o3{Hj|eV!w`#qkssbSGwl z)Kp2BG2DE}v`)Drz2^V)xP13HP3R6Nxx4af=}BE1Ua}a5_4O`AVW5Cm&OS3Jql%El zS9%*xwP|8NSUj!Fz!pybCe2+d*VlR?IAN+Ept9_a@2Cwau+wA(v>e*KC0&#Vv#;Ol zwn{z(RzkldF0kbU!)N__Vde~gj3i;^W!%0TC%1@+n&tGB3#m;}kxn(a=-3}GD+|Av zk9esp{`901+(yU9P^O|ympX-CeNDNbMkx^TxjUS->BJ15AKNZf|6pl+le?H6<*wOt zf9&ElPYIxd_mY^nNEj2%uWr`gD7F(n9M z7VU`PCY_>%ke!O>@p8k9JVChAEA%q=J(n7yhevR6eVyKFeXYsxgO*3hOgF&GvpoLC zA(@UGVYM+a##mWMO(xHQb$D5edc^Ft8}`Im zcVX6&+CE~XupZCd-x0RI+BA4WK1qzqVq*Gawo-fWBA60kP1UywJXcpPO>w8AheoS@ zrJhr6tvug(`s033JGtf~&!aw*4@prFpjCFn}A8R;`X zkiuD)!uQ|a#jg%iE#^k!`(eK0*#e0|Gl?A>1G9N*CoVt~%yQAwB8i zj((Gl8OqPXa}M94A&X?(kV=X_BK|(JT zM;mZhm*}~!GH(l`Ge5%YCwIRbx5l?|8J#-NU>-%#Hflvb%9L%t%EmyuoVi*sXUCuh zU?KVBg!i*!JTnOCP5E~=hn0HI^jA4}M!nV(uKIK--ChWc)W>vL#;NS~Z-n6p?mj7R z;trKjBgbw_;x#zm#`3!Igs;kig4kK3ixC zY_vdcpOVJn+)+%*`}E0C$hImV-Gf~ZCN_f^eeMK;Wl*)`3+44o$I_oaeThBiXvtl0 zO_*Mb_o2;a8umd(kEyW8=7zIhw&e9&rx%zYe6i=^@)#1`i}G*|S*ukK3+7>B;j=_n z;8mpROytD)jDQR|=K~dS3*E(1zfxM{?*_qRGyG9^`Q}6gl?PwUhs@Kg6f1m@iQbM& ze2qI;Q^;QKoxpD;H7t51Jk)zpWSE2~MU7G^p1-*Cwki`&18l$W+Hj_F zoq%tBFShPpFnK_r*%9I1h2>as+G24*H{e9Z9g$u)F%Yv4XWfN$tmGnywg#I6 zy(-VM!68#;Oa2V=o)*+dxHnhQ@dIj?TTHBbcG$C&4vb#;tI^PzK<_3*imEL8tnA?6 zfdAw6(^GYP&BEmU_mj{cWCawQ%)vWZcg8WhG<;d?54OeL-dwAZ-cF{^wP>)THx$!H zV(vaPpbO?+Bd;-ag02A3LAnrnQpufILB8uBS*c#z(4!+~;jpLb;-~qmc~{^(9r!Yr zOL_iK+Xt~gwSnqeQeNY0qfVlu7*jsRcN6Yd!w1#;Nt+35+`yQ(1N-$L`a6o!I9kF$ zvjB=hk$1VN48oB1%C<}D7S@pPft&29jJ-`Vl(z@jC0{}*2f6~&d4l2X^_@iz>}!b9 ztKT2O&9Bj<2pMX|*+k*-3VUm`W1zuu&NQo4wY&zYT_OFs`v+bmQ)jEhM8N+qrFxZ2 zK^)uXb01c4BAn}DRkWbE`$renWsPub4Au}qRGpvBd#-2NBt=)g$n!^1mlTAkBKIqe zhBGTyzpjPPj|j;t9w|(#d>xc_UnsxOsM9?6RQSdIxF91nt|k&!vYUX_{YYn-Kc07g z^tvKrl3ah6dsp&BJYE^YivDd*=F_myvaB`)b*LcUdomSg?#Vylttz|ssUl)VvMK)EiWTDEWFn&T^5Jk+R0a3|tl zxtW%qism&uvOIM9;s(icS8J7Re*0hOCWm_5aO$sN<+3qRmu{5Or-h!!M0B-IZf*8Y z#!qc;c=ghT;B4L+Y}aok(z+(qDLO(=`oRg;7n#pTgKyt&IhjTW-Qa~2It9vE8z+J< zXz}^d=snPK$E(*DEO|sq$rESs7*3@lKI$H|Trsi16duad_#*TE@!SOGBtQgA`^9J)_j7t7kfS}AF8Mac&aHw>Wd=f8b z$UAyb%Zrh4pbt+|Akr58m~bfnfIHKeh30-55%97I5<+?S!lm7=^hXu#j?u9>I3lPt zhflMpe$wj_1=DIk!RM?*F4t~+<0gU_qnI~a<&IrmUbkF}9+1EN@I{S#P|7*{%Ga6? zw}@9+(2>DTJ|OMe0&4CwRy6SX$duMm?}GENS6b@NulsH;O$x%I^Ipc#&$bj7;yx*! zMs4~;LP=A_hh~%P&I+l$d375OSrJOk>G?jC{KlXnwV6Jk6$hvIK{bMY$=^N6X407T z>L9IBx3^Y0>bZuxzqY>_ybrW7CBCp{my5$imEw-gzRj-wjoimHCe5|{?}lH(E97vV z+W!%0aNz9fRSYiiQYizZ#S&`&S$O`z0-)R}bMgJ$zZlK~10(+;EAdt1Q4)9+4OSv* zFd8)R9lxU4fd2QPbp?#ugbAUbW4@q|>-6LD!rSt9bIt|Bq1Ny-^G*~+!lq4|<(DJ> z^#1eC>lL(^t)HIVA_I?A&L0dKN=zB#XeQ=y0NzCZvaYUIIE81n>QS2s(-Y1Ip9fV= zm$`3E#pcqG!c|J;*yzDg&OiAQtk#rulqR7ZIy7iQeD43&l?dj3y%UL{zGVEMJug?c zIv;NLZ6&YZBilOL$POb;ur)^kFY~%+#=66v+roE16Bc-G>D+MV< zywc_vfNAB`vil>4sZO$az{O2-=)Iksbo}>R-d;viLlSg_awTV{P$v=9hYbQG&E%=pynR9CTAV|Kh#x z;Roe<7OSQSys~n*=m}<7wBb~J(n;6K)j9PI^pe7^vEA^6u+prLzGAL^5D>h(prLy5 znnbH$hVQ89tf6xw_If@#(dUM!@1TV^Th3vJ&-3uVHau?>msG`{$q#O)v;I6N>b^YP z>YkH7u-_ZoULR*zc=}@NfIB_-f{^GQ50R;7lMs4#UHXTg;QtN*bt6Uy$j60HZg_Au zHJ(vq?2c^0H%A;_;4m&VZoMoyZ|TNq?F+b6i1!!I%=QfjsHw7rEKh{?>S0asL*=jX zecKbQ;mQlvUfi+^C*Xc(t-hhH!lS_Sv)bYYN3O03e~p*61VMjPQ*xBSV@%!xS?Sn5 zxwg{q0nl~H&S957anwW!>uLXd`)ccaco)ba;8{0MpZ=PJ>-V{CSx=8TWM6x7#S!D3 z5uxJ)VSCnq64W>W4@XYaRc^oZ%8Bx%5MR;Ck4b&N34?X#RBq`T8af?~()ySya=vU( z;(GLK#Iib=(@@Q~ zLSDzgc2!YdD;RHCf@KaaDR1sdjaR*k8ZJO9G6Fyd&V~abekJ*br!}rj7L>YWUPBv&IV_9UOoXOO8EF6zX0h)%6%H%!o> zpVIU^4Jb#SEHAAMqKZu0vlx0kTxFf zOxQw@;U65%fF+~C`YgwwDXXhS76Oj7S61u#jZ9kKqgB?QKXzU=jIPq%UMP>0ii9KJ zHH6fVq_8o@1;FtGjCP?NEw(APqu`ogTiTV08d2uc7JgLg{Hgf8o4;-R#Y8`HUN%s+ zM;wc1tJpzQ zoyFiy*x))ZpAiUo`b-gj79`WV=f{@B$!L)&KbA-WTm~VgQ#zlB)o73!-1#zk@U9dH zzb!hEU=DP(kNrAh`o^`6MKMl=avlnP%8T(d2XbsQ9BpfVwv}dqQGDzD?t{azzu(Vx zw$ag9e%xG7-~*P!`>l7p`N3Ml*dQ~uM@Fnvx_Iozr3D&epIo-T@msu*j_`N0Ge2ho zBb2>BMA_Mg9;vSD?g-7B+$awg-VuH1+7S zW7hPp^{=C{H2Dt=)6>&9E&hD`FzE`UDD+Fc^j5VnKb- zKJl&fq(CC6p#;r+9i@fJ2a(@E9#9vwtvg`jcO?4h>$>nQqErAF;58zr06rt%9bP|g zRiTi57a)bH8T1TojaB_7Hq#37=QujCeju{$9Iv4?hvrAvYO6JxY;PDo_8bbJt3Ww? z_T@xU%T%WDv>TX-92BuA-u=jgfbv`-xIQbdcW!1nrvX?)KETNxGKdno~^36CV2G3u&F-ywu0Ff zpR|8Lc=}MBOH59kRNV>9N&YW zzynRZoZHNVYo^^N)cH!&P90}pb?PrQF{l$Ae{44q_f40vD#*KUbmtVR6XQC|s(giZ zJ=wrzGCxW^TZi^D^JO>L8MYj1H_COR0f{Bi7buP5I~U~Y8TO+={MIycq!(GD zkviNDM^2wdDjVxVV7JRjPYC~RnEh|+Al{2SFa_l+e-xk{#0wd!fQx~my~t7TJ)u~y zU*(qgth>#AT$!sc+tY*ZUB%>S@<|^~hJ>1gPpq3ACqoKmH6nrX&lLBBAxdoCmfR zvLU<8d1BY&TN=_193Bnkm;o*y%q~QZGrROt|{B^+nCWX@;_(w7^q0xYNap2cg zZkF%Wc*@%G-4P zD?o}3BiU#^?&1GpqY6bB5S*bk^>^_~o^;Xu*1JA&kxSQ^;Z1ZL=#B)q@qf{`?*j|j z7rX3czSh;%@wm)oSK@s_ z%4W=pA2Ak7y-nCtj8FImW+qGHuVqZp)LuWky)jDyHKos8*JO@|$-=vUo&ky42cHh5 zds2i_e=)bd5vbfjWiaZJOZNk3IrcXWYfzG}nJquuHJ+q{-0|txLaQ${hzVoPvem-M zWz#Z_Wx}LDUhc3&@hWVEAC5%_?-{%wHgaJOAY9@kOElwkW9!|6=P11(*>}5ewzvMc zl(0#|^3qvlKPVm6bwFdSjilPDntybW-WTW%@29Pb;zUL`)Rk9X-qcewoh!Q&V83Lo zje?Oe-91$guXykKUH};G>gpyVdphcdii47?iIU#|@ON{Fe>9nEl5)p= z2RyOE6vO4=J!VS*;`m5s#a{C9aK5BduJrO(LD+7^$FoEy@q+zl@DySY@>Woue)~4Z z%Y}6WSvC#_Q>8bC`9uavKuJ^Qjhn)mS7*Mm2^%`X5DyX}p`Q-=%VIN0)pAmp8bTzI z0a94ZM=rljZ_**1jGwnc3&&_^zd*!yAQ2N;v}nwTb9nb-g}Dg*82JtgV-(?nP&k9hl}wETndFgDM+YMi7TDVH5YZdGft;Nk2_N zT~tHa_8c*8r7AD4y(kLu<)!$Y-jzZY>z)dK^m38mf>_A}eR;c{@j^6h^DPFA5~298 z`a1u&ZpDb1RjfFKh0RMy`HNr-5Y8ntd3ZnZ78HM)-^;*x*H0+6z5hD-LlUXf_4k5| zZg)W5c)M5(Ka^I9T@zKTqAM6L#AU1}rEPBj!UxmQQhf7R=(Wv$`avuljdTGE&j^9we+Lq+9#fQ_X@!RmM=O~YS|q_eiPTpY9tjhAMMwaHP7O8%lH$1--^ z65NX_g})6wvR~M2T8G!f#a@_+b&`K!g;W9g#caV4)8LzBiozC;XN_r7(?CqEC zMV7@1bQEEa*EI-i4(_Ca(d+?eBR&$9{Fm+;AjTl&REsAxV25Xd0uwrUdkgS;<-UVw zvhI=F!SV>{Le;N`&=3ZrTB*iaxj987aEk58VWa&rS5C<23O+jIl zhdBeVcknT@^U61n?wSsz596{DdHipWd_`$^z#-`x11tBzLppMr}ySQfy= zn~A*q)MfAKfI_cCyR150`a)!q(=*73gk<0(fMYxiCl%+Yr?rAVZ~d%5++$Xw#MenZ zptR#Ie5N0P0BiDVJ4Xd8k}hxV+u>+=d3pBj_ZGqol0CyG>-}wviub(Is3vD7p83Cb zawsS}E~^CyMJ8JnXESld7aszTjMotz8_ir}gbyvSQ{TRS-t}pm%5D*m;B#`LoXx z_;3|e<~o%@n#xyyZOp@@T;{<&`F~;c-`tbQq+osoYf8b`(tOPGac0VvbhGUuG&Q2U zGl>OQOy98vOFyoSqq*YbmeMmF84U`Ii9tmrmW6U}!;Ph^wV|f<^5>6IrcZ6tJ!Jrd-@*Wy|Lo)J_!R`DbEsr&Bf35<^(R?9t)H zvvD4tk$Mgb0>3QyLs;gsjsi$L);#%FC_2voV6QX%NuRi(^T$o}Nl8g}<)g$Z^onRp z{MGnr8e8Kjxd42k_N7;Fq$ zeF)cuqP%DM2Xk~E^y=f!ZT@Oa$2uu4i&LqDX`we=?d&ps(YIy>y8aY?QG&OsceS;$ z1J5)=!rUbr&Hm{*)9u99@xWgd{0`c}CTjj}D$6!OQ?%k!I_+u5A;146u#Ug-rl24U znIC*L-w^!4Y4KQ7S&*MYtgWP0?ML5my4G>toqkuWF)MwPUfNfs@}fGcaikb;yD9Zq zvO%?&=pEH`NmNxU5#tUjxRf+76dPjh%1`36Wmrw(x|&FNY!l<)saJmhjhe2tf1 zu@!L7Jw+en^gqf_Jn(@5h~_T^&2-k z5#O}p=+rK_wzmSV@P#O`@T;R4&F&Rm2*$8K^B}O`du~zm(}tz-!uKke06L`&zs|oj z^!l&QDC1}?Hfp-r9t!}Tcj`1;%ZdG0zV%+{D-=yZ;kt0xzXJx|s_gRSkHT8baC`A# zrt88fx;+fgy#6{or37;5J-oO62ipAK6EiwkO3U8PP9S=8IMv!;<9PYD5}#?2+V;di zH0;EGy8~E?>XYDb*4UUe;0VPj2jlbR{wah1D1F_oe#M*if2m>|JYCxD#ZN?$RtJI~LC@9E|abe+}vi+e|0v=MYnE|NvZTY$8a z@f4|DwPKILgqo}A$yKX$(#3bD9y4`S$F2aqWoT#ATxGtl)fp|JZ+gwJ@gN;P6X5@A z?_9&#%(ggQn!1##5|e7kn6`>^TsnfHl*TCI(rHPFju4ktOQK3cqG()Gt?1~u-#VC7 zDKS)>YD?USpe|8F8;rXWZ4gD`?j1o4_IECQ5LAv8op=8%tje%}O50}}V&usW z3s;vY1Pwlu%r}ySpwtRE2VakFW}QkMo)ujLDDibg6M{%mTtvyxYficwr34A4s{(&U z4aua}K5w?4xU@(4M|@z|FJ>ES(Tu3jQ)8(-PA`>)8RG^unv~spMR-pgX^YW2w0C98 z7P$3>ZVD2*3lE0ar`mWfXba?I^jX30r^oJg_k9`Po1zwAerFS!v6A2~Ny|}+ zR5P;Y<*mKNedr?pYa)q21ycR2c~1Nb__4qpNk3bjp6&ZK&-&&JQ@6%{F*`%#`DX@i5#RiosrxqwTU|81H zdF|XNqn47~8}fxQi}37`Ruakw+Mr~wUi{Ekgl;YY8bxPjcV_z^^!P2s>?_U_dlIgJ z#MYG0p=tB4Z*eobc2|?J(~wc7pypQhIEc@Q+6Ik|BbJbDbNa-IeHq7tCWI$SrvN-h zMxJG?Kr1T6B%s_8oQ}2?oOdc(DsDT{sglmNR;J})sPZq5ZZ6hv{B`ORLos$!%(n@; z33o5awLEd{>4Ep|Y_~fAJLM#`=8NPyxxE}1%$jiHo<&o}-q5@bWbBo_XEItb25Z{6;lxge}SFH8rwj@B1ET)$Pd>la7Q&azDh24EOj1lNsxY76w) zc1Ez?VWA*`^w&z69v9fx2qgEk0}jB<21GX8>LF%?80TH9_hfjhC^3O$>BImUO*;m zT6(0!@kqpRZR=Gu;vS`I;2B5#D>QJ_c@x(u*_gZwY}e=szWnF)$%`|0up*G5b1S?c zpAF;Fac@cz$(?l$DD_ChK-`ZORQkqVHAVW&j5tR~lWx zB|(}MWj8I)lWD)q>i zTO6Fx$^{+wmtE=(!n2Luv^^#@yudc|p@kba2gXNHU3p~;=JP6~^1+;G5h}c}5 zb8EGHWD-`BOYOUkfOce!hw41lX!jwvi7g1dicDn-(rd$Bf1#p8Wvul;Ff*k(3Op=c z-9mVKquWnlnX$)U&v?Nz>%9DSn%9KU6-km0AtOsqSxRg2^UdnEFQ0k|V{Ki3EP`5Z(*YK$${hSGD$(%Ino zy?fQ;Dma%TB^x6ZrqjBMJ}0+*#h+WN)oRc7t2l3;lxW%bM#Jp>OUKc-x59D&MyFY< z?V)Wr8?_h)3N-k05=GLBeZ=uWbOm~iP6Q`%xZ?d&OW+TTE7y4XqbT!yQRi%DwUh@+xMPEudA+sS;KYs_qxohX9gTcwaYbK!9LLoC2nq1$VT=9 zkKx6rKs_Q*p0u=kEihn__lsU)-A&`2gyF+^U-iCH&eMvx_?XF^3f>I)Z@K$vLaSSe zikGWx)=T$ue6l7J)mm(Ft$9A_M?1rQ+enxF7b8Xg;7i^*a5 ziz&4m3f9dVU|ExryVgO}NBtxh%Z&H79A$^j-FTgk)a?IO|2+ioZs}&KC;qoUp_c_> zo05F7vRATDuh=_$d$RsSZ0Xg5_`}&9qih9?u#4nB6W9gMAvXdNPkFNNZO=-fiQPT1 z4a#Tba})Pf<6%YWyZy_1tPGAWzpcfu=kh#yyhq6l03p|chcVF5gQcqc+I=dyJy~pg z0li-!CfO3cVb#Tt(~q9sYF=>tTxt2EA4;@nP^lPc3Ll@St&8y=q39Dj=I-!<5BCGf z(^1!RxLsOyxnsKVXDh@t%9Ubok5pzNB~c#9`nOsDjAVLAl?I@A&h}I_!H#j8?BZt4s&~bJK2d7flGf_ zAU@|=A<@fK2h0+SG$hCM6tK0Av3~w9M(WF;^PuX%6`8`p5P(4jX(Ls=iRA3Y zqjQjdAdY|qa}&bVfgI%l$hQ}q@(0V{hJnYoT0vlTAiq-D`5WTw_J9N!Ne#153|M1w$R3W&k&>G_2%6c6Q6yueWZpx& z?zA3qlo)lCFuD%emK4IWqjUDcd%|N^upO)RuQmU5^qGkshpT1;v(s5PVU^m~*L|+k zB_|k^KJrEqjCMnnYS&}VW+48QZ1l9u(Ww3|XYo%~X_pQrMHY;Y$(Eof1N=xN#}B|y zKeWBT|EnKAU0^bDn+g8zlcx2lCxuEN)LX=}U`2eK24DKH-F>sZeU z(6<`MwDMqVB6NfW>;}C?FHaKIrU2fV?kgVRcEX}9AO$Ce?XifRMcf%;XFBW7Uvzp! zEyow?MgG?%YMu;=XVliDLQ}*C9pN(sF%hjOmIJxB5ylrOcx{;7s zEAcE@y_JR7KP~^IC$iYl_!Hr|LaElsfQ2Rs^EdEtItgt2$)klIE>vBW)B937%jsOk z=%!i}k*3H;cak2~@0xzSG=)0F)J3eiUD$h2!Jmj=_oba$`4y3((N8suQeVhRA4q_u z@VkQT2H63rjA72b20XOTtLGA_-+zvX%-&!Xb(HaAQLTFOZ}jzw zgiOy_Q*G?gO=NX1FJJQ3u&0g*LRI?IU(dZSjgkkATOezXGSqniTrFR?B#(a8FuD%H zmGaVkvS5P_N9qAp5R+VPZD8Z2GxpuySA!L)ryU*SO4e^;RkB{@dufE`lH5aS%jLJJ zD!x}cO1BBD(pql1aHuQIo_iu%cd2hyw(>6%qeadN9(KK+12HGEDb$`o0NMzA+*cUa@1Sz= zZ;9zeDh>xyQ5V%q(Y~;S|I?|?#a w{bSIksD4HJ__9z1UXH|stboLk~^m(z={`Y z8fa5Fe@+Nc=@Sf5qaoJ9`XF5LEz^V*k~k6r9*Gn2l{|TzZxlY2@Rq~C*!5a%n3?F6 zrsK;~3;2bWL9r290$TzZ*I-+k6c{whhof5nCwFU?^u3hvN)W`cXv^T@>$ZE8-! zYDG|cqx1PIj*?Zh{-RfI+qUG;n{HWcs-UF+U2-<4sL7aH9i<;L3wZ=7LW zopR1I%Mmof7^6-S9u3_m8gSa?iwZkuS7XC1rGfhNGi*f*sI0?>S*TT4=V6_WP8NyP z2La)@q7ozntXB#Jyyl`x%d%(A%5a~gl%QmT*ufHhVKEm$``sde~q z#GGEOnQCKq)##kXZ`w55)n&%;*iFn;8tQ7<|LH>8kE8))%PZ;jswLS%Y~8>iTq+X~ zc?cYUWrU}?gfxir46d*EIEgM;=B&bGtvW|4C1}wcw#s{q6v!F4i=y2k=Jht#wt;wc zO+s+MFj;Cltqd4E3o6mTlSnE&3^G+o0GpXgWb5~g4V8iA0zb-8F4rGsgBB2$WragdJLq>KA6&A*;w{Ln&fe^GMwa1!7^(Fr}V&Y(IC<{ z%n=Ls%hAXyLR^F*+dcTyr`8?(W-keVdp)R3#7Y6Aa7I!nL^&^=P+O$2xIV`(WimUJcG+?Ahs^c&Q4gmjjZ1Xu6D8 zCRjAVkj?eoU)%T)bke3ACDjlALLf5MVWq)rB`XPtrrY)@*X)9TZ@r_{>Y>o=aV|hr zZG7|T(Qgj4eIgJh5CWIcvdp~EvWI6IRYt!0gQ8lHD)l{~=~A;iu51b53gb$}w-jhU zVA57@<#F%n!2)(-r!(JERwM-$Rx&s|H8Y|Z;zT-_h5f40Fr=n8bAvp{J=erJLMp%_!iR=1uT*fleMbusiQCY8V#|W+=rEueJu+#Sjx> zWVY)q!A576@(bACD&O9$4`F;r(yx+F3!w)~b2JM@DrZ16w0bAZ@Wl#>sF^QfVIuG4 zgD5tR2eIUU^y&}FSELA9&xQtg+8Cl#c;;WvY17Or`T($S1%x*L(CAmC`A4N*hHe1UMr zvXyv`5Krlbrr_>ea1qYy26}2L5vg4%rPVa`;pz9VtK;OLCSU?>Nv4c>DDBq9GcD{6 z^nzK#a?a|;*vrbug1iuyVu(iF*B(A+X`x`jHrw^XZMK#?>T0c!A1GINU9` z0*HJ*el)`*5f}4?=RDOsMBo$~{}#C9a4Q0S9ATnS6~rWY)h?d_?T8kr-Pu_M2@W{z zsEc@FY~}XXY=^{=t`*#a6|Ivs?tJDBewgX8R3?t5y^hN97;*%7Tkdm<-H=O{&PoIJ7547S5q1KzwFZleJ%F~^ z_<+u5{=EX~l>L3^(o-^vJy&uYcU%GVn_U-)aND&it*cE%bPsvHK+A=@6DL*0BfTEh z{klIFPKu5eQ}91Y;=t`IO6W?Uqvyp^2V(XHwZAQnZ?g2AVdyi0=mTkTNGb;M_r}5cGf-_}SRH5r%1!w1^^NCZvy6KZ#_=#+*143mqp~UMe#F^L-QFQ|zl+_Tshg zXsc4HOseoE1T?@{uA@NK0C@QZNFk|?*~7YThMn+$e9gir=F%EZy>5UTXb58X4M65j z)h8W;iQQ)&gu`)Uj#w5#lz7wOTv0+WLJBnWB3ZtNUWy>**cf=hHO{pSIclIT$hrf$EK$mgcb7#{gI`rvXvVV6krJvEIT>(A*> zQ)tchn~&J2r6CL{P9itz(N64st3O+LU)CP5B-&TjCXaXZ_ zC5c+d1t7VpLN6#>&RIAy2S~SI6$;*Ebu?PH+B@kD8k6E0BzeCVwE-ftK?yrL_FerF z6ST7)yqh&^1{eR}!uT4o5VnobU#XbwSlmpouKj{g?Fp`%Nrpi_7i@Y)73U{ZZH=&;DpxP1i6a2gLI^LYj?HIw# z5CTG9CQ3p*%_+a^F%b18~ErgUJAJxT4JG%($8EjLVQD4cm}r2NjP0=rY0= zqHP%NG@{{;)&s*Si#>Fk7pf`s@7vS{@>+CY%Agf3dfe3CwofelE@J(miqOH1;ShTM zc&!h(>@x?4e}FB(kb*Iu>R_@~-DXz#)H`4^1A2}Zk-eUC>dCC-t@CKU9kX`O%s5Nv znOLdGyj19o9wZgbIokpt{Bm825I)jjAL9H2Lt+SwGwFs>V}QZ4mA zL{m75!LriYe;EhXXs7e{`9u_gl1~(I9MK@yyOY?q2c}Xg@kP-r>T;CsXiy+1C$C&H zubCmRR&gc^&f~Iwuqfp6a8wM2X-;$(lzJE=f8{y1JUFND^feH@8K=(SuiMfHj%_aCE1IZ=#vMOF zB~)>hBE^65S}5hPFv7pOXX2EHnB`15{z48z3gEX!%-4w2GKjX&i<5a{9#pcgVb_I^ zN_($E;7YPNnr7BxG6yy)1a9-Jl-_gt5hE;rBqR0F%^{3!mWRfgq zH8o@cI4b=U7<-7p@xO`U=*yK9VuSk&o^n~FA@oR%#>Wv$STEsSYZ_j>Sdo>P5-@!h z*Pn7`4DZl;0m73NkY6bQ?dgTFbfa6roPA;`4ATkFD-j`zq&|@z>={2#r4n1amlI`%7GNG6m*H{3dQu(ox3%DQQdOxS7=-{2Yt#p_6U)5|J&yG?cYVM4-xR-AioTVvGGN{f&{_c5_BxLpK&LIV| z+~dBy4>XwVG@LY>UTPP3jbOH`kHO?^t~}yJWe4nkwvNG1&WA}Be7;{k^D+8Ej<7Dx zL3+(C#XD!?D*A#%ZYBp{;2rZYc=!$MZ@f>Uh8N)>KnR zV}^kcD5z9bzSXvLOROGyAF4q_ei@chYD^As?YNZnFF{~kP)Zsc&m%neigN$Jnjl7F zUuqAu1qi4k;q!Ka@`WcHWnpoBeA8C*%(nPIVv-GYD_{YxYD1BD^z)^p8fxj#JMhzj zksKMF!cZ&lj_K{vmbs_Jr`*vKPq;eSjm&fjuZI$L3}J&Ut$Q>qrm_+gX10MdJ4Hga zA$;Jh4O1IB+7otTQfUZ!YLZ?rYxKcM06kr8NG$ZM%t$uDks{`!v#kGRuhxjiaESe) zT^tKk@N9P^E%Dtir~wdE%V^(SbmRznOg9Ug`t67?UX9*?>K-KCgE7NT{c&Ad@5DNG z0w!EHqBII33wK_|4Q#i?TL_aLL0K2+EvA*TaaxR z+G71{K#P$0pptgpzmmmyU?ms+2x({0{4F&7C`AMj)WFS1!FPXBKCA|F&@F~a+&hq0 zxyeoM^tZAqV!HnTi01A5P5=5zw6-_-6ecM8neitqY2KYw)zM}+kGKHQ@*Q!?6=z$X zrPos$n8^92_owP%m_nk|g`SK3(8oD1tstK5kW$FTG#h%5>}$*}^$9mdHB8=w_E;Sq zR=FFRk$Kq#&$Lk)#^v!c>#5JyZ8CGDmNC+l@#{L(H|wK+wLM**EiuK z-Jc#p8v*tB^WaU9%vO$!>Vm=l{kEcbnM)LIolX=%=JyR1S`#qLhu1|~D6P%oE#${P z;)C3@v;H^;-4X>nLHTi%kqXeWc}o)Nvn!knKK9r(oy)bekGE-IwboRI?;t1L+spI6 zIK#fehoeTgdWTnQVxH0ovANhn!JpL(ijRMq5RCZOw1F?2@(Svh8Xttil*k$RBD9}$ z?*4p)3>Sd9(<(VIt@|!S*gJR;Q%aOY0ocdQji$k(efeSzxj}=i;zD_xmXVlQemSi} z!mZsx1|uEvlU#*`H9waR`03b(7W=`}8vw~?ysR8mXc0V0CpSKhygxG!0ozL)p^!Dl z{7HFdR659(cI84Q!0XCX)h&7xpt%}8HxibTtdttFUN zNs_I<0awMfuH%O$I8ao^@Y4ZvO%W_dGy4hDQ|&Wl0+i(jZ21*YT*>99P4_;> za4sc^06J5rG!C=(c#M0YgL^}6bgCTu1O-7|kSU+HCFq!Q1;j@U%78d#ZT5uiR1~AC zhND756>f2%j+i=GD6XmN)Q?rfgSby-^lK>Qrh{Ll&@oY%ZsoypkbMVgM(vIgg$Q4@ zw*9=rel3%dFB6}ZvBu|}JS>F?w31-1hcG16+tm~h0A7NbA#M?LufJH0{I0MqbQk>| z6l~g(WL%;M`F|6O#c)#`t;}z?UiOs8Pn-~Tf;x2++<#c{&CA$FlFXe_NiOiK(`>8N zFZ^p(>1%$`+b^o*%M%D_HpY>7#~>fpf*8&5KVxmvlhUF|6KD2mxAYXEgx~imd{8pv zKe~!IObU!h-|oHCA5oYt*JxEI4`j5R!0Nx;eWX{VE9)rw?uA;`r@2Wx6H_NyK3Z}9 zK0z&sELpW0X3)!*3XiiJvR`+86V(+&N38d3zFzN&a4v+q^r0n=Nl~X~eI|`u)1NJnT%J~4No`?uMH!-xm< z*_&codr+d4Y(83KYWAJ{)lLj^I1Q;Vx&d7MZ?i7Ep9iwglV$yIL3Oz!wihV@zrWGfm!IK%J#3x07H^;o zJX+kZZIX9BGg7qP~_3Exrk^&YW~jTL(E4v9diFBq*q|{V%>+% z6^dcbb>n@b2{@Vxv_vWhgKnu?cl(WotBbSb4{*JB!9VwcGgq+AXUfxW(^GwlormIC zk*}pYd@}*mz>)$`xr}^}t@$a=rmhpv$ z+Giv(s#Lpbe1(Z&1BfQql{-pk?6b4!z`TL$1HEgu%7htGi`fcAI~x#1I3q~&D1nwb|h)qSp~IckbK03p0}@kQ8B0`3iPM#S-rZmI>xV14oW zTSNpXc6JsIQZKxFlKZs2Ula~OcT%Phj>bCMMA-j|wc6?J;zPA>ADdBqa2}#pJryS58Y2m?ZVh;gTY>-S#L0yO&95)h*L^ zCp~d?EK2)Etc2D96ari&)5{cv4IUBQVIip#e9f+GdsmJfERSk7B^`|M0C)d)LnG`F&A}TH+TlMUNKX%P+VuWVvQjuS~oZo18&SDl; zJxBoZh78^4Ift-hh`*-p217pEeh|g8!Auz7f}K6+s5fJn0)DnWJ`{F6o1<`}gCHoSSJ1XZS%Suud#EM(TLS zxJmK46hZ|-XdZkmyLP|iJ)`s~sg$Ik9>3mO&WOq}7v{W#O!?K^eI6YsqqA*(lSsXk6;p2{t=NO0VAmH&eug@oZIl&WBFJu{i#xdzq+!^qw? zMIJQDLj$Gw8*CK^$kodGWvzuSpNSdpGI+TY4gJ?t;n5=@A&GdLWz&f}M#d?Ipb&dx zLehvllGB{%h{J*h?{LI63&E+Fk0W@FJ48E-FNNeKKepZ%9-M-@s3c(t+3y*S^5c(c z4*__k)ek1#?1VoL^av@ZzgRyU$1^8)O7>dp!Ed(?Ms?sXU2p#<(_GX5j!6$g=LbvR z#5FiKenLKc=n7%siUsAdC*Q4BoBpK4b|5g~(4F(B5b0fd{>>0LVCuKvGVg*Bcur5V zsABxz#|{uin2hR!dH4Qd4A%XQ9j02~Xum!58vnTU-l*5tF?|&?QNmq$5C7A8?!N$` z!U2&YJZhOvSf~5$A3&MhZI(u8tesUK|F|+0_AF3_AVtK3Xxp99?<%&FU9%+T?mU0d z+G`qJN1FxpP03ELoCeCN;kAkjOV&h#V$l@V0R_&tsv|BYk1Q7Idc0P+zO z`gm#d6)28ZM)EgIZO9FDHaUOn6r=f-lzW}yjTq*g+h9QUOIN=2{!5u=uXqffDWBxm za;j#lb+%#PS-h}CMZYp-%Jn=*#+7j20B7;yR0C$g=4JOne|3+*P&JR0eYi@G>+RLm zb0;{1dJVY}1}gBaP!nQ5YJ5OC|88J>9-9lPnLV=*rFrioTxbc54p|1_ms_G2>JzU) zrFY`BA0C+57KyywesTAl*XR}(+^DpZ%%<$aUH_3i4wLtGh`6B@2P#SAiw%AM&Jk!6|rD<7d6&$B3g4M8BvNSue{=%Uaj&G^+xc6II5G$oWISo+l^Zfyz zVFvi^ZY^+Dw!jP1*phmGxnGturaqy0Xm&1VQ;4?9Sg{aVSf6NX_-Je^yEh{H%5<0> z`s*A>izwUVkPqNaszi0fEeLU!d~L{Qk+xNn0-(+oU7ifLO@$*JBOb%4j5q!wiY^0~ z#l$)|zuF_42Wi73CJXu`VIQq}Dd^dv+_uGj*z9XAbN%^!Yu;mrGrdeLUr-`Twpo1x zf9=Yca!#YNTWERIJ(_Fx*F8J4EsFh>?vv2r^W@3HvyeI<^FPAI@n-b{Qqq3vIFt2L zKy;sVyt_%w%st;Eb<RF^frehea&f=`}tF0p2?@dl8MmkHgk^<$L4i6RTKxML@O)B&f{%`TpWgG&vyk&Lp{UPTZa)cMr<@k5cKnW0L*Qv_oC@xIWEG z^&_YrPKCU#BP~E_P7HOK>h)⩾!yGZE4c`i(M)g0d#49`e0ztyfP1Ot+<6qyC}b zO!LpuleD~Ttlv(;A-=L--MdHh`mVq41=gnEH!-ZTmGHZu4_UEMM;{= zf^4&?+sK3I%Vr`1>F;o-wuD0KOCSeQ?8~*_bYc8@*bb+=zvwB^I%{ExVjoUMXTvvf z3u#_S75hFeGs=VM#U)2wE5-`c=4H@9g`MK#+AboSn&q%ZDf5AnTYt1axZm3bl*X#Y zqv}`&)fJh}NZ+woNm0o2j|URn^fPz#OM$$h$b6*&zydmvBCR~`XX2O);UWV^d`R0z zt7|{@S6*C)9gjA1pOmv#At|=4vn3%e(X-#B1pAO^O$ZH9q!1^#W7K3)(RthVlV{PD zJ|2GV+OlI2z8D(Zl$8?A?jKR+Le+cg>;b6EM$Hp z9Wk8mk5+%I+&g`_9xY(xi<*+#UU&&*JZc+;X%tsP*6GtbKv_l2+bYkN?(r9LSf{Bt_Q!L#~f zZzLRO{gm~`Z+r_ApH zN3tj{ONyUW-{CRPBO+~PUf2#f)Z;d%ftwsW+`NHx4GY`}6-0BBNyXNqJ%&%aHwrX| z%??kW{9|*WA*yYo4@9R)BOyWa3#9l^$OhH&jABTm7l**<#^91?Ok6x(AoqL|3f4%B z@c^r*mAefUkXcvdsbm}U)X+t}Z9vdB(`dS2 z26ARm&{xk$3;XCA@L^E@w3f!Iw^FSz@F~FlJFRc{{!jY5%xCV&{$0K8Rz7N_4{+Yt zLAbuqeZO|)@G+NO8G3a9vJl$+&7)ev?_17k-NdEJkrAz# zuDY!DdeJk7OYx>0RsGwLTm@1Lx-8++u6U`5%9{TyTFG`~sQwz}Jq3;Q=8x7?L4TIi zN|TS|^vP7JWxF|1k%HW27gk=UR>YkWMJLTvPcQGQkBOX&ncVMf-!iGHyCYMVrN({gDfI zD(r}_?(-jv`JuR*0rG$817w!8vx<{Lq)nqgDHLCOU(z zeBlhRWB>*So;gy`4{&z2c;;Zp=mZ^Ga_#jbZ9xNu%*pL$-^F>5Xe47T<#Bf!;>~XZ z9}Ly_FvM;7((EAOQ}?+2v46;1X9|F#TCDU?)UB0K$!A>WbWxiTdoX;A{2(r*Qu4Ku zRBhf4yFULiT~6Vtdh|It8;3D!MzU$Tsr!y&3*6*jarwb2*Te@Lj;jy3Gi&gmS|h)~ zr2u2(7^MEjTz^4{x;rfTwdJ5JR`bvrKAU3ku#nLEQdCd<8h5YljW`pZ3$w@mky(Y7 z(U>&B`%RQt%s|N?X(t={y&TgZ?1q1;nW5lR5>Esv3hzY-;mwRE=;G)W&#|gfIL!%0 z62xoOz{3e4kRJc>%Ne53dU8A9ffnJ%+VXwiN+WrFGuVkm2+}nNmg$!EU|v4NYv620 z>HkL&X`gY&(qT3lg@Vj)>40a}Icoh0U0NwGW&SO9Nh4n=rcYpxHhHdhm)*p-&92-% z>5#GMBA-q7ug?TA2JMck+Am47MB$-Xr)iS8roIkIn73bhys;jA-a>q!;jQ$f-gr1m z295MacH%Ddz=SZXRpESdJqr%7`n@IgC21tcv+j%woF`n*jTGhdJB9<_i;yfwA2LfY znh3t$+fV zn^9CLEbT)OXwh0{NROs-pqc4#(v1Gl*fc{nsqa$*b%K-{zSI+f@5Z(lcJLD z9w|n>v=al{KEri*SSt|xVmIePUOhYd*%o8L&}OZB(>|jw7c^__LX**!8S;#9DES2_|C2%U zA2ns7(h?Q$7ZYE?#nUWMut1_sdfro;qCwxwCxL>V17_7^(w` z>DO#nX~?h9-hgODEDqoV%F65C!D>=7a6LVtr272Lp+T7J@ zcCc&&;tmfB9sLmm`lTABIjBh+txkEIKk%@u4h5x4MTqBXSg0uzrjSbS0dE4`qoZLO?@_4)*lq97LsVV)$j)MB-e#oSvtYQb>s{?7V;n}#^H6k z(h*1Yrym+H1&!u@fjK*S88ue#M6-TzXj4{Iqh046`kMMbZ>;b1AU`itcY%xaHLRtFHGdy5S6olYyst3m{7r8)9Q_FwX$Usg+n7ca4Iv=YU({C zNPeHIk$~~QHJU#)DPIi8FwW@09}B+}sQ)GfKHc-8dN)(+FVM-W_2*LRE`Arv(~!A( zEnJeb+3>RCHM7XZzj%>PbTnPz0kNWk$FaBkW)nD8TX(699`8*(l~o*bdlG1S?yaOZ zGn8zL%%juY5dSO!Cz_r(XFVZ9Ybc50Qw*iyTwvZ-wx3N(6yfgqKJg+X2mf5`*=UbR z-{i!g3(AMrr^S|r1}S5BB!5CQ@^g)sZ7Df6#R8$~UtP(@x`9c~3pXm}QrA!F%r-`- zYW5U;*}87%yRfP`+C_O$Fa*!^_X6*ehMP#(2W{>DLK#81qjAHgS;pZXg!Cod=X4Rn zGDZ@F;!ffGFEpAKIu2*60EtUR)4g?|x;qBpFxY+aBVVAmBU}noN-F=cQL!*eS8R*v z4;xg?S1&zt?cvC!XR4ER)TeCQ;K;L>88tjNXoO$PY(`0~5i_jUV188EiN8pWjnpVw zIs4NTviP0Y2+LC=Q+*vX;$#HzxRu4Kvi#ilRUkSl1-k4A!8>Ag!Ms}{=&PFo+UF)&3E0$M7MR!jgR8vY! z`isN}gjWTq&>>yKtT*j*Ds^$48elRVkj{{emy0sY?azJK`g^H_9EEA;kHuVqAF?=@ zh9BvwBy@^<$&$I=B*&3FnOg>xfcDRTuScoQL5zA;6am61*new|1m!Q4m+x!KBP{Y8 zP&>tM+gN+$fW0FRL*G9TMJ~q}g2yRDzM(&L_RpIQ*+1{J8u{u*!nNCWoj(KPWGVG> z*+d4r;M!o)8Cy_J6yfN);)YwXhm(Jr(eP0hEd%c$9Za^eSDC6u%kqIG&Dq+>6xBqD z+8J(4O*oOfcIin`ELFrOVSjiV%vHHUqO7iNrW)+5DI(!c<@Ct%WiJsF#?0DuJn4*} zL#oR&^7r@IEQX$meEWEUjz4z4MD~s2Kq^V(!LCEz6=!-G@K$Nlmnt15DE`|gLGhB} zw|}WdUZ7Z^k}gOd@LWW6sl#9Y7s?7#>dxMmPw54@dcr>eV=KmKLp9$hejE(HLJO(& zPOT5VW7Q?;fCvU> z2$fiC)V`+{9?aOy1dR%Z2}aN?EXmbhxW4dCA5hp$7qTArFl4h4S*FI$_QWx2Mr`+7 zT3wb!rwI$E)wX*~E|tNd6BoIEZO07UsHY@#WmDM5+D9@aJlaIjbTU&T?d7z!E6J%h zwzbd3%73c%X+i#bV-fKjuW&5Rk-LhvvZSTw3bjqpp{36#)f2pF35iTA+kIcBD zBqT@XSB|{tcN$d1TzW$35tbNc9FjvKJ&0Xi#N!2jOB5^wG%#A}yp^jv{DcmdMP5?s zHNh&((5D1tM{|;f!%3VSKjRks!S^?eM91K3UXHPPv^)Q|!}})M2&j=qo3_Kuo%;#j zPW+=yNXr+N9xii)dIjJ=y!@{W1np3$FX&a|;1hpfEaAObZ?=ryF+FYf5nc~2>L&-h zj`af}6x<_vlcf+*oIpbw;!{nucLp94*fIwWq(PW&8%I(G9HaRO?hNCQmr{oLg{X1I zR~W>_QCb2lobBHhiPBrCdap({BNT0#ix&2z@7g_rXy~6@Pq)*hS$-_j?&LWb{9Sq5U4gW546PqU2b?2LqedeXcACsQZ+|!fW+N_NB zidA1d$GgTW27pTb0h%Da*+jPd<%33NT7jvBM-zm#KLT)9Or#Y8uuE~R#X-bQvGL1#hKNxC^!`cCAwZ(76I%#HZrah*_r1cl@k3cWTt(M z1Msbhtnx>&!%!_i>JhgKD%WN@0mO1!S;HwzZo;+dyQmew)Hxi#LuY>8lh#qoCHL9B z0a?5q-!+LOdUN?#>tGXTRqr_03{Px+fO+7ZyALbzu6zzd2vgkvg*m`(QO0SE1pL6Q zGZrtqm z<`u7lyJZ+7p01EE7^w9dsnCfNC2>}p(Zo1oKl|GLOgdZPZ$H*G!HiQyw58py0BvVy z5ms=!HK9@EY8k-EWYE$#h!#jH^9{O5VV7f@9CQrDJ9i%GQUFekZ3zs>h zJSy_QCb29(QH549em5#uNFsvA=;_)Eb5=-5{d%cph{Krv{BUNljd#XRZ!9N>C8@Dq zX4NZ0uxCq>=D{zBKpei@rb4|4iu6ETj#R{+8#VG>tE)z}qKeXD2S9vL(;B^-sH#;n zQbAO+^pK0F`CuAiGi&$1@K4EsQIJz$)Z8xjm8+Hipumgth|Vv zaBTKSLvm3hE3fRK-3r3_vZd5b!CV+jA*U=jDI6I~c}6z2AQhZIh%4nKNX8uKRCChR zI6OldVpsIpMRmfC&=ceTKxH3+{8cQY)v*2C!Qf<E=QXV{;rPQ= zlg)q-1AAcSS<-=gYk=u503T>hP}q~QLd z(V|z&Qx4blCHrjDpAQ9V`KodGx6?pa7&nL-_O{Eg`cbkptEP2&<7ObP!Bx^iUdxm5 zV}H$nhbjPHuplDdvffGo6RQgvE2#S8nc!7*aSOFC=F+gM09~AgBEM&0IHi`%p5x9H zUEToUJ>mi@HVZ!mkTOV#hp!dv6AHZF&-b~YUr|F33hj!DR!U19ar?vsnNOv?p>jgq z*{uK^a9C%W7PA+pLwidw#Sv%C?AgVcDl*s{6$$Sf-}%NpAYU6~U;`lAuY(Vji6eV1PZnMFR58 zAoN}oTNbg6WRjO~-5gyD^I)pk@Vtgh9_Lc@i5s65t7)T>BTFt z)92pj&;qZUJ~Pb1kAc5^z3D*@wa_%J3KhNJ1nF^k7o$1aHa#{aLV$N}YSAN^yKd3PyQwojeLwp^sBv0hq~W^ zF(2*w$?OeGuMJgUw$v29rNDX{RTQ4bm0RQCNrrY#2i0TYw#uK4LF0aoKzXWhTK(wh z(O6gwIho=etTN2(LegX1?+rQRKzuPHx z#&$A$LoOs;^cQMkhQ0jx+aar0JbpvxI6%Q4zR{i%_{aCAU7!Ol?u3z4 z`$LTA(09gfG1q@@5%1=Qj7b?%5ssu2tpt&Q=i#bWsxUzX9v23H%CMXLA-zUQdyL@7 zl`((TNeG33AVL?{*6;$u&kf4Rj-*0+>Jb_bR>Tt>Q)7k1d@8zD>58;i9uG$7sJu<< zquC;cE!6rCNoys!6;Gyuxu?jB;11*z+%yI$1tr~onNS_Xaf&gq(t&AmJ;59G!vf`O zNEn2Yr)CooePBv0q0A|aLN|2JmHM9j!YS8nVqDYh2c0NflDUjqFXsm>!TEw<=lRk1 zXAl=$h5{ZB`1j`Q)aARvqKB7QCdK=*nqhrv!Bu9*lkJ6y+FV4B z)F(@0_iobNj&1Bq?)Nld+s0V2#85_DWaI0=Y!L%B;vlV(AU6tZfm_iWK7Kw8a~Gof z7kcwX#>6xfCyK(_Yqg4lVE#B!e-FJ2q)UDU=xgq*u@S?5NTL>zP4dRrhGjdES2m0% zMdnj|s8jFaL}>m_K>#r;)X4to46^R}Q@dG<$bFgS1*!7AL6s9xy=MHh;9~(T-;+d{ z2f%UwO?5EyuL%$14cJR~w#yHj0joy@@-m~aR|N+J^V`MRq*JV@lCkhRK*?C>#3tQp z$S&CGW9`Imbtp_E8(;8N0F6S#2zH|$ZrhY2L}L=RL|{E4CsfVQmq3$wWZ zQ)?M{c3ivEc6rJ=3M0)p&LmYU(rKQagz2(|Nm(s+ZbV{sa9!Whw=#38&g>mY165w)o0x4b;R4?*2y<$*=kg6og8lm ztsXpO83*bi!#n(6(r-A`5!dYiCo@5#z#)nhw7iPMquB5?&Ebj9Gb77wg)`(r`WwK< zr0)z6Ttz9Y@zq{fiK91F?THe(#ICdmW1?~iXvshWf)g#TQAr;QdOuR1jS>FkPo*@k z!H|_hz=w-%!D8XG)|Fqf`3E+-Qh~ijW?ZOLJ;mp%IH0%QY@y_a5g@@E~8lC@SXCw}9 zgSe37mj;uuhtC+RE|I@;?BeKwrPe1r5vW08B2D^mD!-TEK8phb)O_ zn)n>18AZU>T4opsJ}X3r!&Ix}1F~qU5>{)DO;uV}8=-z? zhMz)GjoENHp9AHaf8;F5Q<2iwUhRes1>ns z5f*PpnvGR(ZNoCMp0@>ZtF)|BnHby)OY^x5uuO%V(VTvZEm46cQt8-%?!w^pB>eeV z8xKUd8i4Wra*R)?2L7v=M`Zz*^67ClE;$7+Wxu1ZuRt$ut6v|FK&L{c-}APwu?zvQ$^N z7pc-JV50Q3axd;`q#X)4(9%GS47(zRp%icl9|M9~>|cZ>xncgD^fUD zIl*Xq5@A#RT-cnihkw~ME&vH9Jd9gtdMmDg2ON18@wMEnX2e*fx$2?FF?mQYSl2y- ztid*?1qhg9ix{M@KA~eB(9x4k!K6pq1Z~tcFv~5`g|*^ zUBr(cl#06x!DlOBv)fwj$i;jb<^kKX@N=lp5#i<+4Lo~qgYrwv2ONUZ3i3Vh(N$%E z1Wb@pS1t&LFO=QY%6ql(9YHv=U!I8y;DV;m zvslkP$Zwx;9KV9U7B6_X2eF07t8jUg(rk?oMCDLmP1-7FoL77pP`dEf;eZ;i#!-gz zVU2HegYnbWXY>bo=eTF$^3!1u7)p1^2SHV^N&>k^Ajpp)7y1e(MFQB@s1$IDJc@Z~ zhC;qAU_DO@7!#{?jEBw$G9D;vsq{Px;5?Tt;DU}%HiV#qq10{Rv6gJ>UPh?Z6e!^r zaG1f+k%i728QdI{O5B#dGg^8J8S#T4(kmVAVA&&CMkx!Qi~Yinp~tr{Vd=!`yR^XZ zcvVpnD2@b_0xphRZK>i7s1&fRWC~vyvMk>Um<*AR3wnT2z1-jGpy2GsM%?g~eMoT4 zHApK?eRvHI*77XmK7PWkM<`%tj{u4HVhuTat5OCbD5N7PIw)P%xfYrd3hnWWjG#qV ztK2W00C333A{OAt!x^p7J^ti)!3Zq~LW%nsbQ|8RTD zIh#jXfV|Huv+MoT@ICzIn|)i*I7+a@7I5|$;5U$-ia?XAY$~coNx)tLN&(x;r6{!_ zp_O_@UzBlF15qtTA-{a$$h({@V7!&rVB66Gzx^cPG)*2{;3x0}mKS-Z2?WqcCL`iA z;cd>gQi)tGX>29-zdcA;XTVkaOmaN_F9Io41eOJY_s@oZNO%=IBh~#=t1L)-Zw7ft zJgPNBbDU6B)_AIs8>N6t_|zKj*el&{S=LDjSr|24L8r!W!sb7;R&V+ZGO!#o##&XoPa60W0V%aksV!YFZ3LNBqv@og`=TdmYX zE=PaF#mKs-2LlGT-u3`VHe(XCowGVvy&Cc*2QQq3WM3fTufPR5%@dOIw6TJCwgo&y z#7&_Y=qu=6-zj`BDgz6BrEtWHey!(NQUJnwU7(3k3TnYF;pxaZKA>PsRE1MShADQWMj=z7HlegWP+NWdhh$1*eQM@ zqhmS+Q@}=jPQh`$`K~cQu8SA>n87F6*ntTw9BadjNG>b*Dwgm%mCF|y1r$&MRX4i} zIKGMl*NiccIc)}G-N?YS9~%$r5rliZaNDO@aigewhBU4PQowNABB+WTZ$C1=4$EXy zq3k>(lI`*1wd-)$kaYj~F0kV9{qQW(P1~3SCD=AN8q;3DNOf$IoBrqGka%iJ&QHmr z=T#xBHZHWsNz>=sf^xQX>TiA%X!)hLkD24soa?rfi(4tgOxX6T@!b$|;Q%h_2CU(q zh)#Vxn&PsSppH0!H}r|b65>35V7Au%?{h!I%z5GFTMtCE_kr<;mk=~(FJT||(fN}u00qN(Nw13?1m$~3Uet7@fb)`Hj>yUd zoulA#N6X%iiI-C=*Fy6(t^Qd|tbIq7=FY=tz|?|lt6{MaJqW2JU7m>8#j~ixae%7>GGz&DJNW}|q z9p;t^cXV_GaMIk}==(1L=)^%AqjiD27(wah3XkXMJDEtXefYvN#5!I-cnFklwgg_M9&z=f2iKU#8q&c1MdqNdf=K^Z@851?xON&t$+Q+cJ*I`IEq z>xNVDaaf2Ok=YSeZ~X~Snn$I=PJy6ETe2Jz&cT7xfy_Ikq7%s{>~Mufn~w8>3>d;qNXuwoM4a3cStBmQaB!F`%-W zK1(xXOC8rc{>7Tv3V&J?H$J9w{7+IFMXAn;5*{_-&-Q{dI z3@h?0#_JEjI%yl4Vdxa5V0|qxEzb1K=-t!u24&k*jwI*U+dCZrS!q@lsc77u2EfIRw0oNyPS$C#^wRNc;JmBo zBqZ#}NttVAOin;t$LHODF(A0J7poZEjnFiC!@y%aY%WBV=j3bz&HD-yHKl;P#Z+(z zrGN`AQfGn|H3U@(b%>i;q#Jde2-sTVOj)&LOMqr%*SXg;J+$r}K&)_w3fu}krq2qB z#Wmfh;^FRa*tiTOyd|xFn8s=MHa7qWGcw<>X$o3}&}-|DL6PF{BxaB`u#;PZD;UN8 z39DRM?w-8o)R%vx7#Jvr0^W}-R%^LV3;eSh0Fsk;!j}N#6sSK5CdKfNgReVT*PZ&RLh7Y(j5U$t#0Tq2|mHFIgIb|5KpjHp{(=D{` z$TlNq1P&{>C(#UrHZF1P@d>ZICA|W<7tZd<6&ZZa$;8EY5#284Gb?ryvj*M-Od+iF z%x>K7b(|GE8*cTjWLd@kEpoO;8`D3CEM%FYM{*D0fX`tx$*tCl5n>lb#!yO##RHke) z!D|p8DFfUJx(;i@1wr}ZyAY3Nw`neD)r7svzElp^dO-(KZN!uUF6k4aMFr?i=R6`` zy5(@c6@dK`k`UbM1#Zg>au$ED0P+>Y>n7TE;8n>r-+RH4K*XOr8IgRjg5wGps~A(@ zvlq(_?Z0Tg|CW6YZLZa{bZG=*^D|+c$gS99EYrO0O0XD>RNQkKw7S8S zm@UbyTO}!Gj2~W-j#MW}2`B}e6h(0_h9jUIQXK~Fdkz9?p9qXK$`BAqJTI3ax7`2y zCRks;%>_4!!!p-vO!gNM@j*g z=mA+;cmcZd{bKcEOVbRvHu+vJxKZAolLD49Gr9I%d^qm0^^hy_@`in_;o8w5u1&YA zx2xlv{93>=d~X~k7V#}X$O}DSN6(?I4odhlC1=q+bN*9{gt~HpmK)bZ>XTjt|OO+BzptKLTgafOzuJD35yP!Pb z#2c6P(CV~*sQIK$LF-JHSK}t-p0^=nal7v*Abx`Jn~ugr7^x6%IhHC{USQY zH!J{&cwA|~a-IBVVjv==8D?S-s!`rhWn4=+TFN?sg{I%K7QOYq2PnS8IZ3kuy8_6VPb3^r0xsA4Yx z9tyE@?3=u6g+<>6RIn{HobMnWKE<25Jp`U-AWM^JJeZYk(J|=RgylRF>NrqeI4oo# z)`LBjFqG%_oA0`7yZs2gYdb%EwcKuVfH!0|1o&HkHQ^Q6Ek&pyHnyNtIqkryMxBt~ z+=9~ePi#*2wO{t_eeq9LK{`o5DPSkb6tKSg&1q+W+<^ronIIL`cwNpfx&)D0{=_8Z zyj;gO^`RF!;ko!iFnmKQHGkR8YzgbZf|guzLEDmNY%E^zugf;Lp@UA|;NgueLTX8= zgs~&YuF&S)`}X|UuElSi#(5z3b0%G47WNGfzJ#3AMfDJ2Z`UqZsv8jv9c+@8$*{cF zX@fpy0ADxCDQ>`yWG;1hz#q&7PvPVy0i}R*lUpDx%*vMpLI?niy}_qqtIOH|D5DFM z=-jvmBF9m;JowylXsYGD&^cevxN*8KSVth;X9A8P0#K%=BPoXl1n+jqCU~QLZ%0Zc zg}uRx@b?7jGW&-9;GaJ9p@-~GXi6pGat$KE-Oriyh>m6yJ&uWi*O1ZcF~}*Hry!G_ z%{?-6wK3B?l+0K&4xU(H1LCg=M+$2715*(+m{X!2f1UV)!gjR}`-4a z3yqG&&C*Tv5f_lWn{a|s7(^T_lDx@RJAtOoO%vzmGYDalw1RzDIbrz>@UG9f><9J{ zL9nq^CwL$b(U9wu*pY~b_PymPjW4q1LkYit;r5Nx^B+xKe&;_&?t@=-f8xt>e{G?5 z3&>x^;Y^qdgf*}m*|Ul^Q6&poyK*Uqauz;mfFL;4)BSg_6 z#Y|A3iIW^0wh?c}->C&3XmElt7K;6o)%EYwjM)2xFuxI;gu=*00!jhrA~l~l?Ux>K zEI?2837#)l1NNSdM-**$zJzKJKg>)$?}<>$2`_vN>vRvM$_EzcRw!TTKFrI@fiQWg zyaHPuNZJux0iz>e4eysFnf*xMFXtkNT|42v`|?(r-EX&l2JsS$g%xeV`s&J@8FE3y zeC1%5=A4@_4T=Rt<`MI4-~^s@1?|Tb-Y}ow_5JT~!#oRc2S`|0!sqS&-sg{V#Z!$1 zk$_Ub1reh?5ksR(>|2Uci-4?Yh>Cs-)bo0w7Kgj?QpFKApP|M1B|_oL5Le5%+)nn7 zq0#W2LJPwMoWBc^nDz?jNJSp7;B}cpK_J%DVYG>yiJWtbV*+wr>wWkAF$WgXP`WQ! zIv(!fZw*(DWtO>!iP&n~X9jwu;pK2M73;z~a=>#(;S}8Z*RAR9M2pzUsgp$pp5>v~ z>F6bEQo;p&vxvPb(!&x^3b;sLg^rIOdYnb|pC%r87+sW1Awmhy!_kHE#idG$1Yjmk z?+8!GYkP|%hXfsc(fY{0z)gA&*4?pZQ=6I6P41Um@5u?7Fek{R5FG}Y$Tepa;UmCo z2r~VD{@~=Jj{5iXy4>z49A7`iNk=0EYxideGb(w^O*SK>(yxTMDZL{9JIsQDZ68DM z6>Q&G`S+Xy1unMmx`6>?@^VW(`Qh&GeBsC(s4J8J5>N^_fUx|6;>Cug9r|A=Ls%Ea zIm8>D0&cHg{Cwe%rc(Z}b9*u8+j%A93;LFPHulzPh9+YqwT+&0U7q1)CLC4GbMnr=6M*jUN3T_fm3fN6LIkt5!Ipiw5 zO{HrmUf`xX9ZGmJUM9oqZi$xSd}2LnIucL}TQ|k&qbjL#c5KH*;|sDZAT8mp$Zly- z|Dh1F$W@CW6D6`H(aX8;04p8=RBScAmvS9Az}{WDt9eq>`={J`Ye6QJf~+5)f5#vu z@HdNFW&2k)Q3g_AqMUzs{bDBf zP3K*XT)sZq5}9@ZPUE>HUt$`?toP~TzLv6j32XWORa;_Y$D4%7<@jHTA%AR(rTW7fUn&4K-+bSB-@w>ZHGL!*fnR> z=>qU2EkdgsSc^`BxRo`Fv1?1p7|t7Hdc0f> z9y_wOby%C6V%wl9XEucU_|3Ug5Iuio#~_i%dyl<)~mSgU-(yMXbrTto}D7K{nP zAuMK2Z&cfs!e{P;{S@~S;I2`B}eo7`-%CiF14 z(&EYPZQErfG8f@TP*?mQDbD6#YfDL7-j0K<&;}@yZ|uCRHLq<$nL}%R;PL;~7)h6O zP60!32M=V@^m`AhCSQww0uFL`tWU>n@fb%(*eu& zb)N)u64pX4Vk-a-y_Zj+(~!+;Lp~G|xOjfka9uCtlDc<6(0gu&O?Z>Gfe6P|ay_TB zg^7g+p5%sjp!m_Tp8Z1#KbIa-Adf&`2tB& zw7cM>8`#>SN~Q!z@U?uv`Wb`|-Bq~KeE$;jf)ac_6BZN3f0Aw;-Msw1n={+;x$k^)!}Y-bprirM&>>{?wjeIbw?!%hwJ|I9 z1kn8`JlTz1!JNs;X~JYjOx7V6ws&xsM(^O4Kh=pwLF*0k8LpdW1f_`y*C3DdJ;k}z zCchiTpuU`zwdLb!3ChB#fL98*0FSwS$-&29pXuEo<@<%snN-5{PXfjJ2b&|0IIG<^ zKRI?$b!a5Cr?_Rz2jiD@E^hj6p$9PSfydr%z5nrZ@LWFw+WsSB5i*}MFT@M;JGP>! z6T;n45}#_l_ujKx@BZT(IrCn^Mwx0c?~#KYvwtY6JQa( z02<3F6r0;}-+i~_NiPe%W6325xT62cY8a}RBE~;MLziH zeQ@%8*32+V6CfFiZp1_|&{{L+Vh6A5+lxtp^yH)!e9x}vx$d&BrB$k%ydRE0@H-uVkgWGvC34IFe%(NF0aw-2}=i;Wb ziZSli`=4&d+VE#BxNu=j$ZSzg?$TZy19wODV2k;%H+0{F1!Z2@CA_9`;lU?7fPZJxpMn3(}iAw z+X`<(c(3IDm=Mc*Y_-6Q0WTQ?7h%$P81)fUh2!@K*%+#$I?!kYvP|$GJp=MHgugNs z-oaz~Ao!f_z`$N6w1utrb*yt}FGzLA2Tq$UxOO#U_6ZOin;;5L$P37-b$sGJ_^0*~ z;;^zTmH>1%U}EyE$^ZA@-FRjNG^j^7=(_B(xx~fesF^rY>wE>7JjwJirvd^jd6t-_ zhu(p5IM@@?4uJS7W<(a+H4WG8g;2V%O3;OLnKFZwTw2% z1Uw{@U1&m+`7E$Rpb4C`fltDH!(cDhtvP7N@10K+oF?HrA%FW)Hu6*cX7VzQwD7=V zq?7ax_$x9yR=B?oWAHR4Ad7yC@65)p-ukpZ;tI-10)D>SPPn7CzGundO@o^86D|fm zj4xyIMVgmD$XbC(e_I|j`hmLzf7`GnaU0a+cKq7636sCUcz++)taiiD+b8~Z&z?N` z$QgC}mD3vZ{k>B(;?of+I-P5Lx=a`Xf-~_a-NFaquNcK-v5*W#JH~_fc8gi3OAjgk782rZUkUG%DC~~+zIETfGzy^|16kxQ_CM8cm&r)QM__Y z?VrA=Rf}r!Wj!Ov7`YbuC*yVgyc3-No$H}@1)#qh)5$H^fBlx609i&@Xo7b^2wl*Z zT{<(nknOCi+1%S*s6Eeqd6N<1dj|#@nPE+`$Z7>&hv4@-3x9_|fzN`%KLo!WM@5O* z^G@vfm5MyD>joGkwBKunxJcX~I`;-5dZ%fo4bmzdKceNTBnR$4^6|cp2q- z1ZKl4p)_Tng5NK?b5|T$vZ%KXufYaqpt$P_>pJuzUfeHXM!p$l>1IodZf^N)=S~;A zif`(^WYYK;Gml_`9|e2+Sj@1cMIx=byuPDk87p9cwD}akTbjnsX}-H_gHQOp!8$)M zf1Iuv3rsT3!?#RELE8Ge_#CpZty&up-{AJFvhb|}?Z$u?p;;ktPZ-ksOrws{?j^-Z1F z7OEFpnr^(YCka=D7u%Oz+0YPe!1>172}tR20+i>8cz4L}K$u6OE0Jkykmh zL0IR)SV5L=<(nYgDHy$f>hkV_WL!!R_4~Q`<85^fyR~_kcN_<0c`PPa$HLM$3i!_Z zmWi)pJd(8V@x9pvcjsF;<>)Oa=Pi)kTAk>bNq>6iRmc#{r70a3Um(|f|AgM;5=1-G zja?Yi+WzX5RjplsKe5o=A$O*d8bq|{@)SlDP-aYQCw-zpv+;8VK00;T0;zt7AmmhTw zX3}>7Hfs$FQJW8A+uS*R=(xe)&A66RFV>RiVQ#U%HXXY2Ilip;0UK^SRZ?5$rH%8* zqRX(>cMj^g$zIS*L%y^lZg385QM|maX}P!S$E8$Q4!lT>$-6%Co+*|UJqSyR zGw^-GC$EPx!@G~~q!4?B%({J=rFBl8vaOTf{9Me5R;lVcF1p|XcxZl){zl}_X%p5* z!8aWch!4VK!f2?&9M|rdZ-v!+i$>b7n|^Z>6tyY|3G{sDGpAt<`ciyHK8CucM1A5D z%Tg~Q(T1Gd<+Z<;hkJiVeDJ2mtN&bU>|FI0_O0Uieap3q{-|Ec}yt3~lM$)z8~s+;nH6iW~nQwo z@t9An;hyahNV_nTo7_07?M;{Xl`WgLYrj3hFpOU_rDil#uW+n0PgqhRiXyhi_{|Bd za+Q~;QnLn*#8VrlJ}loCE11&cx8MJhAykP@r2g))@XWq34Qa+A<_o<`)Uf#8#V^0@ z!e)to9B|ib&t+sF2~tO>5|e`nwB2+D%$)oY+@+YD(_o|%5?zd72AQ2kIC-9_{GaQe z{ZyLG(g13=h~H1x?sbE%Ft~aZwiMlNx1K#Zc*^qN6H1`!k^1L1m9+DYz5Y6*>M@aj z|EekYg_Ve7`x$K~{_1vcImU_;Xy2r&DrQa?Gi6NOu*3gXr{a5H)U;{gjMS1KBLZSp z+;JfM6Z{|k6~vBra$wfG5VK$5RA*(rZh`N{orCUreO*RjDwx2|!(UPGh%H7@>JxE- zCuZ4kn3JPxpoB+T)c+hC-tTCf_UT_O%4K$r2<)0ve-pN1{T6JT44$l7)fo3Lg*&eJ z*tA2r{8pj9hw7R?+iz8&lJeUoe)~kMNk4@?kB#p)wwZlFde&`8nh0O)b)qt15>1z0 zVv7E=`e$F11J!gWWm(qsU$>0VnD!v5UvHa*5?jvvohO4{EC-}ZQ561WKs5MEdQWpG z9bu)GNd#*s%4py2y~FG%GjD!@IIYXY|K|fArcLX62QfSMds1oZ(y7J@`gEc=zBWF^wDA ziFhC!u1%#rLICy)oG_yu2>TVn5{Ee-<&+n=*Eosc@bOU|W?X zh9Rs6FU3~7zNk&WbdHbtZmpd*O*->r1lT;J zizj9YHM-wwA3}e5{2w;KG%Zsx5ZG}|{dtHcGzV=T;h=q(m;+sKC=mmwv~LTfJwj<<96qDjl+m04n*J=+GrmNRQ-^0Rjm6uN_e3V{DRNmO4YPs zFKznNQDqMi0|2`F+NRM6RrxJE$8bm_CnSr9_3)iss|jJobaZlA%Dw8JhtQJaQ}PdG z0J8<0`F}j@x#w0|(a8M5A<+Kd%}tCLFQdhJOi=dHVEYa{+ow8~fFovhVBZexIX>Q^ zu85)ivO?I@Ap7zoI(bXjg03vH%T@skuh_lY&O4bK212ya^YG#QJN%>L*JQ2#p&XxT zJ-QkMXIm4cEs@GMEj-03+%>Ub06z3$9V+eMigINhwia6|3tQ3~!G*B8o(Sc%7`kjR z9*D(Clz!Oo+WHP#^@~9C)s0`Tq0!jon2CN9bNDMEf(>5?WjWDq8dlrou7Gz9iNc;hGxvRiCA|n%AdP zkQqfsVBe$PtW^vqd{^vEUg9NY+5+m|e@T3n#q|NIcESwJhgp6o`T!qtbf`}Pig`+1C<`M4<@-Tqx8317Y zbqo6Sb&wx}KfGgUlSWuTZ1)+Jmdd1%>cJLHlqQp+q%0bPXY@i$jhfJpJ*vW+>SnyO zCW%PidI*8g@2@&iQ;1N)!u#1)!m`C0Qngo@orS+Z`b4#NG5h_y^hEKcdGjNYQ#-r=FkBj3-aUh7%?q@i9aZ;O@TopaJtFTPE zBTT`2wv89c@jOC}b8+<7IWPkriF+tc_YyTii?JWk&ibCSM!E4J*R^v|iDjyYe%;;p z$2G!vK_-tQaBxl75V32;v#3EywYs8w!Hjd?yOjGKd{C^x;w<Dmt3qtehM7AY`Slh8L4GcE7hpxlM7Ne&p&hP;hX z+#g}k`yov2i-JGc`I#k`JQd_c0rx1~9b+fpXu0oW_INd9w=ylb)-9a;vYP855i8d9 z#0`BOF)|+SH48J}Yg}(!GRzwSJFjUx6I;*b!sNUhrsu4cT(W?aSAE7g6EjX1Qc#?R zm4!k))HGFA;G`t69AhhCmhd@celaM?;@iMriK?nc-AprWERB$ZfFM4mw4 zz{6MFhKJ&*48n9%|ADR8YGlu3&9u3cg(hU+nOMxaKBUw*KQYtljxVN4*zUw^BNXxL zIIsNWstvn7panPxCySTNs-uL6qtpfj^qVna|3^?AXHA)>61^Lyte&IW+;70pTHw&t zWWjg?QKIJ}o^yvqxy<$`0=p&-yM}Pzi*AVr1-6+Q)uY_v2j%$|j4zjG&yRKzRwaBP z?7!>G{}ITA)yH`arPUNoimp;ij6?WAFi5-|Lk=Za&V${etpZ#Z$`Rb%Cz?gjX4`!Q~eD>Bct&y&%rs#FH}wH zSdgxbT*_E3m3!+7C#c<1#{KVseq$???_yz&*oAU;~JOs@YjwTUUpLhwzFV%4ov z@_ng%T}Qki3>%wwz64Ey^?&W>-%D6`z@hTRI-s!vb!{nMAtPqW zL&dL?L`e+JiDWoly&vI>r}?252Y%T7(tz&?%dQhXc4%{ExaaoqWA@bUT!O|P#$SV- zM)O8P)${SS+LSN;IECe6gwET6ZB2b1n;3BXX!x|%tMk>SBQYkp_HNtw(l z0R#r!{`ziw^M0BlDvd3mYDZH_zIg`LeM4KZ?mN=F*fQ7PtlR$Nj>9>N7^hfrGb^C{ zQ|uBrBdO>-^!eq;NduO2zc|R&>^`Z!MHXX#X4@~JpCA7a&I6xPHZ6YE-MAR@6JnVcW)o}XN(H|)k@gJW(s$QMTpZdQ(chH7BR~X zQpI;E_DxDtT0Vg;BySr29hwuKk=ZBi>(0Kic};u?vU*1oxN)O4;oim#oElFfh9)dq zStrgSml76U3I;1+HTu-3irtqrGg;A72xtns1MdE9MD1>&{*DQDg~K)=L_~{@wM!pQ zM_D)!f1l|Z@uzDt(Ik~)u&5oMl$U%fl=H3OP-J=ctU)#%!ID|@-#dP}_ue-4jF`@r zLJDz~r#BM~jE1didp?Cj|Nl62`pRwjY(w%c90HBaZJpS=`0K1&#GntSCJE;wR}(jg zFK33-I2JtOW}X%^K@jScKrgT(4k0@Pd`joC)hR^t$D`tMvc4THj5vaZM)h#5@ z{xDTTO|~0mWGPp=?9Y@F81p5VHQTqeQdnlG?r2vGqVO!kYjRXVv(`7qQiCrDy*TKH z)+Cur94X+Pw_kFm!RTBlpHZ1v7mos)uxqFi6zg7?tmY-@8fQ71Eb}6nWSTVsdbRc> zD(`p70A~fSKp+a&m9LNkXGiC*fH|*$>FhJ7{pz9T+BnQ~a z%LvT2D;%D`&4hJ`*=05keu!Y&opEbhoGe=@jU=z*#ub}iG7_=v+;xZp-c3HO`-Oq` znsdyYYg+;Dxqa*@3TJPFWnE^r9-yGgX#Ce&^CJnLIc(~Z>^^Au5lC`%0fFH&cH$go zGR+YyPBv1bY7|C#xhqPOhOuAP=|{nYix`RWuS^RvXX4u#9RpvF#hFRXDX@0--%!H& z8Vs8^Ds%IWz`!XBvBpZUWi3hz#Xm8 zB8bxx1e$ZKw9!81d*s3JO9LN-)fHWImTW2Dw!gpR1cbp8TQwS-u`ez)WmOje9`TU4 znQ(|W{gs;xB`@zFK$#)7(b-DSC|eO7RqRO=e|bsp;taaN*wC{uLHD(+Fdjn$J7ZN} zj@=7Rb~G6E&)X%Jbftv=24>STNF0Vj{1i&`=nQn`#T>g6SV%0P*h`2%jj^6HM@UY( zu`Zp{#1hK)-q1?}es536eyWTLxb2pU2g37LM64O!U#(=Wd7tn2tAL?e-k=^31uM#W2G*Z0Q~Z_E-qV zs|lI8CQ@(~U5**HeI=h*z3&e1>Rs@Vp4^vV<07L1W(s=(d`_^z%bO$=C%q_vPuvm? zLGK-pa@=ZL@+lAkoV=K!r2{ZK?}uma5gF)QC>Dp5n2d3T4jD=H=uQacY@Rt$sj5{*} zQ)521%Ux{FGOOH2yIvT4uHBkx6>!__I7gD0Mx**@Cx@SIG=mHX}jX&7{^yk z!S^X7_Vf5Wt7}UIOc;L%mE3Uvv1+}b`80szg~b8s`?GBpQ2}RVTUXXAEd=0H@m7+= zAz}p9BCt-@>sd+sY~RAUa+hu5G=vDQaWo}b#`WzqEaHlcvE0ybj2mM5nkOxw#DwpN zOqGBD4PkJqpP6%4DgOxj##NYwZ8=im`J2yJ?CXx&rFdp(2=hyKtV@Y)^A^e?qY$(D z?(i)2vc$5~J5vR`d&=1V1ylOhQprm@2NAN4B1!<}LlSe^6U-boH%$F#z0-Pf*XdRx3G`GT{sSod|Zw;1Eu$YFlq=h~VVi_)S{sDS2p&$ zLV`iqvK;!d4~lCU9InQL2R^m7QVdR4yWsika2JBnN{DXTNy?Tac<@EsbE?F|j5l8X zW1}_ded&0lDBuBp{Vs4Am9A3Bm%%9uXQXb1nQ^<_z98=dfj)Do(;U>{porv7P!_Kof{(aQ4jK!7pdD~N+p2ZiOj zqlP1d4Bb&wAA7v@Gbq>R+oQEl#kqbS)gwpa$ic1lYe~vdM}XeGR4nVQancVgVJw8n z`AEVY9!p3Om5oN7E|G@e*d=^`?08k%?FN)@Y%ylPQw_9H+KXC2|5P#G86E*5^cr zpMtCJk%3w7B1-j(Gd51hon5oZVP@REy5 zz-Bf6+xU^Gk`U0SZ;KcDi;q;DljRZzGBeJ@BSA0B(Jm&s!9`}0RjDA*bDEH~uuz`x zJ~rOR{@%|B36Ha;v@D4+HQJ7Cd3BZ)CI$F*ufrO%9YtwU+cXQhKtt3m&F;xgennSr zFyqvxfcM-sZUS_0&f&Ps8Y!JTTGulpX4N&61r}f}3yvj3$3Jj&>GVNL>J|bDqo2FQ zD_1^@@-K4bVOpQ_1{}8VCCbE_w@V2p-^f&P2*8?RwoKQBwOVWy3aw#kXDwx27ZIkN zlm$3C!b~nfd>p5dwKS~zissZT+BC0%HDAiKvF3$Vx%4Kpl0F4Y7@cfJP?>o&faqD5 zP*yXVE7jX6Ro|;j94foljm;l-VyK&74JjTXh9j?iM{;tUv&uyfdv_8BuAUjVxYHxa z#SQ{hVZVs9X7tcfegH}edty7I*w;H2!7_y6#9St;B$V$5P~2HKQ0WV??JgUIZz?FZ z1kTH_=4)CHnWsi)`BI$5dK9oYZ~+$9?^)HX)B z)1bV{!K35x?Of%~@yn9Hx-pfuIq-5aEH~tdd(Y3=ZctnJN(hs&rHnCDa6mqTC0J2($q& z6ru zMP0#WCe}y2kZQ2TwlU9wD4d6|A5&}1r5&B9S%-VgrCnLK_z2*1`ixsT?MMm7!<9Hk z7`{0AadzmFX`p!L!}HmVnC(#v0+ztz*{;J?vM6!){IJN&{xjk-$iW#f;opeQl-G#} zFsXnO0jZ!n*quBw0~T)h+?ANn8UB3=0&2iR728Czbj%uF+kG&CScZmVarq$tZ+)w+ zoVGhlnbFFW$SF>}oqHR=d$IusLS!>*?q$*Yt0T_z3T!pHus0*;Uv4DmTGoC?)6rmB zfOAcA7QJW;bc5#40j6lZk`3EIfOEq`>zTTHo}8IFZ^W=iIJyj4zatZAQhA6z*%Oo@ zDO=uF1_Cq(1Y`Djxal87VU;2?_&#V^W}Xv zxL~>)Cs@RDEW>L#11`gVs?9E82J9JgDh2}u9Vp|pqw@$4x8Y;9#U@;WOwN8CfK_g! zSSHEN1kHIDXR(#=c82oJI_<`~Zv@JFc-9;aRTSRNFgxcWikCL3%#2*{)Q;Hi0TfvaC$dl!hLk0N&0H_xYybmRRJE`F*GKW-){DQ+sdZuLS~C zp#p#OEaZ^c1xEl%&r8b>_|A~e3JwC{gE{X$2YZF**=XA{76TiyN)ne5z_CD1P0*Vu z4a3oiXTMh$%Op6K@GB_a<8$o}X7FpV{(L(4^!P{R%FnDmXY`ofd{|e>(B2#QYl;A` ztN%dGfK8z(|4gberyrC&gQ6E}Fvt)wYdq)CMKJNDF*8+G@Q45a5eZ2|K~z#X1J3gk z;AQqWh(67;ktjKqHkL6ZJOs)&`=u2qU(BAjU=RDriQKcg+YXOsgni}JS=F{(sehJ^ zmx|=LQQCH^rApXE?>`jXrBr($<<1%b=F_}11IBj-+xd%_aNOV+%AYuc zkW3XSyxgf5XobU7UftJX1Wp~^NidPs7h4)c)ie+`(%VhC}hm6G4XJ9s(;B~gF|JX$ETcf639v8j5u#jPps3^!GNdlda1vQVy+#= z$*Fkuk=ywSOa70(th-ks1N~u<$*7Y)xjvocv{57Wbq4TQj)Z zBw~A$_+(@M(k>C@Wwu5O@T&9FA0iUXZugIc{(W&)!cJ(dPP*VWq=&>@oAK0Q8iqST zKv3L8Ctqil%6+N%#{1c}cwQCfQx8kY!OZ|uN;_x1ESNUPnjMYQ2^O$XBr2~+P5srk zCf<`!RK7=c9i>vwIB|>gDwq1HSmj1ntFMCOkt@Xo=ilnt(u$)ou%i)ibvNhj|9l~ zNDj&6S1r1-&_e5nMt#;b0?I*46MFvX5dph8ezi*alU|}PLU~&Z&Z?IX#kI6J4z_4< zJ7`+p(``L$4kbO^cX-;T9>x8FB>kBRaJDZucH$S8L*r2ZsZI@5k5CZ9~F( z8#F%?y|^JB&$dY-zd&KU31hU_sPdz(Gu~Om=ki5TJA`^cWM30m7TT28X;i-u)Puh{ zPTTITur(=eyvV3I5H;BgTUI8Qh-}_0h=B=(xe0RdzoOy=dc2ZGp?Z`4c8Z^f<6;}j zH9@%0!o>~b1>4@WzSPN_yM2n&E{6xswOgz^T)cqZB@Hzs^5)^+Wl>W=VGIm0%1 z=DQ8sN0Z5XF^*cBW^<*_Cwk=f{yt|(TuM4#HY-mF5f{_6;QY#DFcZ33QX7lAE^Z^r z#cxPutG}Z2Wk_qK^oTRPvIzW}z*KouSdCGqf$}Eesn-Gt+==DpLs~}CN%c|QdN1mc z!D`x)jxHE0KM#IQPGtJP=$%ej+EaE+_!qtJM+?YVl8a@mF}xR$ME0hs+FZj&Es4^m z4K6s@46;~`V*cKHemPBXceS78SEbX8Skpp51&Vp4Ks?Z4^thy%`x#Gw??c=h*F{?* zIh?B8r3Rk=0`P*XzT9v~za_#>udH2U=9K|2)cZ^kQ`EPxRXat8;>lo&6w}4s2%Ch@&+lyrFv_{L2=E0JSd9~)Vkaw zc;qNF);4#f$lR?jHshMA>;0q8GLHtji&~ZRtr{&jvwZFeq$MBKE?9|WLD2Z%5h6(v zE3A9mwV&Db@@WrTd1Y$Ywm5e=SDOALKu!y>+vCg9_pL$Hq|7~gu5hKoT!QwsGMjNc zBhMR$VH+Zzl_?%T>?Lu!lBKp&l+6vvyytZAC6>Cat=^c$7`ek+&dbx@gK6L_+Yq`@ zsL6PL9w8zz7MKocSh3coM4Q5so38`QTde zg1q2%P(o4EmG9I0W=r`eKM}^MCl1ZwCwD+n3yuOe__rXkhief0&;;_pS27FOAG8!}r?bIlvFjzt@?(&}E1Rtf^*m z-?f^-Ma6DX%WtRk)gTRtH-(7PhJ}?C|KD}KI$bPvzp^oY0W03O8qoMsG(29LCU8-x zU$|O9(w}Gr8827G2}5YKnB~-MAYceNat5ilX7a0lM_w$?6rsoUym;(^0Ta_-0-s%+ zClBslqJl1QH5i}Q%~yQ4B{e!;OQtb*OP(Uf1QWI%98-aho!o^L+19M;)=1a4 z%6?B7z9w|ja9Au4Q^6*sze3tYjJ20DVB{s`dtuclf*)kXUWSRWu)o;p7Ubaxp_eCM zZ|%z-;mVTTpD)d)>VD7z)G{f+7Xz^zc5h9SHvbau1U_qe`b8m6OILgD;YBLHBDT1U zB%B|s>UT43YwZKRoL!yc2rKxN$|DvLV={7XH^AZgm(z#V!CmKW1E!lINf90i0WRz(Llrnm9^_Em)=Y2jDEUGM zr{?;`%u5y|Is+GH565u0;(?)+V$wkuLuA&bH{I*0R^#O&hF{&OF*6*QLme16*T+mS z$UJb9jGDG<$5?f&JF3xu1F*HdRiQLIv-mk|>2R(0P)*E%xkbOH$fkn*!uB>+bjwSd zj4arBJ<;o+VBhSX+!0}ahXbNLs4gX{$qy&9kYMPH!C%hpw=zk`(D=!1XUHTV!$6bZ zx!Y96IJaS0&IYY0<)Fp;cRt5~)~^90w|es>iz0p%v)r2>yDgsRQQz=W?*xl9)4f?M z3+JkC4gua&t^~1lk11kJks~>CDGQ^_I%i$vsJ9-)#}eKrXLXsGuP?RbOIb@}!WN%T zpf9vs->rRw`neKVB*}KYENp#m|Cn!hy}B8;sO7fxN*qAj}62Ow(P0VGmY`}Rpl2Zgi>%V(Ej5jXcOGT?_vd1ZCgFv;`k&W529g+&Cxk}EZo~dzNCKot7=Er3k)AC zI-dEzOd5az#&@E8g8IGtIYUtYH9(c6Z&p%AX)Tz8JSpk9#Hjy(zzU&pSUC0QYZg)+ zuIxfNg}Nkm{~28Z*=06%=l+~#=^$0*eOOTARgsG%X~wmyrLFn$(Q<%V@xLFJX6icJ zARxPMHG7*EczduQ{i_a}7k^D;DYLxxWH!Ac(gF%+#BcaWLxr{LG?jKtSvZoX^6lAf za{>&9F1Vp4c*>&b`lq8O>xXJX3czhhpI^1qafr-2n`}M9Lz4r<_rFniYIeP9Q7gBD z)0vqM6cM`>1C5RrhLsY)i>YV&PYbqoH)RkiaXxZUU;nLd4lEXc7TUPss7?G|;5#)|H2XPxQH1F^%7Zg?Y#Hu4qa^|ERZ!{AB z96F*J`^?O$Xnhn?GS~r#*N!D7gEddMZ8!p#fSnY_)wN`r@gXTZGs&9M$4T?DvY`e& zyV7JLZ@_i^#RNFO^c?2&TD1pR2%}+t&Ur0(XfhgbuOZY%)SWct*7Cex;T*eoSpv#! zl^qKeHUjsY4o9XoJrv?W92^DmBo;`(pHp_4n5x*q3x1`H)wo4ax6AB-4`x91vHr?eYZaE$vX%IW>ga1d5qSAl zz#=n7n+ggfjKRu(w8Jf0kAxlvJ2tgglVpnneU%)%4DQPR97cD&;27n{KyZ0KtxNB? zM3OwagBT{3J?b8`b5@-TMR=mFi`)k~mYA^+?yGigabY(g1?<^1qkIn%Dw7AwRitNk zG;4C&R_^0I%?u3GtHgaQPeWZ8-#rHMCMIyX&u5| z-5TKK{>Xn~_6eESUKSP@z2a>)3(k2aKU87L`9mquTnc-JC#`|z{Pjd&yAkO%5BNwJ z@JOPwTFINN64LB|9FAcaT^7a_Y5qBRm^R}IwCP+(ojpijp|m!h{>kCHJaHs4-hjaWH#zqgnD6&U?{7o9wb^9EzqW0SVY|4j% zYs5XPgKHDoPgg9YCw5x=C1=4$6))O(MCGna4M>Sxjy= z0vV@C5q|R0`o^HR3KBZ3f+Eci*S*@iV$6xLq$i#Am9;eNHbCmqt6a-T#5HZnnneaw zS;zg;L6Qbp;eD@D()yJt1)%eORcv*M+`>bHGD=`?NFXZ*16C~eDd%P&e;ceex_@@k z@fZnb;b)j*tBR@g=*S(&Rh%}PKr!eGI^FiC>U2&P)==h%6vD>IioI)jI&n`uCW6~UH z?3Ror16b{?Dv@XvA0)9PhXOk;Oj&`M5&Qx43djohe`+})z None: + """ + Validate provider credentials + if validate failed, raise exception + + :param credentials: provider credentials, credentials form defined in `provider_credential_schema`. + """ + try: + model_instance = self.get_model_instance(ModelType.TEXT_EMBEDDING) + + # Use `mxbai-embed-large-v1` model for validate, + model_instance.validate_credentials(model="mxbai-embed-large-v1", credentials=credentials) + except CredentialsValidateFailedError as ex: + raise ex + except Exception as ex: + logger.exception(f"{self.get_provider_schema().provider} credentials validate failed") + raise ex diff --git a/api/core/model_runtime/model_providers/mixedbread/mixedbread.yaml b/api/core/model_runtime/model_providers/mixedbread/mixedbread.yaml new file mode 100644 index 00000000000000..2f43aea6ade2c6 --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/mixedbread.yaml @@ -0,0 +1,31 @@ +provider: mixedbread +label: + en_US: MixedBread +description: + en_US: Embedding and Rerank Model Supported +icon_small: + en_US: icon_s_en.png +icon_large: + en_US: icon_l_en.png +background: "#EFFDFD" +help: + title: + en_US: Get your API key from MixedBread AI + zh_Hans: 从 MixedBread 获取 API Key + url: + en_US: https://www.mixedbread.ai/ +supported_model_types: + - text-embedding + - rerank +configurate_methods: + - predefined-model +provider_credential_schema: + credential_form_schemas: + - variable: api_key + label: + en_US: API Key + type: secret-input + required: true + placeholder: + zh_Hans: 在此输入您的 API Key + en_US: Enter your API Key diff --git a/api/core/model_runtime/model_providers/mixedbread/rerank/__init__.py b/api/core/model_runtime/model_providers/mixedbread/rerank/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/mixedbread/rerank/mxbai-rerank-large-v1-en.yaml b/api/core/model_runtime/model_providers/mixedbread/rerank/mxbai-rerank-large-v1-en.yaml new file mode 100644 index 00000000000000..beda2199537450 --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/rerank/mxbai-rerank-large-v1-en.yaml @@ -0,0 +1,4 @@ +model: mxbai-rerank-large-v1 +model_type: rerank +model_properties: + context_size: 512 diff --git a/api/core/model_runtime/model_providers/mixedbread/rerank/rerank.py b/api/core/model_runtime/model_providers/mixedbread/rerank/rerank.py new file mode 100644 index 00000000000000..bf3c12fd86dc35 --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/rerank/rerank.py @@ -0,0 +1,125 @@ +from typing import Optional + +import httpx + +from core.model_runtime.entities.common_entities import I18nObject +from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType +from core.model_runtime.entities.rerank_entities import RerankDocument, RerankResult +from core.model_runtime.errors.invoke import ( + InvokeAuthorizationError, + InvokeBadRequestError, + InvokeConnectionError, + InvokeError, + InvokeRateLimitError, + InvokeServerUnavailableError, +) +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.rerank_model import RerankModel + + +class MixedBreadRerankModel(RerankModel): + """ + Model class for MixedBread rerank model. + """ + + def _invoke( + self, + model: str, + credentials: dict, + query: str, + docs: list[str], + score_threshold: Optional[float] = None, + top_n: Optional[int] = None, + user: Optional[str] = None, + ) -> RerankResult: + """ + Invoke rerank model + + :param model: model name + :param credentials: model credentials + :param query: search query + :param docs: docs for reranking + :param score_threshold: score threshold + :param top_n: top n documents to return + :param user: unique user id + :return: rerank result + """ + if len(docs) == 0: + return RerankResult(model=model, docs=[]) + + base_url = credentials.get("base_url", "https://api.mixedbread.ai/v1") + base_url = base_url.removesuffix("/") + + try: + response = httpx.post( + base_url + "/reranking", + json={"model": model, "query": query, "input": docs, "top_k": top_n, "return_input": True}, + headers={"Authorization": f"Bearer {credentials.get('api_key')}", "Content-Type": "application/json"}, + ) + response.raise_for_status() + results = response.json() + + rerank_documents = [] + for result in results["data"]: + rerank_document = RerankDocument( + index=result["index"], + text=result["input"], + score=result["score"], + ) + if score_threshold is None or result["score"] >= score_threshold: + rerank_documents.append(rerank_document) + + return RerankResult(model=model, docs=rerank_documents) + except httpx.HTTPStatusError as e: + raise InvokeServerUnavailableError(str(e)) + + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ + try: + self._invoke( + model=model, + credentials=credentials, + query="What is the capital of the United States?", + docs=[ + "Carson City is the capital city of the American state of Nevada. At the 2010 United States " + "Census, Carson City had a population of 55,274.", + "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific Ocean that " + "are a political division controlled by the United States. Its capital is Saipan.", + ], + score_threshold=0.8, + ) + except Exception as ex: + raise CredentialsValidateFailedError(str(ex)) + + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + """ + Map model invoke error to unified error + """ + return { + InvokeConnectionError: [httpx.ConnectError], + InvokeServerUnavailableError: [httpx.RemoteProtocolError], + InvokeRateLimitError: [], + InvokeAuthorizationError: [httpx.HTTPStatusError], + InvokeBadRequestError: [httpx.RequestError], + } + + def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity: + """ + generate custom model entities from credentials + """ + entity = AIModelEntity( + model=model, + label=I18nObject(en_US=model), + model_type=ModelType.RERANK, + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_properties={ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size", "512"))}, + ) + + return entity diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/__init__.py b/api/core/model_runtime/model_providers/mixedbread/text_embedding/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-2d-large-v1-en.yaml b/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-2d-large-v1-en.yaml new file mode 100644 index 00000000000000..0c3c863d06b89a --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-2d-large-v1-en.yaml @@ -0,0 +1,8 @@ +model: mxbai-embed-2d-large-v1 +model_type: text-embedding +model_properties: + context_size: 512 +pricing: + input: '0.0001' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-large-v1-en.yaml b/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-large-v1-en.yaml new file mode 100644 index 00000000000000..0c5cda2a72a99e --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/text_embedding/mxbai-embed-large-v1-en.yaml @@ -0,0 +1,8 @@ +model: mxbai-embed-large-v1 +model_type: text-embedding +model_properties: + context_size: 512 +pricing: + input: '0.0001' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py new file mode 100644 index 00000000000000..05d9a9a0c69f45 --- /dev/null +++ b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py @@ -0,0 +1,163 @@ +import time +from json import JSONDecodeError, dumps +from typing import Optional + +import requests + +from core.model_runtime.entities.common_entities import I18nObject +from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType +from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult +from core.model_runtime.errors.invoke import ( + InvokeAuthorizationError, + InvokeBadRequestError, + InvokeConnectionError, + InvokeError, + InvokeRateLimitError, + InvokeServerUnavailableError, +) +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel + + +class MixedBreadTextEmbeddingModel(TextEmbeddingModel): + """ + Model class for MixedBread text embedding model. + """ + + api_base: str = "https://api.mixedbread.ai/v1" + + def _invoke( + self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :return: embeddings result + """ + api_key = credentials["api_key"] + if not api_key: + raise CredentialsValidateFailedError("api_key is required") + + base_url = credentials.get("base_url", self.api_base) + base_url = base_url.removesuffix("/") + + url = base_url + "/embeddings" + headers = {"Authorization": "Bearer " + api_key, "Content-Type": "application/json"} + + data = {"model": model, "input": texts} + + try: + response = requests.post(url, headers=headers, data=dumps(data)) + except Exception as e: + raise InvokeConnectionError(str(e)) + + if response.status_code != 200: + try: + resp = response.json() + msg = resp["detail"] + if response.status_code == 401: + raise InvokeAuthorizationError(msg) + elif response.status_code == 429: + raise InvokeRateLimitError(msg) + elif response.status_code == 500: + raise InvokeServerUnavailableError(msg) + else: + raise InvokeBadRequestError(msg) + except JSONDecodeError as e: + raise InvokeServerUnavailableError( + f"Failed to convert response to json: {e} with text: {response.text}" + ) + + try: + resp = response.json() + embeddings = resp["data"] + usage = resp["usage"] + except Exception as e: + raise InvokeServerUnavailableError(f"Failed to convert response to json: {e} with text: {response.text}") + + usage = self._calc_response_usage(model=model, credentials=credentials, tokens=usage["total_tokens"]) + + result = TextEmbeddingResult( + model=model, embeddings=[[float(data) for data in x["embedding"]] for x in embeddings], usage=usage + ) + + return result + + def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :return: + """ + return sum(self._get_num_tokens_by_gpt2(text) for text in texts) + + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ + try: + self._invoke(model=model, credentials=credentials, texts=["ping"]) + except Exception as e: + raise CredentialsValidateFailedError(f"Credentials validation failed: {e}") + + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + return { + InvokeConnectionError: [InvokeConnectionError], + InvokeServerUnavailableError: [InvokeServerUnavailableError], + InvokeRateLimitError: [InvokeRateLimitError], + InvokeAuthorizationError: [InvokeAuthorizationError], + InvokeBadRequestError: [KeyError, InvokeBadRequestError], + } + + def _calc_response_usage(self, model: str, credentials: dict, tokens: int) -> EmbeddingUsage: + """ + Calculate response usage + + :param model: model name + :param credentials: model credentials + :param tokens: input tokens + :return: usage + """ + # get input price info + input_price_info = self.get_price( + model=model, credentials=credentials, price_type=PriceType.INPUT, tokens=tokens + ) + + # transform usage + usage = EmbeddingUsage( + tokens=tokens, + total_tokens=tokens, + unit_price=input_price_info.unit_price, + price_unit=input_price_info.unit, + total_price=input_price_info.total_amount, + currency=input_price_info.currency, + latency=time.perf_counter() - self.started_at, + ) + + return usage + + def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity: + """ + generate custom model entities from credentials + """ + entity = AIModelEntity( + model=model, + label=I18nObject(en_US=model), + model_type=ModelType.TEXT_EMBEDDING, + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_properties={ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size", "512"))}, + ) + + return entity diff --git a/api/pyproject.toml b/api/pyproject.toml index 41244f516c95c4..9e38c0945601cf 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -122,6 +122,7 @@ CODE_EXECUTION_API_KEY = "dify-sandbox" FIRECRAWL_API_KEY = "fc-" TEI_EMBEDDING_SERVER_URL = "http://a.abc.com:11451" TEI_RERANK_SERVER_URL = "http://a.abc.com:11451" +MIXEDBREAD_API_KEY = "mk-aaaaaaaaaaaaaaaaaaaa" [tool.poetry] name = "dify-api" diff --git a/api/tests/integration_tests/model_runtime/mixedbread/__init__.py b/api/tests/integration_tests/model_runtime/mixedbread/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/tests/integration_tests/model_runtime/mixedbread/test_provider.py b/api/tests/integration_tests/model_runtime/mixedbread/test_provider.py new file mode 100644 index 00000000000000..25c9f3ce8dffa9 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/mixedbread/test_provider.py @@ -0,0 +1,28 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.mixedbread.mixedbread import MixedBreadProvider + + +def test_validate_provider_credentials(): + provider = MixedBreadProvider() + + with pytest.raises(CredentialsValidateFailedError): + provider.validate_provider_credentials(credentials={"api_key": "hahahaha"}) + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "usage": {"prompt_tokens": 3, "total_tokens": 3}, + "model": "mixedbread-ai/mxbai-embed-large-v1", + "data": [{"embedding": [0.23333 for _ in range(1024)], "index": 0, "object": "embedding"}], + "object": "list", + "normalized": "true", + "encoding_format": "float", + "dimensions": 1024, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + provider.validate_provider_credentials(credentials={"api_key": os.environ.get("MIXEDBREAD_API_KEY")}) diff --git a/api/tests/integration_tests/model_runtime/mixedbread/test_rerank.py b/api/tests/integration_tests/model_runtime/mixedbread/test_rerank.py new file mode 100644 index 00000000000000..b65aab74aa96d3 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/mixedbread/test_rerank.py @@ -0,0 +1,100 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.entities.rerank_entities import RerankResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.mixedbread.rerank.rerank import MixedBreadRerankModel + + +def test_validate_credentials(): + model = MixedBreadRerankModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials( + model="mxbai-rerank-large-v1", + credentials={"api_key": "invalid_key"}, + ) + with patch("httpx.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "usage": {"prompt_tokens": 86, "total_tokens": 86}, + "model": "mixedbread-ai/mxbai-rerank-large-v1", + "data": [ + { + "index": 0, + "score": 0.06762695, + "input": "Carson City is the capital city of the American state of Nevada. At the 2010 United " + "States Census, Carson City had a population of 55,274.", + "object": "text_document", + }, + { + "index": 1, + "score": 0.057403564, + "input": "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific " + "Ocean that are a political division controlled by the United States. Its capital is " + "Saipan.", + "object": "text_document", + }, + ], + "object": "list", + "top_k": 2, + "return_input": True, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + model.validate_credentials( + model="mxbai-rerank-large-v1", + credentials={ + "api_key": os.environ.get("MIXEDBREAD_API_KEY"), + }, + ) + + +def test_invoke_model(): + model = MixedBreadRerankModel() + with patch("httpx.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "usage": {"prompt_tokens": 56, "total_tokens": 56}, + "model": "mixedbread-ai/mxbai-rerank-large-v1", + "data": [ + { + "index": 0, + "score": 0.6044922, + "input": "Kasumi is a girl name of Japanese origin meaning mist.", + "object": "text_document", + }, + { + "index": 1, + "score": 0.0703125, + "input": "Her music is a kawaii bass, a mix of future bass, pop, and kawaii music and she leads a " + "team named PopiParty.", + "object": "text_document", + }, + ], + "object": "list", + "top_k": 2, + "return_input": "true", + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + result = model.invoke( + model="mxbai-rerank-large-v1", + credentials={ + "api_key": os.environ.get("MIXEDBREAD_API_KEY"), + }, + query="Who is Kasumi?", + docs=[ + "Kasumi is a girl name of Japanese origin meaning mist.", + "Her music is a kawaii bass, a mix of future bass, pop, and kawaii music and she leads a team named " + "PopiParty.", + ], + score_threshold=0.5, + ) + + assert isinstance(result, RerankResult) + assert len(result.docs) == 1 + assert result.docs[0].index == 0 + assert result.docs[0].score >= 0.5 diff --git a/api/tests/integration_tests/model_runtime/mixedbread/test_text_embedding.py b/api/tests/integration_tests/model_runtime/mixedbread/test_text_embedding.py new file mode 100644 index 00000000000000..ca97a1895113f0 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/mixedbread/test_text_embedding.py @@ -0,0 +1,78 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.mixedbread.text_embedding.text_embedding import MixedBreadTextEmbeddingModel + + +def test_validate_credentials(): + model = MixedBreadTextEmbeddingModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials(model="mxbai-embed-large-v1", credentials={"api_key": "invalid_key"}) + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "usage": {"prompt_tokens": 3, "total_tokens": 3}, + "model": "mixedbread-ai/mxbai-embed-large-v1", + "data": [{"embedding": [0.23333 for _ in range(1024)], "index": 0, "object": "embedding"}], + "object": "list", + "normalized": "true", + "encoding_format": "float", + "dimensions": 1024, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + model.validate_credentials( + model="mxbai-embed-large-v1", credentials={"api_key": os.environ.get("MIXEDBREAD_API_KEY")} + ) + + +def test_invoke_model(): + model = MixedBreadTextEmbeddingModel() + + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "usage": {"prompt_tokens": 6, "total_tokens": 6}, + "model": "mixedbread-ai/mxbai-embed-large-v1", + "data": [ + {"embedding": [0.23333 for _ in range(1024)], "index": 0, "object": "embedding"}, + {"embedding": [0.23333 for _ in range(1024)], "index": 1, "object": "embedding"}, + ], + "object": "list", + "normalized": "true", + "encoding_format": "float", + "dimensions": 1024, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + result = model.invoke( + model="mxbai-embed-large-v1", + credentials={ + "api_key": os.environ.get("MIXEDBREAD_API_KEY"), + }, + texts=["hello", "world"], + user="abc-123", + ) + + assert isinstance(result, TextEmbeddingResult) + assert len(result.embeddings) == 2 + assert result.usage.total_tokens == 6 + + +def test_get_num_tokens(): + model = MixedBreadTextEmbeddingModel() + + num_tokens = model.get_num_tokens( + model="mxbai-embed-large-v1", + credentials={ + "api_key": os.environ.get("MIXEDBREAD_API_KEY"), + }, + texts=["ping"], + ) + + assert num_tokens == 1 diff --git a/dev/pytest/pytest_model_runtime.sh b/dev/pytest/pytest_model_runtime.sh index 4c0083a2de3435..b60ff64fdcd901 100755 --- a/dev/pytest/pytest_model_runtime.sh +++ b/dev/pytest/pytest_model_runtime.sh @@ -8,4 +8,5 @@ pytest api/tests/integration_tests/model_runtime/anthropic \ api/tests/integration_tests/model_runtime/huggingface_hub/test_llm.py \ api/tests/integration_tests/model_runtime/upstage \ api/tests/integration_tests/model_runtime/fireworks \ - api/tests/integration_tests/model_runtime/nomic + api/tests/integration_tests/model_runtime/nomic \ + api/tests/integration_tests/model_runtime/mixedbread From aebe5fc68ceeb30c8a8958c644197cc8690c0429 Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Tue, 24 Sep 2024 13:06:21 +0800 Subject: [PATCH 19/79] fix: Remove unsupported parameters in qwen model (#8699) --- .../model_providers/tongyi/llm/farui-plus.yaml | 9 --------- .../tongyi/llm/qwen-coder-turbo-0919.yaml | 9 --------- .../tongyi/llm/qwen-coder-turbo-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-coder-turbo.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-long.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-math-plus-0816.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-math-plus-0919.yaml | 9 --------- .../tongyi/llm/qwen-math-plus-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-math-plus.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-math-turbo-0919.yaml | 9 --------- .../tongyi/llm/qwen-math-turbo-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-math-turbo.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-0107.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-0403.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-0428.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-0919.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-1201.yaml | 6 ------ .../model_providers/tongyi/llm/qwen-max-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-max-longcontext.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-0206.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-0624.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-0723.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-0806.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-0919.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-chat.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-plus-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-turbo-0206.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-turbo-0624.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-turbo-0919.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-turbo-chat.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-turbo-latest.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-vl-max-0809.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-vl-max.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-vl-plus-0201.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-vl-plus-0809.yaml | 9 --------- .../model_providers/tongyi/llm/qwen-vl-plus.yaml | 9 --------- .../tongyi/llm/qwen2-math-1.5b-instruct.yaml | 9 --------- .../tongyi/llm/qwen2-math-72b-instruct.yaml | 9 --------- .../tongyi/llm/qwen2-math-7b-instruct.yaml | 9 --------- .../tongyi/llm/qwen2.5-0.5b-instruct.yaml | 9 --------- .../tongyi/llm/qwen2.5-1.5b-instruct.yaml | 9 --------- .../model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml | 9 --------- .../model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml | 9 --------- .../model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml | 9 --------- .../model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml | 9 --------- .../model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml | 9 --------- .../tongyi/llm/qwen2.5-coder-7b-instruct.yaml | 9 --------- 47 files changed, 420 deletions(-) diff --git a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml index e5de586c1caf16..d0ff44382789b4 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml index 6ab39cde2d052d..d9792e71ee4547 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml index be6d9a0e07272e..0b03505c45084d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml index d2aca4f514ae09..2a6c0408530053 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml index a59a3350f62dda..bad7f4f4725c45 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml @@ -68,15 +68,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml index cab7233c98d691..c14aee1e1eb87f 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml index f82fba0c012aa8..9d74eeca3e8788 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml index e2fb6e0e554d46..b8601a969a7fae 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml index 8803e747e5eb7e..4a948be5976492 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml index 0dc5a066f03001..bffe324a9649c4 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml index 2ac0e4692a701b..0747e966142194 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml index 9a7f1312e97246..dffb5557ffb7c2 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml index c0eef375570d40..8ae159f1bfb13b 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml index c12444bd7b57b1..93fb37254e4c8e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml index 173c55b6b9796f..a5c9d49609041a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml index 692a38140dbed0..e4a6dae63736a3 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml index dc234783cde63b..6fae8a7d38f36e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml @@ -66,12 +66,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml index afd7fb4b77696d..8e2096885957eb 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml index d02ba7af18318a..9bc50c73fcc3ee 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml index 1111298c3712ac..430599300bccb9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml index ef8dd083ade5dd..906995d2b95c86 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml index 87a4417df59b8a..b33e725dd08c39 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml index 967f258fa91626..bb394fad814557 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml index 9d44852ac9cd5e..118e304a9723a3 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml index df9448ae04ddf2..761312bc381236 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml index 32ccb8d6159034..430872fb31a442 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml index bf976b518a9114..2628d824fe8242 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml index 060e7fb4c9d092..8097459bf09061 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml index 97cd34929b4f4d..e43beeb195aec9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml index 8d77ba7a2a698a..c30cb7ca100968 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml @@ -67,15 +67,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml index 4458c706aad40e..e443d6888b7ef6 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml index 12e9e0dd569f93..fd20377002042a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml @@ -69,15 +69,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml index b811fdece48f5a..31a9fb51bbbb54 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml @@ -69,15 +69,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml index 188dea389a7e21..5f90cf48bc999a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml @@ -69,15 +69,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml index bc623e2f032110..97820c0f3a1a21 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml @@ -69,15 +69,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml index 8977e12e4f2d90..6af36cd6f3d630 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml @@ -69,15 +69,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml index de237842af7d8e..158e2c7ee16e5a 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml index 1fda35abaf1fec..e26a6923d12f21 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml index 06fd33c5f47400..589119b26ea3c1 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml index ebf809955337b4..dd608fbf7675af 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml index e9bc99339de0b1..08237b3958b327 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml index 3ed85dade82817..640b01970351ba 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml index 328519c16822ff..3a90ca753273df 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml index d1ed3c2a73edf1..b79755eb9bd844 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml index 0e88c24aa83475..e9dd51a3416243 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml index 35313cd1f70947..04f26cf5fe12b6 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml index 35313cd1f70947..04f26cf5fe12b6 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml @@ -65,15 +65,6 @@ parameter_rules: help: zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. - - name: enable_search - type: boolean - default: false - label: - zh_Hans: 联网搜索 - en_US: Web Search - help: - zh_Hans: 模型内置了互联网搜索服务，该参数控制模型在生成文本时是否参考使用互联网搜索结果。启用互联网搜索，模型会将搜索结果作为文本生成过程中的参考信息，但模型会基于其内部逻辑“自行判断”是否使用互联网搜索结果。 - en_US: The model has a built-in Internet search service. This parameter controls whether the model refers to Internet search results when generating text. When Internet search is enabled, the model will use the search results as reference information in the text generation process, but the model will "judge" whether to use Internet search results based on its internal logic. - name: response_format use_template: response_format pricing: From 4638f99aaabdfef5ad751d725ef5ca26027bd7fe Mon Sep 17 00:00:00 2001 From: Benjamin Date: Tue, 24 Sep 2024 13:26:58 +0800 Subject: [PATCH 20/79] fix: change model provider name issue Ref #8691 (#8710) --- api/core/model_runtime/model_providers/jina/jina.yaml | 2 +- api/core/tools/provider/builtin/jina/jina.yaml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/api/core/model_runtime/model_providers/jina/jina.yaml b/api/core/model_runtime/model_providers/jina/jina.yaml index 2f2d6e6daad73c..4ff6ba0f22c846 100644 --- a/api/core/model_runtime/model_providers/jina/jina.yaml +++ b/api/core/model_runtime/model_providers/jina/jina.yaml @@ -1,4 +1,4 @@ -provider: Jina AI +provider: jina label: en_US: Jina AI description: diff --git a/api/core/tools/provider/builtin/jina/jina.yaml b/api/core/tools/provider/builtin/jina/jina.yaml index 9ce5cbd6d1afa3..346175c41fbe17 100644 --- a/api/core/tools/provider/builtin/jina/jina.yaml +++ b/api/core/tools/provider/builtin/jina/jina.yaml @@ -1,6 +1,6 @@ identity: author: Dify - name: Jina AI + name: jina label: en_US: Jina AI zh_Hans: Jina AI From 64baedb48429bf7332feab61d7a83cf1b7c91cda Mon Sep 17 00:00:00 2001 From: ice yao Date: Tue, 24 Sep 2024 14:04:07 +0800 Subject: [PATCH 21/79] fix: update nomic model provider token calculation (#8705) --- api/core/model_runtime/model_providers/_position.yaml | 1 + .../nomic/text_embedding/text_embedding.py | 10 +--------- 2 files changed, 2 insertions(+), 9 deletions(-) diff --git a/api/core/model_runtime/model_providers/_position.yaml b/api/core/model_runtime/model_providers/_position.yaml index 79ebd007647c85..80db22ea84fe63 100644 --- a/api/core/model_runtime/model_providers/_position.yaml +++ b/api/core/model_runtime/model_providers/_position.yaml @@ -39,3 +39,4 @@ - zhinao - fireworks - mixedbread +- nomic diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py index 6cccff6d4601cf..ccbfd196a95222 100644 --- a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py @@ -77,15 +77,7 @@ def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int :param texts: texts to embed :return: """ - if len(texts) == 0: - return 0 - - _, prompt_tokens, _ = self.embed_text( - model=model, - credentials=credentials, - texts=texts, - ) - return prompt_tokens + return sum(self._get_num_tokens_by_gpt2(text) for text in texts) def validate_credentials(self, model: str, credentials: dict) -> None: """ From f42ef0624dbb8a1286a4fd00b15eaf90ceb70921 Mon Sep 17 00:00:00 2001 From: zxhlyh Date: Tue, 24 Sep 2024 17:23:11 +0800 Subject: [PATCH 22/79] fix: embedded chat on ios (#8718) --- web/public/embed.js | 60 ++++++++++++++++++++--------------------- web/public/embed.min.js | 57 ++++++++++++++++++--------------------- 2 files changed, 56 insertions(+), 61 deletions(-) diff --git a/web/public/embed.js b/web/public/embed.js index 8ed7a67dc8a311..3c2735b6fc8d1f 100644 --- a/web/public/embed.js +++ b/web/public/embed.js @@ -69,38 +69,47 @@ iframe.id = iframeId; iframe.src = iframeUrl; iframe.style.cssText = ` - border: none; position: fixed; flex-direction: column; justify-content: space-between; + border: none; position: absolute; flex-direction: column; justify-content: space-between; box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px; - bottom: 5rem; right: 1rem; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem; + bottom: 55px; right: 0; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem; max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647; overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none; `; - document.body.appendChild(iframe); + return iframe; } // Function to reset the iframe position function resetIframePosition() { + if (window.innerWidth <= 640) + return + const targetIframe = document.getElementById(iframeId); const targetButton = document.getElementById(buttonId); if (targetIframe && targetButton) { const buttonRect = targetButton.getBoundingClientRect(); - const buttonBottom = window.innerHeight - buttonRect.bottom; - const buttonRight = window.innerWidth - buttonRect.right; - const buttonLeft = buttonRect.left; - - // Adjust iframe position to stay within viewport - targetIframe.style.bottom = `${ - buttonBottom + buttonRect.height + 5 + targetIframe.clientHeight > window.innerHeight - ? buttonBottom - targetIframe.clientHeight - 5 - : buttonBottom + buttonRect.height + 5 - }px`; - - targetIframe.style.right = `${ - buttonRight + targetIframe.clientWidth > window.innerWidth - ? window.innerWidth - buttonLeft - targetIframe.clientWidth - : buttonRight - }px`; + + const buttonInBottom = buttonRect.top - 5 > targetIframe.clientHeight + + if (buttonInBottom) { + targetIframe.style.bottom = `${buttonRect.height + 5}px`; + targetIframe.style.top = 'unset'; + } + else { + targetIframe.style.bottom = 'unset'; + targetIframe.style.top = `${buttonRect.height + 5}px`; + } + + const buttonInRight = buttonRect.right > targetIframe.clientWidth; + + if (buttonInRight) { + targetIframe.style.right = '0'; + targetIframe.style.left = 'unset'; + } + else { + targetIframe.style.right = 'unset'; + targetIframe.style.left = 0; + } } } @@ -146,12 +155,6 @@ box-shadow: var(--${containerDiv.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px); cursor: pointer; z-index: 2147483647; - transition: all 0.2s ease-in-out 0s; - } - `); - styleSheet.sheet.insertRule(` - #${containerDiv.id}:hover { - transform: var(--${containerDiv.id}-hover-transform, scale(1.1)); } `); @@ -167,7 +170,7 @@ containerDiv.addEventListener("click", function () { const targetIframe = document.getElementById(iframeId); if (!targetIframe) { - createIframe(); + containerDiv.appendChild(createIframe()); resetIframePosition(); this.title = "Exit (ESC)"; displayDiv.innerHTML = svgIcons.close; @@ -255,9 +258,6 @@ if (!document.getElementById(buttonId)) { createButton(); } - - createIframe(); - document.getElementById(iframeId).style.display = 'none'; } // Add esc Exit keyboard event triggered @@ -279,4 +279,4 @@ } else { document.body.onload = embedChatbot; } -})(); +})(); \ No newline at end of file diff --git a/web/public/embed.min.js b/web/public/embed.min.js index 0e023cb5d19704..eb2085814883b9 100644 --- a/web/public/embed.min.js +++ b/web/public/embed.min.js @@ -1,31 +1,26 @@ -(()=>{let t="difyChatbotConfig",a="dify-chatbot-bubble-button",c="dify-chatbot-bubble-window",h=window[t],p={open:` - - `,close:` - - `};async function e(){if(h&&h.token){var e=new URLSearchParams(await(async()=>{var e=h?.inputs||{};let n={};return await Promise.all(Object.entries(e).map(async([e,t])=>{n[e]=(e=t,e=(new TextEncoder).encode(e),e=new Response(new Blob([e]).stream().pipeThrough(new CompressionStream("gzip"))).arrayBuffer(),e=new Uint8Array(await e),await btoa(String.fromCharCode(...e)))})),n})());let t=`${h.baseUrl||`https://${h.isDev?"dev.":""}udify.app`}/chatbot/${h.token}?`+e;function o(){var e=document.createElement("iframe");e.allow="fullscreen;microphone",e.title="dify chatbot bubble window",e.id=c,e.src=t,e.style.cssText=` - border: none; position: fixed; flex-direction: column; justify-content: space-between; - box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px; - bottom: 5rem; right: 1rem; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem; - max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647; - overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none; - `,document.body.appendChild(e)}function i(){var e,t,n,o=document.getElementById(c),i=document.getElementById(a);o&&i&&(i=i.getBoundingClientRect(),e=window.innerHeight-i.bottom,t=window.innerWidth-i.right,n=i.left,o.style.bottom=`${e+i.height+5+o.clientHeight>window.innerHeight?e-o.clientHeight-5:e+i.height+5}px`,o.style.right=`${t+o.clientWidth>window.innerWidth?window.innerWidth-n-o.clientWidth:t}px`)}function n(){let n=document.createElement("div");Object.entries(h.containerProps||{}).forEach(([e,t])=>{"className"===e?n.classList.add(...t.split(" ")):"style"===e?"object"==typeof t?Object.assign(n.style,t):n.style.cssText=t:"function"==typeof t?n.addEventListener(e.replace(/^on/,"").toLowerCase(),t):n[e]=t}),n.id=a;var e=document.createElement("style");document.head.appendChild(e),e.sheet.insertRule(` - #${n.id} { - position: fixed; - bottom: var(--${n.id}-bottom, 1rem); - right: var(--${n.id}-right, 1rem); - left: var(--${n.id}-left, unset); - top: var(--${n.id}-top, unset); - width: var(--${n.id}-width, 50px); - height: var(--${n.id}-height, 50px); - border-radius: var(--${n.id}-border-radius, 25px); - background-color: var(--${n.id}-bg-color, #155EEF); - box-shadow: var(--${n.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px); - cursor: pointer; - z-index: 2147483647; - transition: all 0.2s ease-in-out 0s; - } - `),e.sheet.insertRule(` - #${n.id}:hover { - transform: var(--${n.id}-hover-transform, scale(1.1)); - } - `);let t=document.createElement("div");if(t.style.cssText="display: flex; align-items: center; justify-content: center; width: 100%; height: 100%; z-index: 2147483647;",t.innerHTML=p.open,n.appendChild(t),document.body.appendChild(n),n.addEventListener("click",function(){var e=document.getElementById(c);e?(e.style.display="none"===e.style.display?"block":"none",t.innerHTML="none"===e.style.display?p.open:p.close,"none"===e.style.display?document.removeEventListener("keydown",d):document.addEventListener("keydown",d),i()):(o(),i(),this.title="Exit (ESC)",t.innerHTML=p.close,document.addEventListener("keydown",d))}),h.draggable){var s=n;var l=h.dragAxis||"both";let i=!1,d,r;s.addEventListener("mousedown",function(e){i=!0,d=e.clientX-s.offsetLeft,r=e.clientY-s.offsetTop}),document.addEventListener("mousemove",function(e){var t,n,o;i&&(s.style.transition="none",s.style.cursor="grabbing",(t=document.getElementById(c))&&(t.style.display="none",s.querySelector("div").innerHTML=p.open),t=e.clientX-d,e=window.innerHeight-e.clientY-r,o=s.getBoundingClientRect(),n=window.innerWidth-o.width,o=window.innerHeight-o.height,"x"!==l&&"both"!==l||s.style.setProperty(`--${a}-left`,Math.max(0,Math.min(t,n))+"px"),"y"!==l&&"both"!==l||s.style.setProperty(`--${a}-bottom`,Math.max(0,Math.min(e,o))+"px"))}),document.addEventListener("mouseup",function(){i=!1,s.style.transition="",s.style.cursor="pointer"})}}2048 + + `,close:` + + `};async function e(){if(p&&p.token){var e=new URLSearchParams(await async function(){var e=p?.inputs||{};const n={};return await Promise.all(Object.entries(e).map(async([e,t])=>{n[e]=(e=t,e=(new TextEncoder).encode(e),e=new Response(new Blob([e]).stream().pipeThrough(new CompressionStream("gzip"))).arrayBuffer(),e=new Uint8Array(await e),await btoa(String.fromCharCode(...e)))})),n}());const i=`${p.baseUrl||`https://${p.isDev?"dev.":""}udify.app`}/chatbot/${p.token}?`+e;function o(){var e,t;window.innerWidth<=640||(e=document.getElementById(c),t=document.getElementById(a),e&&t&&((t=t.getBoundingClientRect()).top-5>e.clientHeight?(e.style.bottom=t.height+5+"px",e.style.top="unset"):(e.style.bottom="unset",e.style.top=t.height+5+"px"),t.right>e.clientWidth?(e.style.right="0",e.style.left="unset"):(e.style.right="unset",e.style.left=0)))}function t(){const n=document.createElement("div");Object.entries(p.containerProps||{}).forEach(([e,t])=>{"className"===e?n.classList.add(...t.split(" ")):"style"===e?"object"==typeof t?Object.assign(n.style,t):n.style.cssText=t:"function"==typeof t?n.addEventListener(e.replace(/^on/,"").toLowerCase(),t):n[e]=t}),n.id=a;var e=document.createElement("style");document.head.appendChild(e),e.sheet.insertRule(` + #${n.id} { + position: fixed; + bottom: var(--${n.id}-bottom, 1rem); + right: var(--${n.id}-right, 1rem); + left: var(--${n.id}-left, unset); + top: var(--${n.id}-top, unset); + width: var(--${n.id}-width, 50px); + height: var(--${n.id}-height, 50px); + border-radius: var(--${n.id}-border-radius, 25px); + background-color: var(--${n.id}-bg-color, #155EEF); + box-shadow: var(--${n.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px); + cursor: pointer; + z-index: 2147483647; + } + `);const t=document.createElement("div");if(t.style.cssText="display: flex; align-items: center; justify-content: center; width: 100%; height: 100%; z-index: 2147483647;",t.innerHTML=h.open,n.appendChild(t),document.body.appendChild(n),n.addEventListener("click",function(){var e=document.getElementById(c);e?(e.style.display="none"===e.style.display?"block":"none",t.innerHTML="none"===e.style.display?h.open:h.close,"none"===e.style.display?document.removeEventListener("keydown",d):document.addEventListener("keydown",d),o()):(n.appendChild(((e=document.createElement("iframe")).allow="fullscreen;microphone",e.title="dify chatbot bubble window",e.id=c,e.src=i,e.style.cssText=` + border: none; position: absolute; flex-direction: column; justify-content: space-between; + box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px; + bottom: 55px; right: 0; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem; + max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647; + overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none; + `,e)),o(),this.title="Exit (ESC)",t.innerHTML=h.close,document.addEventListener("keydown",d))}),p.draggable){var s=n;var l=p.dragAxis||"both";let i=!1,d,r;s.addEventListener("mousedown",function(e){i=!0,d=e.clientX-s.offsetLeft,r=e.clientY-s.offsetTop}),document.addEventListener("mousemove",function(e){var t,n,o;i&&(s.style.transition="none",s.style.cursor="grabbing",(t=document.getElementById(c))&&(t.style.display="none",s.querySelector("div").innerHTML=h.open),t=e.clientX-d,e=window.innerHeight-e.clientY-r,o=s.getBoundingClientRect(),n=window.innerWidth-o.width,o=window.innerHeight-o.height,"x"!==l&&"both"!==l||s.style.setProperty(`--${a}-left`,Math.max(0,Math.min(t,n))+"px"),"y"!==l&&"both"!==l||s.style.setProperty(`--${a}-bottom`,Math.max(0,Math.min(e,o))+"px"))}),document.addEventListener("mouseup",function(){i=!1,s.style.transition="",s.style.cursor="pointer"})}}2048 Date: Tue, 24 Sep 2024 17:33:29 +0800 Subject: [PATCH 23/79] chore: remove windows platform timezone set (#8712) --- api/app.py | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/api/app.py b/api/app.py index 91a49337fccbde..1b58beee158199 100644 --- a/api/app.py +++ b/api/app.py @@ -53,11 +53,9 @@ warnings.simplefilter("ignore", ResourceWarning) -# fix windows platform -if os.name == "nt": - os.system('tzutil /s "UTC"') -else: - os.environ["TZ"] = "UTC" +os.environ["TZ"] = "UTC" +# windows platform not support tzset +if hasattr(time, "tzset"): time.tzset() From 1c7877b048d2131c8c133c26c60c8ccd342f0b0c Mon Sep 17 00:00:00 2001 From: Shota Totsuka <153569547+totsukash@users.noreply.github.com> Date: Tue, 24 Sep 2024 21:53:26 +0900 Subject: [PATCH 24/79] fix: remove harm category setting from vertex ai (#8721) --- .../model_providers/vertex_ai/llm/llm.py | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/llm.py b/api/core/model_runtime/model_providers/vertex_ai/llm/llm.py index da69b7cdf382de..1dd785d5454082 100644 --- a/api/core/model_runtime/model_providers/vertex_ai/llm/llm.py +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/llm.py @@ -2,6 +2,7 @@ import io import json import logging +import time from collections.abc import Generator from typing import Optional, Union, cast @@ -20,7 +21,6 @@ from google.cloud import aiplatform from google.oauth2 import service_account from PIL import Image -from vertexai.generative_models import HarmBlockThreshold, HarmCategory from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta, LLMUsage from core.model_runtime.entities.message_entities import ( @@ -34,6 +34,7 @@ ToolPromptMessage, UserPromptMessage, ) +from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.errors.invoke import ( InvokeAuthorizationError, InvokeBadRequestError, @@ -503,20 +504,12 @@ def _generate( else: history.append(content) - safety_settings = { - HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE, - } - google_model = glm.GenerativeModel(model_name=model, system_instruction=system_instruction) response = google_model.generate_content( contents=history, generation_config=glm.GenerationConfig(**config_kwargs), stream=stream, - safety_settings=safety_settings, tools=self._convert_tools_to_glm_tool(tools) if tools else None, ) From debe5953a814f075aec5f878bd338829cd63ca9e Mon Sep 17 00:00:00 2001 From: Sa Zhang <55871322+Nick17t@users.noreply.github.com> Date: Tue, 24 Sep 2024 21:19:49 +0800 Subject: [PATCH 25/79] Fix/update jina ai products labels and descriptions (#8730) Co-authored-by: sa zhang --- .../builtin/jina/tools/jina_reader.yaml | 42 +++++++++---------- .../builtin/jina/tools/jina_search.yaml | 31 +++++++------- .../builtin/jina/tools/jina_tokenizer.yaml | 16 +++++-- 3 files changed, 50 insertions(+), 39 deletions(-) diff --git a/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml b/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml index 58ad6d8694222d..589bc3433d9478 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml @@ -2,14 +2,14 @@ identity: name: jina_reader author: Dify label: - en_US: JinaReader - zh_Hans: JinaReader - pt_BR: JinaReader + en_US: Fetch Single Page + zh_Hans: 获取单页面 + pt_BR: Fetch Single Page description: human: - en_US: Convert any URL to an LLM-friendly input. Experience improved output for your agent and RAG systems at no cost. - zh_Hans: 将任何 URL 转换为 LLM 友好的输入。无需付费即可体验为您的 Agent 和 RAG 系统提供的改进输出。 - pt_BR: Converta qualquer URL em uma entrada amigável ao LLM. Experimente uma saída aprimorada para seus sistemas de agente e RAG sem custo. + en_US: Fetch the target URL (can be a PDF) and convert it into a LLM-friendly markdown. + zh_Hans: 获取目标网址（可以是 PDF），并将其转换为适合大模型处理的 Markdown 格式。 + pt_BR: Busque a URL de destino (que pode ser um PDF) e converta em um Markdown LLM-friendly. llm: A tool for scraping webpages. Input should be a URL. parameters: - name: url @@ -17,13 +17,13 @@ parameters: required: true label: en_US: URL - zh_Hans: 网页链接 + zh_Hans: 网址 pt_BR: URL human_description: - en_US: used for linking to webpages - zh_Hans: 用于链接到网页 - pt_BR: used for linking to webpages - llm_description: url for scraping + en_US: Web link + zh_Hans: 网页链接 + pt_BR: URL da web + llm_description: url para scraping form: llm - name: request_params type: string @@ -31,14 +31,14 @@ parameters: label: en_US: Request params zh_Hans: 请求参数 - pt_BR: Request params + pt_BR: Parâmetros de solicitação human_description: en_US: | request parameters, format: {"key1": "value1", "key2": "value2"} zh_Hans: | 请求参数，格式：{"key1": "value1", "key2": "value2"} pt_BR: | - request parameters, format: {"key1": "value1", "key2": "value2"} + parâmetros de solicitação, formato: {"key1": "value1", "key2": "value2"} llm_description: request parameters form: llm - name: target_selector @@ -51,7 +51,7 @@ parameters: human_description: en_US: css selector for scraping specific elements zh_Hans: css 选择器用于抓取特定元素 - pt_BR: css selector for scraping specific elements + pt_BR: css selector para scraping de elementos específicos llm_description: css selector of the target element to scrape form: form - name: wait_for_selector @@ -64,7 +64,7 @@ parameters: human_description: en_US: css selector for waiting for specific elements zh_Hans: css 选择器用于等待特定元素 - pt_BR: css selector for waiting for specific elements + pt_BR: css selector para aguardar elementos específicos llm_description: css selector of the target element to wait for form: form - name: image_caption @@ -77,8 +77,8 @@ parameters: pt_BR: Legenda da imagem human_description: en_US: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." - zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签。这允许下游 LLM 在推理和总结等活动中与图像进行交互。" - pt_BR: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." + zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签，以支持下游模型的图像交互。" + pt_BR: "Adiciona legendas a todas as imagens na URL especificada, adicionando 'Imagem [idx]: [legenda]' como uma tag alt para aquelas que não têm uma. Isso permite que os modelos LLM inferiores interajam com as imagens em atividades como raciocínio e resumo." llm_description: Captions all images at the specified URL form: form - name: gather_all_links_at_the_end @@ -91,8 +91,8 @@ parameters: pt_BR: Coletar todos os links ao final human_description: en_US: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. - zh_Hans: 最后会创建一个“按钮和链接”部分。这可以帮助下游 LLM 或 Web 代理浏览页面或采取进一步的行动。 - pt_BR: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. + zh_Hans: 末尾将添加“按钮和链接”部分，方便下游模型或网络代理做页面导航或执行进一步操作。 + pt_BR: Um "Botões & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. llm_description: Gather all links at the end form: form - name: gather_all_images_at_the_end @@ -105,8 +105,8 @@ parameters: pt_BR: Coletar todas as imagens ao final human_description: en_US: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. - zh_Hans: 最后会创建一个“图像”部分。这可以让下游的 LLM 概览页面上的所有视觉效果，从而提高推理能力。 - pt_BR: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. + zh_Hans: 末尾会新增“图片”部分，方便下游模型全面了解页面的视觉内容，提升推理效果。 + pt_BR: Um "Imagens" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. llm_description: Gather all images at the end form: form - name: proxy_server diff --git a/api/core/tools/provider/builtin/jina/tools/jina_search.yaml b/api/core/tools/provider/builtin/jina/tools/jina_search.yaml index 2bc70e1be1934d..e58c639e5690d0 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_search.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_search.yaml @@ -2,13 +2,14 @@ identity: name: jina_search author: Dify label: - en_US: JinaSearch - zh_Hans: JinaSearch - pt_BR: JinaSearch + en_US: Search the web + zh_Hans: 联网搜索 + pt_BR: Search the web description: human: - en_US: Search on the web and get the top 5 results. Useful for grounding using information from the web. - zh_Hans: 在网络上搜索返回前 5 个结果。 + en_US: Search on the public web of a given query and return the top results as LLM-friendly markdown. + zh_Hans: 针对给定的查询在互联网上进行搜索，并以适合大模型处理的 Markdown 格式返回最相关的结果。 + pt_BR: Procurar na web pública de uma consulta fornecida e retornar os melhores resultados como markdown para LLMs. llm: A tool for searching results on the web for grounding. Input should be a simple question. parameters: - name: query @@ -16,11 +17,13 @@ parameters: required: true label: en_US: Question (Query) - zh_Hans: 信息查询 + zh_Hans: 查询 + pt_BR: Pergunta (Consulta) human_description: en_US: used to find information on the web zh_Hans: 在网络上搜索信息 - llm_description: simple question to ask on the web + pt_BR: Usado para encontrar informações na web + llm_description: Pergunta simples para fazer na web form: llm - name: image_caption type: boolean @@ -32,7 +35,7 @@ parameters: pt_BR: Legenda da imagem human_description: en_US: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." - zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签。这允许下游 LLM 在推理和总结等活动中与图像进行交互。" + zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签，以支持下游模型的图像交互。" pt_BR: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." llm_description: Captions all images at the specified URL form: form @@ -46,8 +49,8 @@ parameters: pt_BR: Coletar todos os links ao final human_description: en_US: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. - zh_Hans: 最后会创建一个“按钮和链接”部分。这可以帮助下游 LLM 或 Web 代理浏览页面或采取进一步的行动。 - pt_BR: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. + zh_Hans: 末尾将添加“按钮和链接”部分，汇总页面上的所有链接。方便下游模型或网络代理做页面导航或执行进一步操作。 + pt_BR: Um "Botão & Links" seção será criada no final. Isso ajuda os LLMs ou agentes da web navegando pela página ou executar ações adicionais. llm_description: Gather all links at the end form: form - name: gather_all_images_at_the_end @@ -60,8 +63,8 @@ parameters: pt_BR: Coletar todas as imagens ao final human_description: en_US: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. - zh_Hans: 最后会创建一个“图像”部分。这可以让下游的 LLM 概览页面上的所有视觉效果，从而提高推理能力。 - pt_BR: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. + zh_Hans: 末尾会新增“图片”部分，汇总页面上的所有图片。方便下游模型概览页面的视觉内容，提升推理效果。 + pt_BR: Um "Imagens" seção será criada no final. Isso fornece uma visão geral de todas as imagens na página para os LLMs, que pode melhorar a razão. llm_description: Gather all images at the end form: form - name: proxy_server @@ -74,7 +77,7 @@ parameters: human_description: en_US: Use proxy to access URLs zh_Hans: 利用代理访问 URL - pt_BR: Use proxy to access URLs + pt_BR: Usar proxy para acessar URLs llm_description: Use proxy to access URLs form: form - name: no_cache @@ -83,7 +86,7 @@ parameters: default: false label: en_US: Bypass the Cache - zh_Hans: 绕过缓存 + zh_Hans: 是否绕过缓存 pt_BR: Ignorar o cache human_description: en_US: Bypass the Cache diff --git a/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml b/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml index 62a5c7e7bacd75..74885cdf9a7048 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml @@ -2,11 +2,14 @@ identity: name: jina_tokenizer author: hjlarry label: - en_US: JinaTokenizer + en_US: Segment + zh_Hans: 切分器 + pt_BR: Segment description: human: - en_US: Free API to tokenize text and segment long text into chunks. - zh_Hans: 免费的API可以将文本tokenize，也可以将长文本分割成多个部分。 + en_US: Split long text into chunks and do tokenization. + zh_Hans: 将长文本拆分成小段落，并做分词处理。 + pt_BR: Dividir o texto longo em pedaços e fazer tokenização. llm: Free API to tokenize text and segment long text into chunks. parameters: - name: content @@ -15,6 +18,7 @@ parameters: label: en_US: Content zh_Hans: 内容 + pt_BR: Conteúdo llm_description: the content which need to tokenize or segment form: llm - name: return_tokens @@ -23,18 +27,22 @@ parameters: label: en_US: Return the tokens zh_Hans: 是否返回tokens + pt_BR: Retornar os tokens human_description: en_US: Return the tokens and their corresponding ids in the response. zh_Hans: 返回tokens及其对应的ids。 + pt_BR: Retornar os tokens e seus respectivos ids na resposta. form: form - name: return_chunks type: boolean label: en_US: Return the chunks zh_Hans: 是否分块 + pt_BR: Retornar os chunks human_description: en_US: Chunking the input into semantically meaningful segments while handling a wide variety of text types and edge cases based on common structural cues. - zh_Hans: 将输入分块为具有语义意义的片段，同时根据常见的结构线索处理各种文本类型和边缘情况。 + zh_Hans: 将输入文本分块为语义有意义的片段，同时基于常见的结构线索处理各种文本类型和特殊情况。 + pt_BR: Dividir o texto de entrada em segmentos semanticamente significativos, enquanto lida com uma ampla variedade de tipos de texto e casos de borda com base em pistas estruturais comuns. form: form - name: tokenizer type: select From 4669eb24befe840789de1dc9ae32f631a31a765d Mon Sep 17 00:00:00 2001 From: Jyong <76649700+JohnJyong@users.noreply.github.com> Date: Tue, 24 Sep 2024 21:53:50 +0800 Subject: [PATCH 26/79] add embedding input type parameter (#8724) --- api/core/embedding/cached_embedding.py | 9 ++++++-- api/core/embedding/embedding_constant.py | 10 ++++++++ api/core/model_manager.py | 7 +++++- .../__base/text_embedding_model.py | 23 +++++++++++++++---- .../text_embedding/text_embedding.py | 8 ++++++- .../baichuan/text_embedding/text_embedding.py | 8 ++++++- .../bedrock/text_embedding/text_embedding.py | 8 ++++++- .../cohere/text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../hunyuan/text_embedding/text_embedding.py | 8 ++++++- .../jina/text_embedding/text_embedding.py | 8 ++++++- .../localai/text_embedding/text_embedding.py | 8 ++++++- .../minimax/text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../nomic/text_embedding/text_embedding.py | 2 ++ .../nvidia/text_embedding/text_embedding.py | 8 ++++++- .../oci/text_embedding/text_embedding.py | 9 +++++++- .../ollama/text_embedding/text_embedding.py | 8 ++++++- .../openai/text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../openllm/text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 9 +++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../tongyi/text_embedding/text_embedding.py | 2 ++ .../upstage/text_embedding/text_embedding.py | 10 +++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 8 ++++++- .../wenxin/text_embedding/text_embedding.py | 8 ++++++- .../text_embedding/text_embedding.py | 9 +++++++- .../zhipuai/text_embedding/text_embedding.py | 8 ++++++- 33 files changed, 239 insertions(+), 35 deletions(-) create mode 100644 api/core/embedding/embedding_constant.py diff --git a/api/core/embedding/cached_embedding.py b/api/core/embedding/cached_embedding.py index 8ce12fd59f5761..75219051cd30cd 100644 --- a/api/core/embedding/cached_embedding.py +++ b/api/core/embedding/cached_embedding.py @@ -5,6 +5,7 @@ import numpy as np from sqlalchemy.exc import IntegrityError +from core.embedding.embedding_constant import EmbeddingInputType from core.model_manager import ModelInstance from core.model_runtime.entities.model_entities import ModelPropertyKey from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel @@ -56,7 +57,9 @@ def embed_documents(self, texts: list[str]) -> list[list[float]]: for i in range(0, len(embedding_queue_texts), max_chunks): batch_texts = embedding_queue_texts[i : i + max_chunks] - embedding_result = self._model_instance.invoke_text_embedding(texts=batch_texts, user=self._user) + embedding_result = self._model_instance.invoke_text_embedding( + texts=batch_texts, user=self._user, input_type=EmbeddingInputType.DOCUMENT + ) for vector in embedding_result.embeddings: try: @@ -100,7 +103,9 @@ def embed_query(self, text: str) -> list[float]: redis_client.expire(embedding_cache_key, 600) return list(np.frombuffer(base64.b64decode(embedding), dtype="float")) try: - embedding_result = self._model_instance.invoke_text_embedding(texts=[text], user=self._user) + embedding_result = self._model_instance.invoke_text_embedding( + texts=[text], user=self._user, input_type=EmbeddingInputType.QUERY + ) embedding_results = embedding_result.embeddings[0] embedding_results = (embedding_results / np.linalg.norm(embedding_results)).tolist() diff --git a/api/core/embedding/embedding_constant.py b/api/core/embedding/embedding_constant.py new file mode 100644 index 00000000000000..9b4934646bc0e8 --- /dev/null +++ b/api/core/embedding/embedding_constant.py @@ -0,0 +1,10 @@ +from enum import Enum + + +class EmbeddingInputType(Enum): + """ + Enum for embedding input type. + """ + + DOCUMENT = "document" + QUERY = "query" diff --git a/api/core/model_manager.py b/api/core/model_manager.py index 990efd36c609c2..74b445236268d9 100644 --- a/api/core/model_manager.py +++ b/api/core/model_manager.py @@ -3,6 +3,7 @@ from collections.abc import Callable, Generator, Sequence from typing import IO, Optional, Union, cast +from core.embedding.embedding_constant import EmbeddingInputType from core.entities.provider_configuration import ProviderConfiguration, ProviderModelBundle from core.entities.provider_entities import ModelLoadBalancingConfiguration from core.errors.error import ProviderTokenNotInitError @@ -158,12 +159,15 @@ def get_llm_num_tokens( tools=tools, ) - def invoke_text_embedding(self, texts: list[str], user: Optional[str] = None) -> TextEmbeddingResult: + def invoke_text_embedding( + self, texts: list[str], user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT + ) -> TextEmbeddingResult: """ Invoke large language model :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ if not isinstance(self.model_type_instance, TextEmbeddingModel): @@ -176,6 +180,7 @@ def invoke_text_embedding(self, texts: list[str], user: Optional[str] = None) -> credentials=self.credentials, texts=texts, user=user, + input_type=input_type, ) def get_text_embedding_num_tokens(self, texts: list[str]) -> int: diff --git a/api/core/model_runtime/model_providers/__base/text_embedding_model.py b/api/core/model_runtime/model_providers/__base/text_embedding_model.py index 54a44860236918..a948dca20d69a4 100644 --- a/api/core/model_runtime/model_providers/__base/text_embedding_model.py +++ b/api/core/model_runtime/model_providers/__base/text_embedding_model.py @@ -4,6 +4,7 @@ from pydantic import ConfigDict +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import ModelPropertyKey, ModelType from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult from core.model_runtime.model_providers.__base.ai_model import AIModel @@ -20,35 +21,47 @@ class TextEmbeddingModel(AIModel): model_config = ConfigDict(protected_namespaces=()) def invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ - Invoke large language model + Invoke text embedding model :param model: model name :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ self.started_at = time.perf_counter() try: - return self._invoke(model, credentials, texts, user) + return self._invoke(model, credentials, texts, user, input_type) except Exception as e: raise self._transform_invoke_error(e) @abstractmethod def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ - Invoke large language model + Invoke text embedding model :param model: model name :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ raise NotImplementedError diff --git a/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py index d9cff8ecbbadb9..6b270b65ff00a6 100644 --- a/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py @@ -7,6 +7,7 @@ import tiktoken from openai import AzureOpenAI +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import AIModelEntity, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.validate import CredentialsValidateFailedError @@ -17,7 +18,12 @@ class AzureOpenAITextEmbeddingModel(_CommonAzureOpenAI, TextEmbeddingModel): def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: base_model_name = credentials["base_model_name"] credentials_kwargs = self._to_credential_kwargs(credentials) diff --git a/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py index 779dfbb6081638..210c274bdf5e3e 100644 --- a/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ from requests import post +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -35,7 +36,12 @@ class BaichuanTextEmbeddingModel(TextEmbeddingModel): api_base: str = "http://api.baichuan-ai.com/v1/embeddings" def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py index 251170d1aec492..8c4c50b26958ce 100644 --- a/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py @@ -13,6 +13,7 @@ UnknownServiceError, ) +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -30,7 +31,12 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel): def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py index a1c5e98118e4c6..1f93068a8c9e52 100644 --- a/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py @@ -5,6 +5,7 @@ import numpy as np from cohere.core import RequestOptions +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -25,7 +26,12 @@ class CohereTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py index 4ad96c4233fb1a..cf18f84ac8cb22 100644 --- a/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py @@ -6,6 +6,7 @@ import requests from huggingface_hub import HfApi, InferenceClient +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -18,7 +19,12 @@ class HuggingfaceHubTextEmbeddingModel(_CommonHuggingfaceHub, TextEmbeddingModel): def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: client = InferenceClient(token=credentials["huggingfacehub_api_token"]) diff --git a/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py index 55f3c25804ea09..58baf4933c2414 100644 --- a/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py @@ -1,6 +1,7 @@ import time from typing import Optional +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -23,7 +24,12 @@ class HuggingfaceTeiTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py index 1396e59e188bca..3e14371f897e01 100644 --- a/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py @@ -9,6 +9,7 @@ from tencentcloud.common.profile.http_profile import HttpProfile from tencentcloud.hunyuan.v20230901 import hunyuan_client, models +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -26,7 +27,12 @@ class HunyuanTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py index 6c96699ea2e3f1..9120f26b8d92fa 100644 --- a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ from requests import post +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -60,7 +61,12 @@ def transform_jina_input_text(model, text): return data def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py index 7d258be81e0580..d8878c7be85bc5 100644 --- a/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py @@ -5,6 +5,7 @@ from requests import post from yarl import URL +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -26,7 +27,12 @@ class LocalAITextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py index 76fd1342bdb929..d0d1d2aea1ccf5 100644 --- a/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ from requests import post +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -34,7 +35,12 @@ class MinimaxTextEmbeddingModel(TextEmbeddingModel): api_base: str = "https://api.minimax.chat/v1/embeddings" def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py index 05d9a9a0c69f45..cdc2d58d0c7e3c 100644 --- a/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ import requests +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -27,7 +28,12 @@ class MixedBreadTextEmbeddingModel(TextEmbeddingModel): api_base: str = "https://api.mixedbread.ai/v1" def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py index ccbfd196a95222..a797521576b236 100644 --- a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py @@ -5,6 +5,7 @@ from nomic import embed from nomic import login as nomic_login +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import ( EmbeddingUsage, @@ -46,6 +47,7 @@ def _invoke( credentials: dict, texts: list[str], user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py index 00cec265d5ed98..a4ea28bd1073b1 100644 --- a/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ from requests import post +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -27,7 +28,12 @@ class NvidiaTextEmbeddingModel(TextEmbeddingModel): models: list[str] = ["NV-Embed-QA"] def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/oci/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/oci/text_embedding/text_embedding.py index 80ad2be9f56872..4de9296ccaa92d 100644 --- a/api/core/model_runtime/model_providers/oci/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/oci/text_embedding/text_embedding.py @@ -6,6 +6,7 @@ import numpy as np import oci +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -41,7 +42,12 @@ class OCITextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model @@ -50,6 +56,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ # get model properties diff --git a/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py index b4c61d8a6dc790..0501c8b841666c 100644 --- a/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py @@ -8,6 +8,7 @@ import numpy as np import requests +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import ( AIModelEntity, @@ -38,7 +39,12 @@ class OllamaEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py index 535d8388bc7d7d..7945723636ced1 100644 --- a/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py @@ -6,6 +6,7 @@ import tiktoken from openai import OpenAI +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.validate import CredentialsValidateFailedError @@ -19,7 +20,12 @@ class OpenAITextEmbeddingModel(_CommonOpenAI, TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py index e83cfdf873cb4b..68b5773e16afe0 100644 --- a/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py @@ -7,6 +7,7 @@ import numpy as np import requests +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import ( AIModelEntity, @@ -28,7 +29,12 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py index 00e583cc797dcb..c0a3efbb003dc4 100644 --- a/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py @@ -5,6 +5,7 @@ from requests import post from requests.exceptions import ConnectionError, InvalidSchema, MissingSchema +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import ( @@ -25,7 +26,12 @@ class OpenLLMTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/perfxcloud/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/text_embedding.py index b62a2d2aaf06fe..1e86f351c8ae57 100644 --- a/api/core/model_runtime/model_providers/perfxcloud/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/text_embedding.py @@ -7,6 +7,7 @@ import numpy as np import requests +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import ( AIModelEntity, @@ -28,7 +29,12 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model @@ -37,6 +43,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py index 71b6fb99c4a1b2..b6cf89bcd98683 100644 --- a/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ from replicate import Client as ReplicateClient +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -14,7 +15,12 @@ class ReplicateEmbeddingModel(_CommonReplicate, TextEmbeddingModel): def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: client = ReplicateClient(api_token=credentials["replicate_api_token"], timeout=30) diff --git a/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py index d55144f8a79be9..957f2e5d0f78fe 100644 --- a/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py @@ -6,6 +6,7 @@ import boto3 +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -53,7 +54,12 @@ def _sagemaker_embedding(self, sm_client, endpoint_name, content_list: list[str] return embeddings def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py index 6cdf4933b47c73..c6c681c15d9d40 100644 --- a/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py @@ -1,5 +1,6 @@ from typing import Optional +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult from core.model_runtime.model_providers.openai_api_compatible.text_embedding.text_embedding import ( OAICompatEmbeddingModel, @@ -16,7 +17,12 @@ def validate_credentials(self, model: str, credentials: dict) -> None: super().validate_credentials(model, credentials) def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: self._add_custom_parameters(credentials) return super()._invoke(model, credentials, texts, user) diff --git a/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py index 5783d2e383e2de..0eef0db3e7df03 100644 --- a/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py @@ -4,6 +4,7 @@ import dashscope import numpy as np +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import ( EmbeddingUsage, @@ -27,6 +28,7 @@ def _invoke( credentials: dict, texts: list[str], user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py index edd4a36d98f587..812bf92eeace77 100644 --- a/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py @@ -7,6 +7,7 @@ from openai import OpenAI from tokenizers import Tokenizer +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.validate import CredentialsValidateFailedError @@ -22,7 +23,14 @@ class UpstageTextEmbeddingModel(_CommonUpstage, TextEmbeddingModel): def _get_tokenizer(self) -> Tokenizer: return Tokenizer.from_pretrained("upstage/solar-1-mini-tokenizer") - def _invoke(self, model: str, credentials: dict, texts: list[str], user: str | None = None) -> TextEmbeddingResult: + def _invoke( + self, + model: str, + credentials: dict, + texts: list[str], + user: str | None = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, + ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py index 519373a7f31a35..509b41d9517bf8 100644 --- a/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py @@ -9,6 +9,7 @@ from google.oauth2 import service_account from vertexai.language_models import TextEmbeddingModel as VertexTextEmbeddingModel +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import ( AIModelEntity, @@ -30,7 +31,12 @@ class VertexAiTextEmbeddingModel(_CommonVertexAi, TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py index 9cba2cb8794184..9d800af5f4d38f 100644 --- a/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py @@ -2,6 +2,7 @@ from decimal import Decimal from typing import Optional +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import ( AIModelEntity, @@ -41,7 +42,12 @@ class VolcengineMaaSTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py index 4d6f6dccd0cf72..1b5a0904db9639 100644 --- a/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py @@ -7,6 +7,7 @@ import numpy as np from requests import Response, post +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.invoke import InvokeError @@ -70,7 +71,12 @@ def _create_text_embedding(self, api_key: str, secret_key: str) -> TextEmbedding return WenxinTextEmbedding(api_key, secret_key) def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model diff --git a/api/core/model_runtime/model_providers/xinference/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/xinference/text_embedding/text_embedding.py index 8043af1d6cf11e..16272391320d55 100644 --- a/api/core/model_runtime/model_providers/xinference/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/xinference/text_embedding/text_embedding.py @@ -3,6 +3,7 @@ from xinference_client.client.restful.restful_client import Client, RESTfulEmbeddingModelHandle +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.common_entities import I18nObject from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult @@ -25,7 +26,12 @@ class XinferenceTextEmbeddingModel(TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model @@ -40,6 +46,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ server_url = credentials["server_url"] diff --git a/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py index ee20954381053d..707c08ef1b5f64 100644 --- a/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py @@ -1,6 +1,7 @@ import time from typing import Optional +from core.embedding.embedding_constant import EmbeddingInputType from core.model_runtime.entities.model_entities import PriceType from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult from core.model_runtime.errors.validate import CredentialsValidateFailedError @@ -15,7 +16,12 @@ class ZhipuAITextEmbeddingModel(_CommonZhipuaiAI, TextEmbeddingModel): """ def _invoke( - self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: """ Invoke text embedding model From 91f70d0bd9d96e07c61c5a37cc36d456526758e2 Mon Sep 17 00:00:00 2001 From: ice yao Date: Wed, 25 Sep 2024 08:47:11 +0800 Subject: [PATCH 27/79] Add embedding models in fireworks provider (#8728) --- .../model_providers/fireworks/fireworks.yaml | 1 + .../text_embedding/UAE-Large-V1.yaml | 12 ++ .../fireworks/text_embedding/__init__.py | 0 .../fireworks/text_embedding/gte-base.yaml | 12 ++ .../fireworks/text_embedding/gte-large.yaml | 12 ++ .../text_embedding/nomic-embed-text-v1.5.yaml | 12 ++ .../text_embedding/nomic-embed-text-v1.yaml | 12 ++ .../text_embedding/text_embedding.py | 151 ++++++++++++++++++ .../fireworks/test_text_embedding.py | 54 +++++++ 9 files changed, 266 insertions(+) create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/UAE-Large-V1.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/__init__.py create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/gte-base.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/gte-large.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.5.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/text_embedding/text_embedding.py create mode 100644 api/tests/integration_tests/model_runtime/fireworks/test_text_embedding.py diff --git a/api/core/model_runtime/model_providers/fireworks/fireworks.yaml b/api/core/model_runtime/model_providers/fireworks/fireworks.yaml index f886fa23b5bd82..cdb87a55e94660 100644 --- a/api/core/model_runtime/model_providers/fireworks/fireworks.yaml +++ b/api/core/model_runtime/model_providers/fireworks/fireworks.yaml @@ -15,6 +15,7 @@ help: en_US: https://fireworks.ai/account/api-keys supported_model_types: - llm + - text-embedding configurate_methods: - predefined-model provider_credential_schema: diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/UAE-Large-V1.yaml b/api/core/model_runtime/model_providers/fireworks/text_embedding/UAE-Large-V1.yaml new file mode 100644 index 00000000000000..d7c11691cf9bbc --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/UAE-Large-V1.yaml @@ -0,0 +1,12 @@ +model: WhereIsAI/UAE-Large-V1 +label: + zh_Hans: UAE-Large-V1 + en_US: UAE-Large-V1 +model_type: text-embedding +model_properties: + context_size: 512 + max_chunks: 1 +pricing: + input: '0.008' + unit: '0.000001' + currency: 'USD' diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/__init__.py b/api/core/model_runtime/model_providers/fireworks/text_embedding/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-base.yaml b/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-base.yaml new file mode 100644 index 00000000000000..d09bafb4d312f9 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-base.yaml @@ -0,0 +1,12 @@ +model: thenlper/gte-base +label: + zh_Hans: GTE-base + en_US: GTE-base +model_type: text-embedding +model_properties: + context_size: 512 + max_chunks: 1 +pricing: + input: '0.008' + unit: '0.000001' + currency: 'USD' diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-large.yaml b/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-large.yaml new file mode 100644 index 00000000000000..c41fa2f9d32361 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/gte-large.yaml @@ -0,0 +1,12 @@ +model: thenlper/gte-large +label: + zh_Hans: GTE-large + en_US: GTE-large +model_type: text-embedding +model_properties: + context_size: 512 + max_chunks: 1 +pricing: + input: '0.008' + unit: '0.000001' + currency: 'USD' diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.5.yaml b/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.5.yaml new file mode 100644 index 00000000000000..c9098503d96529 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.5.yaml @@ -0,0 +1,12 @@ +model: nomic-ai/nomic-embed-text-v1.5 +label: + zh_Hans: nomic-embed-text-v1.5 + en_US: nomic-embed-text-v1.5 +model_type: text-embedding +model_properties: + context_size: 8192 + max_chunks: 16 +pricing: + input: '0.008' + unit: '0.000001' + currency: 'USD' diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.yaml b/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.yaml new file mode 100644 index 00000000000000..89078d3ff69f93 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.yaml @@ -0,0 +1,12 @@ +model: nomic-ai/nomic-embed-text-v1 +label: + zh_Hans: nomic-embed-text-v1 + en_US: nomic-embed-text-v1 +model_type: text-embedding +model_properties: + context_size: 8192 + max_chunks: 16 +pricing: + input: '0.008' + unit: '0.000001' + currency: 'USD' diff --git a/api/core/model_runtime/model_providers/fireworks/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/fireworks/text_embedding/text_embedding.py new file mode 100644 index 00000000000000..cdce69ff380338 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/text_embedding/text_embedding.py @@ -0,0 +1,151 @@ +import time +from collections.abc import Mapping +from typing import Optional, Union + +import numpy as np +from openai import OpenAI + +from core.embedding.embedding_constant import EmbeddingInputType +from core.model_runtime.entities.model_entities import PriceType +from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel +from core.model_runtime.model_providers.fireworks._common import _CommonFireworks + + +class FireworksTextEmbeddingModel(_CommonFireworks, TextEmbeddingModel): + """ + Model class for Fireworks text embedding model. + """ + + def _invoke( + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, + ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ + + credentials_kwargs = self._to_credential_kwargs(credentials) + client = OpenAI(**credentials_kwargs) + + extra_model_kwargs = {} + if user: + extra_model_kwargs["user"] = user + + extra_model_kwargs["encoding_format"] = "float" + + context_size = self._get_context_size(model, credentials) + max_chunks = self._get_max_chunks(model, credentials) + + inputs = [] + indices = [] + used_tokens = 0 + + for i, text in enumerate(texts): + # Here token count is only an approximation based on the GPT2 tokenizer + # TODO: Optimize for better token estimation and chunking + num_tokens = self._get_num_tokens_by_gpt2(text) + + if num_tokens >= context_size: + cutoff = int(np.floor(len(text) * (context_size / num_tokens))) + # if num tokens is larger than context length, only use the start + inputs.append(text[0:cutoff]) + else: + inputs.append(text) + indices += [i] + + batched_embeddings = [] + _iter = range(0, len(inputs), max_chunks) + + for i in _iter: + embeddings_batch, embedding_used_tokens = self._embedding_invoke( + model=model, + client=client, + texts=inputs[i : i + max_chunks], + extra_model_kwargs=extra_model_kwargs, + ) + used_tokens += embedding_used_tokens + batched_embeddings += embeddings_batch + + usage = self._calc_response_usage(model=model, credentials=credentials, tokens=used_tokens) + return TextEmbeddingResult(embeddings=batched_embeddings, usage=usage, model=model) + + def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :return: + """ + return sum(self._get_num_tokens_by_gpt2(text) for text in texts) + + def validate_credentials(self, model: str, credentials: Mapping) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ + try: + # transform credentials to kwargs for model instance + credentials_kwargs = self._to_credential_kwargs(credentials) + client = OpenAI(**credentials_kwargs) + + # call embedding model + self._embedding_invoke(model=model, client=client, texts=["ping"], extra_model_kwargs={}) + except Exception as ex: + raise CredentialsValidateFailedError(str(ex)) + + def _embedding_invoke( + self, model: str, client: OpenAI, texts: Union[list[str], str], extra_model_kwargs: dict + ) -> tuple[list[list[float]], int]: + """ + Invoke embedding model + :param model: model name + :param client: model client + :param texts: texts to embed + :param extra_model_kwargs: extra model kwargs + :return: embeddings and used tokens + """ + response = client.embeddings.create(model=model, input=texts, **extra_model_kwargs) + return [data.embedding for data in response.data], response.usage.total_tokens + + def _calc_response_usage(self, model: str, credentials: dict, tokens: int) -> EmbeddingUsage: + """ + Calculate response usage + + :param model: model name + :param credentials: model credentials + :param tokens: input tokens + :return: usage + """ + input_price_info = self.get_price( + model=model, credentials=credentials, tokens=tokens, price_type=PriceType.INPUT + ) + + usage = EmbeddingUsage( + tokens=tokens, + total_tokens=tokens, + unit_price=input_price_info.unit_price, + price_unit=input_price_info.unit, + total_price=input_price_info.total_amount, + currency=input_price_info.currency, + latency=time.perf_counter() - self.started_at, + ) + + return usage diff --git a/api/tests/integration_tests/model_runtime/fireworks/test_text_embedding.py b/api/tests/integration_tests/model_runtime/fireworks/test_text_embedding.py new file mode 100644 index 00000000000000..7bf723b3a93742 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/fireworks/test_text_embedding.py @@ -0,0 +1,54 @@ +import os + +import pytest + +from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.fireworks.text_embedding.text_embedding import FireworksTextEmbeddingModel +from tests.integration_tests.model_runtime.__mock.openai import setup_openai_mock + + +@pytest.mark.parametrize("setup_openai_mock", [["text_embedding"]], indirect=True) +def test_validate_credentials(setup_openai_mock): + model = FireworksTextEmbeddingModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials( + model="nomic-ai/nomic-embed-text-v1.5", credentials={"fireworks_api_key": "invalid_key"} + ) + + model.validate_credentials( + model="nomic-ai/nomic-embed-text-v1.5", credentials={"fireworks_api_key": os.environ.get("FIREWORKS_API_KEY")} + ) + + +@pytest.mark.parametrize("setup_openai_mock", [["text_embedding"]], indirect=True) +def test_invoke_model(setup_openai_mock): + model = FireworksTextEmbeddingModel() + + result = model.invoke( + model="nomic-ai/nomic-embed-text-v1.5", + credentials={ + "fireworks_api_key": os.environ.get("FIREWORKS_API_KEY"), + }, + texts=["hello", "world", " ".join(["long_text"] * 100), " ".join(["another_long_text"] * 100)], + user="foo", + ) + + assert isinstance(result, TextEmbeddingResult) + assert len(result.embeddings) == 4 + assert result.usage.total_tokens == 2 + + +def test_get_num_tokens(): + model = FireworksTextEmbeddingModel() + + num_tokens = model.get_num_tokens( + model="nomic-ai/nomic-embed-text-v1.5", + credentials={ + "fireworks_api_key": os.environ.get("FIREWORKS_API_KEY"), + }, + texts=["hello", "world"], + ) + + assert num_tokens == 2 From 68c7e68a8a0db9efe5c18c85c64be8346d3b2574 Mon Sep 17 00:00:00 2001 From: ybalbert001 <120714773+ybalbert001@users.noreply.github.com> Date: Wed, 25 Sep 2024 09:12:35 +0800 Subject: [PATCH 28/79] Fix Issue: switch LLM of SageMaker endpoint doesn't take effect (#8737) Co-authored-by: Yuanbo Li --- .../model_providers/sagemaker/llm/llm.py | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/api/core/model_runtime/model_providers/sagemaker/llm/llm.py b/api/core/model_runtime/model_providers/sagemaker/llm/llm.py index 04789197eec2a1..97b76920443840 100644 --- a/api/core/model_runtime/model_providers/sagemaker/llm/llm.py +++ b/api/core/model_runtime/model_providers/sagemaker/llm/llm.py @@ -84,8 +84,9 @@ class SageMakerLargeLanguageModel(LargeLanguageModel): Model class for Cohere large language model. """ - sagemaker_client: Any = None + sagemaker_session: Any = None predictor: Any = None + sagemaker_endpoint: str = None def _handle_chat_generate_response( self, @@ -211,7 +212,7 @@ def _invoke( :param user: unique user id :return: full response or stream response chunk generator result """ - if not self.sagemaker_client: + if not self.sagemaker_session: access_key = credentials.get("aws_access_key_id") secret_key = credentials.get("aws_secret_access_key") aws_region = credentials.get("aws_region") @@ -226,11 +227,14 @@ def _invoke( else: boto_session = boto3.Session() - self.sagemaker_client = boto_session.client("sagemaker") - sagemaker_session = Session(boto_session=boto_session, sagemaker_client=self.sagemaker_client) + sagemaker_client = boto_session.client("sagemaker") + self.sagemaker_session = Session(boto_session=boto_session, sagemaker_client=sagemaker_client) + + if self.sagemaker_endpoint != credentials.get("sagemaker_endpoint"): + self.sagemaker_endpoint = credentials.get("sagemaker_endpoint") self.predictor = Predictor( - endpoint_name=credentials.get("sagemaker_endpoint"), - sagemaker_session=sagemaker_session, + endpoint_name=self.sagemaker_endpoint, + sagemaker_session=self.sagemaker_session, serializer=serializers.JSONSerializer(), ) From bf64ff215be9e9094afddbd9523188c78942b039 Mon Sep 17 00:00:00 2001 From: crazywoola <100913391+crazywoola@users.noreply.github.com> Date: Wed, 25 Sep 2024 10:09:20 +0800 Subject: [PATCH 29/79] fix: . is missing in file_extension (#8736) --- api/core/rag/extractor/extract_processor.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/api/core/rag/extractor/extract_processor.py b/api/core/rag/extractor/extract_processor.py index fe7eaa32e62cb6..0ffc89b214c2d9 100644 --- a/api/core/rag/extractor/extract_processor.py +++ b/api/core/rag/extractor/extract_processor.py @@ -124,7 +124,7 @@ def extract( extractor = UnstructuredPPTXExtractor(file_path, unstructured_api_url) elif file_extension == ".xml": extractor = UnstructuredXmlExtractor(file_path, unstructured_api_url) - elif file_extension == "epub": + elif file_extension == ".epub": extractor = UnstructuredEpubExtractor(file_path, unstructured_api_url) else: # txt @@ -146,7 +146,7 @@ def extract( extractor = WordExtractor(file_path, upload_file.tenant_id, upload_file.created_by) elif file_extension == ".csv": extractor = CSVExtractor(file_path, autodetect_encoding=True) - elif file_extension == "epub": + elif file_extension == ".epub": extractor = UnstructuredEpubExtractor(file_path) else: # txt From cb1942c242b17d930111cc97947e78f2efb9d1cd Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Wed, 25 Sep 2024 11:27:17 +0800 Subject: [PATCH 30/79] chore: make url display in the middle of http node (#8741) --- web/app/components/workflow/nodes/http/node.tsx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/web/app/components/workflow/nodes/http/node.tsx b/web/app/components/workflow/nodes/http/node.tsx index 5bbb10fc3a5966..4b7dbea2571ae1 100644 --- a/web/app/components/workflow/nodes/http/node.tsx +++ b/web/app/components/workflow/nodes/http/node.tsx @@ -15,7 +15,7 @@ const Node: FC> = ({
{method}
-
+
Date: Wed, 25 Sep 2024 14:48:06 +0800 Subject: [PATCH 31/79] chore: apply ruff reformat for python-client sdk (#8752) --- sdks/python-client/dify_client/client.py | 137 ++++++++--------------- 1 file changed, 47 insertions(+), 90 deletions(-) diff --git a/sdks/python-client/dify_client/client.py b/sdks/python-client/dify_client/client.py index 2be079bdf381ce..5e42507a42abc7 100644 --- a/sdks/python-client/dify_client/client.py +++ b/sdks/python-client/dify_client/client.py @@ -1,103 +1,80 @@ import json + import requests class DifyClient: - def __init__(self, api_key, base_url: str = 'https://api.dify.ai/v1'): + def __init__(self, api_key, base_url: str = "https://api.dify.ai/v1"): self.api_key = api_key self.base_url = base_url def _send_request(self, method, endpoint, json=None, params=None, stream=False): - headers = { - "Authorization": f"Bearer {self.api_key}", - "Content-Type": "application/json" - } + headers = {"Authorization": f"Bearer {self.api_key}", "Content-Type": "application/json"} url = f"{self.base_url}{endpoint}" response = requests.request(method, url, json=json, params=params, headers=headers, stream=stream) return response - def _send_request_with_files(self, method, endpoint, data, files): - headers = { - "Authorization": f"Bearer {self.api_key}" - } + headers = {"Authorization": f"Bearer {self.api_key}"} url = f"{self.base_url}{endpoint}" response = requests.request(method, url, data=data, headers=headers, files=files) return response - + def message_feedback(self, message_id, rating, user): - data = { - "rating": rating, - "user": user - } + data = {"rating": rating, "user": user} return self._send_request("POST", f"/messages/{message_id}/feedbacks", data) - + def get_application_parameters(self, user): params = {"user": user} return self._send_request("GET", "/parameters", params=params) - + def file_upload(self, user, files): - data = { - "user": user - } + data = {"user": user} return self._send_request_with_files("POST", "/files/upload", data=data, files=files) - def text_to_audio(self, text:str, user:str, streaming:bool=False): - data = { - "text": text, - "user": user, - "streaming": streaming - } + def text_to_audio(self, text: str, user: str, streaming: bool = False): + data = {"text": text, "user": user, "streaming": streaming} return self._send_request("POST", "/text-to-audio", data=data) - - def get_meta(self,user): - params = { "user": user} - return self._send_request("GET", f"/meta", params=params) + + def get_meta(self, user): + params = {"user": user} + return self._send_request("GET", "/meta", params=params) class CompletionClient(DifyClient): def create_completion_message(self, inputs, response_mode, user, files=None): - data = { - "inputs": inputs, - "response_mode": response_mode, - "user": user, - "files": files - } - return self._send_request("POST", "/completion-messages", data, - stream=True if response_mode == "streaming" else False) + data = {"inputs": inputs, "response_mode": response_mode, "user": user, "files": files} + return self._send_request( + "POST", "/completion-messages", data, stream=True if response_mode == "streaming" else False + ) class ChatClient(DifyClient): def create_chat_message(self, inputs, query, user, response_mode="blocking", conversation_id=None, files=None): - data = { - "inputs": inputs, - "query": query, - "user": user, - "response_mode": response_mode, - "files": files - } + data = {"inputs": inputs, "query": query, "user": user, "response_mode": response_mode, "files": files} if conversation_id: data["conversation_id"] = conversation_id - return self._send_request("POST", "/chat-messages", data, - stream=True if response_mode == "streaming" else False) - - def get_suggested(self, message_id, user:str): + return self._send_request( + "POST", "/chat-messages", data, stream=True if response_mode == "streaming" else False + ) + + def get_suggested(self, message_id, user: str): params = {"user": user} return self._send_request("GET", f"/messages/{message_id}/suggested", params=params) - + def stop_message(self, task_id, user): data = {"user": user} - return self._send_request("POST", f"/chat-messages/{task_id}/stop", data) + return self._send_request("POST", f"/chat-messages/{task_id}/stop", data) def get_conversations(self, user, last_id=None, limit=None, pinned=None): params = {"user": user, "last_id": last_id, "limit": limit, "pinned": pinned} return self._send_request("GET", "/conversations", params=params) - + def get_conversation_messages(self, user, conversation_id=None, first_id=None, limit=None): params = {"user": user} @@ -109,15 +86,15 @@ def get_conversation_messages(self, user, conversation_id=None, first_id=None, l params["limit"] = limit return self._send_request("GET", "/messages", params=params) - - def rename_conversation(self, conversation_id, name,auto_generate:bool, user:str): - data = {"name": name, "auto_generate": auto_generate,"user": user} + + def rename_conversation(self, conversation_id, name, auto_generate: bool, user: str): + data = {"name": name, "auto_generate": auto_generate, "user": user} return self._send_request("POST", f"/conversations/{conversation_id}/name", data) def delete_conversation(self, conversation_id, user): data = {"user": user} return self._send_request("DELETE", f"/conversations/{conversation_id}", data) - + def audio_to_text(self, audio_file, user): data = {"user": user} files = {"audio_file": audio_file} @@ -125,10 +102,10 @@ def audio_to_text(self, audio_file, user): class WorkflowClient(DifyClient): - def run(self, inputs:dict, response_mode:str="streaming", user:str="abc-123"): + def run(self, inputs: dict, response_mode: str = "streaming", user: str = "abc-123"): data = {"inputs": inputs, "response_mode": response_mode, "user": user} return self._send_request("POST", "/workflows/run", data) - + def stop(self, task_id, user): data = {"user": user} return self._send_request("POST", f"/workflows/tasks/{task_id}/stop", data) @@ -137,10 +114,8 @@ def get_result(self, workflow_run_id): return self._send_request("GET", f"/workflows/run/{workflow_run_id}") - class KnowledgeBaseClient(DifyClient): - - def __init__(self, api_key, base_url: str = 'https://api.dify.ai/v1', dataset_id: str = None): + def __init__(self, api_key, base_url: str = "https://api.dify.ai/v1", dataset_id: str = None): """ Construct a KnowledgeBaseClient object. @@ -150,10 +125,7 @@ def __init__(self, api_key, base_url: str = 'https://api.dify.ai/v1', dataset_id dataset_id (str, optional): ID of the dataset. Defaults to None. You don't need this if you just want to create a new dataset. or list datasets. otherwise you need to set this. """ - super().__init__( - api_key=api_key, - base_url=base_url - ) + super().__init__(api_key=api_key, base_url=base_url) self.dataset_id = dataset_id def _get_dataset_id(self): @@ -162,10 +134,10 @@ def _get_dataset_id(self): return self.dataset_id def create_dataset(self, name: str, **kwargs): - return self._send_request('POST', '/datasets', {'name': name}, **kwargs) + return self._send_request("POST", "/datasets", {"name": name}, **kwargs) def list_datasets(self, page: int = 1, page_size: int = 20, **kwargs): - return self._send_request('GET', f'/datasets?page={page}&limit={page_size}', **kwargs) + return self._send_request("GET", f"/datasets?page={page}&limit={page_size}", **kwargs) def create_document_by_text(self, name, text, extra_params: dict = None, **kwargs): """ @@ -193,14 +165,7 @@ def create_document_by_text(self, name, text, extra_params: dict = None, **kwarg } :return: Response from the API """ - data = { - 'indexing_technique': 'high_quality', - 'process_rule': { - 'mode': 'automatic' - }, - 'name': name, - 'text': text - } + data = {"indexing_technique": "high_quality", "process_rule": {"mode": "automatic"}, "name": name, "text": text} if extra_params is not None and isinstance(extra_params, dict): data.update(extra_params) url = f"/datasets/{self._get_dataset_id()}/document/create_by_text" @@ -233,10 +198,7 @@ def update_document_by_text(self, document_id, name, text, extra_params: dict = } :return: Response from the API """ - data = { - 'name': name, - 'text': text - } + data = {"name": name, "text": text} if extra_params is not None and isinstance(extra_params, dict): data.update(extra_params) url = f"/datasets/{self._get_dataset_id()}/documents/{document_id}/update_by_text" @@ -269,16 +231,11 @@ def create_document_by_file(self, file_path, original_document_id=None, extra_pa :return: Response from the API """ files = {"file": open(file_path, "rb")} - data = { - 'process_rule': { - 'mode': 'automatic' - }, - 'indexing_technique': 'high_quality' - } + data = {"process_rule": {"mode": "automatic"}, "indexing_technique": "high_quality"} if extra_params is not None and isinstance(extra_params, dict): data.update(extra_params) if original_document_id is not None: - data['original_document_id'] = original_document_id + data["original_document_id"] = original_document_id url = f"/datasets/{self._get_dataset_id()}/document/create_by_file" return self._send_request_with_files("POST", url, {"data": json.dumps(data)}, files) @@ -352,11 +309,11 @@ def list_documents(self, page: int = None, page_size: int = None, keyword: str = """ params = {} if page is not None: - params['page'] = page + params["page"] = page if page_size is not None: - params['limit'] = page_size + params["limit"] = page_size if keyword is not None: - params['keyword'] = keyword + params["keyword"] = keyword url = f"/datasets/{self._get_dataset_id()}/documents" return self._send_request("GET", url, params=params, **kwargs) @@ -383,9 +340,9 @@ def query_segments(self, document_id, keyword: str = None, status: str = None, * url = f"/datasets/{self._get_dataset_id()}/documents/{document_id}/segments" params = {} if keyword is not None: - params['keyword'] = keyword + params["keyword"] = keyword if status is not None: - params['status'] = status + params["status"] = status if "params" in kwargs: params.update(kwargs["params"]) return self._send_request("GET", url, params=params, **kwargs) From d0e0111f88da0fc972b07b1f97893df18cbc0522 Mon Sep 17 00:00:00 2001 From: cherryhuahua <68722306+cherryhuahua@users.noreply.github.com> Date: Wed, 25 Sep 2024 14:51:42 +0800 Subject: [PATCH 32/79] fix:Spark's large language model token calculation error #7911 (#8755) --- api/core/app/apps/base_app_runner.py | 2 +- api/core/model_runtime/model_providers/spark/llm/llm.py | 9 ++++++--- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/api/core/app/apps/base_app_runner.py b/api/core/app/apps/base_app_runner.py index 1b412b86396c25..203aca3384376d 100644 --- a/api/core/app/apps/base_app_runner.py +++ b/api/core/app/apps/base_app_runner.py @@ -309,7 +309,7 @@ def _handle_invoke_result_stream( if not prompt_messages: prompt_messages = result.prompt_messages - if not usage and result.delta.usage: + if result.delta.usage: usage = result.delta.usage if not usage: diff --git a/api/core/model_runtime/model_providers/spark/llm/llm.py b/api/core/model_runtime/model_providers/spark/llm/llm.py index 57193dc0316b72..1181ba699af886 100644 --- a/api/core/model_runtime/model_providers/spark/llm/llm.py +++ b/api/core/model_runtime/model_providers/spark/llm/llm.py @@ -213,18 +213,21 @@ def _handle_generate_stream_response( :param prompt_messages: prompt messages :return: llm response chunk generator result """ + completion = "" for index, content in enumerate(client.subscribe()): if isinstance(content, dict): delta = content["data"] else: delta = content - + completion += delta assistant_prompt_message = AssistantPromptMessage( content=delta or "", ) - + temp_assistant_prompt_message = AssistantPromptMessage( + content=completion, + ) prompt_tokens = self.get_num_tokens(model, credentials, prompt_messages) - completion_tokens = self.get_num_tokens(model, credentials, [assistant_prompt_message]) + completion_tokens = self.get_num_tokens(model, credentials, [temp_assistant_prompt_message]) # transform usage usage = self._calc_response_usage(model, credentials, prompt_tokens, completion_tokens) From b0927c39fb589b9df5c8278dc8bdac115497fe2e Mon Sep 17 00:00:00 2001 From: zhuiyue132 Date: Wed, 25 Sep 2024 15:06:54 +0800 Subject: [PATCH 33/79] fix: expose the configuration of HTTP request node to Docker (#8716) Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com> --- docker/.env.example | 4 ++++ docker/docker-compose.yaml | 2 ++ 2 files changed, 6 insertions(+) diff --git a/docker/.env.example b/docker/.env.example index 7eaaceb92834dc..f7479791cea320 100644 --- a/docker/.env.example +++ b/docker/.env.example @@ -568,6 +568,10 @@ WORKFLOW_MAX_EXECUTION_STEPS=500 WORKFLOW_MAX_EXECUTION_TIME=1200 WORKFLOW_CALL_MAX_DEPTH=5 +# HTTP request node in workflow configuration +HTTP_REQUEST_NODE_MAX_BINARY_SIZE=10485760 +HTTP_REQUEST_NODE_MAX_TEXT_SIZE=1048576 + # SSRF Proxy server HTTP URL SSRF_PROXY_HTTP_URL=http://ssrf_proxy:3128 # SSRF Proxy server HTTPS URL diff --git a/docker/docker-compose.yaml b/docker/docker-compose.yaml index 16bef279bcc688..414919063a739f 100644 --- a/docker/docker-compose.yaml +++ b/docker/docker-compose.yaml @@ -207,6 +207,8 @@ x-shared-env: &shared-api-worker-env WORKFLOW_CALL_MAX_DEPTH: ${WORKFLOW_MAX_EXECUTION_TIME:-5} SSRF_PROXY_HTTP_URL: ${SSRF_PROXY_HTTP_URL:-http://ssrf_proxy:3128} SSRF_PROXY_HTTPS_URL: ${SSRF_PROXY_HTTPS_URL:-http://ssrf_proxy:3128} + HTTP_REQUEST_NODE_MAX_BINARY_SIZE: ${HTTP_REQUEST_NODE_MAX_BINARY_SIZE:-10485760} + HTTP_REQUEST_NODE_MAX_TEXT_SIZE: ${HTTP_REQUEST_NODE_MAX_TEXT_SIZE:-1048576} services: # API service From 2ef8b187fad67c248fc9ce20e440503ba24c0a42 Mon Sep 17 00:00:00 2001 From: Hash Brown Date: Wed, 25 Sep 2024 15:50:51 +0800 Subject: [PATCH 34/79] Add GitHub Actions Workflow for Web Tests (#8753) --- .github/workflows/web-tests.yml | 46 +++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) create mode 100644 .github/workflows/web-tests.yml diff --git a/.github/workflows/web-tests.yml b/.github/workflows/web-tests.yml new file mode 100644 index 00000000000000..5aee64b8e6da02 --- /dev/null +++ b/.github/workflows/web-tests.yml @@ -0,0 +1,46 @@ +name: Web Tests + +on: + pull_request: + branches: + - main + paths: + - web/** + +concurrency: + group: web-tests-${{ github.head_ref || github.run_id }} + cancel-in-progress: true + +jobs: + test: + name: Web Tests + runs-on: ubuntu-latest + defaults: + run: + working-directory: ./web + + steps: + - name: Checkout code + uses: actions/checkout@v4 + + - name: Check changed files + id: changed-files + uses: tj-actions/changed-files@v45 + with: + files: web/** + + - name: Setup Node.js + uses: actions/setup-node@v4 + if: steps.changed-files.outputs.any_changed == 'true' + with: + node-version: 20 + cache: yarn + cache-dependency-path: ./web/package.json + + - name: Install dependencies + if: steps.changed-files.outputs.any_changed == 'true' + run: yarn install --frozen-lockfile + + - name: Run tests + if: steps.changed-files.outputs.any_changed == 'true' + run: yarn test From ef47f68e4ad5713b4694d9e0d0c646fd47efc55d Mon Sep 17 00:00:00 2001 From: NFish Date: Wed, 25 Sep 2024 18:25:06 +0800 Subject: [PATCH 35/79] fix: the translation result may cause a different meaning (#8763) --- web/i18n/es-ES/common.ts | 2 +- web/i18n/zh-Hans/common.ts | 2 +- web/i18n/zh-Hant/common.ts | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/web/i18n/es-ES/common.ts b/web/i18n/es-ES/common.ts index 2ba907361fe8c8..59a05f63d8a3e2 100644 --- a/web/i18n/es-ES/common.ts +++ b/web/i18n/es-ES/common.ts @@ -202,7 +202,7 @@ const translation = { invitationLink: 'Enlace de invitación', failedInvitationEmails: 'Los siguientes usuarios no fueron invitados exitosamente', ok: 'OK', - removeFromTeam: 'Eliminar del equipo', + removeFromTeam: 'Eliminar del espacio de trabajo', removeFromTeamTip: 'Se eliminará el acceso al equipo', setAdmin: 'Establecer como administrador', setMember: 'Establecer como miembro ordinario', diff --git a/web/i18n/zh-Hans/common.ts b/web/i18n/zh-Hans/common.ts index 52ab7d6f024168..7947d32f251f19 100644 --- a/web/i18n/zh-Hans/common.ts +++ b/web/i18n/zh-Hans/common.ts @@ -200,7 +200,7 @@ const translation = { invitationLink: '邀请链接', failedInvitationEmails: '邀请以下邮箱失败', ok: '好的', - removeFromTeam: '移除团队', + removeFromTeam: '移出团队', removeFromTeamTip: '将取消团队访问', setAdmin: '设为管理员', setMember: '设为普通成员', diff --git a/web/i18n/zh-Hant/common.ts b/web/i18n/zh-Hant/common.ts index c1f3ed2b2b53e6..8cd51b1991f1a2 100644 --- a/web/i18n/zh-Hant/common.ts +++ b/web/i18n/zh-Hant/common.ts @@ -194,7 +194,7 @@ const translation = { invitationLink: '邀請連結', failedInvitationEmails: '邀請以下郵箱失敗', ok: '好的', - removeFromTeam: '移除團隊', + removeFromTeam: '移出團隊', removeFromTeamTip: '將取消團隊訪問', setAdmin: '設為管理員', setMember: '設為普通成員', From 02ff6cca70421b6f2d482989a581a98117b64e7d Mon Sep 17 00:00:00 2001 From: "Pan, Wen-Ming" Date: Wed, 25 Sep 2024 21:27:26 +0800 Subject: [PATCH 36/79] feat: add support for Vertex AI Gemini 1.5 002 and experimental models (#8767) --- ...5-flash.yaml => gemini-1.5-flash-001.yaml} | 2 +- .../vertex_ai/llm/gemini-1.5-flash-002.yaml | 37 +++++++++++++++++++ ...i-1.5-pro.yaml => gemini-1.5-pro-001.yaml} | 2 +- .../vertex_ai/llm/gemini-1.5-pro-002.yaml | 37 +++++++++++++++++++ .../llm/gemini-flash-experimental.yaml | 37 +++++++++++++++++++ .../llm/gemini-pro-experimental.yaml | 37 +++++++++++++++++++ 6 files changed, 150 insertions(+), 2 deletions(-) rename api/core/model_runtime/model_providers/vertex_ai/llm/{gemini-1.5-flash.yaml => gemini-1.5-flash-001.yaml} (96%) create mode 100644 api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-002.yaml rename api/core/model_runtime/model_providers/vertex_ai/llm/{gemini-1.5-pro.yaml => gemini-1.5-pro-001.yaml} (96%) create mode 100644 api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-002.yaml create mode 100644 api/core/model_runtime/model_providers/vertex_ai/llm/gemini-flash-experimental.yaml create mode 100644 api/core/model_runtime/model_providers/vertex_ai/llm/gemini-pro-experimental.yaml diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-001.yaml similarity index 96% rename from api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash.yaml rename to api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-001.yaml index c308f0a322fddd..f5386be06da6be 100644 --- a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash.yaml +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-001.yaml @@ -1,6 +1,6 @@ model: gemini-1.5-flash-001 label: - en_US: Gemini 1.5 Flash + en_US: Gemini 1.5 Flash 001 model_type: llm features: - agent-thought diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-002.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-002.yaml new file mode 100644 index 00000000000000..97bd44f06b5145 --- /dev/null +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-002.yaml @@ -0,0 +1,37 @@ +model: gemini-1.5-flash-002 +label: + en_US: Gemini 1.5 Flash 002 +model_type: llm +features: + - agent-thought + - vision +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + en_US: Top k + type: int + help: + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: presence_penalty + use_template: presence_penalty + - name: frequency_penalty + use_template: frequency_penalty + - name: max_output_tokens + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-001.yaml similarity index 96% rename from api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro.yaml rename to api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-001.yaml index 744863e7731e15..5e08f2294e2ebf 100644 --- a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro.yaml +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-001.yaml @@ -1,6 +1,6 @@ model: gemini-1.5-pro-001 label: - en_US: Gemini 1.5 Pro + en_US: Gemini 1.5 Pro 001 model_type: llm features: - agent-thought diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-002.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-002.yaml new file mode 100644 index 00000000000000..8f327ea2f3d37e --- /dev/null +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-002.yaml @@ -0,0 +1,37 @@ +model: gemini-1.5-pro-002 +label: + en_US: Gemini 1.5 Pro 002 +model_type: llm +features: + - agent-thought + - vision +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + en_US: Top k + type: int + help: + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: presence_penalty + use_template: presence_penalty + - name: frequency_penalty + use_template: frequency_penalty + - name: max_output_tokens + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-flash-experimental.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-flash-experimental.yaml new file mode 100644 index 00000000000000..0f5eb34c0cdf03 --- /dev/null +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-flash-experimental.yaml @@ -0,0 +1,37 @@ +model: gemini-flash-experimental +label: + en_US: Gemini Flash Experimental +model_type: llm +features: + - agent-thought + - vision +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + en_US: Top k + type: int + help: + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: presence_penalty + use_template: presence_penalty + - name: frequency_penalty + use_template: frequency_penalty + - name: max_output_tokens + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-pro-experimental.yaml b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-pro-experimental.yaml new file mode 100644 index 00000000000000..fa31cabb85abb0 --- /dev/null +++ b/api/core/model_runtime/model_providers/vertex_ai/llm/gemini-pro-experimental.yaml @@ -0,0 +1,37 @@ +model: gemini-pro-experimental +label: + en_US: Gemini Pro Experimental +model_type: llm +features: + - agent-thought + - vision +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + en_US: Top k + type: int + help: + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: presence_penalty + use_template: presence_penalty + - name: frequency_penalty + use_template: frequency_penalty + - name: max_output_tokens + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD From a8b837c4a9143385718f87fdd4863844607daae3 Mon Sep 17 00:00:00 2001 From: Bowen Liang Date: Wed, 25 Sep 2024 22:55:24 +0800 Subject: [PATCH 37/79] dep: bump ElasticSearch from 8.14.x to 8.15.x (#8197) --- api/poetry.lock | 13 ++++++++----- api/pyproject.toml | 2 +- docker/docker-compose.yaml | 4 ++-- 3 files changed, 11 insertions(+), 8 deletions(-) diff --git a/api/poetry.lock b/api/poetry.lock index 184cdb9e81e57d..bce21fb547c8ca 100644 --- a/api/poetry.lock +++ b/api/poetry.lock @@ -2333,13 +2333,13 @@ develop = ["aiohttp", "furo", "httpx", "opentelemetry-api", "opentelemetry-sdk", [[package]] name = "elasticsearch" -version = "8.14.0" +version = "8.15.1" description = "Python client for Elasticsearch" optional = false -python-versions = ">=3.7" +python-versions = ">=3.8" files = [ - {file = "elasticsearch-8.14.0-py3-none-any.whl", hash = "sha256:cef8ef70a81af027f3da74a4f7d9296b390c636903088439087b8262a468c130"}, - {file = "elasticsearch-8.14.0.tar.gz", hash = "sha256:aa2490029dd96f4015b333c1827aa21fd6c0a4d223b00dfb0fe933b8d09a511b"}, + {file = "elasticsearch-8.15.1-py3-none-any.whl", hash = "sha256:02a0476e98768a30d7926335fc0d305c04fdb928eea1354c6e6040d8c2814569"}, + {file = "elasticsearch-8.15.1.tar.gz", hash = "sha256:40c0d312f8adf8bdc81795bc16a0b546ddf544cb1f90e829a244e4780c4dbfd8"}, ] [package.dependencies] @@ -2347,7 +2347,10 @@ elastic-transport = ">=8.13,<9" [package.extras] async = ["aiohttp (>=3,<4)"] +dev = ["aiohttp", "black", "build", "coverage", "isort", "jinja2", "mapbox-vector-tile", "nox", "numpy", "orjson", "pandas", "pyarrow", "pytest", "pytest-asyncio", "pytest-cov", "python-dateutil", "pyyaml (>=5.4)", "requests (>=2,<3)", "simsimd", "twine", "unasync"] +docs = ["sphinx", "sphinx-autodoc-typehints", "sphinx-rtd-theme (>=2.0)"] orjson = ["orjson (>=3)"] +pyarrow = ["pyarrow (>=1)"] requests = ["requests (>=2.4.0,!=2.32.2,<3.0.0)"] vectorstore-mmr = ["numpy (>=1)", "simsimd (>=3)"] @@ -10498,4 +10501,4 @@ cffi = ["cffi (>=1.11)"] [metadata] lock-version = "2.0" python-versions = ">=3.10,<3.13" -content-hash = "17c4108d92c415d987f8b437ea3e0484c5601a05bfe175339a8546c93c159bc5" +content-hash = "69b42bb1ff033f14e199fee8335356275099421d72bbd7037b7a991ea65cae08" diff --git a/api/pyproject.toml b/api/pyproject.toml index 9e38c0945601cf..f004865d5f3438 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -253,7 +253,7 @@ alibabacloud_gpdb20160503 = "~3.8.0" alibabacloud_tea_openapi = "~0.3.9" chromadb = "0.5.1" clickhouse-connect = "~0.7.16" -elasticsearch = "8.14.0" +elasticsearch = "~8.15.1" oracledb = "~2.2.1" pgvecto-rs = { version = "~0.2.1", extras = ['sqlalchemy'] } pgvector = "0.2.5" diff --git a/docker/docker-compose.yaml b/docker/docker-compose.yaml index 414919063a739f..95e271a0e9e4fb 100644 --- a/docker/docker-compose.yaml +++ b/docker/docker-compose.yaml @@ -630,7 +630,7 @@ services: # https://www.elastic.co/guide/en/elasticsearch/reference/current/settings.html # https://www.elastic.co/guide/en/elasticsearch/reference/current/docker.html#docker-prod-prerequisites elasticsearch: - image: docker.elastic.co/elasticsearch/elasticsearch:8.14.3 + image: docker.elastic.co/elasticsearch/elasticsearch:8.15.1 container_name: elasticsearch profiles: - elasticsearch @@ -657,7 +657,7 @@ services: # https://www.elastic.co/guide/en/kibana/current/docker.html # https://www.elastic.co/guide/en/kibana/current/settings.html kibana: - image: docker.elastic.co/kibana/kibana:8.14.3 + image: docker.elastic.co/kibana/kibana:8.15.1 container_name: kibana profiles: - elasticsearch From fefbc43fb035fca2171d949ac5b4d261c343d8b1 Mon Sep 17 00:00:00 2001 From: Qun <51054082+QunBB@users.noreply.github.com> Date: Thu, 26 Sep 2024 08:18:13 +0800 Subject: [PATCH 38/79] chore: fix comfyui tool doc url (#8775) --- api/core/tools/provider/builtin/comfyui/comfyui.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/api/core/tools/provider/builtin/comfyui/comfyui.yaml b/api/core/tools/provider/builtin/comfyui/comfyui.yaml index 066fd853082817..3891eebf3ac7e4 100644 --- a/api/core/tools/provider/builtin/comfyui/comfyui.yaml +++ b/api/core/tools/provider/builtin/comfyui/comfyui.yaml @@ -39,4 +39,4 @@ credentials_for_provider: en_US: The checkpoint name of the ComfyUI server, e.g. xxx.safetensors zh_Hans: ComfyUI服务器的模型名称, 比如 xxx.safetensors pt_BR: The checkpoint name of the ComfyUI server, e.g. xxx.safetensors - url: https://docs.dify.ai/tutorials/tool-configuration/comfyui + url: https://github.com/comfyanonymous/ComfyUI#installing From 5ba19d64e90a9d5c94b8d4e6cd93cd4043fc8943 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Thu, 26 Sep 2024 11:22:18 +0800 Subject: [PATCH 39/79] fix: TavilySearch tool get api link (#8780) --- api/core/tools/provider/builtin/tavily/tavily.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/api/core/tools/provider/builtin/tavily/tavily.yaml b/api/core/tools/provider/builtin/tavily/tavily.yaml index 7b25a8184857ca..95820f4d18b051 100644 --- a/api/core/tools/provider/builtin/tavily/tavily.yaml +++ b/api/core/tools/provider/builtin/tavily/tavily.yaml @@ -28,4 +28,4 @@ credentials_for_provider: en_US: Get your Tavily API key from Tavily zh_Hans: 从 TavilyApi 获取您的 Tavily API key pt_BR: Get your Tavily API key from Tavily - url: https://docs.tavily.com/docs/tavily-api/introduction + url: https://docs.tavily.com/docs/welcome From ac737637264f4caf10866ab3953cd5c203961939 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Thu, 26 Sep 2024 11:23:09 +0800 Subject: [PATCH 40/79] chore: add input_type param desc for the _invoke method of text_embedding (#8778) --- .../azure_openai/text_embedding/text_embedding.py | 10 ++++++++++ .../baichuan/text_embedding/text_embedding.py | 1 + .../bedrock/text_embedding/text_embedding.py | 1 + .../cohere/text_embedding/text_embedding.py | 1 + .../huggingface_hub/text_embedding/text_embedding.py | 10 ++++++++++ .../huggingface_tei/text_embedding/text_embedding.py | 1 + .../hunyuan/text_embedding/text_embedding.py | 1 + .../jina/text_embedding/text_embedding.py | 1 + .../localai/text_embedding/text_embedding.py | 3 ++- .../minimax/text_embedding/text_embedding.py | 1 + .../mixedbread/text_embedding/text_embedding.py | 1 + .../nomic/text_embedding/text_embedding.py | 1 + .../nvidia/text_embedding/text_embedding.py | 1 + .../ollama/text_embedding/text_embedding.py | 1 + .../openai/text_embedding/text_embedding.py | 1 + .../text_embedding/text_embedding.py | 1 + .../openllm/text_embedding/text_embedding.py | 1 + .../replicate/text_embedding/text_embedding.py | 10 ++++++++++ .../sagemaker/text_embedding/text_embedding.py | 1 + .../siliconflow/text_embedding/text_embedding.py | 10 ++++++++++ .../tongyi/text_embedding/text_embedding.py | 1 + .../upstage/text_embedding/text_embedding.py | 1 + .../vertex_ai/text_embedding/text_embedding.py | 2 ++ .../volcengine_maas/text_embedding/text_embedding.py | 1 + .../wenxin/text_embedding/text_embedding.py | 1 + .../zhipuai/text_embedding/text_embedding.py | 1 + 26 files changed, 64 insertions(+), 1 deletion(-) diff --git a/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py index 6b270b65ff00a6..8701a3805002eb 100644 --- a/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py @@ -25,6 +25,16 @@ def _invoke( user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ base_model_name = credentials["base_model_name"] credentials_kwargs = self._to_credential_kwargs(credentials) client = AzureOpenAI(**credentials_kwargs) diff --git a/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py index 210c274bdf5e3e..56b9be1c365340 100644 --- a/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py @@ -50,6 +50,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ api_key = credentials["api_key"] diff --git a/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py index 8c4c50b26958ce..d9c57265921133 100644 --- a/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py @@ -45,6 +45,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ client_config = Config(region_name=credentials["aws_region"]) diff --git a/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py index 1f93068a8c9e52..4da20806904ba0 100644 --- a/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py @@ -40,6 +40,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ # get model properties diff --git a/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py index cf18f84ac8cb22..b2e6d1b6520c72 100644 --- a/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py @@ -26,6 +26,16 @@ def _invoke( user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ client = InferenceClient(token=credentials["huggingfacehub_api_token"]) execute_model = model diff --git a/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py index 58baf4933c2414..b8ff3ca549a63b 100644 --- a/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py @@ -44,6 +44,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ server_url = credentials["server_url"] diff --git a/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py index 3e14371f897e01..75701ebc54a749 100644 --- a/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py @@ -41,6 +41,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py index 9120f26b8d92fa..c7b729c14eaacb 100644 --- a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py @@ -75,6 +75,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ api_key = credentials["api_key"] diff --git a/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py index d8878c7be85bc5..ab8ca76c2f0971 100644 --- a/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py @@ -23,7 +23,7 @@ class LocalAITextEmbeddingModel(TextEmbeddingModel): """ - Model class for Jina text embedding model. + Model class for LocalAI text embedding model. """ def _invoke( @@ -41,6 +41,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ if len(texts) != 1: diff --git a/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py index d0d1d2aea1ccf5..74d2a221d1b57e 100644 --- a/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py @@ -49,6 +49,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ api_key = credentials["minimax_api_key"] diff --git a/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py index cdc2d58d0c7e3c..68b7b448bfec75 100644 --- a/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py @@ -42,6 +42,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ api_key = credentials["api_key"] diff --git a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py index a797521576b236..857dfb5f41e2f5 100644 --- a/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py @@ -56,6 +56,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ embeddings, prompt_tokens, total_tokens = self.embed_text( diff --git a/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py index a4ea28bd1073b1..936ceb8dd2c60e 100644 --- a/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py @@ -42,6 +42,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ api_key = credentials["api_key"] diff --git a/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py index 0501c8b841666c..5cf3f1c6fa87f3 100644 --- a/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py @@ -53,6 +53,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py index 7945723636ced1..16f1a0cfa1117b 100644 --- a/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py @@ -34,6 +34,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ # transform credentials to kwargs for model instance diff --git a/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py index 68b5773e16afe0..64fa6aaa3c5a71 100644 --- a/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py @@ -43,6 +43,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py index c0a3efbb003dc4..c5d43309127822 100644 --- a/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py @@ -40,6 +40,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ server_url = credentials["server_url"] diff --git a/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py index b6cf89bcd98683..9f724a77ac040b 100644 --- a/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py @@ -22,6 +22,16 @@ def _invoke( user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ client = ReplicateClient(api_token=credentials["replicate_api_token"], timeout=30) if "model_version" in credentials: diff --git a/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py index 957f2e5d0f78fe..8f993ce6722522 100644 --- a/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py @@ -68,6 +68,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ # get model properties diff --git a/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py index c6c681c15d9d40..c5dcc126107aa2 100644 --- a/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py @@ -24,6 +24,16 @@ def _invoke( user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ self._add_custom_parameters(credentials) return super()._invoke(model, credentials, texts, user) diff --git a/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py index 0eef0db3e7df03..736cd44df8888f 100644 --- a/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py @@ -37,6 +37,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ credentials_kwargs = self._to_credential_kwargs(credentials) diff --git a/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py index 812bf92eeace77..b6509cd26cfa28 100644 --- a/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py @@ -38,6 +38,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py index 509b41d9517bf8..fce9544df0a414 100644 --- a/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py @@ -44,6 +44,8 @@ def _invoke( :param model: model name :param credentials: model credentials :param texts: texts to embed + :param user: unique user id + :param input_type: input type :return: embeddings result """ service_account_info = json.loads(base64.b64decode(credentials["vertex_service_account_key"])) diff --git a/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py index 9d800af5f4d38f..0dd4037c958567 100644 --- a/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py @@ -56,6 +56,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ if ArkClientV3.is_legacy(credentials): diff --git a/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py index 1b5a0904db9639..c21d0c055277f7 100644 --- a/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py @@ -85,6 +85,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ diff --git a/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py index 707c08ef1b5f64..14a529dddf82d1 100644 --- a/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py @@ -30,6 +30,7 @@ def _invoke( :param credentials: model credentials :param texts: texts to embed :param user: unique user id + :param input_type: input type :return: embeddings result """ credentials_kwargs = self._to_credential_kwargs(credentials) From 0c96f0aa51e65d970d26f043be1f592f89dba507 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Thu, 26 Sep 2024 11:24:03 +0800 Subject: [PATCH 41/79] fix: credential *** should be string (#8785) --- api/core/tools/provider/tool_provider.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/api/core/tools/provider/tool_provider.py b/api/core/tools/provider/tool_provider.py index 05c88b904e4a6d..321b21201414e8 100644 --- a/api/core/tools/provider/tool_provider.py +++ b/api/core/tools/provider/tool_provider.py @@ -153,6 +153,9 @@ def validate_credentials_format(self, credentials: dict[str, Any]) -> None: # check type credential_schema = credentials_need_to_validate[credential_name] + if not credential_schema.required and credentials[credential_name] is None: + continue + if credential_schema.type in { ToolProviderCredentials.CredentialsType.SECRET_INPUT, ToolProviderCredentials.CredentialsType.TEXT_INPUT, From 4c9ef6e8301b2986653f11ffad837cf9e45f3284 Mon Sep 17 00:00:00 2001 From: Aaron Ji <127167174+DresAaron@users.noreply.github.com> Date: Thu, 26 Sep 2024 11:29:35 +0800 Subject: [PATCH 42/79] fix: update usage for Jina Embeddings v3 (#8771) --- .../model_providers/jina/jina.yaml | 43 ------------------- .../jina/text_embedding/text_embedding.py | 20 +++------ 2 files changed, 6 insertions(+), 57 deletions(-) diff --git a/api/core/model_runtime/model_providers/jina/jina.yaml b/api/core/model_runtime/model_providers/jina/jina.yaml index 4ff6ba0f22c846..970b22965b5d29 100644 --- a/api/core/model_runtime/model_providers/jina/jina.yaml +++ b/api/core/model_runtime/model_providers/jina/jina.yaml @@ -67,46 +67,3 @@ model_credential_schema: required: false type: text-input default: '8192' - - variable: task - label: - zh_Hans: 下游任务 - en_US: Downstream task - placeholder: - zh_Hans: 选择将使用向量模型的下游任务。模型将返回针对该任务优化的向量。 - en_US: Select the downstream task for which the embeddings will be used. The model will return the optimized embeddings for that task. - required: false - type: select - options: - - value: retrieval.query - label: - en_US: retrieval.query - - value: retrieval.passage - label: - en_US: retrieval.passage - - value: separation - label: - en_US: separation - - value: classification - label: - en_US: classification - - value: text-matching - label: - en_US: text-matching - - variable: dimensions - label: - zh_Hans: 输出维度 - en_US: Output dimensions - placeholder: - zh_Hans: 输入您的输出维度 - en_US: Enter output dimensions - required: false - type: text-input - - variable: late_chunking - label: - zh_Hans: 后期分块 - en_US: Late chunking - placeholder: - zh_Hans: 应用后期分块技术来利用模型的长上下文功能来生成上下文块向量化。 - en_US: Apply the late chunking technique to leverage the model's long-context capabilities for generating contextual chunk embeddings. - required: false - type: switch diff --git a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py index c7b729c14eaacb..b39712951256c8 100644 --- a/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py +++ b/api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py @@ -28,7 +28,7 @@ class JinaTextEmbeddingModel(TextEmbeddingModel): api_base: str = "https://api.jina.ai/v1" - def _to_payload(self, model: str, texts: list[str], credentials: dict) -> dict: + def _to_payload(self, model: str, texts: list[str], credentials: dict, input_type: EmbeddingInputType) -> dict: """ Parse model credentials @@ -45,18 +45,10 @@ def transform_jina_input_text(model, text): data = {"model": model, "input": [transform_jina_input_text(model, text) for text in texts]} - task = credentials.get("task") - dimensions = credentials.get("dimensions") - late_chunking = credentials.get("late_chunking") - - if task is not None: - data["task"] = task - - if dimensions is not None: - data["dimensions"] = int(dimensions) - - if late_chunking is not None: - data["late_chunking"] = late_chunking + # model specific parameters + if model == "jina-embeddings-v3": + # set `task` type according to input type for the best performance + data["task"] = "retrieval.query" if input_type == EmbeddingInputType.QUERY else "retrieval.passage" return data @@ -88,7 +80,7 @@ def _invoke( url = base_url + "/embeddings" headers = {"Authorization": "Bearer " + api_key, "Content-Type": "application/json"} - data = self._to_payload(model=model, texts=texts, credentials=credentials) + data = self._to_payload(model=model, texts=texts, credentials=credentials, input_type=input_type) try: response = post(url, headers=headers, data=dumps(data)) From a0b0809b1c1598db7e8762d78c6d9f6d34070168 Mon Sep 17 00:00:00 2001 From: Shenghang Tsai Date: Thu, 26 Sep 2024 11:29:53 +0800 Subject: [PATCH 43/79] Add more models for SiliconFlow (#8779) --- .../siliconflow/llm/_position.yaml | 39 ++++++++++++------- 1 file changed, 26 insertions(+), 13 deletions(-) diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml index 43db4aed11916b..a3e5d0981f7436 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml @@ -1,25 +1,38 @@ -- Qwen/Qwen2.5-7B-Instruct -- Qwen/Qwen2.5-14B-Instruct -- Qwen/Qwen2.5-32B-Instruct - Qwen/Qwen2.5-72B-Instruct +- Qwen/Qwen2.5-Math-72B-Instruct +- Qwen/Qwen2.5-32B-Instruct +- Qwen/Qwen2.5-14B-Instruct +- Qwen/Qwen2.5-7B-Instruct +- Qwen/Qwen2.5-Coder-7B-Instruct +- deepseek-ai/DeepSeek-V2.5 - Qwen/Qwen2-72B-Instruct - Qwen/Qwen2-57B-A14B-Instruct - Qwen/Qwen2-7B-Instruct - Qwen/Qwen2-1.5B-Instruct -- 01-ai/Yi-1.5-34B-Chat -- 01-ai/Yi-1.5-9B-Chat-16K -- 01-ai/Yi-1.5-6B-Chat -- THUDM/glm-4-9b-chat -- deepseek-ai/DeepSeek-V2.5 - deepseek-ai/DeepSeek-V2-Chat - deepseek-ai/DeepSeek-Coder-V2-Instruct +- THUDM/glm-4-9b-chat +- THUDM/chatglm3-6b +- 01-ai/Yi-1.5-34B-Chat-16K +- 01-ai/Yi-1.5-9B-Chat-16K +- 01-ai/Yi-1.5-6B-Chat +- internlm/internlm2_5-20b-chat - internlm/internlm2_5-7b-chat -- google/gemma-2-27b-it -- google/gemma-2-9b-it -- meta-llama/Meta-Llama-3-70B-Instruct -- meta-llama/Meta-Llama-3-8B-Instruct - meta-llama/Meta-Llama-3.1-405B-Instruct - meta-llama/Meta-Llama-3.1-70B-Instruct - meta-llama/Meta-Llama-3.1-8B-Instruct -- mistralai/Mixtral-8x7B-Instruct-v0.1 +- meta-llama/Meta-Llama-3-70B-Instruct +- meta-llama/Meta-Llama-3-8B-Instruct +- google/gemma-2-27b-it +- google/gemma-2-9b-it - mistralai/Mistral-7B-Instruct-v0.2 +- Pro/Qwen/Qwen2-7B-Instruct +- Pro/Qwen/Qwen2-1.5B-Instruct +- Pro/THUDM/glm-4-9b-chat +- Pro/THUDM/chatglm3-6b +- Pro/01-ai/Yi-1.5-9B-Chat-16K +- Pro/01-ai/Yi-1.5-6B-Chat +- Pro/internlm/internlm2_5-7b-chat +- Pro/meta-llama/Meta-Llama-3.1-8B-Instruct +- Pro/meta-llama/Meta-Llama-3-8B-Instruct +- Pro/google/gemma-2-9b-it From d1173a69f8ecf6680b78d194f8c266cada2042ba Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Thu, 26 Sep 2024 13:48:06 +0800 Subject: [PATCH 44/79] fix: the Image-1X tool (#8787) --- .../tools/provider/builtin/stepfun/stepfun.py | 2 +- .../provider/builtin/stepfun/stepfun.yaml | 17 ++--------- .../provider/builtin/stepfun/tools/image.py | 13 ++------- .../provider/builtin/stepfun/tools/image.yaml | 29 ------------------- 4 files changed, 5 insertions(+), 56 deletions(-) diff --git a/api/core/tools/provider/builtin/stepfun/stepfun.py b/api/core/tools/provider/builtin/stepfun/stepfun.py index b24f730c95a7c2..239db85b1118b0 100644 --- a/api/core/tools/provider/builtin/stepfun/stepfun.py +++ b/api/core/tools/provider/builtin/stepfun/stepfun.py @@ -16,7 +16,7 @@ def _validate_credentials(self, credentials: dict[str, Any]) -> None: user_id="", tool_parameters={ "prompt": "cute girl, blue eyes, white hair, anime style", - "size": "1024x1024", + "size": "256x256", "n": 1, }, ) diff --git a/api/core/tools/provider/builtin/stepfun/stepfun.yaml b/api/core/tools/provider/builtin/stepfun/stepfun.yaml index 1f841ec369b5c3..e8139a4d7d6cfd 100644 --- a/api/core/tools/provider/builtin/stepfun/stepfun.yaml +++ b/api/core/tools/provider/builtin/stepfun/stepfun.yaml @@ -4,11 +4,9 @@ identity: label: en_US: Image-1X zh_Hans: 阶跃星辰绘画 - pt_BR: Image-1X description: en_US: Image-1X zh_Hans: 阶跃星辰绘画 - pt_BR: Image-1X icon: icon.png tags: - image @@ -20,27 +18,16 @@ credentials_for_provider: label: en_US: Stepfun API key zh_Hans: 阶跃星辰API key - pt_BR: Stepfun API key - help: - en_US: Please input your stepfun API key - zh_Hans: 请输入你的阶跃星辰 API key - pt_BR: Please input your stepfun API key placeholder: - en_US: Please input your stepfun API key + en_US: Please input your Stepfun API key zh_Hans: 请输入你的阶跃星辰 API key - pt_BR: Please input your stepfun API key + url: https://platform.stepfun.com/interface-key stepfun_base_url: type: text-input required: false label: en_US: Stepfun base URL zh_Hans: 阶跃星辰 base URL - pt_BR: Stepfun base URL - help: - en_US: Please input your Stepfun base URL - zh_Hans: 请输入你的阶跃星辰 base URL - pt_BR: Please input your Stepfun base URL placeholder: en_US: Please input your Stepfun base URL zh_Hans: 请输入你的阶跃星辰 base URL - pt_BR: Please input your Stepfun base URL diff --git a/api/core/tools/provider/builtin/stepfun/tools/image.py b/api/core/tools/provider/builtin/stepfun/tools/image.py index 0b92b122bf59bc..eb55dae5183c41 100644 --- a/api/core/tools/provider/builtin/stepfun/tools/image.py +++ b/api/core/tools/provider/builtin/stepfun/tools/image.py @@ -1,4 +1,3 @@ -import random from typing import Any, Union from openai import OpenAI @@ -19,7 +18,7 @@ def _invoke( """ invoke tools """ - base_url = self.runtime.credentials.get("stepfun_base_url", "https://api.stepfun.com") + base_url = self.runtime.credentials.get("stepfun_base_url") or "https://api.stepfun.com" base_url = str(URL(base_url) / "v1") client = OpenAI( @@ -28,9 +27,7 @@ def _invoke( ) extra_body = {} - model = tool_parameters.get("model", "step-1x-medium") - if not model: - return self.create_text_message("Please input model name") + model = "step-1x-medium" # prompt prompt = tool_parameters.get("prompt", "") if not prompt: @@ -67,9 +64,3 @@ def _invoke( ) ) return result - - @staticmethod - def _generate_random_id(length=8): - characters = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789" - random_id = "".join(random.choices(characters, k=length)) - return random_id diff --git a/api/core/tools/provider/builtin/stepfun/tools/image.yaml b/api/core/tools/provider/builtin/stepfun/tools/image.yaml index dcc5bd2db2f5ba..8d7c9b6586b6cd 100644 --- a/api/core/tools/provider/builtin/stepfun/tools/image.yaml +++ b/api/core/tools/provider/builtin/stepfun/tools/image.yaml @@ -29,35 +29,6 @@ parameters: pt_BR: Image prompt, you can check the official documentation of step-1x llm_description: Image prompt of step-1x you should describe the image you want to generate as a list of words as possible as detailed form: llm - - name: model - type: select - required: false - human_description: - en_US: used for selecting the model name - zh_Hans: 用于选择模型的名字 - pt_BR: used for selecting the model name - label: - en_US: Model Name - zh_Hans: 模型名字 - pt_BR: Model Name - form: form - options: - - value: step-1x-turbo - label: - en_US: turbo - zh_Hans: turbo - pt_BR: turbo - - value: step-1x-medium - label: - en_US: medium - zh_Hans: medium - pt_BR: medium - - value: step-1x-large - label: - en_US: large - zh_Hans: large - pt_BR: large - default: step-1x-medium - name: size type: select required: false From 62406991df10d6eaa721df1b56e8df957802759b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Thu, 26 Sep 2024 16:28:20 +0800 Subject: [PATCH 45/79] fix: start node input config modal raise 'variable name is required' (#8793) --- .../configuration/config-var/config-modal/index.tsx | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/web/app/components/app/configuration/config-var/config-modal/index.tsx b/web/app/components/app/configuration/config-var/config-modal/index.tsx index 606280653eeec2..ac1d86c1a2c2da 100644 --- a/web/app/components/app/configuration/config-var/config-modal/index.tsx +++ b/web/app/components/app/configuration/config-var/config-modal/index.tsx @@ -1,6 +1,6 @@ 'use client' import type { FC } from 'react' -import React, { useCallback, useState } from 'react' +import React, { useCallback, useEffect, useRef, useState } from 'react' import { useTranslation } from 'react-i18next' import { useContext } from 'use-context-selector' import ModalFoot from '../modal-foot' @@ -40,6 +40,12 @@ const ConfigModal: FC = ({ const { t } = useTranslation() const [tempPayload, setTempPayload] = useState(payload || getNewVarInWorkflow('') as any) const { type, label, variable, options, max_length } = tempPayload + const modalRef = useRef(null) + useEffect(() => { + // To fix the first input element auto focus, then directly close modal will raise error + if (isShow) + modalRef.current?.focus() + }, [isShow]) const isStringInput = type === InputVarType.textInput || type === InputVarType.paragraph const checkVariableName = useCallback((value: string) => { @@ -135,7 +141,7 @@ const ConfigModal: FC = ({ isShow={isShow} onClose={onClose} > -
+
From 128a66f7fe8383cb30dec5e5e3896d75a225c77e Mon Sep 17 00:00:00 2001 From: cx <88480957+free-cx@users.noreply.github.com> Date: Thu, 26 Sep 2024 16:34:40 +0800 Subject: [PATCH 46/79] =?UTF-8?q?fix:=20Ollama=20modelfeature=20set=20visi?= =?UTF-8?q?on,=20and=20an=20exception=20occurred=20at=20the=E2=80=A6=20(#8?= =?UTF-8?q?783)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../model_providers/ollama/llm/llm.py | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) diff --git a/api/core/model_runtime/model_providers/ollama/llm/llm.py b/api/core/model_runtime/model_providers/ollama/llm/llm.py index ff732e69259e2b..a7ea53e0e99c5f 100644 --- a/api/core/model_runtime/model_providers/ollama/llm/llm.py +++ b/api/core/model_runtime/model_providers/ollama/llm/llm.py @@ -364,14 +364,21 @@ def create_final_llm_result_chunk( if chunk_json["done"]: # calculate num tokens - if "prompt_eval_count" in chunk_json and "eval_count" in chunk_json: - # transform usage + if "prompt_eval_count" in chunk_json: prompt_tokens = chunk_json["prompt_eval_count"] - completion_tokens = chunk_json["eval_count"] else: - # calculate num tokens - prompt_tokens = self._get_num_tokens_by_gpt2(prompt_messages[0].content) - completion_tokens = self._get_num_tokens_by_gpt2(full_text) + prompt_message_content = prompt_messages[0].content + if isinstance(prompt_message_content, str): + prompt_tokens = self._get_num_tokens_by_gpt2(prompt_message_content) + else: + content_text = "" + for message_content in prompt_message_content: + if message_content.type == PromptMessageContentType.TEXT: + message_content = cast(TextPromptMessageContent, message_content) + content_text += message_content.data + prompt_tokens = self._get_num_tokens_by_gpt2(content_text) + + completion_tokens = chunk_json.get("eval_count", self._get_num_tokens_by_gpt2(full_text)) # transform usage usage = self._calc_response_usage(model, credentials, prompt_tokens, completion_tokens) From 008e0efeb00faecd44877245ec48f08fda6a048f Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Thu, 26 Sep 2024 16:36:21 +0800 Subject: [PATCH 47/79] refactor: update delete method as an abstract method (#8794) --- api/core/rag/datasource/vdb/vector_base.py | 1 + 1 file changed, 1 insertion(+) diff --git a/api/core/rag/datasource/vdb/vector_base.py b/api/core/rag/datasource/vdb/vector_base.py index 1a0dc7f48b8031..22e191340d3a47 100644 --- a/api/core/rag/datasource/vdb/vector_base.py +++ b/api/core/rag/datasource/vdb/vector_base.py @@ -45,6 +45,7 @@ def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Doc def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]: raise NotImplementedError + @abstractmethod def delete(self) -> None: raise NotImplementedError From 6df14e50b2c71e2fc9b0601943d37a78f47e69c2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Thu, 26 Sep 2024 17:50:36 +0800 Subject: [PATCH 48/79] fix: workflow as tool always outdated (#8798) --- web/app/components/tools/workflow-tool/configure-button.tsx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/web/app/components/tools/workflow-tool/configure-button.tsx b/web/app/components/tools/workflow-tool/configure-button.tsx index d2c5142f53c10e..6521410daea017 100644 --- a/web/app/components/tools/workflow-tool/configure-button.tsx +++ b/web/app/components/tools/workflow-tool/configure-button.tsx @@ -65,7 +65,7 @@ const WorkflowToolConfigureButton = ({ else { if (item.type === 'paragraph' && param.type !== 'string') return true - if (param.type !== item.type && !(param.type === 'string' && item.type === 'paragraph')) + if (item.type === 'text-input' && param.type !== 'string') return true } } From 3d2cb25a6704ecd53b132deb4d7ea77b34a43d0d Mon Sep 17 00:00:00 2001 From: Joel Date: Thu, 26 Sep 2024 17:53:11 +0800 Subject: [PATCH 49/79] fix: change wrong company name (#8801) --- web/app/activate/page.tsx | 2 +- web/app/forgot-password/page.tsx | 2 +- web/app/install/page.tsx | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/web/app/activate/page.tsx b/web/app/activate/page.tsx index 90874f50cefe0a..0f1854433552db 100644 --- a/web/app/activate/page.tsx +++ b/web/app/activate/page.tsx @@ -22,7 +22,7 @@ const Activate = () => {
- © {new Date().getFullYear()} Dify, Inc. All rights reserved. + © {new Date().getFullYear()} LangGenius, Inc. All rights reserved.
diff --git a/web/app/forgot-password/page.tsx b/web/app/forgot-password/page.tsx index fa44d1a20c689a..bb46011c06302b 100644 --- a/web/app/forgot-password/page.tsx +++ b/web/app/forgot-password/page.tsx @@ -28,7 +28,7 @@ const ForgotPassword = () => {
{token ? : }
- © {new Date().getFullYear()} Dify, Inc. All rights reserved. + © {new Date().getFullYear()} LangGenius, Inc. All rights reserved.
diff --git a/web/app/install/page.tsx b/web/app/install/page.tsx index 9fa38dd15e2a8c..395fae34ec1690 100644 --- a/web/app/install/page.tsx +++ b/web/app/install/page.tsx @@ -22,7 +22,7 @@ const Install = () => {
- © {new Date().getFullYear()} Dify, Inc. All rights reserved. + © {new Date().getFullYear()} LangGenius, Inc. All rights reserved.
From 03edfbe6f530edfc4fb7781adda3852d6ab643a2 Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Thu, 26 Sep 2024 19:04:25 +0800 Subject: [PATCH 50/79] feat: add qwen to add custom model parameters (#8759) --- .../model_providers/tongyi/llm/llm.py | 102 +++++++++--------- .../model_providers/tongyi/tongyi.yaml | 43 +++++++- 2 files changed, 92 insertions(+), 53 deletions(-) diff --git a/api/core/model_runtime/model_providers/tongyi/llm/llm.py b/api/core/model_runtime/model_providers/tongyi/llm/llm.py index f90c7f075fe86c..3e3585b30ae33d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/llm.py +++ b/api/core/model_runtime/model_providers/tongyi/llm/llm.py @@ -18,7 +18,7 @@ UnsupportedModel, ) -from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta +from core.model_runtime.entities.llm_entities import LLMMode, LLMResult, LLMResultChunk, LLMResultChunkDelta from core.model_runtime.entities.message_entities import ( AssistantPromptMessage, ImagePromptMessageContent, @@ -35,6 +35,7 @@ FetchFrom, I18nObject, ModelFeature, + ModelPropertyKey, ModelType, ParameterRule, ParameterType, @@ -97,6 +98,11 @@ def get_num_tokens( :param tools: tools for tool calling :return: """ + # Check if the model was added via get_customizable_model_schema + if self.get_customizable_model_schema(model, credentials) is not None: + # For custom models, tokens are not calculated. + return 0 + if model in {"qwen-turbo-chat", "qwen-plus-chat"}: model = model.replace("-chat", "") if model == "farui-plus": @@ -537,55 +543,51 @@ def get_customizable_model_schema(self, model: str, credentials: dict) -> AIMode :param credentials: model credentials :return: AIModelEntity or None """ - rules = [ - ParameterRule( - name="temperature", - type=ParameterType.FLOAT, - use_template="temperature", - label=I18nObject(zh_Hans="温度", en_US="Temperature"), - ), - ParameterRule( - name="top_p", - type=ParameterType.FLOAT, - use_template="top_p", - label=I18nObject(zh_Hans="Top P", en_US="Top P"), - ), - ParameterRule( - name="top_k", - type=ParameterType.INT, - min=0, - max=99, - label=I18nObject(zh_Hans="top_k", en_US="top_k"), - ), - ParameterRule( - name="max_tokens", - type=ParameterType.INT, - min=1, - max=128000, - default=1024, - label=I18nObject(zh_Hans="最大生成长度", en_US="Max Tokens"), - ), - ParameterRule( - name="seed", - type=ParameterType.INT, - default=1234, - label=I18nObject(zh_Hans="随机种子", en_US="Random Seed"), - ), - ParameterRule( - name="repetition_penalty", - type=ParameterType.FLOAT, - default=1.1, - label=I18nObject(zh_Hans="重复惩罚", en_US="Repetition Penalty"), - ), - ] - - entity = AIModelEntity( + return AIModelEntity( model=model, - label=I18nObject(en_US=model), - fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + label=I18nObject(en_US=model, zh_Hans=model), model_type=ModelType.LLM, - model_properties={}, - parameter_rules=rules, + features=[ModelFeature.TOOL_CALL, ModelFeature.MULTI_TOOL_CALL, ModelFeature.STREAM_TOOL_CALL] + if credentials.get("function_calling_type") == "tool_call" + else [], + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_properties={ + ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size", 8000)), + ModelPropertyKey.MODE: LLMMode.CHAT.value, + }, + parameter_rules=[ + ParameterRule( + name="temperature", + use_template="temperature", + label=I18nObject(en_US="Temperature", zh_Hans="温度"), + type=ParameterType.FLOAT, + ), + ParameterRule( + name="max_tokens", + use_template="max_tokens", + default=512, + min=1, + max=int(credentials.get("max_tokens", 1024)), + label=I18nObject(en_US="Max Tokens", zh_Hans="最大标记"), + type=ParameterType.INT, + ), + ParameterRule( + name="top_p", + use_template="top_p", + label=I18nObject(en_US="Top P", zh_Hans="Top P"), + type=ParameterType.FLOAT, + ), + ParameterRule( + name="top_k", + use_template="top_k", + label=I18nObject(en_US="Top K", zh_Hans="Top K"), + type=ParameterType.FLOAT, + ), + ParameterRule( + name="frequency_penalty", + use_template="frequency_penalty", + label=I18nObject(en_US="Frequency Penalty", zh_Hans="重复惩罚"), + type=ParameterType.FLOAT, + ), + ], ) - - return entity diff --git a/api/core/model_runtime/model_providers/tongyi/tongyi.yaml b/api/core/model_runtime/model_providers/tongyi/tongyi.yaml index fabe6d90e6b206..1a09c20fd93f61 100644 --- a/api/core/model_runtime/model_providers/tongyi/tongyi.yaml +++ b/api/core/model_runtime/model_providers/tongyi/tongyi.yaml @@ -37,14 +37,51 @@ model_credential_schema: en_US: Model Name zh_Hans: 模型名称 placeholder: - en_US: Enter full model name - zh_Hans: 输入模型全称 + en_US: Enter your model name + zh_Hans: 输入模型名称 credential_form_schemas: - variable: dashscope_api_key - required: true label: en_US: API Key type: secret-input + required: true placeholder: zh_Hans: 在此输入您的 API Key en_US: Enter your API Key + - variable: context_size + label: + zh_Hans: 模型上下文长度 + en_US: Model context size + required: true + type: text-input + default: '4096' + placeholder: + zh_Hans: 在此输入您的模型上下文长度 + en_US: Enter your Model context size + - variable: max_tokens + label: + zh_Hans: 最大 token 上限 + en_US: Upper bound for max tokens + default: '4096' + type: text-input + show_on: + - variable: __model_type + value: llm + - variable: function_calling_type + label: + en_US: Function calling + type: select + required: false + default: no_call + options: + - value: no_call + label: + en_US: Not Support + zh_Hans: 不支持 + - value: function_call + label: + en_US: Support + zh_Hans: 支持 + show_on: + - variable: __model_type + value: llm From 9a4b53a212054901f9c597ac64c0eb97e461e641 Mon Sep 17 00:00:00 2001 From: AAEE86 <33052466+AAEE86@users.noreply.github.com> Date: Thu, 26 Sep 2024 19:08:59 +0800 Subject: [PATCH 51/79] feat: add stream for Gemini (#8678) --- .../google/llm/gemini-1.5-flash-8b-exp-0827.yaml | 9 +++++++++ .../google/llm/gemini-1.5-flash-exp-0827.yaml | 9 +++++++++ .../google/llm/gemini-1.5-flash-latest.yaml | 9 +++++++++ .../google/llm/gemini-1.5-pro-exp-0801.yaml | 9 +++++++++ .../google/llm/gemini-1.5-pro-exp-0827.yaml | 9 +++++++++ .../google/llm/gemini-1.5-pro-latest.yaml | 9 +++++++++ .../model_providers/google/llm/gemini-pro-vision.yaml | 9 +++++++++ .../model_providers/google/llm/gemini-pro.yaml | 9 +++++++++ 8 files changed, 72 insertions(+) diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0827.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0827.yaml index bbc697e934e055..4e0209890a336a 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0827.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0827.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-exp-0827.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-exp-0827.yaml index c5695e5dda8eb0..faabc5e4d13a73 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-exp-0827.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-exp-0827.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml index 24b1c5af8a3fd8..6a0344699a78b7 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0801.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0801.yaml index 0a918e0d7b1ac3..97c68f7a18d91e 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0801.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0801.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0827.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0827.yaml index 7452ce46e7dcb6..860e4816a163cc 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0827.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-exp-0827.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml index b3e1ecf3aff379..92cd6b310d07ed 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml @@ -32,6 +32,15 @@ parameter_rules: max: 8192 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-pro-vision.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-pro-vision.yaml index 075e484e469308..2d213d56adb9c7 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-pro-vision.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-pro-vision.yaml @@ -27,6 +27,15 @@ parameter_rules: default: 4096 min: 1 max: 4096 + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-pro.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-pro.yaml index 4e9f59e7da94f0..e2f487c1ee9219 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-pro.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-pro.yaml @@ -31,6 +31,15 @@ parameter_rules: max: 2048 - name: response_format use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false pricing: input: '0.00' output: '0.00' From 3dfbc348e3f53b26a3d385e83dea10bf0ef73736 Mon Sep 17 00:00:00 2001 From: Hash Brown Date: Thu, 26 Sep 2024 19:41:59 +0800 Subject: [PATCH 52/79] feat: improved SVG output UX (#8765) --- .../base/chat/chat/answer/index.tsx | 13 ++ web/app/components/base/markdown.tsx | 119 ++++++++++-------- web/app/components/base/svg-gallery/index.tsx | 18 ++- .../workflow/panel/chat-record/index.tsx | 6 +- 4 files changed, 95 insertions(+), 61 deletions(-) diff --git a/web/app/components/base/chat/chat/answer/index.tsx b/web/app/components/base/chat/chat/answer/index.tsx index 705cd73ddf18d9..8184967edc9820 100644 --- a/web/app/components/base/chat/chat/answer/index.tsx +++ b/web/app/components/base/chat/chat/answer/index.tsx @@ -85,6 +85,19 @@ const Answer: FC = ({ getContentWidth() }, [responding]) + // Recalculate contentWidth when content changes (e.g., SVG preview/source toggle) + useEffect(() => { + if (!containerRef.current) + return + const resizeObserver = new ResizeObserver(() => { + getContentWidth() + }) + resizeObserver.observe(containerRef.current) + return () => { + resizeObserver.disconnect() + } + }, []) + return (
diff --git a/web/app/components/base/markdown.tsx b/web/app/components/base/markdown.tsx index 443ee3410c4de4..39a399cc9f73c9 100644 --- a/web/app/components/base/markdown.tsx +++ b/web/app/components/base/markdown.tsx @@ -116,59 +116,80 @@ const CodeBlock: CodeComponent = memo(({ inline, className, children, ...props } const match = /language-(\w+)/.exec(className || '') const language = match?.[1] const languageShowName = getCorrectCapitalizationLanguageName(language || '') - let chartData = JSON.parse(String('{"title":{"text":"ECharts error - Wrong JSON format."}}').replace(/\n$/, '')) - if (language === 'echarts') { - try { - chartData = JSON.parse(String(children).replace(/\n$/, '')) - } - catch (error) { + const chartData = useMemo(() => { + if (language === 'echarts') { + try { + return JSON.parse(String(children).replace(/\n$/, '')) + } + catch (error) {} } - } + return JSON.parse('{"title":{"text":"ECharts error - Wrong JSON format."}}') + }, [language, children]) - // Use `useMemo` to ensure that `SyntaxHighlighter` only re-renders when necessary - return useMemo(() => { - return (!inline && match) - ? ( -
-
-
{languageShowName}
-
- {language === 'mermaid' && } - -
-
- {(language === 'mermaid' && isSVG) - ? () - : (language === 'echarts' - ? (
) - : (language === 'svg' - ? () - : ( - {String(children).replace(/\n$/, '')} - )))} + const renderCodeContent = useMemo(() => { + const content = String(children).replace(/\n$/, '') + if (language === 'mermaid' && isSVG) { + return + } + else if (language === 'echarts') { + return ( +
+ + +
) - : ({children}) - }, [chartData, children, className, inline, isSVG, language, languageShowName, match, props]) + } + else if (language === 'svg' && isSVG) { + return ( + + + + ) + } + else { + return ( + + {content} + + ) + } + }, [language, match, props, children, chartData, isSVG]) + + if (inline || !match) + return {children} + + return ( +
+
+
{languageShowName}
+
+ {(['mermaid', 'svg']).includes(language!) && } + +
+
+ {renderCodeContent} +
+ ) }) CodeBlock.displayName = 'CodeBlock' diff --git a/web/app/components/base/svg-gallery/index.tsx b/web/app/components/base/svg-gallery/index.tsx index 81e8e876550097..4368df00e9d38d 100644 --- a/web/app/components/base/svg-gallery/index.tsx +++ b/web/app/components/base/svg-gallery/index.tsx @@ -29,7 +29,7 @@ export const SVGRenderer = ({ content }: { content: string }) => { if (svgRef.current) { try { svgRef.current.innerHTML = '' - const draw = SVG().addTo(svgRef.current).size('100%', '100%') + const draw = SVG().addTo(svgRef.current) const parser = new DOMParser() const svgDoc = parser.parseFromString(content, 'image/svg+xml') @@ -40,13 +40,11 @@ export const SVGRenderer = ({ content }: { content: string }) => { const originalWidth = parseInt(svgElement.getAttribute('width') || '400', 10) const originalHeight = parseInt(svgElement.getAttribute('height') || '600', 10) - const scale = Math.min(windowSize.width / originalWidth, windowSize.height / originalHeight, 1) - const scaledWidth = originalWidth * scale - const scaledHeight = originalHeight * scale - draw.size(scaledWidth, scaledHeight) + draw.viewbox(0, 0, originalWidth, originalHeight) + + svgRef.current.style.width = `${Math.min(originalWidth, 298)}px` const rootElement = draw.svg(content) - rootElement.scale(scale) rootElement.click(() => { setImagePreview(svgToDataURL(svgElement as Element)) @@ -54,7 +52,7 @@ export const SVGRenderer = ({ content }: { content: string }) => { } catch (error) { if (svgRef.current) - svgRef.current.innerHTML = 'Error rendering SVG. Wait for the image content to complete.' + svgRef.current.innerHTML = 'Error rendering SVG. Wait for the image content to complete.' } } }, [content, windowSize]) @@ -62,14 +60,14 @@ export const SVGRenderer = ({ content }: { content: string }) => { return ( <>
{imagePreview && ( setImagePreview('')} />)} diff --git a/web/app/components/workflow/panel/chat-record/index.tsx b/web/app/components/workflow/panel/chat-record/index.tsx index 1bcfd6474d7c88..16d2c304a73c5e 100644 --- a/web/app/components/workflow/panel/chat-record/index.tsx +++ b/web/app/components/workflow/panel/chat-record/index.tsx @@ -90,7 +90,7 @@ const ChatRecord = () => { return (
{ supportCitationHitInfo: true, } as any} chatList={chatList} - chatContainerClassName='px-4' + chatContainerClassName='px-3' chatContainerInnerClassName='pt-6 w-full max-w-full mx-auto' chatFooterClassName='px-4 rounded-b-2xl' chatFooterInnerClassName='pb-4 w-full max-w-full mx-auto' @@ -129,6 +129,8 @@ const ChatRecord = () => { noChatInput allToolIcons={{}} showPromptLog + noSpacing + chatAnswerContainerInner='!pr-2' />
From 063474f4083c6e73402c8dd8b9862ec8dabd14b8 Mon Sep 17 00:00:00 2001 From: ice yao Date: Thu, 26 Sep 2024 22:21:01 +0800 Subject: [PATCH 53/79] Add llama3.2 model in fireworks provider (#8809) --- .../llm/llama-v3p2-11b-vision-instruct.yaml | 46 +++++++++++++++++++ .../fireworks/llm/llama-v3p2-1b-instruct.yaml | 46 +++++++++++++++++++ .../fireworks/llm/llama-v3p2-3b-instruct.yaml | 46 +++++++++++++++++++ .../llm/llama-v3p2-90b-vision-instruct.yaml | 46 +++++++++++++++++++ 4 files changed, 184 insertions(+) create mode 100644 api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-11b-vision-instruct.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-1b-instruct.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-3b-instruct.yaml create mode 100644 api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-90b-vision-instruct.yaml diff --git a/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-11b-vision-instruct.yaml b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-11b-vision-instruct.yaml new file mode 100644 index 00000000000000..31415a24fa8b7e --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-11b-vision-instruct.yaml @@ -0,0 +1,46 @@ +model: accounts/fireworks/models/llama-v3p2-11b-vision-instruct +label: + zh_Hans: Llama 3.2 11B Vision Instruct + en_US: Llama 3.2 11B Vision Instruct +model_type: llm +features: + - agent-thought + - tool-call +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + - name: max_tokens + use_template: max_tokens + - name: context_length_exceeded_behavior + default: None + label: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + help: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + type: string + options: + - None + - truncate + - error + - name: response_format + use_template: response_format +pricing: + input: '0.2' + output: '0.2' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-1b-instruct.yaml b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-1b-instruct.yaml new file mode 100644 index 00000000000000..c2fd77d2568d29 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-1b-instruct.yaml @@ -0,0 +1,46 @@ +model: accounts/fireworks/models/llama-v3p2-1b-instruct +label: + zh_Hans: Llama 3.2 1B Instruct + en_US: Llama 3.2 1B Instruct +model_type: llm +features: + - agent-thought + - tool-call +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + - name: max_tokens + use_template: max_tokens + - name: context_length_exceeded_behavior + default: None + label: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + help: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + type: string + options: + - None + - truncate + - error + - name: response_format + use_template: response_format +pricing: + input: '0.1' + output: '0.1' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-3b-instruct.yaml b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-3b-instruct.yaml new file mode 100644 index 00000000000000..4b3c459c7bf2fc --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-3b-instruct.yaml @@ -0,0 +1,46 @@ +model: accounts/fireworks/models/llama-v3p2-3b-instruct +label: + zh_Hans: Llama 3.2 3B Instruct + en_US: Llama 3.2 3B Instruct +model_type: llm +features: + - agent-thought + - tool-call +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + - name: max_tokens + use_template: max_tokens + - name: context_length_exceeded_behavior + default: None + label: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + help: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + type: string + options: + - None + - truncate + - error + - name: response_format + use_template: response_format +pricing: + input: '0.1' + output: '0.1' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-90b-vision-instruct.yaml b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-90b-vision-instruct.yaml new file mode 100644 index 00000000000000..0aece7455d6254 --- /dev/null +++ b/api/core/model_runtime/model_providers/fireworks/llm/llama-v3p2-90b-vision-instruct.yaml @@ -0,0 +1,46 @@ +model: accounts/fireworks/models/llama-v3p2-90b-vision-instruct +label: + zh_Hans: Llama 3.2 90B Vision Instruct + en_US: Llama 3.2 90B Vision Instruct +model_type: llm +features: + - agent-thought + - tool-call +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + - name: max_tokens + use_template: max_tokens + - name: context_length_exceeded_behavior + default: None + label: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + help: + zh_Hans: 上下文长度超出行为 + en_US: Context Length Exceeded Behavior + type: string + options: + - None + - truncate + - error + - name: response_format + use_template: response_format +pricing: + input: '0.9' + output: '0.9' + unit: '0.000001' + currency: USD From ecc951609d4988424bfba58094d6e256ae86144e Mon Sep 17 00:00:00 2001 From: wenmeng zhou Date: Thu, 26 Sep 2024 22:32:33 +0800 Subject: [PATCH 54/79] add more detailed doc for models of qwen series (#8799) Co-authored-by: crazywoola <427733928@qq.com> --- .../model_runtime/model_providers/tongyi/llm/farui-plus.yaml | 1 + .../model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml | 1 + .../model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml | 1 + .../model_providers/tongyi/llm/qwen-coder-turbo.yaml | 1 + .../model_runtime/model_providers/tongyi/llm/qwen-long.yaml | 2 +- .../model_providers/tongyi/llm/qwen-math-plus-0816.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-plus-0919.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-plus-latest.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-plus.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-turbo-0919.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-turbo-latest.yaml | 1 + .../model_providers/tongyi/llm/qwen-math-turbo.yaml | 1 + .../model_providers/tongyi/llm/qwen-max-0107.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-0403.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-0428.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-0919.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-1201.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-latest.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-max-longcontext.yaml | 2 ++ .../model_runtime/model_providers/tongyi/llm/qwen-max.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-0206.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-0624.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-0723.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-0806.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-0919.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-chat.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-plus-latest.yaml | 2 ++ .../model_runtime/model_providers/tongyi/llm/qwen-plus.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-turbo-0206.yaml | 3 +++ .../model_providers/tongyi/llm/qwen-turbo-0624.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-turbo-0919.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-turbo-chat.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-turbo-latest.yaml | 2 ++ .../model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml | 2 ++ .../model_providers/tongyi/llm/qwen-vl-max-0201.yaml | 1 + .../model_providers/tongyi/llm/qwen-vl-max-0809.yaml | 1 + .../model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml | 1 + .../model_providers/tongyi/llm/qwen-vl-plus-0201.yaml | 1 + .../model_providers/tongyi/llm/qwen-vl-plus-0809.yaml | 1 + .../model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml | 1 + .../model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml | 1 + .../model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml | 1 + .../tongyi/text_embedding/text-embedding-v1.yaml | 1 + .../tongyi/text_embedding/text-embedding-v2.yaml | 1 + .../tongyi/text_embedding/text-embedding-v3.yaml | 1 + 54 files changed, 77 insertions(+), 1 deletion(-) diff --git a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml index d0ff44382789b4..34a57d1fc0c9a5 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/farui-plus.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: farui-plus label: en_US: farui-plus diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml index d9792e71ee4547..64a3f331336bc0 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-0919.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-coder-turbo-0919 label: en_US: qwen-coder-turbo-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml index 0b03505c45084d..a4c93f7047ff58 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo-latest.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-coder-turbo-latest label: en_US: qwen-coder-turbo-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml index 2a6c0408530053..ff68faed80810b 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-coder-turbo.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-coder-turbo label: en_US: qwen-coder-turbo diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml index bad7f4f4725c45..c3dbb3616fb961 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-long.yaml @@ -1,4 +1,4 @@ -# model docs: https://help.aliyun.com/zh/model-studio/getting-started/models#27b2b3a15d5c6 +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-long label: en_US: qwen-long diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml index c14aee1e1eb87f..42fe1f68623bc4 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0816.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-plus-0816 label: en_US: qwen-math-plus-0816 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml index 9d74eeca3e8788..9b6567b8cda4d7 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-0919.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-plus-0919 label: en_US: qwen-math-plus-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml index b8601a969a7fae..b2a2393b365fcb 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus-latest.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-plus-latest label: en_US: qwen-math-plus-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml index 4a948be5976492..63f4b7ff0a0879 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-plus.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-plus label: en_US: qwen-math-plus diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml index bffe324a9649c4..4da90eec3eddfd 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-0919.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-turbo-0919 label: en_US: qwen-math-turbo-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml index 0747e966142194..d29f8851dd3909 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo-latest.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-turbo-latest label: en_US: qwen-math-turbo-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml index dffb5557ffb7c2..2a8f7f725e9366 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-math-turbo.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-math-turbo label: en_US: qwen-math-turbo diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml index 8ae159f1bfb13b..ef1841b5173bc5 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0107.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-0107 label: en_US: qwen-max-0107 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml index 93fb37254e4c8e..a2ea5df130f379 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0403.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max-0403, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-0403 label: en_US: qwen-max-0403 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml index a5c9d49609041a..a467665f118a68 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0428.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max-0428, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-0428 label: en_US: qwen-max-0428 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml index e4a6dae63736a3..78661eaea065f2 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-0919.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max-0919, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-0919 label: en_US: qwen-max-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml index 6fae8a7d38f36e..6f4674576b4426 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-1201.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-1201 label: en_US: qwen-max-1201 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml index 8e2096885957eb..8b5f0054733455 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-latest.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-latest label: en_US: qwen-max-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml index 9bc50c73fcc3ee..098494ff95012d 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max-longcontext.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max-longcontext label: en_US: qwen-max-longcontext diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml index c6a64dc507497e..9d0d3f8db39c23 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-max.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-max, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#cf6cc4aa2aokf) model: qwen-max label: en_US: qwen-max diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml index 430599300bccb9..0b1a6f81df80c0 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0206.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-0206, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-0206 label: en_US: qwen-plus-0206 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml index 906995d2b95c86..7706005bb535cd 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0624.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-0624, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-0624 label: en_US: qwen-plus-0624 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml index b33e725dd08c39..348276fc08f98c 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0723.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-0723, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-0723 label: en_US: qwen-plus-0723 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml index bb394fad814557..29f125135eaa3f 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0806.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-0806, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-0806 label: en_US: qwen-plus-0806 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml index 118e304a9723a3..905fa1e1028bbf 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-0919.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-0919, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-0919 label: en_US: qwen-plus-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml index 761312bc381236..c7a3549727ce8e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-chat.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-chat label: en_US: qwen-plus-chat diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml index 430872fb31a442..608f52c2964ea3 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus-latest.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus-latest, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus-latest label: en_US: qwen-plus-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml index f3fce30209e48f..9089e57255bb70 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-plus.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-plus, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#bb0ffee88bwnk) model: qwen-plus label: en_US: qwen-plus diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml index 2628d824fe8242..7ee0d44f2f2834 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0206.yaml @@ -1,3 +1,6 @@ +# this model corresponds to qwen-turbo-0206, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) + model: qwen-turbo-0206 label: en_US: qwen-turbo-0206 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml index 8097459bf09061..20a3f7eb6460f3 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0624.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-turbo-0624, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) model: qwen-turbo-0624 label: en_US: qwen-turbo-0624 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml index e43beeb195aec9..ba73dec3631fb5 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-0919.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-turbo-0919, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) model: qwen-turbo-0919 label: en_US: qwen-turbo-0919 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml index c30cb7ca100968..d785b7fe857878 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-chat.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-turbo, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) model: qwen-turbo-chat label: en_US: qwen-turbo-chat diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml index e443d6888b7ef6..fe38a4283c2d1e 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo-latest.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-turbo-latest, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) model: qwen-turbo-latest label: en_US: qwen-turbo-latest diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml index 33f05967c2a462..215c9ec5fc96aa 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-turbo.yaml @@ -1,3 +1,5 @@ +# this model corresponds to qwen-turbo, for more details +# please refer to (https://help.aliyun.com/zh/model-studio/getting-started/models#ff492e2c10lub) model: qwen-turbo label: en_US: qwen-turbo diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0201.yaml index 63b6074d0d7c1c..d80168ffc3fb55 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0201.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-max-0201 label: en_US: qwen-vl-max-0201 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml index fd20377002042a..50e10226a5f5c4 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max-0809.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-max-0809 label: en_US: qwen-vl-max-0809 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml index 31a9fb51bbbb54..21b127f56c47d9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-max.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-max label: en_US: qwen-vl-max diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml index 5f90cf48bc999a..03cb039d15a7dd 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0201.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-plus-0201 label: en_US: qwen-vl-plus-0201 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml index 97820c0f3a1a21..67b2b2ebddc616 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus-0809.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-plus-0809 label: en_US: qwen-vl-plus-0809 diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml index 6af36cd6f3d630..f55764c6c05500 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen-vl-plus.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen-vl-plus label: en_US: qwen-vl-plus diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml index 158e2c7ee16e5a..ea157f42ded914 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-1.5b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2-math-1.5b-instruct label: en_US: qwen2-math-1.5b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml index e26a6923d12f21..37052a923317d9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-72b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2-math-72b-instruct label: en_US: qwen2-math-72b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml index 589119b26ea3c1..e182f1c27f7ea9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2-math-7b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2-math-7b-instruct label: en_US: qwen2-math-7b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml index dd608fbf7675af..9e75ccc1f210d9 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-0.5b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-0.5b-instruct label: en_US: qwen2.5-0.5b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml index 08237b3958b327..67c9d312432af7 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-1.5b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-1.5b-instruct label: en_US: qwen2.5-1.5b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml index 640b01970351ba..2a38be921cf3fd 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-14b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-14b-instruct label: en_US: qwen2.5-14b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml index 3a90ca753273df..e6e4fbf97808be 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-32b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-32b-instruct label: en_US: qwen2.5-32b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml index b79755eb9bd844..8f250379a788ab 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-3b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-3b-instruct label: en_US: qwen2.5-3b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml index e9dd51a3416243..bb3cdd6141f1ea 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-72b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-72b-instruct label: en_US: qwen2.5-72b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml index 04f26cf5fe12b6..fdcd3d42757edb 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-7b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-7b-instruct label: en_US: qwen2.5-7b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml index 04f26cf5fe12b6..fdcd3d42757edb 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models model: qwen2.5-7b-instruct label: en_US: qwen2.5-7b-instruct diff --git a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v1.yaml b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v1.yaml index f4303c53d38b80..52e35d8b50afd8 100644 --- a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v1.yaml +++ b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v1.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models#3383780daf8hw model: text-embedding-v1 model_type: text-embedding model_properties: diff --git a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v2.yaml b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v2.yaml index f6be3544ed8f65..5bb6a8f4243d53 100644 --- a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v2.yaml +++ b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v2.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models#3383780daf8hw model: text-embedding-v2 model_type: text-embedding model_properties: diff --git a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v3.yaml b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v3.yaml index 171a379ee23f77..d8af0e2b63565d 100644 --- a/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v3.yaml +++ b/api/core/model_runtime/model_providers/tongyi/text_embedding/text-embedding-v3.yaml @@ -1,3 +1,4 @@ +# for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models#3383780daf8hw model: text-embedding-v3 model_type: text-embedding model_properties: From e5efd09ebbd3e327efdc4c96075a4723c17927a0 Mon Sep 17 00:00:00 2001 From: CXwudi Date: Thu, 26 Sep 2024 21:14:33 -0400 Subject: [PATCH 55/79] chore: massive update of the Gemini models based on latest documentation (#8822) --- .../google/llm/gemini-1.5-flash-001.yaml | 48 +++++++++++++++++++ .../google/llm/gemini-1.5-flash-002.yaml | 48 +++++++++++++++++++ .../llm/gemini-1.5-flash-8b-exp-0924.yaml | 48 +++++++++++++++++++ .../google/llm/gemini-1.5-flash-latest.yaml | 2 +- .../google/llm/gemini-1.5-flash.yaml | 48 +++++++++++++++++++ .../google/llm/gemini-1.5-pro-001.yaml | 48 +++++++++++++++++++ .../google/llm/gemini-1.5-pro-002.yaml | 48 +++++++++++++++++++ .../google/llm/gemini-1.5-pro-latest.yaml | 2 +- .../google/llm/gemini-1.5-pro.yaml | 48 +++++++++++++++++++ 9 files changed, 338 insertions(+), 2 deletions(-) create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-001.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-002.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0924.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-001.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-002.yaml create mode 100644 api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro.yaml diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-001.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-001.yaml new file mode 100644 index 00000000000000..d84e9937e0b661 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-001.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-flash-001 +label: + en_US: Gemini 1.5 Flash 001 +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-002.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-002.yaml new file mode 100644 index 00000000000000..2ff70564b2a951 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-002.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-flash-002 +label: + en_US: Gemini 1.5 Flash 002 +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0924.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0924.yaml new file mode 100644 index 00000000000000..2aea8149f4c794 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-8b-exp-0924.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-flash-8b-exp-0924 +label: + en_US: Gemini 1.5 Flash 8B 0924 +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml index 6a0344699a78b7..a22fcca9419b91 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash-latest.yaml @@ -1,6 +1,6 @@ model: gemini-1.5-flash-latest label: - en_US: Gemini 1.5 Flash + en_US: Gemini 1.5 Flash Latest model_type: llm features: - agent-thought diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash.yaml new file mode 100644 index 00000000000000..dfd55c3a949c97 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-flash.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-flash +label: + en_US: Gemini 1.5 Flash +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 1048576 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-001.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-001.yaml new file mode 100644 index 00000000000000..a1feff171d48c2 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-001.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-pro-001 +label: + en_US: Gemini 1.5 Pro 001 +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 2097152 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-002.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-002.yaml new file mode 100644 index 00000000000000..9ae07a06c5118b --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-002.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-pro-002 +label: + en_US: Gemini 1.5 Pro 002 +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 2097152 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml index 92cd6b310d07ed..d1bf7d269de765 100644 --- a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro-latest.yaml @@ -1,6 +1,6 @@ model: gemini-1.5-pro-latest label: - en_US: Gemini 1.5 Pro + en_US: Gemini 1.5 Pro Latest model_type: llm features: - agent-thought diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro.yaml new file mode 100644 index 00000000000000..bdd70b34a21dc5 --- /dev/null +++ b/api/core/model_runtime/model_providers/google/llm/gemini-1.5-pro.yaml @@ -0,0 +1,48 @@ +model: gemini-1.5-pro +label: + en_US: Gemini 1.5 Pro +model_type: llm +features: + - agent-thought + - vision + - tool-call + - stream-tool-call +model_properties: + mode: chat + context_size: 2097152 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: top_k + label: + zh_Hans: 取样数量 + en_US: Top k + type: int + help: + zh_Hans: 仅从每个后续标记的前 K 个选项中采样。 + en_US: Only sample from the top K options for each subsequent token. + required: false + - name: max_tokens_to_sample + use_template: max_tokens + required: true + default: 8192 + min: 1 + max: 8192 + - name: response_format + use_template: response_format + - name: stream + label: + zh_Hans: 流式输出 + en_US: Stream + type: boolean + help: + zh_Hans: 流式输出允许模型在生成文本的过程中逐步返回结果，而不是一次性生成全部结果后再返回。 + en_US: Streaming output allows the model to return results incrementally as it generates text, rather than generating all the results at once. + default: false +pricing: + input: '0.00' + output: '0.00' + unit: '0.000001' + currency: USD From a36117e12d06707e63de5b8dc51d97bf060b4e7c Mon Sep 17 00:00:00 2001 From: Shai Perednik Date: Thu, 26 Sep 2024 21:15:33 -0400 Subject: [PATCH 56/79] Updated the YouTube channel to Dify's (#8817) --- api/core/tools/provider/builtin/youtube/youtube.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/api/core/tools/provider/builtin/youtube/youtube.py b/api/core/tools/provider/builtin/youtube/youtube.py index aad876491c85dc..07e430bcbf27e1 100644 --- a/api/core/tools/provider/builtin/youtube/youtube.py +++ b/api/core/tools/provider/builtin/youtube/youtube.py @@ -13,7 +13,7 @@ def _validate_credentials(self, credentials: dict) -> None: ).invoke( user_id="", tool_parameters={ - "channel": "TOKYO GIRLS COLLECTION", + "channel": "UC2JZCsZSOudXA08cMMRCL9g", "start_date": "2020-01-01", "end_date": "2024-12-31", }, From 6fbaabc1bc4b8b7830d55bac7609afdd703242f9 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Fri, 27 Sep 2024 11:13:29 +0800 Subject: [PATCH 57/79] feat: add pgvecto-rs and analyticdb in docker/.env.example (#8823) --- docker/.env.example | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) diff --git a/docker/.env.example b/docker/.env.example index f7479791cea320..d43c3edc7e722d 100644 --- a/docker/.env.example +++ b/docker/.env.example @@ -346,7 +346,7 @@ VOLCENGINE_TOS_REGION=your-region # ------------------------------ # The type of vector store to use. -# Supported values are `weaviate`, `qdrant`, `milvus`, `myscale`, `relyt`, `pgvector`, `chroma`, `opensearch`, `tidb_vector`, `oracle`, `tencent`, `elasticsearch`. +# Supported values are `weaviate`, `qdrant`, `milvus`, `myscale`, `relyt`, `pgvector`, `pgvecto-rs`, ``chroma`, `opensearch`, `tidb_vector`, `oracle`, `tencent`, `elasticsearch`, `analyticdb`. VECTOR_STORE=weaviate # The Weaviate endpoint URL. Only available when VECTOR_STORE is `weaviate`. @@ -385,13 +385,30 @@ MYSCALE_PASSWORD= MYSCALE_DATABASE=dify MYSCALE_FTS_PARAMS= -# pgvector configurations, only available when VECTOR_STORE is `pgvecto-rs or pgvector` +# pgvector configurations, only available when VECTOR_STORE is `pgvector` PGVECTOR_HOST=pgvector PGVECTOR_PORT=5432 PGVECTOR_USER=postgres PGVECTOR_PASSWORD=difyai123456 PGVECTOR_DATABASE=dify +# pgvecto-rs configurations, only available when VECTOR_STORE is `pgvecto-rs` +PGVECTO_RS_HOST=pgvecto-rs +PGVECTO_RS_PORT=5432 +PGVECTO_RS_USER=postgres +PGVECTO_RS_PASSWORD=difyai123456 +PGVECTO_RS_DATABASE=dify + +# analyticdb configurations, only available when VECTOR_STORE is `analyticdb` +ANALYTICDB_KEY_ID=your-ak +ANALYTICDB_KEY_SECRET=your-sk +ANALYTICDB_REGION_ID=cn-hangzhou +ANALYTICDB_INSTANCE_ID=gp-ab123456 +ANALYTICDB_ACCOUNT=testaccount +ANALYTICDB_PASSWORD=testpassword +ANALYTICDB_NAMESPACE=dify +ANALYTICDB_NAMESPACE_PASSWORD=difypassword + # TiDB vector configurations, only available when VECTOR_STORE is `tidb` TIDB_VECTOR_HOST=tidb TIDB_VECTOR_PORT=4000 From d6b9587a979573fc22b380a13b80770883fc08e0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=9D=9E=E6=B3=95=E6=93=8D=E4=BD=9C?= Date: Fri, 27 Sep 2024 11:13:40 +0800 Subject: [PATCH 58/79] fix: close log status option raise error (#8826) --- web/app/components/app/log/filter.tsx | 2 ++ web/app/components/app/workflow-log/filter.tsx | 2 ++ 2 files changed, 4 insertions(+) diff --git a/web/app/components/app/log/filter.tsx b/web/app/components/app/log/filter.tsx index 0552b44d16b3cb..b8d7ca5a36b783 100644 --- a/web/app/components/app/log/filter.tsx +++ b/web/app/components/app/log/filter.tsx @@ -55,6 +55,8 @@ const Filter: FC = ({ isChatMode, appId, queryParams, setQueryPara className='!w-[300px]' onSelect={ (item) => { + if (!item.value) + return setQueryParams({ ...queryParams, annotation_status: item.value as string }) } } diff --git a/web/app/components/app/workflow-log/filter.tsx b/web/app/components/app/workflow-log/filter.tsx index d239c39d2cf618..58b5252c07446b 100644 --- a/web/app/components/app/workflow-log/filter.tsx +++ b/web/app/components/app/workflow-log/filter.tsx @@ -23,6 +23,8 @@ const Filter: FC = ({ queryParams, setQueryParams }: IFilterProps) className='!min-w-[100px]' onSelect={ (item) => { + if (!item.value) + return setQueryParams({ ...queryParams, status: item.value as string }) } } From 4c1063e1c589d239df7287350464c4a7d729cab2 Mon Sep 17 00:00:00 2001 From: 8bitpd <51897400+lpdink@users.noreply.github.com> Date: Fri, 27 Sep 2024 12:05:21 +0800 Subject: [PATCH 59/79] fix: AnalyticdbVector retrieval scores (#8803) --- .../vdb/analyticdb/analyticdb_vector.py | 19 ++++++------------- 1 file changed, 6 insertions(+), 13 deletions(-) diff --git a/api/core/rag/datasource/vdb/analyticdb/analyticdb_vector.py b/api/core/rag/datasource/vdb/analyticdb/analyticdb_vector.py index 612542dab1df11..6dcd98dcfd14f6 100644 --- a/api/core/rag/datasource/vdb/analyticdb/analyticdb_vector.py +++ b/api/core/rag/datasource/vdb/analyticdb/analyticdb_vector.py @@ -40,19 +40,8 @@ def to_analyticdb_client_params(self): class AnalyticdbVector(BaseVector): - _instance = None - _init = False - - def __new__(cls, *args, **kwargs): - if cls._instance is None: - cls._instance = super().__new__(cls) - return cls._instance - def __init__(self, collection_name: str, config: AnalyticdbConfig): - # collection_name must be updated every time self._collection_name = collection_name.lower() - if AnalyticdbVector._init: - return try: from alibabacloud_gpdb20160503.client import Client from alibabacloud_tea_openapi import models as open_api_models @@ -62,7 +51,6 @@ def __init__(self, collection_name: str, config: AnalyticdbConfig): self._client_config = open_api_models.Config(user_agent="dify", **config.to_analyticdb_client_params()) self._client = Client(self._client_config) self._initialize() - AnalyticdbVector._init = True def _initialize(self) -> None: cache_key = f"vector_indexing_{self.config.instance_id}" @@ -257,11 +245,14 @@ def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Doc documents = [] for match in response.body.matches.match: if match.score > score_threshold: + metadata = json.loads(match.metadata.get("metadata_")) + metadata["score"] = match.score doc = Document( page_content=match.metadata.get("page_content"), - metadata=json.loads(match.metadata.get("metadata_")), + metadata=metadata, ) documents.append(doc) + documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True) return documents def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]: @@ -286,12 +277,14 @@ def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]: for match in response.body.matches.match: if match.score > score_threshold: metadata = json.loads(match.metadata.get("metadata_")) + metadata["score"] = match.score doc = Document( page_content=match.metadata.get("page_content"), vector=match.metadata.get("vector"), metadata=metadata, ) documents.append(doc) + documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True) return documents def delete(self) -> None: From 29275c7447c4afbe2a92fe23e31fbebf79b1541e Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Fri, 27 Sep 2024 12:11:56 +0800 Subject: [PATCH 60/79] feat: deprecate mistral model for siliconflow (#8828) --- .../siliconflow/llm/mistral-7b-instruct-v0.2.yaml | 1 + .../siliconflow/llm/mistral-8x7b-instruct-v0.1.yaml | 1 + 2 files changed, 2 insertions(+) diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/mistral-7b-instruct-v0.2.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/mistral-7b-instruct-v0.2.yaml index 27664eab6c817a..89fb153ba09ac9 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/mistral-7b-instruct-v0.2.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/mistral-7b-instruct-v0.2.yaml @@ -28,3 +28,4 @@ pricing: output: '0' unit: '0.000001' currency: RMB +deprecated: true diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/mistral-8x7b-instruct-v0.1.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/mistral-8x7b-instruct-v0.1.yaml index fd7aada42848aa..2785e7496fb060 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/mistral-8x7b-instruct-v0.1.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/mistral-8x7b-instruct-v0.1.yaml @@ -28,3 +28,4 @@ pricing: output: '1.26' unit: '0.000001' currency: RMB +deprecated: true From bb781764b829b83471a0546cbced7c842e826231 Mon Sep 17 00:00:00 2001 From: HowardChan Date: Fri, 27 Sep 2024 12:13:00 +0800 Subject: [PATCH 61/79] Add Llama3.2 models in Groq provider (#8831) --- .../groq/llm/llama-3.2-11b-text-preview.yaml | 25 +++++++++++++++++++ .../groq/llm/llama-3.2-1b-preview.yaml | 25 +++++++++++++++++++ .../groq/llm/llama-3.2-3b-preview.yaml | 25 +++++++++++++++++++ .../groq/llm/llama-3.2-90b-text-preview.yaml | 25 +++++++++++++++++++ 4 files changed, 100 insertions(+) create mode 100644 api/core/model_runtime/model_providers/groq/llm/llama-3.2-11b-text-preview.yaml create mode 100644 api/core/model_runtime/model_providers/groq/llm/llama-3.2-1b-preview.yaml create mode 100644 api/core/model_runtime/model_providers/groq/llm/llama-3.2-3b-preview.yaml create mode 100644 api/core/model_runtime/model_providers/groq/llm/llama-3.2-90b-text-preview.yaml diff --git a/api/core/model_runtime/model_providers/groq/llm/llama-3.2-11b-text-preview.yaml b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-11b-text-preview.yaml new file mode 100644 index 00000000000000..019d45372361d3 --- /dev/null +++ b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-11b-text-preview.yaml @@ -0,0 +1,25 @@ +model: llama-3.2-11b-text-preview +label: + zh_Hans: Llama 3.2 11B Text (Preview) + en_US: Llama 3.2 11B Text (Preview) +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: max_tokens + use_template: max_tokens + default: 512 + min: 1 + max: 8192 +pricing: + input: '0.05' + output: '0.1' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/groq/llm/llama-3.2-1b-preview.yaml b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-1b-preview.yaml new file mode 100644 index 00000000000000..a44e4ff508eb82 --- /dev/null +++ b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-1b-preview.yaml @@ -0,0 +1,25 @@ +model: llama-3.2-1b-preview +label: + zh_Hans: Llama 3.2 1B Text (Preview) + en_US: Llama 3.2 1B Text (Preview) +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: max_tokens + use_template: max_tokens + default: 512 + min: 1 + max: 8192 +pricing: + input: '0.05' + output: '0.1' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/groq/llm/llama-3.2-3b-preview.yaml b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-3b-preview.yaml new file mode 100644 index 00000000000000..f2fdd0a05e027a --- /dev/null +++ b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-3b-preview.yaml @@ -0,0 +1,25 @@ +model: llama-3.2-3b-preview +label: + zh_Hans: Llama 3.2 3B Text (Preview) + en_US: Llama 3.2 3B Text (Preview) +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: max_tokens + use_template: max_tokens + default: 512 + min: 1 + max: 8192 +pricing: + input: '0.05' + output: '0.1' + unit: '0.000001' + currency: USD diff --git a/api/core/model_runtime/model_providers/groq/llm/llama-3.2-90b-text-preview.yaml b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-90b-text-preview.yaml new file mode 100644 index 00000000000000..3b34e7c07996bd --- /dev/null +++ b/api/core/model_runtime/model_providers/groq/llm/llama-3.2-90b-text-preview.yaml @@ -0,0 +1,25 @@ +model: llama-3.2-90b-text-preview +label: + zh_Hans: Llama 3.2 90B Text (Preview) + en_US: Llama 3.2 90B Text (Preview) +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + - name: top_p + use_template: top_p + - name: max_tokens + use_template: max_tokens + default: 512 + min: 1 + max: 8192 +pricing: + input: '0.05' + output: '0.1' + unit: '0.000001' + currency: USD From 0603359e2dea807f8cac2cfe5ac68aeb081627e4 Mon Sep 17 00:00:00 2001 From: CXwudi Date: Fri, 27 Sep 2024 01:49:03 -0400 Subject: [PATCH 62/79] fix: delete harm catalog settings for gemini (#8829) --- .../model_providers/google/llm/llm.py | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) diff --git a/api/core/model_runtime/model_providers/google/llm/llm.py b/api/core/model_runtime/model_providers/google/llm/llm.py index 3fc6787a444e41..e686ad08d9d355 100644 --- a/api/core/model_runtime/model_providers/google/llm/llm.py +++ b/api/core/model_runtime/model_providers/google/llm/llm.py @@ -9,8 +9,8 @@ import google.generativeai as genai import requests from google.api_core import exceptions -from google.generativeai import client -from google.generativeai.types import ContentType, GenerateContentResponse, HarmBlockThreshold, HarmCategory +from google.generativeai.client import _ClientManager +from google.generativeai.types import ContentType, GenerateContentResponse from google.generativeai.types.content_types import to_part from PIL import Image @@ -200,24 +200,16 @@ def _generate( history.append(content) # Create a new ClientManager with tenant's API key - new_client_manager = client._ClientManager() + new_client_manager = _ClientManager() new_client_manager.configure(api_key=credentials["google_api_key"]) new_custom_client = new_client_manager.make_client("generative") google_model._client = new_custom_client - safety_settings = { - HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE, - HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE, - } - response = google_model.generate_content( contents=history, generation_config=genai.types.GenerationConfig(**config_kwargs), stream=stream, - safety_settings=safety_settings, tools=self._convert_tools_to_glm_tool(tools) if tools else None, request_options={"timeout": 600}, ) From c828a5dfdf6c4a6caf7230ed879ae688608eb373 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E8=B5=B0=E5=9C=A8=E4=BF=AE=E8=A1=8C=E7=9A=84=E5=A4=A7?= =?UTF-8?q?=E8=A1=97=E4=B8=8A?= Date: Fri, 27 Sep 2024 17:31:45 +0800 Subject: [PATCH 63/79] feat(Tools): add feishu tools (#8800) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-authored-by: 黎斌 --- api/core/tools/provider/_position.yaml | 4 + .../builtin/feishu_calendar/_assets/icon.png | Bin 0 -> 5488 bytes .../feishu_calendar/feishu_calendar.py | 7 + .../feishu_calendar/feishu_calendar.yaml | 36 ++ .../tools/add_event_attendees.py | 20 + .../tools/add_event_attendees.yaml | 54 +++ .../feishu_calendar/tools/create_event.py | 26 ++ .../feishu_calendar/tools/create_event.yaml | 119 ++++++ .../feishu_calendar/tools/delete_event.py | 19 + .../feishu_calendar/tools/delete_event.yaml | 38 ++ .../tools/get_primary_calendar.py | 18 + .../tools/get_primary_calendar.yaml | 37 ++ .../feishu_calendar/tools/list_events.py | 21 + .../feishu_calendar/tools/list_events.yaml | 62 +++ .../feishu_calendar/tools/search_events.py | 23 + .../feishu_calendar/tools/search_events.yaml | 100 +++++ .../feishu_calendar/tools/update_event.py | 24 ++ .../feishu_calendar/tools/update_event.yaml | 100 +++++ .../feishu_document/feishu_document.py | 12 +- .../feishu_document/feishu_document.yaml | 8 +- .../tools/create_document.yaml | 9 +- .../tools/get_document_content.py | 4 +- .../tools/get_document_content.yaml | 35 +- .../tools/list_document_blocks.py | 5 +- .../tools/list_document_blocks.yaml | 4 +- .../feishu_document/tools/write_document.py | 2 +- .../feishu_document/tools/write_document.yaml | 30 +- .../builtin/feishu_message/feishu_message.py | 12 +- .../feishu_message/feishu_message.yaml | 8 +- .../feishu_message/tools/get_chat_messages.py | 23 + .../tools/get_chat_messages.yaml | 96 +++++ .../tools/get_thread_messages.py | 21 + .../tools/get_thread_messages.yaml | 72 ++++ .../tools/send_bot_message.yaml | 114 +++-- .../tools/send_webhook_message.yaml | 46 +- .../feishu_spreadsheet/_assets/icon.png | Bin 0 -> 5417 bytes .../feishu_spreadsheet/feishu_spreadsheet.py | 7 + .../feishu_spreadsheet.yaml | 36 ++ .../feishu_spreadsheet/tools/add_cols.py | 22 + .../feishu_spreadsheet/tools/add_cols.yaml | 72 ++++ .../feishu_spreadsheet/tools/add_rows.py | 22 + .../feishu_spreadsheet/tools/add_rows.yaml | 72 ++++ .../tools/create_spreadsheet.py | 19 + .../tools/create_spreadsheet.yaml | 35 ++ .../tools/get_spreadsheet.py | 19 + .../tools/get_spreadsheet.yaml | 49 +++ .../tools/list_spreadsheet_sheets.py | 18 + .../tools/list_spreadsheet_sheets.yaml | 23 + .../feishu_spreadsheet/tools/read_cols.py | 23 + .../feishu_spreadsheet/tools/read_cols.yaml | 97 +++++ .../feishu_spreadsheet/tools/read_rows.py | 23 + .../feishu_spreadsheet/tools/read_rows.yaml | 97 +++++ .../feishu_spreadsheet/tools/read_table.py | 23 + .../feishu_spreadsheet/tools/read_table.yaml | 122 ++++++ .../builtin/feishu_task/_assets/icon.png | Bin 0 -> 3917 bytes .../builtin/feishu_task/feishu_task.py | 7 + .../builtin/feishu_task/feishu_task.yaml | 36 ++ .../builtin/feishu_task/tools/add_members.py | 20 + .../feishu_task/tools/add_members.yaml | 58 +++ .../builtin/feishu_task/tools/create_task.py | 22 + .../feishu_task/tools/create_task.yaml | 74 ++++ .../builtin/feishu_task/tools/delete_task.py | 18 + .../feishu_task/tools/delete_task.yaml | 24 ++ .../builtin/feishu_task/tools/update_task.py | 23 + .../feishu_task/tools/update_task.yaml | 89 ++++ .../builtin/feishu_wiki/_assets/icon.png | Bin 0 -> 4252 bytes .../builtin/feishu_wiki/feishu_wiki.py | 7 + .../builtin/feishu_wiki/feishu_wiki.yaml | 36 ++ .../feishu_wiki/tools/get_wiki_nodes.py | 21 + .../feishu_wiki/tools/get_wiki_nodes.yaml | 63 +++ api/core/tools/utils/feishu_api_utils.py | 401 +++++++++++++++++- 71 files changed, 2755 insertions(+), 132 deletions(-) create mode 100644 api/core/tools/provider/builtin/feishu_calendar/_assets/icon.png create mode 100644 api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/create_event.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/create_event.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/list_events.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/list_events.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/search_events.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/search_events.yaml create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/update_event.py create mode 100644 api/core/tools/provider/builtin/feishu_calendar/tools/update_event.yaml create mode 100644 api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.py create mode 100644 api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.yaml create mode 100644 api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.py create mode 100644 api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/_assets/icon.png create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/feishu_spreadsheet.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/feishu_spreadsheet.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.yaml create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.py create mode 100644 api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.yaml create mode 100644 api/core/tools/provider/builtin/feishu_task/_assets/icon.png create mode 100644 api/core/tools/provider/builtin/feishu_task/feishu_task.py create mode 100644 api/core/tools/provider/builtin/feishu_task/feishu_task.yaml create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/add_members.py create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/add_members.yaml create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/create_task.py create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/create_task.yaml create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/delete_task.py create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/delete_task.yaml create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/update_task.py create mode 100644 api/core/tools/provider/builtin/feishu_task/tools/update_task.yaml create mode 100644 api/core/tools/provider/builtin/feishu_wiki/_assets/icon.png create mode 100644 api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.py create mode 100644 api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.yaml create mode 100644 api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.py create mode 100644 api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.yaml diff --git a/api/core/tools/provider/_position.yaml b/api/core/tools/provider/_position.yaml index 40c3356116770b..e0a8e7511e59ba 100644 --- a/api/core/tools/provider/_position.yaml +++ b/api/core/tools/provider/_position.yaml @@ -34,5 +34,9 @@ - feishu_base - feishu_document - feishu_message +- feishu_wiki +- feishu_task +- feishu_calendar +- feishu_spreadsheet - slack - tianditu diff --git a/api/core/tools/provider/builtin/feishu_calendar/_assets/icon.png b/api/core/tools/provider/builtin/feishu_calendar/_assets/icon.png new file mode 100644 index 0000000000000000000000000000000000000000..2a934747a98c6680065941bcd31d2400da1eaf23 GIT binary patch literal 5488 zcmd^@S6CClyT+3cB}hx?J%AJ?fE0mHMQMV8&_R?YO?nelY7hlNk>0Cxkg60F3{fc} zH9!DCT0jJ80tr>&_S<*on|*iYn;+KLNRROn&m{l=z^H#)+mzb= z`Og8lNIh#4qnG3;S|irW=|=l}|xX42D3f ziFX7+o4)i%ZwLZ_Wy~1N7x0X<1J061*y29gfDyJ2fUTeVMQPHq7@KLU>F~{ov1emn zTTR^lX{pNZoq%B_m`)+II}6z;t*n3|MoU1jS;dIL*?`4jeJEOe)(8nc{{SYk($W#)H9s-F>-9L)q>vds%qH zkcrltA7vCoxnI4*96v@UvWzE-7Cx7P{&@x%&*6D0Sj%Vv92rJtf5apSYHAh%#><`< z2e{f;;;o92hA{8!WL}P&cvjHhvd7BO#hQoR;KbjEOqS{00syPkr2de>038=`YCj_O zPd5aDqo-cUk32YjA+$q>s&#{_U4>P|VUB0oV{Y#n(#1N^fI)(I^SU~-b7vfzi7wtM z^w6|?o(Z6K3k;+Kp}J2e-ZCfW^Q6Q6wEFUk_k^MfZ-zo4F__Xr;-SE%Op=u@zWCtx zh*)#))OFs6a3XAAOzW2 zaQQ(p)m{RktgoK`-$50pJaB2=3&M;Z< z9qKSBQXoHQsl&j0SDWDz3kDNys17>(|0S{QY(fc-ee#?BF<;wCO^An-3EEzIBkWv} zKsIk^KKf7o0v7 zo)OJwp1u-Fzm8t+m$NYP78%(2mN{nOOaPn7rk52RKXip9_VJ&iy2 z(`YX=^0A&q8oV%!9lT#qgwn1%Cj_27!_4MD;@oHK;g{)Qou*`7VPiE&kMQBhO8^r17_RKNU12a>1wxuj`0t0s!bY)4s! zD?#4%fVSZ!q_)Gvm~T(j8fp+Ga=%Bt#I3m-oc&6^U8Hr;PRv-1eVCti15c z82eYGwqp7}`)hNLUtR=Dj$X5rBizuen46llf69Agt=t;*q;xz|RhMGh9D57TaiOlK zW54w4=tz9_xY_l<09!q@*Q*o7el@U{4iA>_x$Ur8rgLtd1}wROt^d})#PqKdN^EzJ z7h0^03z&m4`l2JG1D0xSRrOoZb~{KblW?80uK9ILNkR>xjR`b&#oVou@Cmm6V#ukF4sbg!;3bqq=X%cB7Tqeui z9NUQemDNQ#3o>0InrU5C`{tK$c6a^jxOY*iE@dc1HVm6dXZbe&l(yRQe9rZ8#aZZ^ zw6z&J(neQ_c6Wlb6zOxXR{3k(gvGzbShj_u7a8vdgN?)Gg|#jWL%d$hipGw#^159) zqu^pbdyA};j?!MfU@>0yNb)JcavNAAk|JC>{Yv@aK_w@vn<`L0BfkXk49muLbB}i?kf}5yCmO@z)g%j_r`dtUNCy{`O2m4Bp};pnCLqyyd=Jo}x_XVBrx^@8}}67`{sD|1|051pQ4fgjz{ z37jS;^1aQzysFD%sxq~W{vU)KfX}hm6@2}C7_E=!#mXw}$z_AVTN^$9LQ=ct=!sU( zuSqmPdWU6%TT(Rkiv}7Jk>bHrt%5g1z{g|ItEvD= z&CY);`g;{Wtx^Mmwlh#Ghh;4r7i%m$_j#>HO{r3ue5tQ-+PnW|wV65V(kEG?>=#a5 zZ^@jEZ+%y9ROh%3mK@ueb=Cg;(Tdg~{7CX>JD=m-e2)k&u!ziJnD5%qZZyAg8W|eHJ=)BJ@en|8qf1zrNdV9JR(N58HEGA z5spvQ>TQu~9oAo>J*ex}I>_7SGSXdv6HEDxX!>EArh?9@3GhfiC31nXHaY5>MU~w% zJ5PZibWvcY>CH680FGhbf8?IUW7<~$VM8QrKLtS?;$d$sO-<8GI6&twn^qBJ8P zRT?$m=--sZ5OJr+$~${ zo&TPY0`_c{1}^{~`_y`&Y(#fLb{N_PA7+LHiH48nhxfhLK&boooRx_b0f2XmTO7U+ z^HPYDdWf4Tcl zsw9+!RONFo0O@N-Ml)v9E%!g&DM}xV?9?Hi{}_9;)_7uN=Km`){jBTj&u=n3AP)WZ z_pmAPdJbrcinuatr-OGK-a0dmu9%We3Gw)<1`iotEL<8s*QJJjZLOhW7OUqFPC=3~ zx}>Li@^{PPLqu^hr#@b3T$Lv#jr=%A0kBt^b8g!<+GZ#)1v#UG7S4^+J3pvXVwU;n zVU=vQrF8D-(d!cmnU-E~ovcUM*s2<}qG!f;MP3FIi0uL7T-PNFY&OiE`r?be5iJRKDTVtwq<2%6lC6V8e-6cGm?YS${34;8=(N+NdW}0mjFw;&v{V8oR>uqO1c-?Anj4^3)C*f zT#==b66e{+N-En+9>_wbg@5Q&h8MJipzxr3eZ7aiqN>h$_8<^0BCi&X^Vp6zq=z;c zexUaXRdJ`}V5T~6SHQ!Z!Pmb#Y)hpVr?|o!ll_&+k-8mGi*DtKHRJ>S>1BO{J8&Zz z>Dr6)#$})#Ze9R00L8&!MJ*v=kbi2KcJk_{w_UFm9|vowb@b1n$i6R<<1Tbk~~DSqDm?*#!Y znZI?X0(UB$X(t!Ti4K-o$1D)@KR?Ih3x9j81xU9{!mnNO+S~0e7;=`|Sh>Msh<06W zY=1LH$!%s{U2t*>R6U%7`uRK>P->BL*D-(UX4}_@mhuTrCO?b7zShW?ES9TOyJ&t@ zqLVDkKKb|c?(*`8kr}0o>6}hqj}FNDD+CLqc1dbx;Cn{QiVvUu`NZsE(d^6$m$43c z1;E~ZRzNM;DuqO9s|)6Ch5hORd)-w80tY~CvhF?h^FHerJ$H+JSi+yG!nv9c7mxck zk~Yb@WdGxDV`i)&r`3Np1>%clUC8-6PM_k#lM9(0D=ndZV)qgAUQS~M!yV)-)CgoO zVcpw%O<;p@iUM?Iu1c#9oRH`*yta{8z^&l~Da4s6eH>3UJId>QHDdTdMz=NLIN2`M zIw8LojX0Ds`}uI=U_|jHkbRoV@j*$`E|gs{=^K~lTXr*`85*9+^dpmZtSu$hjL6Au z-QcE)Pk#F-E!LCB35vn+!ekooGu87X3g(_(LT_=Q|8iB?3;i!p8{KXr%L12MoUo>d zdamz=FNS;v0Bv5PJBn>ta~?hdAfMqYL;O&`%Lc2S0pm?dxY*oUxJnc4qcTk)hh?%8 zi|2*s{kj_+=dtGKE{2Ex2vqB9ah1sPUq4AP7?%*R#GgGr(S!J`{xMTT*y93;o#3P4 zbi@T_;J_nB$P<^B)MEH{WW}UD=lQOf#ll?1*=B`NK*TU7#99mi^$UC%1cS4vUu-!x zBHIV)Y~(kklJf(kp8S-$#gLY@^T-w>Q-y-P|*7Qy_VPhFZNW+)qF9L zU+U-9f2zC$pYe{z4}%GR2Hr-R95YC{S0J|8cHP{n4?5CwN4PlxKI6E}9k_dgfE#o? zhu1zHOgbx=3V>|BhR{85bhdhfYKm}y?Qb^lqqqf@a({}yhHT>bZ()T#Q z<9k;)CH$`B>K52hH*p_Ph-a5xP?3p2C9Eo&3P`Sv$Cum^q7s%g_wIF3>Q6Y(B3x}u z-3PvM=J5~F2Y?Rb>LC(rAn5<11lz*q+0{n*B&mUIrvVMm)e;^z(~*3Y=mSCMC7`Ro zj`C+-)G|Bl&zESzdfdZ~8u^FJpjS|b%4Kyu(v;zU?nJ|NLg`UH1rToohOL|U0zhDr z-HtL(F%@VCFXk#ejH4$Qnz@Y%fI&j|T8_pI6@&KwC^iL?2vHBy*eJOw&kqC_)q$%4 zhcJPRI5QM}oVIk?fFM9uMTHSzVkJ5AJ1Q(YDRCw!JeTgBavDV{ppfl}{^*$E@5h3gQ z7ZH`L$n)OW5Wg>M2`^*0L%BHZ9}ft8&$7>PkJKuMZRL`Q88-@2Mk2H@nWTFPt5S+F zLsHzguCWJ@m*dt_$Y?9MPR9%2ZivX+A@%WHx}bu}?mB6wufL%mOccc8b=|TB+Q{GA zZhnp!1zYQl3})ZmWtz2Umtudiq(x~`fywQMn0h!I_(C6z^7u&~X}5H%u_J;?jT1)h z)JcCcc(qkZcuVYIsjsrR4uv(klxqC5pVaib0cm-rseQuh8gxy%x4l-Ld@rBX+@0j! zff+|)Lp1?l9x4YU$o~%vPtYMYop$Ob0Hh25UxJ56k@&OI*n~FT87e{r=<671H{5b~ F_Fv-s=P3XH literal 0 HcmV?d00001 diff --git a/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.py b/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.py new file mode 100644 index 00000000000000..a46a9fa9e80cab --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.py @@ -0,0 +1,7 @@ +from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController +from core.tools.utils.feishu_api_utils import auth + + +class FeishuCalendarProvider(BuiltinToolProviderController): + def _validate_credentials(self, credentials: dict) -> None: + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.yaml b/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.yaml new file mode 100644 index 00000000000000..db5bab5c1081d9 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/feishu_calendar.yaml @@ -0,0 +1,36 @@ +identity: + author: Doug Lea + name: feishu_calendar + label: + en_US: Feishu Calendar + zh_Hans: 飞书日历 + description: + en_US: | + Feishu calendar, requires the following permissions: calendar:calendar:read、calendar:calendar、contact:user.id:readonly. + zh_Hans: | + 飞书日历，需要开通以下权限: calendar:calendar:read、calendar:calendar、contact:user.id:readonly。 + icon: icon.png + tags: + - social + - productivity +credentials_for_provider: + app_id: + type: text-input + required: true + label: + en_US: APP ID + placeholder: + en_US: Please input your feishu app id + zh_Hans: 请输入你的飞书 app id + help: + en_US: Get your app_id and app_secret from Feishu + zh_Hans: 从飞书获取您的 app_id 和 app_secret + url: https://open.larkoffice.com/app + app_secret: + type: secret-input + required: true + label: + en_US: APP Secret + placeholder: + en_US: Please input your app secret + zh_Hans: 请输入你的飞书 app secret diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.py b/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.py new file mode 100644 index 00000000000000..8f83aea5abbe3d --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.py @@ -0,0 +1,20 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class AddEventAttendeesTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + event_id = tool_parameters.get("event_id") + attendee_phone_or_email = tool_parameters.get("attendee_phone_or_email") + need_notification = tool_parameters.get("need_notification", True) + + res = client.add_event_attendees(event_id, attendee_phone_or_email, need_notification) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.yaml new file mode 100644 index 00000000000000..b7744499b07344 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/add_event_attendees.yaml @@ -0,0 +1,54 @@ +identity: + name: add_event_attendees + author: Doug Lea + label: + en_US: Add Event Attendees + zh_Hans: 添加日程参会人 +description: + human: + en_US: Add Event Attendees + zh_Hans: 添加日程参会人 + llm: A tool for adding attendees to events in Feishu. (在飞书中添加日程参会人) +parameters: + - name: event_id + type: string + required: true + label: + en_US: Event ID + zh_Hans: 日程 ID + human_description: + en_US: | + The ID of the event, which will be returned when the event is created. For example: fb2a6406-26d6-4c8d-a487-6f0246c94d2f_0. + zh_Hans: | + 创建日程时会返回日程 ID。例如: fb2a6406-26d6-4c8d-a487-6f0246c94d2f_0。 + llm_description: | + 日程 ID，创建日程时会返回日程 ID。例如: fb2a6406-26d6-4c8d-a487-6f0246c94d2f_0。 + form: llm + + - name: need_notification + type: boolean + required: false + default: true + label: + en_US: Need Notification + zh_Hans: 是否需要通知 + human_description: + en_US: | + Whether to send a Bot notification to attendees. true: send, false: do not send. + zh_Hans: | + 是否给参与人发送 Bot 通知，true: 发送，false: 不发送。 + llm_description: | + 是否给参与人发送 Bot 通知，true: 发送，false: 不发送。 + form: form + + - name: attendee_phone_or_email + type: string + required: true + label: + en_US: Attendee Phone or Email + zh_Hans: 参会人电话或邮箱 + human_description: + en_US: The list of attendee emails or phone numbers, separated by commas. + zh_Hans: 日程参会人邮箱或者手机号列表，使用逗号分隔。 + llm_description: 日程参会人邮箱或者手机号列表，使用逗号分隔。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.py b/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.py new file mode 100644 index 00000000000000..8820bebdbed922 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.py @@ -0,0 +1,26 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class CreateEventTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + summary = tool_parameters.get("summary") + description = tool_parameters.get("description") + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + attendee_ability = tool_parameters.get("attendee_ability") + need_notification = tool_parameters.get("need_notification", True) + auto_record = tool_parameters.get("auto_record", False) + + res = client.create_event( + summary, description, start_time, end_time, attendee_ability, need_notification, auto_record + ) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.yaml new file mode 100644 index 00000000000000..f0784221ce7965 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/create_event.yaml @@ -0,0 +1,119 @@ +identity: + name: create_event + author: Doug Lea + label: + en_US: Create Event + zh_Hans: 创建日程 +description: + human: + en_US: Create Event + zh_Hans: 创建日程 + llm: A tool for creating events in Feishu.(创建飞书日程) +parameters: + - name: summary + type: string + required: false + label: + en_US: Summary + zh_Hans: 日程标题 + human_description: + en_US: The title of the event. If not filled, the event title will display (No Subject). + zh_Hans: 日程标题，若不填则日程标题显示 (无主题)。 + llm_description: 日程标题，若不填则日程标题显示 (无主题)。 + form: llm + + - name: description + type: string + required: false + label: + en_US: Description + zh_Hans: 日程描述 + human_description: + en_US: The description of the event. + zh_Hans: 日程描述。 + llm_description: 日程描述。 + form: llm + + - name: need_notification + type: boolean + required: false + default: true + label: + en_US: Need Notification + zh_Hans: 是否发送通知 + human_description: + en_US: | + Whether to send a bot message when the event is created, true: send, false: do not send. + zh_Hans: 创建日程时是否发送 bot 消息，true：发送，false：不发送。 + llm_description: 创建日程时是否发送 bot 消息，true：发送，false：不发送。 + form: form + + - name: start_time + type: string + required: true + label: + en_US: Start Time + zh_Hans: 开始时间 + human_description: + en_US: | + The start time of the event, format: 2006-01-02 15:04:05. + zh_Hans: 日程开始时间，格式：2006-01-02 15:04:05。 + llm_description: 日程开始时间，格式：2006-01-02 15:04:05。 + form: llm + + - name: end_time + type: string + required: true + label: + en_US: End Time + zh_Hans: 结束时间 + human_description: + en_US: | + The end time of the event, format: 2006-01-02 15:04:05. + zh_Hans: 日程结束时间，格式：2006-01-02 15:04:05。 + llm_description: 日程结束时间，格式：2006-01-02 15:04:05。 + form: llm + + - name: attendee_ability + type: select + required: false + options: + - value: none + label: + en_US: none + zh_Hans: 无 + - value: can_see_others + label: + en_US: can_see_others + zh_Hans: 可以查看参与人列表 + - value: can_invite_others + label: + en_US: can_invite_others + zh_Hans: 可以邀请其它参与人 + - value: can_modify_event + label: + en_US: can_modify_event + zh_Hans: 可以编辑日程 + default: "none" + label: + en_US: attendee_ability + zh_Hans: 参会人权限 + human_description: + en_US: Attendee ability, optional values are none, can_see_others, can_invite_others, can_modify_event, with a default value of none. + zh_Hans: 参会人权限，可选值有无、可以查看参与人列表、可以邀请其它参与人、可以编辑日程，默认值为无。 + llm_description: 参会人权限，可选值有无、可以查看参与人列表、可以邀请其它参与人、可以编辑日程，默认值为无。 + form: form + + - name: auto_record + type: boolean + required: false + default: false + label: + en_US: Auto Record + zh_Hans: 自动录制 + human_description: + en_US: | + Whether to enable automatic recording, true: enabled, automatically record when the meeting starts; false: not enabled. + zh_Hans: 是否开启自动录制，true：开启，会议开始后自动录制；false：不开启。 + llm_description: 是否开启自动录制，true：开启，会议开始后自动录制；false：不开启。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.py b/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.py new file mode 100644 index 00000000000000..144889692f9055 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.py @@ -0,0 +1,19 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class DeleteEventTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + event_id = tool_parameters.get("event_id") + need_notification = tool_parameters.get("need_notification", True) + + res = client.delete_event(event_id, need_notification) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.yaml new file mode 100644 index 00000000000000..54fdb04acc3371 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/delete_event.yaml @@ -0,0 +1,38 @@ +identity: + name: delete_event + author: Doug Lea + label: + en_US: Delete Event + zh_Hans: 删除日程 +description: + human: + en_US: Delete Event + zh_Hans: 删除日程 + llm: A tool for deleting events in Feishu.(在飞书中删除日程) +parameters: + - name: event_id + type: string + required: true + label: + en_US: Event ID + zh_Hans: 日程 ID + human_description: + en_US: | + The ID of the event, for example: e8b9791c-39ae-4908-8ad8-66b13159b9fb_0. + zh_Hans: 日程 ID，例如：e8b9791c-39ae-4908-8ad8-66b13159b9fb_0。 + llm_description: 日程 ID，例如：e8b9791c-39ae-4908-8ad8-66b13159b9fb_0。 + form: llm + + - name: need_notification + type: boolean + required: false + default: true + label: + en_US: Need Notification + zh_Hans: 是否需要通知 + human_description: + en_US: | + Indicates whether to send bot notifications to event participants upon deletion. true: send, false: do not send. + zh_Hans: 删除日程是否给日程参与人发送 bot 通知，true：发送，false：不发送。 + llm_description: 删除日程是否给日程参与人发送 bot 通知，true：发送，false：不发送。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.py b/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.py new file mode 100644 index 00000000000000..a2cd5a8b17d0af --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.py @@ -0,0 +1,18 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class GetPrimaryCalendarTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + user_id_type = tool_parameters.get("user_id_type", "open_id") + + res = client.get_primary_calendar(user_id_type) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.yaml new file mode 100644 index 00000000000000..3440c85d4a9733 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/get_primary_calendar.yaml @@ -0,0 +1,37 @@ +identity: + name: get_primary_calendar + author: Doug Lea + label: + en_US: Get Primary Calendar + zh_Hans: 查询主日历信息 +description: + human: + en_US: Get Primary Calendar + zh_Hans: 查询主日历信息 + llm: A tool for querying primary calendar information in Feishu.(在飞书中查询主日历信息) +parameters: + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.py b/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.py new file mode 100644 index 00000000000000..8815b4c9c871cd --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.py @@ -0,0 +1,21 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class ListEventsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + page_token = tool_parameters.get("page_token") + page_size = tool_parameters.get("page_size") + + res = client.list_events(start_time, end_time, page_token, page_size) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.yaml new file mode 100644 index 00000000000000..f4a5bfe6bab948 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/list_events.yaml @@ -0,0 +1,62 @@ +identity: + name: list_events + author: Doug Lea + label: + en_US: List Events + zh_Hans: 获取日程列表 +description: + human: + en_US: List Events + zh_Hans: 获取日程列表 + llm: A tool for listing events in Feishu.(在飞书中获取日程列表) +parameters: + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 开始时间 + human_description: + en_US: | + The start time, defaults to 0:00 of the current day if not provided, format: 2006-01-02 15:04:05. + zh_Hans: 开始时间，不传值时默认当天 0 点时间，格式为:2006-01-02 15:04:05。 + llm_description: 开始时间，不传值时默认当天 0 点时间，格式为:2006-01-02 15:04:05。 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 结束时间 + human_description: + en_US: | + The end time, defaults to 23:59 of the current day if not provided, format: 2006-01-02 15:04:05. + zh_Hans: 结束时间，不传值时默认当天 23:59 分时间，格式为:2006-01-02 15:04:05。 + llm_description: 结束时间，不传值时默认当天 23:59 分时间，格式为:2006-01-02 15:04:05。 + form: llm + + - name: page_size + type: number + required: false + default: 50 + label: + en_US: Page Size + zh_Hans: 分页大小 + human_description: + en_US: The page size, i.e., the number of data entries returned in a single request. The default value is 50, and the value range is [50,1000]. + zh_Hans: 分页大小，即单次请求所返回的数据条目数。默认值为 50，取值范围为 [50,1000]。 + llm_description: 分页大小，即单次请求所返回的数据条目数。默认值为 50，取值范围为 [50,1000]。 + form: llm + + - name: page_token + type: string + required: false + label: + en_US: Page Token + zh_Hans: 分页标记 + human_description: + en_US: The pagination token. Leave it blank for the first request, indicating to start traversing from the beginning; when the pagination query result has more items, a new page_token will be returned simultaneously, which can be used to obtain the query result in the next traversal. + zh_Hans: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + llm_description: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.py b/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.py new file mode 100644 index 00000000000000..dc365205a4cffa --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class SearchEventsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + query = tool_parameters.get("query") + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + page_token = tool_parameters.get("page_token") + user_id_type = tool_parameters.get("user_id_type", "open_id") + page_size = tool_parameters.get("page_size", 20) + + res = client.search_events(query, start_time, end_time, page_token, user_id_type, page_size) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.yaml new file mode 100644 index 00000000000000..e92a282091efcc --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/search_events.yaml @@ -0,0 +1,100 @@ +identity: + name: search_events + author: Doug Lea + label: + en_US: Search Events + zh_Hans: 搜索日程 +description: + human: + en_US: Search Events + zh_Hans: 搜索日程 + llm: A tool for searching events in Feishu.(在飞书中搜索日程) +parameters: + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form + + - name: query + type: string + required: true + label: + en_US: Query + zh_Hans: 搜索关键字 + human_description: + en_US: The search keyword used for fuzzy searching event names, with a maximum input of 200 characters. + zh_Hans: 用于模糊查询日程名称的搜索关键字，最大输入 200 字符。 + llm_description: 用于模糊查询日程名称的搜索关键字，最大输入 200 字符。 + form: llm + + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 开始时间 + human_description: + en_US: | + The start time, defaults to 0:00 of the current day if not provided, format: 2006-01-02 15:04:05. + zh_Hans: 开始时间，不传值时默认当天 0 点时间，格式为:2006-01-02 15:04:05。 + llm_description: 开始时间，不传值时默认当天 0 点时间，格式为:2006-01-02 15:04:05。 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 结束时间 + human_description: + en_US: | + The end time, defaults to 23:59 of the current day if not provided, format: 2006-01-02 15:04:05. + zh_Hans: 结束时间，不传值时默认当天 23:59 分时间，格式为:2006-01-02 15:04:05。 + llm_description: 结束时间，不传值时默认当天 23:59 分时间，格式为:2006-01-02 15:04:05。 + form: llm + + - name: page_size + type: number + required: false + default: 20 + label: + en_US: Page Size + zh_Hans: 分页大小 + human_description: + en_US: The page size, i.e., the number of data entries returned in a single request. The default value is 20, and the value range is [10,100]. + zh_Hans: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [10,100]。 + llm_description: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [10,100]。 + form: llm + + - name: page_token + type: string + required: false + label: + en_US: Page Token + zh_Hans: 分页标记 + human_description: + en_US: The pagination token. Leave it blank for the first request, indicating to start traversing from the beginning; when the pagination query result has more items, a new page_token will be returned simultaneously, which can be used to obtain the query result in the next traversal. + zh_Hans: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + llm_description: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.py b/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.py new file mode 100644 index 00000000000000..85bcb1d3f63847 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.py @@ -0,0 +1,24 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class UpdateEventTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + event_id = tool_parameters.get("event_id") + summary = tool_parameters.get("summary") + description = tool_parameters.get("description") + need_notification = tool_parameters.get("need_notification", True) + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + auto_record = tool_parameters.get("auto_record", False) + + res = client.update_event(event_id, summary, description, need_notification, start_time, end_time, auto_record) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.yaml b/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.yaml new file mode 100644 index 00000000000000..4d60dbf8c8e1b0 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_calendar/tools/update_event.yaml @@ -0,0 +1,100 @@ +identity: + name: update_event + author: Doug Lea + label: + en_US: Update Event + zh_Hans: 更新日程 +description: + human: + en_US: Update Event + zh_Hans: 更新日程 + llm: A tool for updating events in Feishu.(更新飞书中的日程) +parameters: + - name: event_id + type: string + required: true + label: + en_US: Event ID + zh_Hans: 日程 ID + human_description: + en_US: | + The ID of the event, for example: e8b9791c-39ae-4908-8ad8-66b13159b9fb_0. + zh_Hans: 日程 ID，例如：e8b9791c-39ae-4908-8ad8-66b13159b9fb_0。 + llm_description: 日程 ID，例如：e8b9791c-39ae-4908-8ad8-66b13159b9fb_0。 + form: llm + + - name: summary + type: string + required: false + label: + en_US: Summary + zh_Hans: 日程标题 + human_description: + en_US: The title of the event. + zh_Hans: 日程标题。 + llm_description: 日程标题。 + form: llm + + - name: description + type: string + required: false + label: + en_US: Description + zh_Hans: 日程描述 + human_description: + en_US: The description of the event. + zh_Hans: 日程描述。 + llm_description: 日程描述。 + form: llm + + - name: need_notification + type: boolean + required: false + label: + en_US: Need Notification + zh_Hans: 是否发送通知 + human_description: + en_US: | + Whether to send a bot message when the event is updated, true: send, false: do not send. + zh_Hans: 更新日程时是否发送 bot 消息，true：发送，false：不发送。 + llm_description: 更新日程时是否发送 bot 消息，true：发送，false：不发送。 + form: form + + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 开始时间 + human_description: + en_US: | + The start time of the event, format: 2006-01-02 15:04:05. + zh_Hans: 日程开始时间，格式：2006-01-02 15:04:05。 + llm_description: 日程开始时间，格式：2006-01-02 15:04:05。 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 结束时间 + human_description: + en_US: | + The end time of the event, format: 2006-01-02 15:04:05. + zh_Hans: 日程结束时间，格式：2006-01-02 15:04:05。 + llm_description: 日程结束时间，格式：2006-01-02 15:04:05。 + form: llm + + - name: auto_record + type: boolean + required: false + label: + en_US: Auto Record + zh_Hans: 自动录制 + human_description: + en_US: | + Whether to enable automatic recording, true: enabled, automatically record when the meeting starts; false: not enabled. + zh_Hans: 是否开启自动录制，true：开启，会议开始后自动录制；false：不开启。 + llm_description: 是否开启自动录制，true：开启，会议开始后自动录制；false：不开启。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_document/feishu_document.py b/api/core/tools/provider/builtin/feishu_document/feishu_document.py index b0a1e393eb8116..217ae52082b82c 100644 --- a/api/core/tools/provider/builtin/feishu_document/feishu_document.py +++ b/api/core/tools/provider/builtin/feishu_document/feishu_document.py @@ -1,15 +1,7 @@ -from core.tools.errors import ToolProviderCredentialValidationError from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController -from core.tools.utils.feishu_api_utils import FeishuRequest +from core.tools.utils.feishu_api_utils import auth class FeishuDocumentProvider(BuiltinToolProviderController): def _validate_credentials(self, credentials: dict) -> None: - app_id = credentials.get("app_id") - app_secret = credentials.get("app_secret") - if not app_id or not app_secret: - raise ToolProviderCredentialValidationError("app_id and app_secret is required") - try: - assert FeishuRequest(app_id, app_secret).tenant_access_token is not None - except Exception as e: - raise ToolProviderCredentialValidationError(str(e)) + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_document/feishu_document.yaml b/api/core/tools/provider/builtin/feishu_document/feishu_document.yaml index 8eaa6b27049c6b..8f9afa6149445c 100644 --- a/api/core/tools/provider/builtin/feishu_document/feishu_document.yaml +++ b/api/core/tools/provider/builtin/feishu_document/feishu_document.yaml @@ -5,8 +5,10 @@ identity: en_US: Lark Cloud Document zh_Hans: 飞书云文档 description: - en_US: Lark Cloud Document - zh_Hans: 飞书云文档 + en_US: | + Lark cloud document, requires the following permissions: docx:document、drive:drive、docs:document.content:read. + zh_Hans: | + 飞书云文档，需要开通以下权限: docx:document、drive:drive、docs:document.content:read。 icon: icon.svg tags: - social @@ -23,7 +25,7 @@ credentials_for_provider: help: en_US: Get your app_id and app_secret from Feishu zh_Hans: 从飞书获取您的 app_id 和 app_secret - url: https://open.feishu.cn + url: https://open.larkoffice.com/app app_secret: type: secret-input required: true diff --git a/api/core/tools/provider/builtin/feishu_document/tools/create_document.yaml b/api/core/tools/provider/builtin/feishu_document/tools/create_document.yaml index ddf2729f0e4b5c..85382e9d8e8d1f 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/create_document.yaml +++ b/api/core/tools/provider/builtin/feishu_document/tools/create_document.yaml @@ -7,7 +7,7 @@ identity: description: human: en_US: Create Lark document - zh_Hans: 创建飞书文档，支持创建空文档和带内容的文档，支持 markdown 语法创建。 + zh_Hans: 创建飞书文档，支持创建空文档和带内容的文档，支持 markdown 语法创建。应用需要开启机器人能力(https://open.feishu.cn/document/faq/trouble-shooting/how-to-enable-bot-ability)。 llm: A tool for creating Feishu documents. parameters: - name: title @@ -41,7 +41,8 @@ parameters: en_US: folder_token zh_Hans: 文档所在文件夹的 Token human_description: - en_US: The token of the folder where the document is located. If it is not passed or is empty, it means the root directory. - zh_Hans: 文档所在文件夹的 Token，不传或传空表示根目录。 - llm_description: 文档所在文件夹的 Token，不传或传空表示根目录。 + en_US: | + The token of the folder where the document is located. If it is not passed or is empty, it means the root directory. For Example: https://svi136aogf123.feishu.cn/drive/folder/JgR9fiG9AlPt8EdsSNpcGjIInbf + zh_Hans: 文档所在文件夹的 Token，不传或传空表示根目录。例如：https://svi136aogf123.feishu.cn/drive/folder/JgR9fiG9AlPt8EdsSNpcGjIInbf。 + llm_description: 文档所在文件夹的 Token，不传或传空表示根目录。例如：https://svi136aogf123.feishu.cn/drive/folder/JgR9fiG9AlPt8EdsSNpcGjIInbf。 form: llm diff --git a/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.py b/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.py index c94a5f70ed7e34..e67a017facc8d4 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.py +++ b/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.py @@ -12,8 +12,8 @@ def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMe client = FeishuRequest(app_id, app_secret) document_id = tool_parameters.get("document_id") - mode = tool_parameters.get("mode") - lang = tool_parameters.get("lang", 0) + mode = tool_parameters.get("mode", "markdown") + lang = tool_parameters.get("lang", "0") res = client.get_document_content(document_id, mode, lang) return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.yaml b/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.yaml index 51eda73a60095c..15e827cde91ee6 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.yaml +++ b/api/core/tools/provider/builtin/feishu_document/tools/get_document_content.yaml @@ -23,8 +23,18 @@ parameters: form: llm - name: mode - type: string + type: select required: false + options: + - value: text + label: + en_US: text + zh_Hans: text + - value: markdown + label: + en_US: markdown + zh_Hans: markdown + default: "markdown" label: en_US: mode zh_Hans: 文档返回格式 @@ -32,18 +42,29 @@ parameters: en_US: Format of the document return, optional values are text, markdown, can be empty, default is markdown. zh_Hans: 文档返回格式，可选值有 text、markdown，可以为空，默认值为 markdown。 llm_description: 文档返回格式，可选值有 text、markdown，可以为空，默认值为 markdown。 - form: llm + form: form - name: lang - type: number + type: select required: false - default: 0 + options: + - value: "0" + label: + en_US: User's default name + zh_Hans: 用户的默认名称 + - value: "1" + label: + en_US: User's English name + zh_Hans: 用户的英文名称 + default: "0" label: en_US: lang zh_Hans: 指定@用户的语言 human_description: en_US: | Specifies the language for MentionUser, optional values are [0, 1]. 0: User's default name, 1: User's English name, default is 0. - zh_Hans: 指定返回的 MentionUser，即 @用户 的语言，可选值有 [0,1]。0：该用户的默认名称，1：该用户的英文名称，默认值为 0。 - llm_description: 指定返回的 MentionUser，即 @用户 的语言，可选值有 [0,1]。0：该用户的默认名称，1：该用户的英文名称，默认值为 0。 - form: llm + zh_Hans: | + 指定返回的 MentionUser，即@用户的语言，可选值有 [0,1]。0: 该用户的默认名称，1: 该用户的英文名称，默认值为 0。 + llm_description: | + 指定返回的 MentionUser，即@用户的语言，可选值有 [0,1]。0: 该用户的默认名称，1: 该用户的英文名称，默认值为 0。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.py b/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.py index 572a7abf284193..dd57c6870d0ba9 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.py +++ b/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.py @@ -12,8 +12,9 @@ def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMe client = FeishuRequest(app_id, app_secret) document_id = tool_parameters.get("document_id") - page_size = tool_parameters.get("page_size", 500) page_token = tool_parameters.get("page_token", "") + user_id_type = tool_parameters.get("user_id_type", "open_id") + page_size = tool_parameters.get("page_size", 500) - res = client.list_document_blocks(document_id, page_token, page_size) + res = client.list_document_blocks(document_id, page_token, user_id_type, page_size) return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.yaml b/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.yaml index 019ac983906ff1..5b8ef7d53c23f4 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.yaml +++ b/api/core/tools/provider/builtin/feishu_document/tools/list_document_blocks.yaml @@ -46,12 +46,12 @@ parameters: en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 - form: llm + form: form - name: page_size type: number required: false - default: "500" + default: 500 label: en_US: page_size zh_Hans: 分页大小 diff --git a/api/core/tools/provider/builtin/feishu_document/tools/write_document.py b/api/core/tools/provider/builtin/feishu_document/tools/write_document.py index 6061250e48e136..59f08f53dc68de 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/write_document.py +++ b/api/core/tools/provider/builtin/feishu_document/tools/write_document.py @@ -13,7 +13,7 @@ def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMe document_id = tool_parameters.get("document_id") content = tool_parameters.get("content") - position = tool_parameters.get("position") + position = tool_parameters.get("position", "end") res = client.write_document(document_id, content, position) return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_document/tools/write_document.yaml b/api/core/tools/provider/builtin/feishu_document/tools/write_document.yaml index 4282e3dcf3977f..de70f4e7726a28 100644 --- a/api/core/tools/provider/builtin/feishu_document/tools/write_document.yaml +++ b/api/core/tools/provider/builtin/feishu_document/tools/write_document.yaml @@ -35,25 +35,23 @@ parameters: form: llm - name: position - type: string + type: select required: false - label: - en_US: position - zh_Hans: 添加位置 - human_description: - en_US: | - Enumeration values: start or end. Use 'start' to add content at the beginning of the document, and 'end' to add content at the end. The default value is 'end'. - zh_Hans: 枚举值：start 或 end。使用 'start' 在文档开头添加内容，使用 'end' 在文档结尾添加内容，默认值为 'end'。 - llm_description: | - 枚举值 start、end，start: 在文档开头添加内容；end: 在文档结尾添加内容，默认值为 end。 - form: llm options: - value: start label: - en_US: start - zh_Hans: 在文档开头添加内容 + en_US: document start + zh_Hans: 文档开始 - value: end label: - en_US: end - zh_Hans: 在文档结尾添加内容 - default: start + en_US: document end + zh_Hans: 文档结束 + default: "end" + label: + en_US: position + zh_Hans: 内容添加位置 + human_description: + en_US: Content insertion position, optional values are start, end. 'start' means adding content at the beginning of the document; 'end' means adding content at the end of the document. The default value is end. + zh_Hans: 内容添加位置，可选值有 start、end。start 表示在文档开头添加内容；end 表示在文档结尾添加内容，默认值为 end。 + llm_description: 内容添加位置，可选值有 start、end。start 表示在文档开头添加内容；end 表示在文档结尾添加内容，默认值为 end。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_message/feishu_message.py b/api/core/tools/provider/builtin/feishu_message/feishu_message.py index 7b3adb9293750b..a3b54737691c9c 100644 --- a/api/core/tools/provider/builtin/feishu_message/feishu_message.py +++ b/api/core/tools/provider/builtin/feishu_message/feishu_message.py @@ -1,15 +1,7 @@ -from core.tools.errors import ToolProviderCredentialValidationError from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController -from core.tools.utils.feishu_api_utils import FeishuRequest +from core.tools.utils.feishu_api_utils import auth class FeishuMessageProvider(BuiltinToolProviderController): def _validate_credentials(self, credentials: dict) -> None: - app_id = credentials.get("app_id") - app_secret = credentials.get("app_secret") - if not app_id or not app_secret: - raise ToolProviderCredentialValidationError("app_id and app_secret is required") - try: - assert FeishuRequest(app_id, app_secret).tenant_access_token is not None - except Exception as e: - raise ToolProviderCredentialValidationError(str(e)) + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_message/feishu_message.yaml b/api/core/tools/provider/builtin/feishu_message/feishu_message.yaml index 1bd8953dddcb24..56683ec1680f40 100644 --- a/api/core/tools/provider/builtin/feishu_message/feishu_message.yaml +++ b/api/core/tools/provider/builtin/feishu_message/feishu_message.yaml @@ -5,8 +5,10 @@ identity: en_US: Lark Message zh_Hans: 飞书消息 description: - en_US: Lark Message - zh_Hans: 飞书消息 + en_US: | + Lark message, requires the following permissions: im:message、im:message.group_msg. + zh_Hans: | + 飞书消息，需要开通以下权限: im:message、im:message.group_msg。 icon: icon.svg tags: - social @@ -23,7 +25,7 @@ credentials_for_provider: help: en_US: Get your app_id and app_secret from Feishu zh_Hans: 从飞书获取您的 app_id 和 app_secret - url: https://open.feishu.cn + url: https://open.larkoffice.com/app app_secret: type: secret-input required: true diff --git a/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.py b/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.py new file mode 100644 index 00000000000000..7eb29230b2ceb0 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class GetChatMessagesTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + container_id = tool_parameters.get("container_id") + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + page_token = tool_parameters.get("page_token") + sort_type = tool_parameters.get("sort_type", "ByCreateTimeAsc") + page_size = tool_parameters.get("page_size", 20) + + res = client.get_chat_messages(container_id, start_time, end_time, page_token, sort_type, page_size) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.yaml b/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.yaml new file mode 100644 index 00000000000000..153c8c80e58db4 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_message/tools/get_chat_messages.yaml @@ -0,0 +1,96 @@ +identity: + name: get_chat_messages + author: Doug Lea + label: + en_US: Get Chat Messages + zh_Hans: 获取指定单聊、群聊的消息历史 +description: + human: + en_US: Get Chat Messages + zh_Hans: 获取指定单聊、群聊的消息历史 + llm: A tool for getting chat messages from specific one-on-one chats or group chats.(获取指定单聊、群聊的消息历史) +parameters: + - name: container_id + type: string + required: true + label: + en_US: Container Id + zh_Hans: 群聊或单聊的 ID + human_description: + en_US: The ID of the group chat or single chat. Refer to the group ID description for how to obtain it. https://open.feishu.cn/document/server-docs/group/chat/chat-id-description + zh_Hans: 群聊或单聊的 ID，获取方式参见群 ID 说明。https://open.feishu.cn/document/server-docs/group/chat/chat-id-description + llm_description: 群聊或单聊的 ID，获取方式参见群 ID 说明。https://open.feishu.cn/document/server-docs/group/chat/chat-id-description + form: llm + + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 起始时间 + human_description: + en_US: The start time for querying historical messages, formatted as "2006-01-02 15:04:05". + zh_Hans: 待查询历史信息的起始时间，格式为 "2006-01-02 15:04:05"。 + llm_description: 待查询历史信息的起始时间，格式为 "2006-01-02 15:04:05"。 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 结束时间 + human_description: + en_US: The end time for querying historical messages, formatted as "2006-01-02 15:04:05". + zh_Hans: 待查询历史信息的结束时间，格式为 "2006-01-02 15:04:05"。 + llm_description: 待查询历史信息的结束时间，格式为 "2006-01-02 15:04:05"。 + form: llm + + - name: sort_type + type: select + required: false + options: + - value: ByCreateTimeAsc + label: + en_US: ByCreateTimeAsc + zh_Hans: ByCreateTimeAsc + - value: ByCreateTimeDesc + label: + en_US: ByCreateTimeDesc + zh_Hans: ByCreateTimeDesc + default: "ByCreateTimeAsc" + label: + en_US: Sort Type + zh_Hans: 排序方式 + human_description: + en_US: | + The message sorting method. Optional values are ByCreateTimeAsc: sorted in ascending order by message creation time; ByCreateTimeDesc: sorted in descending order by message creation time. The default value is ByCreateTimeAsc. Note: When using page_token for pagination requests, the sorting method (sort_type) is consistent with the first request and cannot be changed midway. + zh_Hans: | + 消息排序方式，可选值有 ByCreateTimeAsc：按消息创建时间升序排列；ByCreateTimeDesc：按消息创建时间降序排列。默认值为：ByCreateTimeAsc。注意：使用 page_token 分页请求时，排序方式（sort_type）均与第一次请求一致，不支持中途改换排序方式。 + llm_description: 消息排序方式，可选值有 ByCreateTimeAsc：按消息创建时间升序排列；ByCreateTimeDesc：按消息创建时间降序排列。默认值为：ByCreateTimeAsc。注意：使用 page_token 分页请求时，排序方式（sort_type）均与第一次请求一致，不支持中途改换排序方式。 + form: form + + - name: page_size + type: number + required: false + default: 20 + label: + en_US: Page Size + zh_Hans: 分页大小 + human_description: + en_US: The page size, i.e., the number of data entries returned in a single request. The default value is 20, and the value range is [1,50]. + zh_Hans: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [1,50]。 + llm_description: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [1,50]。 + form: llm + + - name: page_token + type: string + required: false + label: + en_US: Page Token + zh_Hans: 分页标记 + human_description: + en_US: The pagination token. Leave it blank for the first request, indicating to start traversing from the beginning; when the pagination query result has more items, a new page_token will be returned simultaneously, which can be used to obtain the query result in the next traversal. + zh_Hans: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + llm_description: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.py b/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.py new file mode 100644 index 00000000000000..3b14f46e0048a8 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.py @@ -0,0 +1,21 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class GetChatMessagesTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + container_id = tool_parameters.get("container_id") + page_token = tool_parameters.get("page_token") + sort_type = tool_parameters.get("sort_type", "ByCreateTimeAsc") + page_size = tool_parameters.get("page_size", 20) + + res = client.get_thread_messages(container_id, page_token, sort_type, page_size) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.yaml b/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.yaml new file mode 100644 index 00000000000000..8d5fed9d0bba24 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_message/tools/get_thread_messages.yaml @@ -0,0 +1,72 @@ +identity: + name: get_thread_messages + author: Doug Lea + label: + en_US: Get Thread Messages + zh_Hans: 获取指定话题的消息历史 +description: + human: + en_US: Get Thread Messages + zh_Hans: 获取指定话题的消息历史 + llm: A tool for getting chat messages from specific threads.(获取指定话题的消息历史) +parameters: + - name: container_id + type: string + required: true + label: + en_US: Thread Id + zh_Hans: 话题 ID + human_description: + en_US: The ID of the thread. Refer to the thread overview on how to obtain the thread_id. https://open.feishu.cn/document/uAjLw4CM/ukTMukTMukTM/reference/im-v1/message/thread-introduction + zh_Hans: 话题 ID，获取方式参见话题概述的如何获取 thread_id 章节。https://open.feishu.cn/document/uAjLw4CM/ukTMukTMukTM/reference/im-v1/message/thread-introduction + llm_description: 话题 ID，获取方式参见话题概述的如何获取 thread_id 章节。https://open.feishu.cn/document/uAjLw4CM/ukTMukTMukTM/reference/im-v1/message/thread-introduction + form: llm + + - name: sort_type + type: select + required: false + options: + - value: ByCreateTimeAsc + label: + en_US: ByCreateTimeAsc + zh_Hans: ByCreateTimeAsc + - value: ByCreateTimeDesc + label: + en_US: ByCreateTimeDesc + zh_Hans: ByCreateTimeDesc + default: "ByCreateTimeAsc" + label: + en_US: Sort Type + zh_Hans: 排序方式 + human_description: + en_US: | + The message sorting method. Optional values are ByCreateTimeAsc: sorted in ascending order by message creation time; ByCreateTimeDesc: sorted in descending order by message creation time. The default value is ByCreateTimeAsc. Note: When using page_token for pagination requests, the sorting method (sort_type) is consistent with the first request and cannot be changed midway. + zh_Hans: | + 消息排序方式，可选值有 ByCreateTimeAsc：按消息创建时间升序排列；ByCreateTimeDesc：按消息创建时间降序排列。默认值为：ByCreateTimeAsc。注意：使用 page_token 分页请求时，排序方式（sort_type）均与第一次请求一致，不支持中途改换排序方式。 + llm_description: 消息排序方式，可选值有 ByCreateTimeAsc：按消息创建时间升序排列；ByCreateTimeDesc：按消息创建时间降序排列。默认值为：ByCreateTimeAsc。注意：使用 page_token 分页请求时，排序方式（sort_type）均与第一次请求一致，不支持中途改换排序方式。 + form: form + + - name: page_size + type: number + required: false + default: 20 + label: + en_US: Page Size + zh_Hans: 分页大小 + human_description: + en_US: The page size, i.e., the number of data entries returned in a single request. The default value is 20, and the value range is [1,50]. + zh_Hans: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [1,50]。 + llm_description: 分页大小，即单次请求所返回的数据条目数。默认值为 20，取值范围为 [1,50]。 + form: llm + + - name: page_token + type: string + required: false + label: + en_US: Page Token + zh_Hans: 分页标记 + human_description: + en_US: The pagination token. Leave it blank for the first request, indicating to start traversing from the beginning; when the pagination query result has more items, a new page_token will be returned simultaneously, which can be used to obtain the query result in the next traversal. + zh_Hans: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + llm_description: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_message/tools/send_bot_message.yaml b/api/core/tools/provider/builtin/feishu_message/tools/send_bot_message.yaml index 6e398b18ab3aee..4f7f65a8a74fc0 100644 --- a/api/core/tools/provider/builtin/feishu_message/tools/send_bot_message.yaml +++ b/api/core/tools/provider/builtin/feishu_message/tools/send_bot_message.yaml @@ -10,53 +10,53 @@ description: zh_Hans: 发送飞书应用消息 llm: A tool for sending Feishu application messages. parameters: + - name: receive_id + type: string + required: true + label: + en_US: receive_id + zh_Hans: 消息接收者的 ID + human_description: + en_US: The ID of the message receiver, the ID type is consistent with the value of the query parameter receive_id_type. + zh_Hans: 消息接收者的 ID，ID 类型与查询参数 receive_id_type 的取值一致。 + llm_description: 消息接收者的 ID，ID 类型与查询参数 receive_id_type 的取值一致。 + form: llm + - name: receive_id_type type: select required: true options: - value: open_id label: - en_US: open id - zh_Hans: open id + en_US: open_id + zh_Hans: open_id - value: union_id label: - en_US: union id - zh_Hans: union id + en_US: union_id + zh_Hans: union_id - value: user_id label: - en_US: user id - zh_Hans: user id + en_US: user_id + zh_Hans: user_id - value: email label: en_US: email zh_Hans: email - value: chat_id label: - en_US: chat id - zh_Hans: chat id + en_US: chat_id + zh_Hans: chat_id label: - en_US: User ID Type - zh_Hans: 用户 ID 类型 + en_US: receive_id_type + zh_Hans: 消息接收者的 ID 类型 human_description: - en_US: User ID Type - zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id、email、chat_id。 - llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id、email、chat_id。 - form: llm - - - name: receive_id - type: string - required: true - label: - en_US: Receive Id - zh_Hans: 消息接收者的 ID - human_description: - en_US: The ID of the message receiver. The ID type should correspond to the query parameter receive_id_type. - zh_Hans: 消息接收者的 ID，ID 类型应与查询参数 receive_id_type 对应。 - llm_description: 消息接收者的 ID，ID 类型应与查询参数 receive_id_type 对应。 - form: llm + en_US: The ID type of the message receiver, optional values are open_id, union_id, user_id, email, chat_id, with a default value of open_id. + zh_Hans: 消息接收者的 ID 类型，可选值有 open_id、union_id、user_id、email、chat_id，默认值为 open_id。 + llm_description: 消息接收者的 ID 类型，可选值有 open_id、union_id、user_id、email、chat_id，默认值为 open_id。 + form: form - name: msg_type - type: string + type: select required: true options: - value: text @@ -65,27 +65,61 @@ parameters: zh_Hans: 文本 - value: interactive label: - en_US: message card - zh_Hans: 消息卡片 + en_US: interactive + zh_Hans: 卡片 + - value: post + label: + en_US: post + zh_Hans: 富文本 + - value: image + label: + en_US: image + zh_Hans: 图片 + - value: file + label: + en_US: file + zh_Hans: 文件 + - value: audio + label: + en_US: audio + zh_Hans: 语音 + - value: media + label: + en_US: media + zh_Hans: 视频 + - value: sticker + label: + en_US: sticker + zh_Hans: 表情包 + - value: share_chat + label: + en_US: share_chat + zh_Hans: 分享群名片 + - value: share_user + label: + en_US: share_user + zh_Hans: 分享个人名片 + - value: system + label: + en_US: system + zh_Hans: 系统消息 label: - en_US: Message type + en_US: msg_type zh_Hans: 消息类型 human_description: - en_US: Message type, optional values are, text (text), interactive (message card). - zh_Hans: 消息类型，可选值有：text(文本)、interactive(消息卡片)。 - llm_description: 消息类型，可选值有：text(文本)、interactive(消息卡片)。 - form: llm + en_US: Message type. Optional values are text, post, image, file, audio, media, sticker, interactive, share_chat, share_user, system. For detailed introduction of different message types, refer to the message content(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json). + zh_Hans: 消息类型。可选值有：text、post、image、file、audio、media、sticker、interactive、share_chat、share_user、system。不同消息类型的详细介绍，参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + llm_description: 消息类型。可选值有：text、post、image、file、audio、media、sticker、interactive、share_chat、share_user、system。不同消息类型的详细介绍，参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + form: form - name: content type: string required: true label: - en_US: Message content + en_US: content zh_Hans: 消息内容 human_description: - en_US: Message content - zh_Hans: | - 消息内容，JSON 结构序列化后的字符串。不同 msg_type 对应不同内容， - 具体格式说明参考：https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json - llm_description: 消息内容，JSON 结构序列化后的字符串。不同 msg_type 对应不同内容。 + en_US: Message content, a JSON structure serialized string. The value of this parameter corresponds to msg_type. For example, if msg_type is text, this parameter needs to pass in text type content. To understand the format and usage limitations of different message types, refer to the message content(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json). + zh_Hans: 消息内容，JSON 结构序列化后的字符串。该参数的取值与 msg_type 对应，例如 msg_type 取值为 text，则该参数需要传入文本类型的内容。了解不同类型的消息内容格式、使用限制，可参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + llm_description: 消息内容，JSON 结构序列化后的字符串。该参数的取值与 msg_type 对应，例如 msg_type 取值为 text，则该参数需要传入文本类型的内容。了解不同类型的消息内容格式、使用限制，可参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 form: llm diff --git a/api/core/tools/provider/builtin/feishu_message/tools/send_webhook_message.yaml b/api/core/tools/provider/builtin/feishu_message/tools/send_webhook_message.yaml index 8b39ce4874d506..eeeae8b29cd935 100644 --- a/api/core/tools/provider/builtin/feishu_message/tools/send_webhook_message.yaml +++ b/api/core/tools/provider/builtin/feishu_message/tools/send_webhook_message.yaml @@ -15,15 +15,18 @@ parameters: required: true label: en_US: webhook - zh_Hans: webhook 的地址 + zh_Hans: webhook human_description: - en_US: The address of the webhook - zh_Hans: webhook 的地址 - llm_description: webhook 的地址 + en_US: | + The address of the webhook, the format of the webhook address corresponding to the bot is as follows: https://open.feishu.cn/open-apis/bot/v2/hook/xxxxxxxxxxxxxxxxx. For details, please refer to: Feishu Custom Bot Usage Guide(https://open.larkoffice.com/document/client-docs/bot-v3/add-custom-bot) + zh_Hans: | + webhook 的地址，机器人对应的 webhook 地址格式如下: https://open.feishu.cn/open-apis/bot/v2/hook/xxxxxxxxxxxxxxxxx，详情可参考: 飞书自定义机器人使用指南(https://open.larkoffice.com/document/client-docs/bot-v3/add-custom-bot) + llm_description: | + webhook 的地址，机器人对应的 webhook 地址格式如下: https://open.feishu.cn/open-apis/bot/v2/hook/xxxxxxxxxxxxxxxxx，详情可参考: 飞书自定义机器人使用指南(https://open.larkoffice.com/document/client-docs/bot-v3/add-custom-bot) form: llm - name: msg_type - type: string + type: select required: true options: - value: text @@ -32,27 +35,34 @@ parameters: zh_Hans: 文本 - value: interactive label: - en_US: message card - zh_Hans: 消息卡片 + en_US: interactive + zh_Hans: 卡片 + - value: image + label: + en_US: image + zh_Hans: 图片 + - value: share_chat + label: + en_US: share_chat + zh_Hans: 分享群名片 label: - en_US: Message type + en_US: msg_type zh_Hans: 消息类型 human_description: - en_US: Message type, optional values are, text (text), interactive (message card). - zh_Hans: 消息类型，可选值有：text(文本)、interactive(消息卡片)。 - llm_description: 消息类型，可选值有：text(文本)、interactive(消息卡片)。 - form: llm + en_US: Message type. Optional values are text, image, interactive, share_chat. For detailed introduction of different message types, refer to the message content(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json). + zh_Hans: 消息类型。可选值有：text、image、interactive、share_chat。不同消息类型的详细介绍，参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + llm_description: 消息类型。可选值有：text、image、interactive、share_chat。不同消息类型的详细介绍，参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + form: form + - name: content type: string required: true label: - en_US: Message content + en_US: content zh_Hans: 消息内容 human_description: - en_US: Message content - zh_Hans: | - 消息内容，JSON 结构序列化后的字符串。不同 msg_type 对应不同内容， - 具体格式说明参考：https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json - llm_description: 消息内容，JSON 结构序列化后的字符串。不同 msg_type 对应不同内容。 + en_US: Message content, a JSON structure serialized string. The value of this parameter corresponds to msg_type. For example, if msg_type is text, this parameter needs to pass in text type content. To understand the format and usage limitations of different message types, refer to the message content(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json). + zh_Hans: 消息内容，JSON 结构序列化后的字符串。该参数的取值与 msg_type 对应，例如 msg_type 取值为 text，则该参数需要传入文本类型的内容。了解不同类型的消息内容格式、使用限制，可参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 + llm_description: 消息内容，JSON 结构序列化后的字符串。该参数的取值与 msg_type 对应，例如 msg_type 取值为 text，则该参数需要传入文本类型的内容。了解不同类型的消息内容格式、使用限制，可参见发送消息内容(https://open.larkoffice.com/document/server-docs/im-v1/message-content-description/create_json)。 form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/_assets/icon.png b/api/core/tools/provider/builtin/feishu_spreadsheet/_assets/icon.png new file mode 100644 index 0000000000000000000000000000000000000000..258b361261d4e3366251613141efaf200cd492db GIT binary patch literal 5417 zcmd^DS5%YRwoZN$O6VAxB7`D>B2{_|NCfFcC5m9^B1Ni1YN8;ag8~B5K}5ljNR=At zDowYWE@T6ufDnpw#PfU4%XzwQcZ_?-eMtT>zS-xRYp%IgytSnsL zdRI`?%X-mSlvNKO|Ie?xLgnG*o=vL-$74~@Z_^H692>&gwz#AF8KDeDRMbw;%S7@^ z0lEPo#d95ysLibL=R-lQaee?Fs}d9;GP^(x<}0F2F-h^b07#9l0oO$-H1Q3K391Vw z<4=0p`{@z^2;xpdlh~~3JnE^>9z);@q9`KIu_BtdpQzbsD575%4Twvli5I4HvMd2) zb~piMXifm)8t=*F!((F(ty-{~Wz)Z}Ndy@Eb!#j;_(=_NxxD1SZ z`^;W2^5auNzYn@A2>xJh;sP2HSC+;woK59g(<;NCo^!c#Ehuot?D4Rd)w8!(U0P`& z%VC9Q-96X{A0A-3$C;#tp*U^SO+lV&Hxa7rN-a66lFyi_>RrN%R|1Vq#cqxM66~=V zqiED^xRLX95ANScBy+VLmy`yfkVWj>n$$_jO&|;@RMZxW`Vz z$O_A_ZC28OHeIiVe5Q^a@Eh@@xS_-0ekSH|OjbBisUizBYwacbqfha(|5(|VJ2}WB zHpATzRjsPoJQBe4PIt1hgrI7Y50}-!}5gFUy?Rg9_E&^py-is{-p;% zTvMdnGy%=Wqn{@HrmR@{jrI>dJKTelw2a|nx8U=k7~^*xb#q+marab_8vS2fSE5+Y zn~5{LDooF9tC$SGL27*G@;{=nn$Xky=U;8P+72;ajTwrUrC1Tnxpj$-`QP(e{CvV9 zh3pN9onxvac@}^z>)z{0S+aM9#p?S7+MZA%@a=)}H%cBs2SJLPD0zo*4!;C}*GG6$ zTMl`6(oQ&vd#E0UZeZkChdQ~D8u|whYR0Tkk_54k46T{#q1zuLx1St1$09Z4@?G-x zP;12H+5MiS`8ypOtP=3S-g>%7u{bLB#@8}4dCm`w!J@u0gpTDCfl`36#Ce`k^1%{YRQF_6`gSF9DyZ3Bm zxr|mHO4@>BaQa#*$Cj2S)LE_&-nSsJv3#y!5=1E}zv^mBz1y3!vXp@@&b|hGYttj^ z?UFMVQR53YFP6~^{CEd_wTWBGlH@xTzP_*PtPwumo>==enAF|ol)S&NnfqRA-FZ#G zf5A^UVYdF1#whcR86TO2%c|Bm2kp)#9~26{2Ge@9z!sOtO3(_ASdG)a5Jouh*GPceIQ8i;V2@|!6kT2)T zP3GTrKXy&4^~U@c_IEJKMZl%eR=P~7mddg1f~Y}6Wy#^*OPN3UfmQK^3@itZpKbd5@0 zc3B?kLSn(>gN`_VSnGc2%ub#NtO@@Vkcuw1I zE&gLUmn!~7K;3W8aMa%NswfiMeh!H(yq2vZu#y-1l|`&3hz-OPVp0Cr>%V@TkU!%> zig$7Eo6aFiCaPc7AQg38k{ODANjtgts?q(!v%Aw&3~?*`#9gB9y+Zv%yYux6M}lWg zZx}2Y?UvriIKFi_e_K9Y{*4iad-AWN{so#fZ7hA_A!jW=?o>IKy zVmr>iln9}Jc@{D>cn$kyL;F=r9*Z3=5=?1+js+lA8D{cPQAguYGbtg}#)-6p{L3|* z@z$zI?_#(<*R6L3#H|ws#J_XKWhkN21!JDd0#QqszIwaZd#vflMa8h(9b$wT5L8*< zBX{8NtDYXEG^ec#34>iz2?zd{3O{5DZf>LtQt;n=e$*{$Wdmn-7EK!_{#(~HLVkHiwO`?8f|1cmvooXKl(DSNxarLW);a=FHJAic)+i^G z(e}q=C|W8Y@LuK-uv76%eROl8~k7sr_FPLKx5)D|=j5MXweA zsi4aJnHaAJ-_CgY`b`pY7?BvwX)Ah}o;`g;d}iT<<{|np=0|Z$nPE>jWATfMx145R;)u$P_vwvE1OY&0~oC274T(v@nP)MM42l z#rob;2tLAOLgc?^gTq+sb;?jg(TJ+PMBwcG4dIK#Er=}+;)jzkBQbA6MY987*4wFXPvD(6q#>};6+{c!gE zuh?~G=IFzw59PEky!+o7(m;57&_TyPg|dS<(VkwmfpJ;)jR3<_=3?xKgTG9sg|mT8 zmbR_XHYesG&?uhyDlEiy%>7r}!@aL&dpZ>1HoqDf{|1sCM-k|%3`JO`Qx1$Y1q`W# z=265^17<`61QLL!MCE=k9QvOWA$*wMEB*<)^nabuVB#n)s@pcn@^cAx%#VAfr_bI)ekLa(&LqeVj$|@{;Y-L}+dGMsG-d-d4iEN~E?1J;#SxCn*=oaRy&&#~_7uBTMaU0*MMYxCUXn$sPyZ5ihR%j-)C1apw>-l9JRzB?mJ5v?YFm*DN3ll)!1?9R}6Z3b@E^)gymONw#o8R29`3SRVA zsi#m1t=0|{3Ol$DlC?xn@?&-b+n4<(hoge6@h^J8p`^7HP29r&!qo2Y?k_>r=O&21`&6h_Sx#E;kq2Ot#^)I6TxR^2Lm5BW3iP4X-x4dph*G zm6EL0WgHQStu6)~e{wh2tBkuxN}e`;DpR8pkKD&Fm)Mta1e}IKjl)|Et;Qyf)CwjV zkPeeGRL>U?*#_-`_0#OH+8^nA;z!>_N-_;=+cGoY(3B$D=}mL@>(pjfZCk%evCrcJ z&uIHP?BLXa=%?HV6Lp6PoN{`SM8K5px4 z06A{wSth;olH1c3?@~tp92xRpW^+pix#O-i$m?O6I<%QK@99QH^!G_a5mi<-Wa`gE zzEx$`fchvk6%$k<3><;_n^f0_-b<{r5(4(e|ICEo-FaWJS52Vs7-pVVc5U=O(B%;n zv7{^76nDI(5CdMo4L}z>=DnZt{YJ#!;~_mwrpfX30Z_0@QJch%LX1fVD8|}N$ z9N>nSWT$qR^Q{>WfTx|6jt&uOCFxfnj6UV(edaeo{}m!Cd`%DeX#7VPoz>WcK<6LQ z^-L@MNLKzL>o41FvMqpzAe(q&>L>YY@!!*<)@KY!myRyFfXqBaH0EJT(-*!;(fo!w zbJ0>MXdoA4mZsRnjJ{g5U${6KbuK2G9(^7p90v)^Re!~rYW$rXub|1*azcCG!`X<- z0JsfOjz3>5x!8VHt5HK<(3JiZaY8YsjE z4#P-|aP9d+&hJ+~x3) zRoe}iD|SBMZC((X;TavAVgj57V|X*prE0P64W*hxeg&y88J~hekQ$+rMdeC*R;&Sn zgx6C89qHlcp@_WZ?QnH`UrBv4bP7TGG&Rtb9xeyMog-$X+=pG(yy-#ELtWJFls{L% zPG7OCf6C;n*PmYab_Xq25TBgp_tTPf3g;+@O>2dfO)Z!zl~Sem?k8imS&3FK{d%ffu4JAj zAjB0MS+yF#;WI_b8j*swkZ%2~-+wRp)VGRi>^J?MGjd<>nRdAr7r7&Dh`-Pv-FK>h zIQo;wPAv-XS-wED56lI^%4ogV%jy{%p)hIaAjY;Y9uVr$e&h4y)Uwd^W~3#Sm}JL6e>&pB)4%2Uz#8r-4KnR5f_ww?1ls1cbrIlQ*~F)F~(XH z5fO%=Ejw1~Pfjtv;bPV{EI26c+pUV(# e^8bzLQ2*ztuUWBv%zNP5C*%^w(uiz`yZ None: + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/feishu_spreadsheet.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/feishu_spreadsheet.yaml new file mode 100644 index 00000000000000..29e448d730f745 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/feishu_spreadsheet.yaml @@ -0,0 +1,36 @@ +identity: + author: Doug Lea + name: feishu_spreadsheet + label: + en_US: Feishu Spreadsheet + zh_Hans: 飞书电子表格 + description: + en_US: | + Feishu Spreadsheet, requires the following permissions: sheets:spreadsheet. + zh_Hans: | + 飞书电子表格，需要开通以下权限: sheets:spreadsheet。 + icon: icon.png + tags: + - social + - productivity +credentials_for_provider: + app_id: + type: text-input + required: true + label: + en_US: APP ID + placeholder: + en_US: Please input your feishu app id + zh_Hans: 请输入你的飞书 app id + help: + en_US: Get your app_id and app_secret from Feishu + zh_Hans: 从飞书获取您的 app_id 和 app_secret + url: https://open.larkoffice.com/app + app_secret: + type: secret-input + required: true + label: + en_US: APP Secret + placeholder: + en_US: Please input your app secret + zh_Hans: 请输入你的飞书 app secret diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.py new file mode 100644 index 00000000000000..44d062f9bdded2 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.py @@ -0,0 +1,22 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class AddColsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + sheet_id = tool_parameters.get("sheet_id") + sheet_name = tool_parameters.get("sheet_name") + length = tool_parameters.get("length") + values = tool_parameters.get("values") + + res = client.add_cols(spreadsheet_token, sheet_id, sheet_name, length, values) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.yaml new file mode 100644 index 00000000000000..ef457f8e009b2c --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_cols.yaml @@ -0,0 +1,72 @@ +identity: + name: add_cols + author: Doug Lea + label: + en_US: Add Cols + zh_Hans: 新增多列至工作表最后 +description: + human: + en_US: Add Cols + zh_Hans: 新增多列至工作表最后 + llm: A tool for adding multiple columns to the end of a spreadsheet. (新增多列至工作表最后) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: spreadsheet_token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 url。 + llm_description: 电子表格 token，支持输入电子表格 url。 + form: llm + + - name: sheet_id + type: string + required: false + label: + en_US: sheet_id + zh_Hans: 工作表 ID + human_description: + en_US: Sheet ID, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表 ID，与 sheet_name 二者其一必填。 + llm_description: 工作表 ID，与 sheet_name 二者其一必填。 + form: llm + + - name: sheet_name + type: string + required: false + label: + en_US: sheet_name + zh_Hans: 工作表名称 + human_description: + en_US: Sheet name, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表名称，与 sheet_id 二者其一必填。 + llm_description: 工作表名称，与 sheet_id 二者其一必填。 + form: llm + + - name: length + type: number + required: true + label: + en_US: length + zh_Hans: 要增加的列数 + human_description: + en_US: Number of columns to add, range (0-5000]. + zh_Hans: 要增加的列数，范围(0-5000]。 + llm_description: 要增加的列数，范围(0-5000]。 + form: llm + + - name: values + type: string + required: false + label: + en_US: values + zh_Hans: 新增列的单元格内容 + human_description: + en_US: | + Content of the new columns, array of objects in string format, each array represents a row of table data, format like: [ [ "ID","Name","Age" ],[ 1,"Zhang San",10 ],[ 2,"Li Si",11 ] ]. + zh_Hans: 新增列的单元格内容，数组对象字符串，每个数组一行表格数据，格式：[["编号","姓名","年龄"],[1,"张三",10],[2,"李四",11]]。 + llm_description: 新增列的单元格内容，数组对象字符串，每个数组一行表格数据，格式：[["编号","姓名","年龄"],[1,"张三",10],[2,"李四",11]]。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.py new file mode 100644 index 00000000000000..3a85b7b46ccb93 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.py @@ -0,0 +1,22 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class AddRowsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + sheet_id = tool_parameters.get("sheet_id") + sheet_name = tool_parameters.get("sheet_name") + length = tool_parameters.get("length") + values = tool_parameters.get("values") + + res = client.add_rows(spreadsheet_token, sheet_id, sheet_name, length, values) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.yaml new file mode 100644 index 00000000000000..37653325aeb16b --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/add_rows.yaml @@ -0,0 +1,72 @@ +identity: + name: add_rows + author: Doug Lea + label: + en_US: Add Rows + zh_Hans: 新增多行至工作表最后 +description: + human: + en_US: Add Rows + zh_Hans: 新增多行至工作表最后 + llm: A tool for adding multiple rows to the end of a spreadsheet. (新增多行至工作表最后) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: spreadsheet_token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 url。 + llm_description: 电子表格 token，支持输入电子表格 url。 + form: llm + + - name: sheet_id + type: string + required: false + label: + en_US: sheet_id + zh_Hans: 工作表 ID + human_description: + en_US: Sheet ID, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表 ID，与 sheet_name 二者其一必填。 + llm_description: 工作表 ID，与 sheet_name 二者其一必填。 + form: llm + + - name: sheet_name + type: string + required: false + label: + en_US: sheet_name + zh_Hans: 工作表名称 + human_description: + en_US: Sheet name, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表名称，与 sheet_id 二者其一必填。 + llm_description: 工作表名称，与 sheet_id 二者其一必填。 + form: llm + + - name: length + type: number + required: true + label: + en_US: length + zh_Hans: 要增加行数 + human_description: + en_US: Number of rows to add, range (0-5000]. + zh_Hans: 要增加行数，范围(0-5000]。 + llm_description: 要增加行数，范围(0-5000]。 + form: llm + + - name: values + type: string + required: false + label: + en_US: values + zh_Hans: 新增行的表格内容 + human_description: + en_US: | + Content of the new rows, array of objects in string format, each array represents a row of table data, format like: [ [ "ID","Name","Age" ],[ 1,"Zhang San",10 ],[ 2,"Li Si",11 ] ]. + zh_Hans: 新增行的表格内容，数组对象字符串，每个数组一行表格数据，格式,如：[["编号","姓名","年龄"],[1,"张三",10],[2,"李四",11]]。 + llm_description: 新增行的表格内容，数组对象字符串，每个数组一行表格数据，格式,如：[["编号","姓名","年龄"],[1,"张三",10],[2,"李四",11]]。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.py new file mode 100644 index 00000000000000..647364fab0a966 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.py @@ -0,0 +1,19 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class CreateSpreadsheetTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + title = tool_parameters.get("title") + folder_token = tool_parameters.get("folder_token") + + res = client.create_spreadsheet(title, folder_token) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.yaml new file mode 100644 index 00000000000000..931310e63172d4 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/create_spreadsheet.yaml @@ -0,0 +1,35 @@ +identity: + name: create_spreadsheet + author: Doug Lea + label: + en_US: Create Spreadsheet + zh_Hans: 创建电子表格 +description: + human: + en_US: Create Spreadsheet + zh_Hans: 创建电子表格 + llm: A tool for creating spreadsheets. (创建电子表格) +parameters: + - name: title + type: string + required: false + label: + en_US: Spreadsheet Title + zh_Hans: 电子表格标题 + human_description: + en_US: The title of the spreadsheet + zh_Hans: 电子表格的标题 + llm_description: 电子表格的标题 + form: llm + + - name: folder_token + type: string + required: false + label: + en_US: Folder Token + zh_Hans: 文件夹 token + human_description: + en_US: The token of the folder, supports folder URL input, e.g., https://bytedance.larkoffice.com/drive/folder/CxHEf4DCSlNkL2dUTCJcPRgentg + zh_Hans: 文件夹 token，支持文件夹 URL 输入，如：https://bytedance.larkoffice.com/drive/folder/CxHEf4DCSlNkL2dUTCJcPRgentg + llm_description: 文件夹 token，支持文件夹 URL 输入，如：https://bytedance.larkoffice.com/drive/folder/CxHEf4DCSlNkL2dUTCJcPRgentg + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.py new file mode 100644 index 00000000000000..dda8c59daffabf --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.py @@ -0,0 +1,19 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class GetSpreadsheetTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + user_id_type = tool_parameters.get("user_id_type", "open_id") + + res = client.get_spreadsheet(spreadsheet_token, user_id_type) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.yaml new file mode 100644 index 00000000000000..c519938617ba8c --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/get_spreadsheet.yaml @@ -0,0 +1,49 @@ +identity: + name: get_spreadsheet + author: Doug Lea + label: + en_US: Get Spreadsheet + zh_Hans: 获取电子表格信息 +description: + human: + en_US: Get Spreadsheet + zh_Hans: 获取电子表格信息 + llm: A tool for getting information from spreadsheets. (获取电子表格信息) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: Spreadsheet Token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 URL。 + llm_description: 电子表格 token，支持输入电子表格 URL。 + form: llm + + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.py new file mode 100644 index 00000000000000..98497791c0fa1e --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.py @@ -0,0 +1,18 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class ListSpreadsheetSheetsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + + res = client.list_spreadsheet_sheets(spreadsheet_token) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.yaml new file mode 100644 index 00000000000000..c6a7ef45d46589 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/list_spreadsheet_sheets.yaml @@ -0,0 +1,23 @@ +identity: + name: list_spreadsheet_sheets + author: Doug Lea + label: + en_US: List Spreadsheet Sheets + zh_Hans: 列出电子表格所有工作表 +description: + human: + en_US: List Spreadsheet Sheets + zh_Hans: 列出电子表格所有工作表 + llm: A tool for listing all sheets in a spreadsheet. (列出电子表格所有工作表) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: Spreadsheet Token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 URL。 + llm_description: 电子表格 token，支持输入电子表格 URL。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.py new file mode 100644 index 00000000000000..ebe3f619d091d1 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class ReadColsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + sheet_id = tool_parameters.get("sheet_id") + sheet_name = tool_parameters.get("sheet_name") + start_col = tool_parameters.get("start_col") + num_cols = tool_parameters.get("num_cols") + user_id_type = tool_parameters.get("user_id_type", "open_id") + + res = client.read_cols(spreadsheet_token, sheet_id, sheet_name, start_col, num_cols, user_id_type) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.yaml new file mode 100644 index 00000000000000..3273857b709bf9 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_cols.yaml @@ -0,0 +1,97 @@ +identity: + name: read_cols + author: Doug Lea + label: + en_US: Read Cols + zh_Hans: 读取工作表列数据 +description: + human: + en_US: Read Cols + zh_Hans: 读取工作表列数据 + llm: A tool for reading column data from a spreadsheet. (读取工作表列数据) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: spreadsheet_token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 url。 + llm_description: 电子表格 token，支持输入电子表格 url。 + form: llm + + - name: sheet_id + type: string + required: false + label: + en_US: sheet_id + zh_Hans: 工作表 ID + human_description: + en_US: Sheet ID, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表 ID，与 sheet_name 二者其一必填。 + llm_description: 工作表 ID，与 sheet_name 二者其一必填。 + form: llm + + - name: sheet_name + type: string + required: false + label: + en_US: sheet_name + zh_Hans: 工作表名称 + human_description: + en_US: Sheet name, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表名称，与 sheet_id 二者其一必填。 + llm_description: 工作表名称，与 sheet_id 二者其一必填。 + form: llm + + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form + + - name: start_col + type: number + required: false + label: + en_US: start_col + zh_Hans: 起始列号 + human_description: + en_US: Starting column number, starting from 1. + zh_Hans: 起始列号，从 1 开始。 + llm_description: 起始列号，从 1 开始。 + form: llm + + - name: num_cols + type: number + required: true + label: + en_US: num_cols + zh_Hans: 读取列数 + human_description: + en_US: Number of columns to read. + zh_Hans: 读取列数 + llm_description: 读取列数 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.py new file mode 100644 index 00000000000000..86b91b104b7029 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class ReadRowsTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + sheet_id = tool_parameters.get("sheet_id") + sheet_name = tool_parameters.get("sheet_name") + start_row = tool_parameters.get("start_row") + num_rows = tool_parameters.get("num_rows") + user_id_type = tool_parameters.get("user_id_type", "open_id") + + res = client.read_rows(spreadsheet_token, sheet_id, sheet_name, start_row, num_rows, user_id_type) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.yaml new file mode 100644 index 00000000000000..3e9206e8ef70fe --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_rows.yaml @@ -0,0 +1,97 @@ +identity: + name: read_rows + author: Doug Lea + label: + en_US: Read Rows + zh_Hans: 读取工作表行数据 +description: + human: + en_US: Read Rows + zh_Hans: 读取工作表行数据 + llm: A tool for reading row data from a spreadsheet. (读取工作表行数据) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: spreadsheet_token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 url。 + llm_description: 电子表格 token，支持输入电子表格 url。 + form: llm + + - name: sheet_id + type: string + required: false + label: + en_US: sheet_id + zh_Hans: 工作表 ID + human_description: + en_US: Sheet ID, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表 ID，与 sheet_name 二者其一必填。 + llm_description: 工作表 ID，与 sheet_name 二者其一必填。 + form: llm + + - name: sheet_name + type: string + required: false + label: + en_US: sheet_name + zh_Hans: 工作表名称 + human_description: + en_US: Sheet name, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表名称，与 sheet_id 二者其一必填。 + llm_description: 工作表名称，与 sheet_id 二者其一必填。 + form: llm + + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form + + - name: start_row + type: number + required: false + label: + en_US: start_row + zh_Hans: 起始行号 + human_description: + en_US: Starting row number, starting from 1. + zh_Hans: 起始行号，从 1 开始。 + llm_description: 起始行号，从 1 开始。 + form: llm + + - name: num_rows + type: number + required: true + label: + en_US: num_rows + zh_Hans: 读取行数 + human_description: + en_US: Number of rows to read. + zh_Hans: 读取行数 + llm_description: 读取行数 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.py b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.py new file mode 100644 index 00000000000000..ddd607d87838f4 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class ReadTableTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + spreadsheet_token = tool_parameters.get("spreadsheet_token") + sheet_id = tool_parameters.get("sheet_id") + sheet_name = tool_parameters.get("sheet_name") + num_range = tool_parameters.get("num_range") + query = tool_parameters.get("query") + user_id_type = tool_parameters.get("user_id_type", "open_id") + + res = client.read_table(spreadsheet_token, sheet_id, sheet_name, num_range, query, user_id_type) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.yaml b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.yaml new file mode 100644 index 00000000000000..e3dc80e1eb0b4f --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_spreadsheet/tools/read_table.yaml @@ -0,0 +1,122 @@ +identity: + name: read_table + author: Doug Lea + label: + en_US: Read Table + zh_Hans: 自定义读取电子表格行列数据 +description: + human: + en_US: Read Table + zh_Hans: 自定义读取电子表格行列数据 + llm: A tool for custom reading of row and column data from a spreadsheet. (自定义读取电子表格行列数据) +parameters: + - name: spreadsheet_token + type: string + required: true + label: + en_US: spreadsheet_token + zh_Hans: 电子表格 token + human_description: + en_US: Spreadsheet token, supports input of spreadsheet URL. + zh_Hans: 电子表格 token，支持输入电子表格 url。 + llm_description: 电子表格 token，支持输入电子表格 url。 + form: llm + + - name: sheet_id + type: string + required: false + label: + en_US: sheet_id + zh_Hans: 工作表 ID + human_description: + en_US: Sheet ID, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表 ID，与 sheet_name 二者其一必填。 + llm_description: 工作表 ID，与 sheet_name 二者其一必填。 + form: llm + + - name: sheet_name + type: string + required: false + label: + en_US: sheet_name + zh_Hans: 工作表名称 + human_description: + en_US: Sheet name, either sheet_id or sheet_name must be filled. + zh_Hans: 工作表名称，与 sheet_id 二者其一必填。 + llm_description: 工作表名称，与 sheet_id 二者其一必填。 + form: llm + + - name: user_id_type + type: select + required: false + options: + - value: open_id + label: + en_US: open_id + zh_Hans: open_id + - value: union_id + label: + en_US: union_id + zh_Hans: union_id + - value: user_id + label: + en_US: user_id + zh_Hans: user_id + default: "open_id" + label: + en_US: user_id_type + zh_Hans: 用户 ID 类型 + human_description: + en_US: User ID type, optional values are open_id, union_id, user_id, with a default value of open_id. + zh_Hans: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + llm_description: 用户 ID 类型，可选值有 open_id、union_id、user_id，默认值为 open_id。 + form: form + + - name: start_row + type: number + required: false + label: + en_US: start_row + zh_Hans: 起始行号 + human_description: + en_US: Starting row number, starting from 1. + zh_Hans: 起始行号，从 1 开始。 + llm_description: 起始行号，从 1 开始。 + form: llm + + - name: num_rows + type: number + required: false + label: + en_US: num_rows + zh_Hans: 读取行数 + human_description: + en_US: Number of rows to read. + zh_Hans: 读取行数 + llm_description: 读取行数 + form: llm + + - name: range + type: string + required: false + label: + en_US: range + zh_Hans: 取数范围 + human_description: + en_US: | + Data range, format like: A1:B2, can be empty when query=all. + zh_Hans: 取数范围,格式如:A1:B2，query=all 时可为空。 + llm_description: 取数范围,格式如:A1:B2，query=all 时可为空。 + form: llm + + - name: query + type: string + required: false + label: + en_US: query + zh_Hans: 查询 + human_description: + en_US: Pass "all" to query all data in the table, but no more than 100 columns. + zh_Hans: 传 all，表示查询表格所有数据，但最多查询 100 列数据。 + llm_description: 传 all，表示查询表格所有数据，但最多查询 100 列数据。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_task/_assets/icon.png b/api/core/tools/provider/builtin/feishu_task/_assets/icon.png new file mode 100644 index 0000000000000000000000000000000000000000..3485be0d0fbd85444995dca78b51634899807537 GIT binary patch literal 3917 zcmeHK`8!nY|35Rv$S@chMwYTYii(M{N3v9;%}$eNRLVAE$sB}1Y7mMPO;kkLL)K0@ zG`2^jO=OvlD0z^vt1O@MT+jC(`2P60uFntW+}CyA@7L@7y3Tno_dD^ZgDsJ;i2wiq z@!$cg+S5TCNFY4gu9+HkUNu91z`Dra zn3cS--JiTCJttvR&menYC;#MG7C&+c`aK=BZR+>*>g%6^Z=3aOmiJ@TulZyD_>>XC zl-`2D=pMGfifRJLrR4Uv2LX5<5LLeeuw`~C+iGHf_d@E|m{`D}9#qW10rryDELubX z!{IMMq0)e^*?%Q)LHc``oXM3Y*0N()fBE&rG+tIJ`KiB1yhwnX$8zY}=>BIz^Um1E zw40?jys$$>?HMvvv*na*Sv5@?T3lv3{Uy}*adNd`M3UZU541ZHUw_#hYw)ni^o;vd ziE7wioKS-QS~fY-d*O4*lB}%5-fg-~iz-058GR3j8f#Ffd||*Qg6?1U;b*DoroI#f zzpjZox`jA$yvLgKrZ;rLwubzdKI_Re^q}`RG7{ zo7^>1+8G`Y=)A5ORr3jPvJyugcr+Sa>6WOFM#xKB^?ju45K|bVvW+jBm{Jk=xjM=4J5@_5f7yZgA;O zeYV}#4zCm(>|t!}*UGOJ8dx7#DY-j+hlsp$+tsXr5jY=`gfr>h>AqN`Oi@$vzP4?< zRVEIe*qBeg{^?=-!_|bJ8C|BA!v0F{JR<`zKJjuDx(5!tCRJG!psd?!b23Uv*>$x; z&&^pY&HNsa)Ma!^woeBmD4uWgp$Dw54N6X@xD#P5`RWcn4la2k;Z)nR0;k=~vR9H2_*d;8uYk~OcyStEIgnF=EO;P4d1DYk(q z>@@QObS2Ug_6|J7fn>`IJOwi^7if#~xm~$)A7%2e{6`qSVd2Wa^`J4QjuWDcC+wP; zIM@5*Ryc@Yr1d35nt2l3H~H}*a-oY6QoD=)9q3Jweh1UPw`aCKba4!1pZ*IA$rxO{GBN!wV9IVbEb~alZacgzy~#zfdK5ZY96h zgYLsS;Br5O+hFyIK~~yWd`qXsj>tG674I@U&1Hkn24hdOa!u%h!nC?DxM<`RnF`Lq z9@>bf;M-*i^b;u~G6TVoed2sb@DXe10u`y@8PQ-5W-QgfiHE$Bsn9|Sk;6n3)mnnD zKM$Vx$8@F`2-M>Dk>U+vpx10}s(~{P3Bc1dVPt4Xy>jKM7|4yU+vO!l4U|?}eBIyX zyFN9cW9oP=|D7GXF7$Nwkqs-KzE!y#9p9{Xab;qm$ho9PVGZ*6MAX~-eEUbw^n5m=jy0*VZAL@O{_Gmcf-@3gf35 z&)=MYECgB)xVrp9?DhHFXTB*pR8^&fOt-pjO<8X z&7#T=m*JN4{rraZPb~^m^gq)L^IS?h(&e+V@5XoqZhJCNt%1rtdlr@3G*q2)tGB?X zm3D%o|5W4e`M$LLeo_$>T^N{|u@bvc`eqSp&Q>QLIFI!aBJ=G|| zkC%Qlf9$W}lgq6&%}+=nJrtutsUQhES5jnGdHK9fG#GktBo1{%FJ$qjPHSP59h#zA z9h&;&?YTFz@s+NAqq*Yqd}FK*n5=$OUYB#XRgABi^DWZ;8;dq(B3Q%{pUcGe8oX5f zy&K2)Dp$F9w(lqp*%8;dl!YI55*mcByhj#@)g95q7bS%HeVk?wT*fq&$SaYk$8ee_ z6wY2>-IP=(3(KR?MzU{e*QYQ7*m_ptw9mT9lf6u+2K#yg6-@S;HN!*PM3bm%`3GJ} z$0UnF)=86*;^18knhA1gn)j~{WGFLT4i;7Ymx{gAzNP^=TM-60p1Ioz!|L=u(DKq-Er;*Q&A{W zC{==)f=XyO?0aVe?IANC=PC4s>^Idu5w_Z=fVglyNF%_`b6{nt{>~L1piAt6(+I^ zs#O=8Dnjap8bm z1zA{-S|1a<_SVE^Qc56N;nZ(R%V2{eX zM(f?xAD=|nycC?Et!zUZM%RM^V_DcEXjafbKNVX?XBxi;pv)1jTvOwqbT;X_OK4sI ztc4J1rvagQcbT|4_rSlZ0ccg~SIavO)p&We+1tTwKX-`m&-oINvg1(DaM$~Q^9xp& zB$6pPW@fA9cM!77N-cxP0K*n_Pe|^Ijn1W@{|tVm(?# z3cr6e$w(5*_|PLP;LNf??TrGPlv`u4$c=|HoMr_x8j!Z_m^WIOMIm74(ZV(rJR`Q8 z1g2Q7_a-3WDCxsP zh>i{QwZcp|!qG;Y&vr`o^8-Z6ZCSoSvmC|jJ#F@K%lm59*TFnfMStWWH)_rA{;ehF3z)m(XvrB-%^G}rq6q4v7#a#v?TM`~9P7M-17VXsJOCOSRs l9H-vG{cq?0|1|{WeK*e(NZd%(N9SYUptXZlxusX!e*o|W00aO4 literal 0 HcmV?d00001 diff --git a/api/core/tools/provider/builtin/feishu_task/feishu_task.py b/api/core/tools/provider/builtin/feishu_task/feishu_task.py new file mode 100644 index 00000000000000..6df05968d8f176 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/feishu_task.py @@ -0,0 +1,7 @@ +from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController +from core.tools.utils.feishu_api_utils import auth + + +class FeishuTaskProvider(BuiltinToolProviderController): + def _validate_credentials(self, credentials: dict) -> None: + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_task/feishu_task.yaml b/api/core/tools/provider/builtin/feishu_task/feishu_task.yaml new file mode 100644 index 00000000000000..88736f79a02e87 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/feishu_task.yaml @@ -0,0 +1,36 @@ +identity: + author: Doug Lea + name: feishu_task + label: + en_US: Feishu Task + zh_Hans: 飞书任务 + description: + en_US: | + Feishu Task, requires the following permissions: task:task:write、contact:user.id:readonly. + zh_Hans: | + 飞书任务，需要开通以下权限: task:task:write、contact:user.id:readonly。 + icon: icon.png + tags: + - social + - productivity +credentials_for_provider: + app_id: + type: text-input + required: true + label: + en_US: APP ID + placeholder: + en_US: Please input your feishu app id + zh_Hans: 请输入你的飞书 app id + help: + en_US: Get your app_id and app_secret from Feishu + zh_Hans: 从飞书获取您的 app_id 和 app_secret + url: https://open.larkoffice.com/app + app_secret: + type: secret-input + required: true + label: + en_US: APP Secret + placeholder: + en_US: Please input your app secret + zh_Hans: 请输入你的飞书 app secret diff --git a/api/core/tools/provider/builtin/feishu_task/tools/add_members.py b/api/core/tools/provider/builtin/feishu_task/tools/add_members.py new file mode 100644 index 00000000000000..e58ed22e0f4797 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/add_members.py @@ -0,0 +1,20 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class AddMembersTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + task_guid = tool_parameters.get("task_guid") + member_phone_or_email = tool_parameters.get("member_phone_or_email") + member_role = tool_parameters.get("member_role", "follower") + + res = client.add_members(task_guid, member_phone_or_email, member_role) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_task/tools/add_members.yaml b/api/core/tools/provider/builtin/feishu_task/tools/add_members.yaml new file mode 100644 index 00000000000000..063c0f7f04956c --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/add_members.yaml @@ -0,0 +1,58 @@ +identity: + name: add_members + author: Doug Lea + label: + en_US: Add Members + zh_Hans: 添加任务成员 +description: + human: + en_US: Add Members + zh_Hans: 添加任务成员 + llm: A tool for adding members to a Feishu task.(添加任务成员) +parameters: + - name: task_guid + type: string + required: true + label: + en_US: Task GUID + zh_Hans: 任务 GUID + human_description: + en_US: | + The GUID of the task to be added, supports passing either the Task ID or the Task link URL. Example of Task ID: 8b5425ec-9f2a-43bd-a3ab-01912f50282b; Example of Task link URL: https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + zh_Hans: 要添加的任务的 GUID，支持传任务 ID 和任务链接 URL。任务 ID 示例：8b5425ec-9f2a-43bd-a3ab-01912f50282b；任务链接 URL 示例：https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + llm_description: 要添加的任务的 GUID，支持传任务 ID 和任务链接 URL。任务 ID 示例：8b5425ec-9f2a-43bd-a3ab-01912f50282b；任务链接 URL 示例：https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + form: llm + + - name: member_phone_or_email + type: string + required: true + label: + en_US: Task Member Phone Or Email + zh_Hans: 任务成员的电话或邮箱 + human_description: + en_US: A list of member emails or phone numbers, separated by commas. + zh_Hans: 任务成员邮箱或者手机号列表，使用逗号分隔。 + llm_description: 任务成员邮箱或者手机号列表，使用逗号分隔。 + form: llm + + - name: member_role + type: select + required: true + options: + - value: assignee + label: + en_US: assignee + zh_Hans: 负责人 + - value: follower + label: + en_US: follower + zh_Hans: 关注人 + default: "follower" + label: + en_US: member_role + zh_Hans: 成员的角色 + human_description: + en_US: Member role, optional values are "assignee" (responsible person) and "follower" (observer), with a default value of "assignee". + zh_Hans: 成员的角色，可选值有 "assignee"（负责人）和 "follower"（关注人），默认值为 "assignee"。 + llm_description: 成员的角色，可选值有 "assignee"（负责人）和 "follower"（关注人），默认值为 "assignee"。 + form: form diff --git a/api/core/tools/provider/builtin/feishu_task/tools/create_task.py b/api/core/tools/provider/builtin/feishu_task/tools/create_task.py new file mode 100644 index 00000000000000..96cdcd71f6d2ec --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/create_task.py @@ -0,0 +1,22 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class CreateTaskTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + summary = tool_parameters.get("summary") + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + completed_time = tool_parameters.get("completed_time") + description = tool_parameters.get("description") + + res = client.create_task(summary, start_time, end_time, completed_time, description) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_task/tools/create_task.yaml b/api/core/tools/provider/builtin/feishu_task/tools/create_task.yaml new file mode 100644 index 00000000000000..7eb4af168bf740 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/create_task.yaml @@ -0,0 +1,74 @@ +identity: + name: create_task + author: Doug Lea + label: + en_US: Create Task + zh_Hans: 创建飞书任务 +description: + human: + en_US: Create Feishu Task + zh_Hans: 创建飞书任务 + llm: A tool for creating tasks in Feishu.(创建飞书任务) +parameters: + - name: summary + type: string + required: true + label: + en_US: Task Title + zh_Hans: 任务标题 + human_description: + en_US: The title of the task. + zh_Hans: 任务标题 + llm_description: 任务标题 + form: llm + + - name: description + type: string + required: false + label: + en_US: Task Description + zh_Hans: 任务备注 + human_description: + en_US: The description or notes for the task. + zh_Hans: 任务备注 + llm_description: 任务备注 + form: llm + + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 任务开始时间 + human_description: + en_US: | + The start time of the task, in the format: 2006-01-02 15:04:05 + zh_Hans: 任务开始时间，格式为：2006-01-02 15:04:05 + llm_description: 任务开始时间，格式为：2006-01-02 15:04:05 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 任务结束时间 + human_description: + en_US: | + The end time of the task, in the format: 2006-01-02 15:04:05 + zh_Hans: 任务结束时间，格式为：2006-01-02 15:04:05 + llm_description: 任务结束时间，格式为：2006-01-02 15:04:05 + form: llm + + - name: completed_time + type: string + required: false + label: + en_US: Completed Time + zh_Hans: 任务完成时间 + human_description: + en_US: | + The completion time of the task, in the format: 2006-01-02 15:04:05. Leave empty to create an incomplete task; fill in a specific time to create a completed task. + zh_Hans: 任务完成时间，格式为：2006-01-02 15:04:05，不填写表示创建一个未完成任务；填写一个具体的时间表示创建一个已完成任务。 + llm_description: 任务完成时间，格式为：2006-01-02 15:04:05，不填写表示创建一个未完成任务；填写一个具体的时间表示创建一个已完成任务。 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_task/tools/delete_task.py b/api/core/tools/provider/builtin/feishu_task/tools/delete_task.py new file mode 100644 index 00000000000000..dee036fee5203a --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/delete_task.py @@ -0,0 +1,18 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class UpdateTaskTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + task_guid = tool_parameters.get("task_guid") + + res = client.delete_task(task_guid) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_task/tools/delete_task.yaml b/api/core/tools/provider/builtin/feishu_task/tools/delete_task.yaml new file mode 100644 index 00000000000000..d3f97413676624 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/delete_task.yaml @@ -0,0 +1,24 @@ +identity: + name: delete_task + author: Doug Lea + label: + en_US: Delete Task + zh_Hans: 删除飞书任务 +description: + human: + en_US: Delete Task + zh_Hans: 删除飞书任务 + llm: A tool for deleting tasks in Feishu.(删除飞书任务) +parameters: + - name: task_guid + type: string + required: true + label: + en_US: Task GUID + zh_Hans: 任务 GUID + human_description: + en_US: | + The GUID of the task to be deleted, supports passing either the Task ID or the Task link URL. Example of Task ID: 8b5425ec-9f2a-43bd-a3ab-01912f50282b; Example of Task link URL: https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + zh_Hans: 要删除的任务的 GUID，支持传任务 ID 和任务链接 URL。任务 ID 示例：8b5425ec-9f2a-43bd-a3ab-01912f50282b；任务链接 URL 示例：https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + llm_description: 要删除的任务的 GUID，支持传任务 ID 和任务链接 URL。任务 ID 示例：8b5425ec-9f2a-43bd-a3ab-01912f50282b；任务链接 URL 示例：https://applink.feishu-pre.net/client/todo/detail?guid=8c6bf822-e4da-449a-b82a-dc44020f9be9&suite_entity_num=t21587362 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_task/tools/update_task.py b/api/core/tools/provider/builtin/feishu_task/tools/update_task.py new file mode 100644 index 00000000000000..4a48cd283abf1d --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/update_task.py @@ -0,0 +1,23 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class UpdateTaskTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + task_guid = tool_parameters.get("task_guid") + summary = tool_parameters.get("summary") + start_time = tool_parameters.get("start_time") + end_time = tool_parameters.get("end_time") + completed_time = tool_parameters.get("completed_time") + description = tool_parameters.get("description") + + res = client.update_task(task_guid, summary, start_time, end_time, completed_time, description) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_task/tools/update_task.yaml b/api/core/tools/provider/builtin/feishu_task/tools/update_task.yaml new file mode 100644 index 00000000000000..83c9bcb1c443ac --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_task/tools/update_task.yaml @@ -0,0 +1,89 @@ +identity: + name: update_task + author: Doug Lea + label: + en_US: Update Task + zh_Hans: 更新飞书任务 +description: + human: + en_US: Update Feishu Task + zh_Hans: 更新飞书任务 + llm: A tool for updating tasks in Feishu.(更新飞书任务) +parameters: + - name: task_guid + type: string + required: true + label: + en_US: Task GUID + zh_Hans: 任务 GUID + human_description: + en_US: | + The task ID, supports inputting either the Task ID or the Task link URL. Example of Task ID: 42cad8a0-f8c8-4344-9be2-d1d7e8e91b64; Example of Task link URL: https://applink.feishu-pre.net/client/todo/detail?guid=42cad8a0-f8c8-4344-9be2-d1d7e8e91b64&suite_entity_num=t21700217 + zh_Hans: | + 任务ID，支持传入任务 ID 和任务链接 URL。任务 ID 示例: 42cad8a0-f8c8-4344-9be2-d1d7e8e91b64；任务链接 URL 示例: https://applink.feishu-pre.net/client/todo/detail?guid=42cad8a0-f8c8-4344-9be2-d1d7e8e91b64&suite_entity_num=t21700217 + llm_description: | + 任务ID，支持传入任务 ID 和任务链接 URL。任务 ID 示例: 42cad8a0-f8c8-4344-9be2-d1d7e8e91b64；任务链接 URL 示例: https://applink.feishu-pre.net/client/todo/detail?guid=42cad8a0-f8c8-4344-9be2-d1d7e8e91b64&suite_entity_num=t21700217 + form: llm + + - name: summary + type: string + required: true + label: + en_US: Task Title + zh_Hans: 任务标题 + human_description: + en_US: The title of the task. + zh_Hans: 任务标题 + llm_description: 任务标题 + form: llm + + - name: description + type: string + required: false + label: + en_US: Task Description + zh_Hans: 任务备注 + human_description: + en_US: The description or notes for the task. + zh_Hans: 任务备注 + llm_description: 任务备注 + form: llm + + - name: start_time + type: string + required: false + label: + en_US: Start Time + zh_Hans: 任务开始时间 + human_description: + en_US: | + The start time of the task, in the format: 2006-01-02 15:04:05 + zh_Hans: 任务开始时间，格式为：2006-01-02 15:04:05 + llm_description: 任务开始时间，格式为：2006-01-02 15:04:05 + form: llm + + - name: end_time + type: string + required: false + label: + en_US: End Time + zh_Hans: 任务结束时间 + human_description: + en_US: | + The end time of the task, in the format: 2006-01-02 15:04:05 + zh_Hans: 任务结束时间，格式为：2006-01-02 15:04:05 + llm_description: 任务结束时间，格式为：2006-01-02 15:04:05 + form: llm + + - name: completed_time + type: string + required: false + label: + en_US: Completed Time + zh_Hans: 任务完成时间 + human_description: + en_US: | + The completion time of the task, in the format: 2006-01-02 15:04:05 + zh_Hans: 任务完成时间，格式为：2006-01-02 15:04:05 + llm_description: 任务完成时间，格式为：2006-01-02 15:04:05 + form: llm diff --git a/api/core/tools/provider/builtin/feishu_wiki/_assets/icon.png b/api/core/tools/provider/builtin/feishu_wiki/_assets/icon.png new file mode 100644 index 0000000000000000000000000000000000000000..878672c9ae5a511f246d3211863ec786596481eb GIT binary patch literal 4252 zcmeHLdo+|?+t=Mqw_}5n^I?!j4%5*oatPzljD$2|kVK9XVVn&i#W-|O8RUs3$w(3z z!-NtsJD3Cnw*9V-s%0`KfZsy_5JmIYrSi|f9!kjeeLV_yRLOz*WPRIJL4=G zDJmo@#K*@cYHMS8j@Pq)H%Nf@J>|m6<8`~j&mk@NC=7WPA0O680MMDe|5vk0p+;|VF?MNmt_+19WtP?o z)~Kxr3F51S!y1^ShKNEJ&l$Dum@1#!w~h2-|0=A+y}e8huwqgbm*ook6m2hl&(RxM zxT?>tiDyMbE94GMzUNmcH2ITa zmLUy4C#YX=jWXmE=*(O_$yHr?l-rjk!OZhH3AV86?uc;;#dI07_efFGAHeO7uiFVj zLtSJqC4cgZ7jE!fi>nPJF`^+Y=PPCGCcW=(Zhu>81A5fsl4Nd;AW*Yp)X)0ig@)_H z*}vu!DoS;WV^PP$Smvaq-tE+Ah0e+T264jTC@ZR-J>u%-8pFwH76o`h#Fj7XIo1aA zZIoZpJsNv#IG!{5hk~OS(}=K)jTyJy%vQo7oxb(npvP^as*efX8nLfhEsM8bPWgh@ zE$bIIC<_n|ECd-G+8P2ouzeR@E?&>L@3l}0;aRtqW6CFnnj{^ls+wRnUDG*3xzmfv7EwX&VG;%WAAX2PW-o!_+<-HDK6)lfLWE$Ic?}6jGlOCQ5 z1_srwnXfl*|CrVA`gL8FFdNEu-ljMSlOb-fH4RU>0^t`|=aTyi;(HYWuXKcB`3cPr z#Z{FLE5jBfH@~NwtYM;ucN4Cnp*HdL?EE#IuGys2c3p7;3O+Nv%gxTRoAAXvhGGg%6#!^96V2xj$<_cy6k4Sk?_ zgN%~Yk0mYu)@Z8m3fj1b8>|k7)UIe3?I!4%@~~nGmSUBX{|c;d3(Vr5*0i^fb4bVI9W(wm6=B2&G4 z2=!pYUM-2Pe!h@Yqmo1)xtnKc_dm{aa_Na{PG6&*RS+ZnoQg=2n7r#&y}lhDc`H5a zolQdxkLVB?LGG(!+2XUV7R=7SOUnhT#HSWadCAVh?oVIaOnTW$gkyA?#@iM& z*Lw}4id$3HJ@F1#plL(m1Rn%-Yny96pL8DWBrm})W7SqUW`x^zB8+~7H>R(Q;$HaH;Y zQeIU*#v7CTmr~b6xn%z0z@}>gwx|ngY(Zlb)98-%j}+u1zttg3o*gR=(nekei2+AJ zV#@%7Sra}RvC`6Nq+i#RG4Xa@Ejri*Bu-`XAfs~f=*rd*E~6}cX*`Q(tpjEtt<8>h>%XFOaC(A>x0W#pSRyDs>WEj zF9{$|$Nnd!qyZrJeOtIvau~36**hAQ66(XEjI^+hnv=F*$JcVs)0hD1a|cK5)80*bXDj3I!>YHHc=e|Go7K=DPD)I|?tZte+DTgT;*nItCtk2& z=OfD!c#dgdAL^j_Ou1_mkjbnPCi}b8pRTMH}OUZWDLCiCD!4< zoM_&nRtAPXS|mDX<5VirKA1CQ9}`dcCIHKZ_?3_g9l21Wgk4q^Pj#U@IB;|8BDJu( zjOO!II4f7%SA|z7EG{U9k6kp`e%4=nt@pL&m<-4Z!mS5ndUi8+D7J}{()Xn!G`?|U z--5_L>#WO(iIe>z<5@FgIG-#3*~)b|9CBCb8gRHrrW5QrUBqsM1Bw;<9DFOA94ViZ z==)pTx12YG8do3AJk=iyFO6N+Iq;^kNb;CF@2q_qqvA<;`@qHFi3-ld9Z5P7N}!S> z(b=eVcCsmzeN$m+qZL*U&lxy#r;h%K%3ee6ndqI2@H^4fS;S934VWhK^xvSeP-oIS z9xQe~f6Rf^12+dyxy|$xkM$J>DlFlHA;qTZO%x1Y3u>7(i4gJ^g&h{$nh-A*&)<6a zvV)g^cb{Me%EgAY@u&SPJUX^UodqR?1qi79FXRb^_b5+)!s=y-`@7HGr<3fenu1~Y zRS_*xi8f((JujrQV{MO26y7T2g*=B{5qDB~S$;HMgKaj&QEc|>T5Y9ykDtFahXje> z`9=0oenp$OD>J)&LfCf)cds601V-`_@Hvbh+w`o~=hM=bcN3IQ7MwxXQXlwO|BByw zYW?78CjQC^MFQPMeh3T6w^vN}KyjjnvoBI!48ZWdBw5*1x_+{v`*7Z3Cu`fm6i!XM zmD7I|3VQAiaW}$5LXK=C7@=a@5Wr}mg<;46t90}59Q*$XB*k^MlsXtyEsApj;bPPH)h}l@5IR2PNi!}y*GID$8%WZS)XeLU z3$(d}Ybv-xEX6-$hb4DCeKUR)ZTuaJ=@H*A(n-l}o?h!~U_9!es@K`ZH%H@d6?P+z zfE0RK4;kXbd&;{G(RcWsy-Z0jI4eEJ``y52YlXI?Sa>D>19*6ea{vGU literal 0 HcmV?d00001 diff --git a/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.py b/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.py new file mode 100644 index 00000000000000..6c5fccb1a31d0d --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.py @@ -0,0 +1,7 @@ +from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController +from core.tools.utils.feishu_api_utils import auth + + +class FeishuWikiProvider(BuiltinToolProviderController): + def _validate_credentials(self, credentials: dict) -> None: + auth(credentials) diff --git a/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.yaml b/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.yaml new file mode 100644 index 00000000000000..1fb5f71cbc5169 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_wiki/feishu_wiki.yaml @@ -0,0 +1,36 @@ +identity: + author: Doug Lea + name: feishu_wiki + label: + en_US: Feishu Wiki + zh_Hans: 飞书知识库 + description: + en_US: | + Feishu Wiki, requires the following permissions: wiki:wiki:readonly. + zh_Hans: | + 飞书知识库，需要开通以下权限: wiki:wiki:readonly。 + icon: icon.png + tags: + - social + - productivity +credentials_for_provider: + app_id: + type: text-input + required: true + label: + en_US: APP ID + placeholder: + en_US: Please input your feishu app id + zh_Hans: 请输入你的飞书 app id + help: + en_US: Get your app_id and app_secret from Feishu + zh_Hans: 从飞书获取您的 app_id 和 app_secret + url: https://open.larkoffice.com/app + app_secret: + type: secret-input + required: true + label: + en_US: APP Secret + placeholder: + en_US: Please input your app secret + zh_Hans: 请输入你的飞书 app secret diff --git a/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.py b/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.py new file mode 100644 index 00000000000000..374b4c9a7d1492 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.py @@ -0,0 +1,21 @@ +from typing import Any + +from core.tools.entities.tool_entities import ToolInvokeMessage +from core.tools.tool.builtin_tool import BuiltinTool +from core.tools.utils.feishu_api_utils import FeishuRequest + + +class GetWikiNodesTool(BuiltinTool): + def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage: + app_id = self.runtime.credentials.get("app_id") + app_secret = self.runtime.credentials.get("app_secret") + client = FeishuRequest(app_id, app_secret) + + space_id = tool_parameters.get("space_id") + parent_node_token = tool_parameters.get("parent_node_token") + page_token = tool_parameters.get("page_token") + page_size = tool_parameters.get("page_size") + + res = client.get_wiki_nodes(space_id, parent_node_token, page_token, page_size) + + return self.create_json_message(res) diff --git a/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.yaml b/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.yaml new file mode 100644 index 00000000000000..7d6ac3c8248772 --- /dev/null +++ b/api/core/tools/provider/builtin/feishu_wiki/tools/get_wiki_nodes.yaml @@ -0,0 +1,63 @@ +identity: + name: get_wiki_nodes + author: Doug Lea + label: + en_US: Get Wiki Nodes + zh_Hans: 获取知识空间子节点列表 +description: + human: + en_US: | + Get the list of child nodes in Wiki, make sure the app/bot is a member of the wiki space. See How to add an app as a wiki base administrator (member). https://open.feishu.cn/document/server-docs/docs/wiki-v2/wiki-qa + zh_Hans: | + 获取知识库全部子节点列表，请确保应用/机器人为知识空间成员。参阅如何将应用添加为知识库管理员（成员）。https://open.feishu.cn/document/server-docs/docs/wiki-v2/wiki-qa + llm: A tool for getting all sub-nodes of a knowledge base.(获取知识空间子节点列表) +parameters: + - name: space_id + type: string + required: true + label: + en_US: Space Id + zh_Hans: 知识空间 ID + human_description: + en_US: | + The ID of the knowledge space. Supports space link URL, for example: https://svi136aogf123.feishu.cn/wiki/settings/7166950623940706332 + zh_Hans: 知识空间 ID，支持空间链接 URL，例如：https://svi136aogf123.feishu.cn/wiki/settings/7166950623940706332 + llm_description: 知识空间 ID，支持空间链接 URL，例如：https://svi136aogf123.feishu.cn/wiki/settings/7166950623940706332 + form: llm + + - name: page_size + type: number + required: false + default: 10 + label: + en_US: Page Size + zh_Hans: 分页大小 + human_description: + en_US: The size of each page, with a maximum value of 50. + zh_Hans: 分页大小，最大值 50。 + llm_description: 分页大小，最大值 50。 + form: llm + + - name: page_token + type: string + required: false + label: + en_US: Page Token + zh_Hans: 分页标记 + human_description: + en_US: The pagination token. Leave empty for the first request to start from the beginning; if the paginated query result has more items, a new page_token will be returned, which can be used to get the next set of results. + zh_Hans: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + llm_description: 分页标记，第一次请求不填，表示从头开始遍历；分页查询结果还有更多项时会同时返回新的 page_token，下次遍历可采用该 page_token 获取查询结果。 + form: llm + + - name: parent_node_token + type: string + required: false + label: + en_US: Parent Node Token + zh_Hans: 父节点 token + human_description: + en_US: The token of the parent node. + zh_Hans: 父节点 token + llm_description: 父节点 token + form: llm diff --git a/api/core/tools/utils/feishu_api_utils.py b/api/core/tools/utils/feishu_api_utils.py index ffdb06498fd519..ce1fd7dc19e808 100644 --- a/api/core/tools/utils/feishu_api_utils.py +++ b/api/core/tools/utils/feishu_api_utils.py @@ -1,9 +1,23 @@ import httpx +from core.tools.errors import ToolProviderCredentialValidationError from extensions.ext_redis import redis_client +def auth(credentials): + app_id = credentials.get("app_id") + app_secret = credentials.get("app_secret") + if not app_id or not app_secret: + raise ToolProviderCredentialValidationError("app_id and app_secret is required") + try: + assert FeishuRequest(app_id, app_secret).tenant_access_token is not None + except Exception as e: + raise ToolProviderCredentialValidationError(str(e)) + + class FeishuRequest: + API_BASE_URL = "https://lark-plugin-api.solutionsuite.cn/lark-plugin" + def __init__(self, app_id: str, app_secret: str): self.app_id = app_id self.app_secret = app_secret @@ -42,7 +56,7 @@ def get_tenant_access_token(self, app_id: str, app_secret: str) -> dict: "expire": 7200 } """ - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/access_token/get_tenant_access_token" + url = f"{self.API_BASE_URL}/access_token/get_tenant_access_token" payload = {"app_id": app_id, "app_secret": app_secret} res = self._send_request(url, require_token=False, payload=payload) return res @@ -63,7 +77,7 @@ def create_document(self, title: str, content: str, folder_token: str) -> dict: "msg": "创建飞书文档成功，请查看" } """ - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/document/create_document" + url = f"{self.API_BASE_URL}/document/create_document" payload = { "title": title, "content": content, @@ -72,13 +86,13 @@ def create_document(self, title: str, content: str, folder_token: str) -> dict: res = self._send_request(url, payload=payload) return res.get("data") - def write_document(self, document_id: str, content: str, position: str = "start") -> dict: - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/document/write_document" + def write_document(self, document_id: str, content: str, position: str = "end") -> dict: + url = f"{self.API_BASE_URL}/document/write_document" payload = {"document_id": document_id, "content": content, "position": position} res = self._send_request(url, payload=payload) return res - def get_document_content(self, document_id: str, mode: str, lang: int = 0) -> dict: + def get_document_content(self, document_id: str, mode: str = "markdown", lang: str = "0") -> dict: """ API url: https://open.larkoffice.com/document/server-docs/docs/docs/docx-v1/document/raw_content Example Response: @@ -95,45 +109,404 @@ def get_document_content(self, document_id: str, mode: str, lang: int = 0) -> di "mode": mode, "lang": lang, } - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/document/get_document_content" - res = self._send_request(url, method="get", params=params) + url = f"{self.API_BASE_URL}/document/get_document_content" + res = self._send_request(url, method="GET", params=params) return res.get("data").get("content") - def list_document_blocks(self, document_id: str, page_token: str, page_size: int = 500) -> dict: + def list_document_blocks( + self, document_id: str, page_token: str, user_id_type: str = "open_id", page_size: int = 500 + ) -> dict: """ API url: https://open.larkoffice.com/document/server-docs/docs/docs/docx-v1/document/list """ - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/document/list_document_blocks" params = { + "user_id_type": user_id_type, "document_id": document_id, "page_size": page_size, "page_token": page_token, } - res = self._send_request(url, method="get", params=params) + url = f"{self.API_BASE_URL}/document/list_document_blocks" + res = self._send_request(url, method="GET", params=params) return res.get("data") def send_bot_message(self, receive_id_type: str, receive_id: str, msg_type: str, content: str) -> dict: """ API url: https://open.larkoffice.com/document/server-docs/im-v1/message/create """ - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/message/send_bot_message" + url = f"{self.API_BASE_URL}/message/send_bot_message" params = { "receive_id_type": receive_id_type, } payload = { "receive_id": receive_id, "msg_type": msg_type, - "content": content, + "content": content.strip('"').replace(r"\"", '"').replace(r"\\", "\\"), } res = self._send_request(url, params=params, payload=payload) return res.get("data") def send_webhook_message(self, webhook: str, msg_type: str, content: str) -> dict: - url = "https://lark-plugin-api.solutionsuite.cn/lark-plugin/message/send_webhook_message" + url = f"{self.API_BASE_URL}/message/send_webhook_message" payload = { "webhook": webhook, "msg_type": msg_type, - "content": content, + "content": content.strip('"').replace(r"\"", '"').replace(r"\\", "\\"), } res = self._send_request(url, require_token=False, payload=payload) return res + + def get_chat_messages( + self, + container_id: str, + start_time: str, + end_time: str, + page_token: str, + sort_type: str = "ByCreateTimeAsc", + page_size: int = 20, + ) -> dict: + """ + API url: https://open.larkoffice.com/document/server-docs/im-v1/message/list + """ + url = f"{self.API_BASE_URL}/message/get_chat_messages" + params = { + "container_id": container_id, + "start_time": start_time, + "end_time": end_time, + "sort_type": sort_type, + "page_token": page_token, + "page_size": page_size, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def get_thread_messages( + self, container_id: str, page_token: str, sort_type: str = "ByCreateTimeAsc", page_size: int = 20 + ) -> dict: + """ + API url: https://open.larkoffice.com/document/server-docs/im-v1/message/list + """ + url = f"{self.API_BASE_URL}/message/get_thread_messages" + params = { + "container_id": container_id, + "sort_type": sort_type, + "page_token": page_token, + "page_size": page_size, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def create_task(self, summary: str, start_time: str, end_time: str, completed_time: str, description: str) -> dict: + # 创建任务 + url = f"{self.API_BASE_URL}/task/create_task" + payload = { + "summary": summary, + "start_time": start_time, + "end_time": end_time, + "completed_at": completed_time, + "description": description, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def update_task( + self, task_guid: str, summary: str, start_time: str, end_time: str, completed_time: str, description: str + ) -> dict: + # 更新任务 + url = f"{self.API_BASE_URL}/task/update_task" + payload = { + "task_guid": task_guid, + "summary": summary, + "start_time": start_time, + "end_time": end_time, + "completed_time": completed_time, + "description": description, + } + res = self._send_request(url, method="PATCH", payload=payload) + return res.get("data") + + def delete_task(self, task_guid: str) -> dict: + # 删除任务 + url = f"{self.API_BASE_URL}/task/delete_task" + payload = { + "task_guid": task_guid, + } + res = self._send_request(url, method="DELETE", payload=payload) + return res + + def add_members(self, task_guid: str, member_phone_or_email: str, member_role: str) -> dict: + # 删除任务 + url = f"{self.API_BASE_URL}/task/add_members" + payload = { + "task_guid": task_guid, + "member_phone_or_email": member_phone_or_email, + "member_role": member_role, + } + res = self._send_request(url, payload=payload) + return res + + def get_wiki_nodes(self, space_id: str, parent_node_token: str, page_token: str, page_size: int = 20) -> dict: + # 获取知识库全部子节点列表 + url = f"{self.API_BASE_URL}/wiki/get_wiki_nodes" + payload = { + "space_id": space_id, + "parent_node_token": parent_node_token, + "page_token": page_token, + "page_size": page_size, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def get_primary_calendar(self, user_id_type: str = "open_id") -> dict: + url = f"{self.API_BASE_URL}/calendar/get_primary_calendar" + params = { + "user_id_type": user_id_type, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def create_event( + self, + summary: str, + description: str, + start_time: str, + end_time: str, + attendee_ability: str, + need_notification: bool = True, + auto_record: bool = False, + ) -> dict: + url = f"{self.API_BASE_URL}/calendar/create_event" + payload = { + "summary": summary, + "description": description, + "need_notification": need_notification, + "start_time": start_time, + "end_time": end_time, + "auto_record": auto_record, + "attendee_ability": attendee_ability, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def update_event( + self, + event_id: str, + summary: str, + description: str, + need_notification: bool, + start_time: str, + end_time: str, + auto_record: bool, + ) -> dict: + url = f"{self.API_BASE_URL}/calendar/update_event/{event_id}" + payload = {} + if summary: + payload["summary"] = summary + if description: + payload["description"] = description + if start_time: + payload["start_time"] = start_time + if end_time: + payload["end_time"] = end_time + if need_notification: + payload["need_notification"] = need_notification + if auto_record: + payload["auto_record"] = auto_record + res = self._send_request(url, method="PATCH", payload=payload) + return res + + def delete_event(self, event_id: str, need_notification: bool = True) -> dict: + url = f"{self.API_BASE_URL}/calendar/delete_event/{event_id}" + params = { + "need_notification": need_notification, + } + res = self._send_request(url, method="DELETE", params=params) + return res + + def list_events(self, start_time: str, end_time: str, page_token: str, page_size: int = 50) -> dict: + url = f"{self.API_BASE_URL}/calendar/list_events" + params = { + "start_time": start_time, + "end_time": end_time, + "page_token": page_token, + "page_size": page_size, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def search_events( + self, + query: str, + start_time: str, + end_time: str, + page_token: str, + user_id_type: str = "open_id", + page_size: int = 20, + ) -> dict: + url = f"{self.API_BASE_URL}/calendar/search_events" + payload = { + "query": query, + "start_time": start_time, + "end_time": end_time, + "page_token": page_token, + "user_id_type": user_id_type, + "page_size": page_size, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def add_event_attendees(self, event_id: str, attendee_phone_or_email: str, need_notification: bool = True) -> dict: + # 参加日程参会人 + url = f"{self.API_BASE_URL}/calendar/add_event_attendees" + payload = { + "event_id": event_id, + "attendee_phone_or_email": attendee_phone_or_email, + "need_notification": need_notification, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def create_spreadsheet( + self, + title: str, + folder_token: str, + ) -> dict: + # 创建电子表格 + url = f"{self.API_BASE_URL}/spreadsheet/create_spreadsheet" + payload = { + "title": title, + "folder_token": folder_token, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def get_spreadsheet( + self, + spreadsheet_token: str, + user_id_type: str = "open_id", + ) -> dict: + # 获取电子表格信息 + url = f"{self.API_BASE_URL}/spreadsheet/get_spreadsheet" + params = { + "spreadsheet_token": spreadsheet_token, + "user_id_type": user_id_type, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def list_spreadsheet_sheets( + self, + spreadsheet_token: str, + ) -> dict: + # 列出电子表格的所有工作表 + url = f"{self.API_BASE_URL}/spreadsheet/list_spreadsheet_sheets" + params = { + "spreadsheet_token": spreadsheet_token, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def add_rows( + self, + spreadsheet_token: str, + sheet_id: str, + sheet_name: str, + length: int, + values: str, + ) -> dict: + # 增加行,在工作表最后添加 + url = f"{self.API_BASE_URL}/spreadsheet/add_rows" + payload = { + "spreadsheet_token": spreadsheet_token, + "sheet_id": sheet_id, + "sheet_name": sheet_name, + "length": length, + "values": values, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def add_cols( + self, + spreadsheet_token: str, + sheet_id: str, + sheet_name: str, + length: int, + values: str, + ) -> dict: + # 增加列,在工作表最后添加 + url = f"{self.API_BASE_URL}/spreadsheet/add_cols" + payload = { + "spreadsheet_token": spreadsheet_token, + "sheet_id": sheet_id, + "sheet_name": sheet_name, + "length": length, + "values": values, + } + res = self._send_request(url, payload=payload) + return res.get("data") + + def read_rows( + self, + spreadsheet_token: str, + sheet_id: str, + sheet_name: str, + start_row: int, + num_rows: int, + user_id_type: str = "open_id", + ) -> dict: + # 读取工作表行数据 + url = f"{self.API_BASE_URL}/spreadsheet/read_rows" + params = { + "spreadsheet_token": spreadsheet_token, + "sheet_id": sheet_id, + "sheet_name": sheet_name, + "start_row": start_row, + "num_rows": num_rows, + "user_id_type": user_id_type, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def read_cols( + self, + spreadsheet_token: str, + sheet_id: str, + sheet_name: str, + start_col: int, + num_cols: int, + user_id_type: str = "open_id", + ) -> dict: + # 读取工作表列数据 + url = f"{self.API_BASE_URL}/spreadsheet/read_cols" + params = { + "spreadsheet_token": spreadsheet_token, + "sheet_id": sheet_id, + "sheet_name": sheet_name, + "start_col": start_col, + "num_cols": num_cols, + "user_id_type": user_id_type, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") + + def read_table( + self, + spreadsheet_token: str, + sheet_id: str, + sheet_name: str, + num_range: str, + query: str, + user_id_type: str = "open_id", + ) -> dict: + # 自定义读取行列数据 + url = f"{self.API_BASE_URL}/spreadsheet/read_table" + params = { + "spreadsheet_token": spreadsheet_token, + "sheet_id": sheet_id, + "sheet_name": sheet_name, + "range": num_range, + "query": query, + "user_id_type": user_id_type, + } + res = self._send_request(url, method="GET", params=params) + return res.get("data") From 55e6123db93ae7fbda06596d6cd0a163747e6559 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Fri, 27 Sep 2024 18:16:20 +0800 Subject: [PATCH 64/79] feat: add min-connection and max-connection for pgvector (#8841) --- api/.env.example | 2 ++ api/configs/middleware/vdb/pgvector_config.py | 10 ++++++++++ api/core/rag/datasource/vdb/pgvector/pgvector.py | 14 ++++++++++++-- .../vdb/pgvector/test_pgvector.py | 2 ++ docker/.env.example | 2 ++ 5 files changed, 28 insertions(+), 2 deletions(-) diff --git a/api/.env.example b/api/.env.example index a3db406aeabdce..7e5c636ff89758 100644 --- a/api/.env.example +++ b/api/.env.example @@ -162,6 +162,8 @@ PGVECTOR_PORT=5433 PGVECTOR_USER=postgres PGVECTOR_PASSWORD=postgres PGVECTOR_DATABASE=postgres +PGVECTOR_MIN_CONNECTION=1 +PGVECTOR_MAX_CONNECTION=5 # Tidb Vector configuration TIDB_VECTOR_HOST=xxx.eu-central-1.xxx.aws.tidbcloud.com diff --git a/api/configs/middleware/vdb/pgvector_config.py b/api/configs/middleware/vdb/pgvector_config.py index 395dcaa4208a93..85f5dca7e23222 100644 --- a/api/configs/middleware/vdb/pgvector_config.py +++ b/api/configs/middleware/vdb/pgvector_config.py @@ -33,3 +33,13 @@ class PGVectorConfig(BaseSettings): description="Name of the PostgreSQL database to connect to", default=None, ) + + PGVECTOR_MIN_CONNECTION: PositiveInt = Field( + description="Min connection of the PostgreSQL database", + default=1, + ) + + PGVECTOR_MAX_CONNECTION: PositiveInt = Field( + description="Max connection of the PostgreSQL database", + default=5, + ) diff --git a/api/core/rag/datasource/vdb/pgvector/pgvector.py b/api/core/rag/datasource/vdb/pgvector/pgvector.py index 79879d4f6394b1..d90707ebcd5cbd 100644 --- a/api/core/rag/datasource/vdb/pgvector/pgvector.py +++ b/api/core/rag/datasource/vdb/pgvector/pgvector.py @@ -23,6 +23,8 @@ class PGVectorConfig(BaseModel): user: str password: str database: str + min_connection: int + max_connection: int @model_validator(mode="before") @classmethod @@ -37,6 +39,12 @@ def validate_config(cls, values: dict) -> dict: raise ValueError("config PGVECTOR_PASSWORD is required") if not values["database"]: raise ValueError("config PGVECTOR_DATABASE is required") + if not values["min_connection"]: + raise ValueError("config PGVECTOR_MIN_CONNECTION is required") + if not values["max_connection"]: + raise ValueError("config PGVECTOR_MAX_CONNECTION is required") + if values["min_connection"] > values["max_connection"]: + raise ValueError("config PGVECTOR_MIN_CONNECTION should less than PGVECTOR_MAX_CONNECTION") return values @@ -61,8 +69,8 @@ def get_type(self) -> str: def _create_connection_pool(self, config: PGVectorConfig): return psycopg2.pool.SimpleConnectionPool( - 1, - 5, + config.min_connection, + config.max_connection, host=config.host, port=config.port, user=config.user, @@ -213,5 +221,7 @@ def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings user=dify_config.PGVECTOR_USER, password=dify_config.PGVECTOR_PASSWORD, database=dify_config.PGVECTOR_DATABASE, + min_connection=dify_config.PGVECTOR_MIN_CONNECTION, + max_connection=dify_config.PGVECTOR_MAX_CONNECTION, ), ) diff --git a/api/tests/integration_tests/vdb/pgvector/test_pgvector.py b/api/tests/integration_tests/vdb/pgvector/test_pgvector.py index c5a986b7479259..72efdc2780eca0 100644 --- a/api/tests/integration_tests/vdb/pgvector/test_pgvector.py +++ b/api/tests/integration_tests/vdb/pgvector/test_pgvector.py @@ -18,6 +18,8 @@ def __init__(self): user="postgres", password="difyai123456", database="dify", + min_connection=1, + max_connection=5, ), ) diff --git a/docker/.env.example b/docker/.env.example index d43c3edc7e722d..eb05f7aa4f0b25 100644 --- a/docker/.env.example +++ b/docker/.env.example @@ -391,6 +391,8 @@ PGVECTOR_PORT=5432 PGVECTOR_USER=postgres PGVECTOR_PASSWORD=difyai123456 PGVECTOR_DATABASE=dify +PGVECTOR_MIN_CONNECTION=1 +PGVECTOR_MAX_CONNECTION=5 # pgvecto-rs configurations, only available when VECTOR_STORE is `pgvecto-rs` PGVECTO_RS_HOST=pgvecto-rs From 27e33fb15c01befb47d02afe257746778b1164ed Mon Sep 17 00:00:00 2001 From: ice yao Date: Sat, 28 Sep 2024 10:54:04 +0800 Subject: [PATCH 65/79] chore: fix wrong VectorType match case (#8857) --- api/controllers/console/datasets/datasets.py | 3 ++- api/core/rag/datasource/vdb/tencent/tencent_vector.py | 2 +- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/api/controllers/console/datasets/datasets.py b/api/controllers/console/datasets/datasets.py index 2c4e5ac60789e7..5a763b3457c695 100644 --- a/api/controllers/console/datasets/datasets.py +++ b/api/controllers/console/datasets/datasets.py @@ -563,10 +563,10 @@ def get(self): case ( VectorType.MILVUS | VectorType.RELYT - | VectorType.PGVECTOR | VectorType.TIDB_VECTOR | VectorType.CHROMA | VectorType.TENCENT + | VectorType.PGVECTO_RS ): return {"retrieval_method": [RetrievalMethod.SEMANTIC_SEARCH.value]} case ( @@ -577,6 +577,7 @@ def get(self): | VectorType.MYSCALE | VectorType.ORACLE | VectorType.ELASTICSEARCH + | VectorType.PGVECTOR ): return { "retrieval_method": [ diff --git a/api/core/rag/datasource/vdb/tencent/tencent_vector.py b/api/core/rag/datasource/vdb/tencent/tencent_vector.py index faa373017becf9..39e3a7f6cfea42 100644 --- a/api/core/rag/datasource/vdb/tencent/tencent_vector.py +++ b/api/core/rag/datasource/vdb/tencent/tencent_vector.py @@ -56,7 +56,7 @@ def _init_database(self): return self._client.create_database(database_name=self._client_config.database) def get_type(self) -> str: - return "tencent" + return VectorType.TENCENT def to_index_struct(self) -> dict: return {"type": self.get_type(), "vector_store": {"class_prefix": self._collection_name}} From a2e2f8a8c98e188dd3d0a4669b4775260756a106 Mon Sep 17 00:00:00 2001 From: Kevin9703 <51311316+Kevin9703@users.noreply.github.com> Date: Sat, 28 Sep 2024 10:54:50 +0800 Subject: [PATCH 66/79] =?UTF-8?q?fix(workflow/nodes/knowledge-retrieval/us?= =?UTF-8?q?e-config):=20Preserve=20rerankin=E2=80=A6=20(#8842)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../components/workflow/nodes/knowledge-retrieval/use-config.ts | 2 ++ 1 file changed, 2 insertions(+) diff --git a/web/app/components/workflow/nodes/knowledge-retrieval/use-config.ts b/web/app/components/workflow/nodes/knowledge-retrieval/use-config.ts index 994bf4f2052dff..6d856003d41109 100644 --- a/web/app/components/workflow/nodes/knowledge-retrieval/use-config.ts +++ b/web/app/components/workflow/nodes/knowledge-retrieval/use-config.ts @@ -136,6 +136,8 @@ const useConfig = (id: string, payload: KnowledgeRetrievalNodeType) => { top_k: multipleRetrievalConfig?.top_k || DATASET_DEFAULT.top_k, score_threshold: multipleRetrievalConfig?.score_threshold, reranking_model: multipleRetrievalConfig?.reranking_model, + reranking_mode: multipleRetrievalConfig?.reranking_mode, + weights: multipleRetrievalConfig?.weights, } }) setInputs(newInput) From 6cd22f3bca2509c19f7b4dcfca4b4785b139d677 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sat, 28 Sep 2024 15:01:27 +0800 Subject: [PATCH 67/79] fix: update qwen2.5-coder-7b model name (#8861) --- .../model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml index fdcd3d42757edb..7ebeec395393c7 100644 --- a/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/tongyi/llm/qwen2.5-coder-7b-instruct.yaml @@ -1,7 +1,7 @@ # for more details, please refer to https://help.aliyun.com/zh/model-studio/getting-started/models -model: qwen2.5-7b-instruct +model: qwen2.5-coder-7b-instruct label: - en_US: qwen2.5-7b-instruct + en_US: qwen2.5-coder-7b-instruct model_type: llm features: - agent-thought From 49af18fbd6549e1ce756544529158a5ca8a7c338 Mon Sep 17 00:00:00 2001 From: takatost Date: Sat, 28 Sep 2024 15:54:26 +0800 Subject: [PATCH 68/79] fix: customize model credentials were invalid despite the provider credentials being active (#8864) --- api/core/entities/provider_configuration.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/api/core/entities/provider_configuration.py b/api/core/entities/provider_configuration.py index 4797b69b8596bb..807f09598c7607 100644 --- a/api/core/entities/provider_configuration.py +++ b/api/core/entities/provider_configuration.py @@ -119,7 +119,7 @@ def get_current_credentials(self, model_type: ModelType, model: str) -> Optional credentials = model_configuration.credentials break - if self.custom_configuration.provider: + if not credentials and self.custom_configuration.provider: credentials = self.custom_configuration.provider.credentials return credentials From 61c89a9168e54f79be54866a1e88171da533f91a Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sat, 28 Sep 2024 16:31:02 +0800 Subject: [PATCH 69/79] feat: add internlm2.5-20b and qwen2.5-coder-7b model (#8862) --- .../siliconflow/llm/_position.yaml | 16 +--- .../siliconflow/llm/internlm2_5-20b-chat.yaml | 30 ++++++++ .../llm/qwen2.5-coder-7b-instruct.yaml | 74 +++++++++++++++++++ .../llm/qwen2.5-math-72b-instruct.yaml | 74 +++++++++++++++++++ 4 files changed, 181 insertions(+), 13 deletions(-) create mode 100644 api/core/model_runtime/model_providers/siliconflow/llm/internlm2_5-20b-chat.yaml create mode 100644 api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-coder-7b-instruct.yaml create mode 100644 api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-math-72b-instruct.yaml diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml index a3e5d0981f7436..8d1df82140b79f 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml @@ -1,18 +1,17 @@ - Qwen/Qwen2.5-72B-Instruct -- Qwen/Qwen2.5-Math-72B-Instruct - Qwen/Qwen2.5-32B-Instruct - Qwen/Qwen2.5-14B-Instruct - Qwen/Qwen2.5-7B-Instruct - Qwen/Qwen2.5-Coder-7B-Instruct -- deepseek-ai/DeepSeek-V2.5 +- Qwen/Qwen2.5-Math-72B-Instruct - Qwen/Qwen2-72B-Instruct - Qwen/Qwen2-57B-A14B-Instruct - Qwen/Qwen2-7B-Instruct - Qwen/Qwen2-1.5B-Instruct +- deepseek-ai/DeepSeek-V2.5 - deepseek-ai/DeepSeek-V2-Chat - deepseek-ai/DeepSeek-Coder-V2-Instruct - THUDM/glm-4-9b-chat -- THUDM/chatglm3-6b - 01-ai/Yi-1.5-34B-Chat-16K - 01-ai/Yi-1.5-9B-Chat-16K - 01-ai/Yi-1.5-6B-Chat @@ -26,13 +25,4 @@ - google/gemma-2-27b-it - google/gemma-2-9b-it - mistralai/Mistral-7B-Instruct-v0.2 -- Pro/Qwen/Qwen2-7B-Instruct -- Pro/Qwen/Qwen2-1.5B-Instruct -- Pro/THUDM/glm-4-9b-chat -- Pro/THUDM/chatglm3-6b -- Pro/01-ai/Yi-1.5-9B-Chat-16K -- Pro/01-ai/Yi-1.5-6B-Chat -- Pro/internlm/internlm2_5-7b-chat -- Pro/meta-llama/Meta-Llama-3.1-8B-Instruct -- Pro/meta-llama/Meta-Llama-3-8B-Instruct -- Pro/google/gemma-2-9b-it +- mistralai/Mixtral-8x7B-Instruct-v0.1 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/internlm2_5-20b-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/internlm2_5-20b-chat.yaml new file mode 100644 index 00000000000000..d9663582e5ca26 --- /dev/null +++ b/api/core/model_runtime/model_providers/siliconflow/llm/internlm2_5-20b-chat.yaml @@ -0,0 +1,30 @@ +model: internlm/internlm2_5-20b-chat +label: + en_US: internlm/internlm2_5-20b-chat +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 32768 +parameter_rules: + - name: temperature + use_template: temperature + - name: max_tokens + use_template: max_tokens + type: int + default: 512 + min: 1 + max: 4096 + help: + zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。 + en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter. + - name: top_p + use_template: top_p + - name: frequency_penalty + use_template: frequency_penalty +pricing: + input: '1' + output: '1' + unit: '0.000001' + currency: RMB diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-coder-7b-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-coder-7b-instruct.yaml new file mode 100644 index 00000000000000..76526200ccdccc --- /dev/null +++ b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-coder-7b-instruct.yaml @@ -0,0 +1,74 @@ +model: Qwen/Qwen2.5-Coder-7B-Instruct +label: + en_US: Qwen/Qwen2.5-Coder-7B-Instruct +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 131072 +parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. + - name: max_tokens + use_template: max_tokens + type: int + default: 8192 + min: 1 + max: 8192 + help: + zh_Hans: 用于指定模型在生成内容时token的最大数量，它定义了生成的上限，但不保证每次都会生成到这个数量。 + en_US: It is used to specify the maximum number of tokens when the model generates content. It defines the upper limit of generation, but does not guarantee that this number will be generated every time. + - name: top_p + use_template: top_p + type: float + default: 0.8 + min: 0.1 + max: 0.9 + help: + zh_Hans: 生成过程中核采样方法概率阈值，例如，取值为0.8时，仅保留概率加起来大于等于0.8的最可能token的最小集合作为候选集。取值范围为（0,1.0)，取值越大，生成的随机性越高；取值越低，生成的确定性越高。 + en_US: The probability threshold of the kernel sampling method during the generation process. For example, when the value is 0.8, only the smallest set of the most likely tokens with a sum of probabilities greater than or equal to 0.8 is retained as the candidate set. The value range is (0,1.0). The larger the value, the higher the randomness generated; the lower the value, the higher the certainty generated. + - name: top_k + type: int + min: 0 + max: 99 + label: + zh_Hans: 取样数量 + en_US: Top k + help: + zh_Hans: 生成时，采样候选集的大小。例如，取值为50时，仅将单次生成中得分最高的50个token组成随机采样的候选集。取值越大，生成的随机性越高；取值越小，生成的确定性越高。 + en_US: The size of the sample candidate set when generated. For example, when the value is 50, only the 50 highest-scoring tokens in a single generation form a randomly sampled candidate set. The larger the value, the higher the randomness generated; the smaller the value, the higher the certainty generated. + - name: seed + required: false + type: int + default: 1234 + label: + zh_Hans: 随机种子 + en_US: Random seed + help: + zh_Hans: 生成时使用的随机数种子，用户控制模型生成内容的随机性。支持无符号64位整数，默认值为 1234。在使用seed时，模型将尽可能生成相同或相似的结果，但目前不保证每次生成的结果完全相同。 + en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: response_format + use_template: response_format +pricing: + input: '0' + output: '0' + unit: '0.000001' + currency: RMB diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-math-72b-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-math-72b-instruct.yaml new file mode 100644 index 00000000000000..90afa0cfd5b96a --- /dev/null +++ b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2.5-math-72b-instruct.yaml @@ -0,0 +1,74 @@ +model: Qwen/Qwen2.5-Math-72B-Instruct +label: + en_US: Qwen/Qwen2.5-Math-72B-Instruct +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 4096 +parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 用于控制随机性和多样性的程度。具体来说，temperature值控制了生成文本时对每个候选词的概率分布进行平滑的程度。较高的temperature值会降低概率分布的峰值，使得更多的低概率词被选择，生成结果更加多样化；而较低的temperature值则会增强概率分布的峰值，使得高概率词更容易被选择，生成结果更加确定。 + en_US: Used to control the degree of randomness and diversity. Specifically, the temperature value controls the degree to which the probability distribution of each candidate word is smoothed when generating text. A higher temperature value will reduce the peak value of the probability distribution, allowing more low-probability words to be selected, and the generated results will be more diverse; while a lower temperature value will enhance the peak value of the probability distribution, making it easier for high-probability words to be selected. , the generated results are more certain. + - name: max_tokens + use_template: max_tokens + type: int + default: 2000 + min: 1 + max: 2000 + help: + zh_Hans: 用于指定模型在生成内容时token的最大数量，它定义了生成的上限，但不保证每次都会生成到这个数量。 + en_US: It is used to specify the maximum number of tokens when the model generates content. It defines the upper limit of generation, but does not guarantee that this number will be generated every time. + - name: top_p + use_template: top_p + type: float + default: 0.8 + min: 0.1 + max: 0.9 + help: + zh_Hans: 生成过程中核采样方法概率阈值，例如，取值为0.8时，仅保留概率加起来大于等于0.8的最可能token的最小集合作为候选集。取值范围为（0,1.0)，取值越大，生成的随机性越高；取值越低，生成的确定性越高。 + en_US: The probability threshold of the kernel sampling method during the generation process. For example, when the value is 0.8, only the smallest set of the most likely tokens with a sum of probabilities greater than or equal to 0.8 is retained as the candidate set. The value range is (0,1.0). The larger the value, the higher the randomness generated; the lower the value, the higher the certainty generated. + - name: top_k + type: int + min: 0 + max: 99 + label: + zh_Hans: 取样数量 + en_US: Top k + help: + zh_Hans: 生成时，采样候选集的大小。例如，取值为50时，仅将单次生成中得分最高的50个token组成随机采样的候选集。取值越大，生成的随机性越高；取值越小，生成的确定性越高。 + en_US: The size of the sample candidate set when generated. For example, when the value is 50, only the 50 highest-scoring tokens in a single generation form a randomly sampled candidate set. The larger the value, the higher the randomness generated; the smaller the value, the higher the certainty generated. + - name: seed + required: false + type: int + default: 1234 + label: + zh_Hans: 随机种子 + en_US: Random seed + help: + zh_Hans: 生成时使用的随机数种子，用户控制模型生成内容的随机性。支持无符号64位整数，默认值为 1234。在使用seed时，模型将尽可能生成相同或相似的结果，但目前不保证每次生成的结果完全相同。 + en_US: The random number seed used when generating, the user controls the randomness of the content generated by the model. Supports unsigned 64-bit integers, default value is 1234. When using seed, the model will try its best to generate the same or similar results, but there is currently no guarantee that the results will be exactly the same every time. + - name: repetition_penalty + required: false + type: float + default: 1.1 + label: + zh_Hans: 重复惩罚 + en_US: Repetition penalty + help: + zh_Hans: 用于控制模型生成时的重复度。提高repetition_penalty时可以降低模型生成的重复度。1.0表示不做惩罚。 + en_US: Used to control the repeatability when generating models. Increasing repetition_penalty can reduce the duplication of model generation. 1.0 means no punishment. + - name: response_format + use_template: response_format +pricing: + input: '4.13' + output: '4.13' + unit: '0.000001' + currency: RMB From 850492dafaa8ef801a332d5a787cc9818c7b6072 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sat, 28 Sep 2024 21:40:27 +0800 Subject: [PATCH 70/79] feat: deprecate gte-Qwen2-7B-instruct embedding model (#8866) --- .../perfxcloud/llm/_position.yaml | 21 +++++++++---------- .../perfxcloud/text_embedding/_position.yaml | 4 ++++ .../text_embedding/gte-Qwen2-7B-instruct.yaml | 1 + 3 files changed, 15 insertions(+), 11 deletions(-) create mode 100644 api/core/model_runtime/model_providers/perfxcloud/text_embedding/_position.yaml diff --git a/api/core/model_runtime/model_providers/perfxcloud/llm/_position.yaml b/api/core/model_runtime/model_providers/perfxcloud/llm/_position.yaml index 37bf400f1e3475..c6930e54f50aa4 100644 --- a/api/core/model_runtime/model_providers/perfxcloud/llm/_position.yaml +++ b/api/core/model_runtime/model_providers/perfxcloud/llm/_position.yaml @@ -1,24 +1,23 @@ - Qwen2.5-72B-Instruct - Qwen2.5-7B-Instruct +- Qwen2-72B-Instruct +- Qwen2-72B-Instruct-AWQ-int4 +- Qwen2-72B-Instruct-GPTQ-Int4 +- Qwen2-7B-Instruct +- Qwen2-7B +- Qwen1.5-110B-Chat-GPTQ-Int4 +- Qwen1.5-72B-Chat-GPTQ-Int4 +- Qwen1.5-7B +- Qwen-14B-Chat-Int4 - Yi-Coder-1.5B-Chat - Yi-Coder-9B-Chat -- Qwen2-72B-Instruct-AWQ-int4 - Yi-1_5-9B-Chat-16K -- Qwen2-7B-Instruct - Reflection-Llama-3.1-70B -- Qwen2-72B-Instruct - Meta-Llama-3.1-8B-Instruct - - Meta-Llama-3.1-405B-Instruct-AWQ-INT4 - Meta-Llama-3-70B-Instruct-GPTQ-Int4 -- chatglm3-6b - Meta-Llama-3-8B-Instruct - Llama3-Chinese_v2 - deepseek-v2-lite-chat -- Qwen2-72B-Instruct-GPTQ-Int4 -- Qwen2-7B -- Qwen-14B-Chat-Int4 -- Qwen1.5-72B-Chat-GPTQ-Int4 -- Qwen1.5-7B -- Qwen1.5-110B-Chat-GPTQ-Int4 - deepseek-v2-chat +- chatglm3-6b diff --git a/api/core/model_runtime/model_providers/perfxcloud/text_embedding/_position.yaml b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/_position.yaml new file mode 100644 index 00000000000000..99163d42931b16 --- /dev/null +++ b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/_position.yaml @@ -0,0 +1,4 @@ +- gte-Qwen2-7B-instruct +- BAAI/bge-large-en-v1.5 +- BAAI/bge-large-zh-v1.5 +- BAAI/bge-m3 diff --git a/api/core/model_runtime/model_providers/perfxcloud/text_embedding/gte-Qwen2-7B-instruct.yaml b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/gte-Qwen2-7B-instruct.yaml index 03db0d8bce8500..161d5ea9a2657e 100644 --- a/api/core/model_runtime/model_providers/perfxcloud/text_embedding/gte-Qwen2-7B-instruct.yaml +++ b/api/core/model_runtime/model_providers/perfxcloud/text_embedding/gte-Qwen2-7B-instruct.yaml @@ -2,3 +2,4 @@ model: gte-Qwen2-7B-instruct model_type: text-embedding model_properties: context_size: 2048 +deprecated: true From f97607370a2dae9eff8bfeece505b417d232df76 Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sat, 28 Sep 2024 21:41:02 +0800 Subject: [PATCH 71/79] refactor: update Callback to an abstract class (#8868) --- api/core/model_runtime/callbacks/base_callback.py | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/api/core/model_runtime/callbacks/base_callback.py b/api/core/model_runtime/callbacks/base_callback.py index 92da53c9a464df..6bd9325785a2da 100644 --- a/api/core/model_runtime/callbacks/base_callback.py +++ b/api/core/model_runtime/callbacks/base_callback.py @@ -1,3 +1,4 @@ +from abc import ABC, abstractmethod from typing import Optional from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk @@ -13,7 +14,7 @@ } -class Callback: +class Callback(ABC): """ Base class for callbacks. Only for LLM. @@ -21,6 +22,7 @@ class Callback: raise_error: bool = False + @abstractmethod def on_before_invoke( self, llm_instance: AIModel, @@ -48,6 +50,7 @@ def on_before_invoke( """ raise NotImplementedError() + @abstractmethod def on_new_chunk( self, llm_instance: AIModel, @@ -77,6 +80,7 @@ def on_new_chunk( """ raise NotImplementedError() + @abstractmethod def on_after_invoke( self, llm_instance: AIModel, @@ -106,6 +110,7 @@ def on_after_invoke( """ raise NotImplementedError() + @abstractmethod def on_invoke_error( self, llm_instance: AIModel, From 74f58f29f9a3070152b0bb097eb2bebb9f9cd19e Mon Sep 17 00:00:00 2001 From: Bowen Liang Date: Sun, 29 Sep 2024 00:29:59 +0800 Subject: [PATCH 72/79] chore: bump ruff to 0.6.8 for fixing violation in SIM910 (#8869) --- .../zhipuai/zhipuai_sdk/core/_utils/_utils.py | 2 +- .../rag/datasource/keyword/jieba/jieba.py | 2 +- .../event_handlers/create_document_index.py | 2 +- api/poetry.lock | 40 +++++++++---------- api/pyproject.toml | 2 +- .../integration_tests/tools/__mock/http.py | 2 +- .../workflow/nodes/__mock/http.py | 4 +- 7 files changed, 27 insertions(+), 27 deletions(-) diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_utils/_utils.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_utils/_utils.py index ce5e7786aa2937..3a7b234ab0c067 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_utils/_utils.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_utils/_utils.py @@ -248,7 +248,7 @@ def inner(func: CallableT) -> CallableT: @functools.wraps(func) def wrapper(*args: object, **kwargs: object) -> object: given_params: set[str] = set() - for i, _ in enumerate(args): + for i in range(len(args)): try: given_params.add(positional[i]) except IndexError: diff --git a/api/core/rag/datasource/keyword/jieba/jieba.py b/api/core/rag/datasource/keyword/jieba/jieba.py index 3073100746d360..a0153c1e58a1a8 100644 --- a/api/core/rag/datasource/keyword/jieba/jieba.py +++ b/api/core/rag/datasource/keyword/jieba/jieba.py @@ -45,7 +45,7 @@ def add_texts(self, texts: list[Document], **kwargs): keyword_table_handler = JiebaKeywordTableHandler() keyword_table = self._get_dataset_keyword_table() - keywords_list = kwargs.get("keywords_list", None) + keywords_list = kwargs.get("keywords_list") for i in range(len(texts)): text = texts[i] if keywords_list: diff --git a/api/events/event_handlers/create_document_index.py b/api/events/event_handlers/create_document_index.py index 54f6a76e167231..5af45e1e5026df 100644 --- a/api/events/event_handlers/create_document_index.py +++ b/api/events/event_handlers/create_document_index.py @@ -14,7 +14,7 @@ @document_index_created.connect def handle(sender, **kwargs): dataset_id = sender - document_ids = kwargs.get("document_ids", None) + document_ids = kwargs.get("document_ids") documents = [] start_at = time.perf_counter() for document_id in document_ids: diff --git a/api/poetry.lock b/api/poetry.lock index bce21fb547c8ca..85c68cd75f1292 100644 --- a/api/poetry.lock +++ b/api/poetry.lock @@ -8074,29 +8074,29 @@ pyasn1 = ">=0.1.3" [[package]] name = "ruff" -version = "0.6.5" +version = "0.6.8" description = "An extremely fast Python linter and code formatter, written in Rust." optional = false python-versions = ">=3.7" files = [ - {file = "ruff-0.6.5-py3-none-linux_armv6l.whl", hash = "sha256:7e4e308f16e07c95fc7753fc1aaac690a323b2bb9f4ec5e844a97bb7fbebd748"}, - {file = "ruff-0.6.5-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:932cd69eefe4daf8c7d92bd6689f7e8182571cb934ea720af218929da7bd7d69"}, - {file = "ruff-0.6.5-py3-none-macosx_11_0_arm64.whl", hash = "sha256:3a8d42d11fff8d3143ff4da41742a98f8f233bf8890e9fe23077826818f8d680"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:a50af6e828ee692fb10ff2dfe53f05caecf077f4210fae9677e06a808275754f"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:794ada3400a0d0b89e3015f1a7e01f4c97320ac665b7bc3ade24b50b54cb2972"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:381413ec47f71ce1d1c614f7779d88886f406f1fd53d289c77e4e533dc6ea200"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:52e75a82bbc9b42e63c08d22ad0ac525117e72aee9729a069d7c4f235fc4d276"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:09c72a833fd3551135ceddcba5ebdb68ff89225d30758027280968c9acdc7810"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:800c50371bdcb99b3c1551d5691e14d16d6f07063a518770254227f7f6e8c178"}, - {file = "ruff-0.6.5-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8e25ddd9cd63ba1f3bd51c1f09903904a6adf8429df34f17d728a8fa11174253"}, - {file = "ruff-0.6.5-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:7291e64d7129f24d1b0c947ec3ec4c0076e958d1475c61202497c6aced35dd19"}, - {file = "ruff-0.6.5-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:9ad7dfbd138d09d9a7e6931e6a7e797651ce29becd688be8a0d4d5f8177b4b0c"}, - {file = "ruff-0.6.5-py3-none-musllinux_1_2_i686.whl", hash = "sha256:005256d977021790cc52aa23d78f06bb5090dc0bfbd42de46d49c201533982ae"}, - {file = "ruff-0.6.5-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:482c1e6bfeb615eafc5899127b805d28e387bd87db38b2c0c41d271f5e58d8cc"}, - {file = "ruff-0.6.5-py3-none-win32.whl", hash = "sha256:cf4d3fa53644137f6a4a27a2b397381d16454a1566ae5335855c187fbf67e4f5"}, - {file = "ruff-0.6.5-py3-none-win_amd64.whl", hash = "sha256:3e42a57b58e3612051a636bc1ac4e6b838679530235520e8f095f7c44f706ff9"}, - {file = "ruff-0.6.5-py3-none-win_arm64.whl", hash = "sha256:51935067740773afdf97493ba9b8231279e9beef0f2a8079188c4776c25688e0"}, - {file = "ruff-0.6.5.tar.gz", hash = "sha256:4d32d87fab433c0cf285c3683dd4dae63be05fd7a1d65b3f5bf7cdd05a6b96fb"}, + {file = "ruff-0.6.8-py3-none-linux_armv6l.whl", hash = "sha256:77944bca110ff0a43b768f05a529fecd0706aac7bcce36d7f1eeb4cbfca5f0f2"}, + {file = "ruff-0.6.8-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:27b87e1801e786cd6ede4ada3faa5e254ce774de835e6723fd94551464c56b8c"}, + {file = "ruff-0.6.8-py3-none-macosx_11_0_arm64.whl", hash = "sha256:cd48f945da2a6334f1793d7f701725a76ba93bf3d73c36f6b21fb04d5338dcf5"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:677e03c00f37c66cea033274295a983c7c546edea5043d0c798833adf4cf4c6f"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:9f1476236b3eacfacfc0f66aa9e6cd39f2a624cb73ea99189556015f27c0bdeb"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:6f5a2f17c7d32991169195d52a04c95b256378bbf0de8cb98478351eb70d526f"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:5fd0d4b7b1457c49e435ee1e437900ced9b35cb8dc5178921dfb7d98d65a08d0"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:f8034b19b993e9601f2ddf2c517451e17a6ab5cdb1c13fdff50c1442a7171d87"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:6cfb227b932ba8ef6e56c9f875d987973cd5e35bc5d05f5abf045af78ad8e098"}, + {file = "ruff-0.6.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6ef0411eccfc3909269fed47c61ffebdcb84a04504bafa6b6df9b85c27e813b0"}, + {file = "ruff-0.6.8-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:007dee844738c3d2e6c24ab5bc7d43c99ba3e1943bd2d95d598582e9c1b27750"}, + {file = "ruff-0.6.8-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:ce60058d3cdd8490e5e5471ef086b3f1e90ab872b548814e35930e21d848c9ce"}, + {file = "ruff-0.6.8-py3-none-musllinux_1_2_i686.whl", hash = "sha256:1085c455d1b3fdb8021ad534379c60353b81ba079712bce7a900e834859182fa"}, + {file = "ruff-0.6.8-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:70edf6a93b19481affd287d696d9e311388d808671bc209fb8907b46a8c3af44"}, + {file = "ruff-0.6.8-py3-none-win32.whl", hash = "sha256:792213f7be25316f9b46b854df80a77e0da87ec66691e8f012f887b4a671ab5a"}, + {file = "ruff-0.6.8-py3-none-win_amd64.whl", hash = "sha256:ec0517dc0f37cad14a5319ba7bba6e7e339d03fbf967a6d69b0907d61be7a263"}, + {file = "ruff-0.6.8-py3-none-win_arm64.whl", hash = "sha256:8d3bb2e3fbb9875172119021a13eed38849e762499e3cfde9588e4b4d70968dc"}, + {file = "ruff-0.6.8.tar.gz", hash = "sha256:a5bf44b1aa0adaf6d9d20f86162b34f7c593bfedabc51239953e446aefc8ce18"}, ] [[package]] @@ -10501,4 +10501,4 @@ cffi = ["cffi (>=1.11)"] [metadata] lock-version = "2.0" python-versions = ">=3.10,<3.13" -content-hash = "69b42bb1ff033f14e199fee8335356275099421d72bbd7037b7a991ea65cae08" +content-hash = "c4580c22e2b220c8c80dbc3f765060a09e14874ed29b690c13a533bf0365e789" diff --git a/api/pyproject.toml b/api/pyproject.toml index f004865d5f3438..64b35621b2ec60 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -286,4 +286,4 @@ optional = true [tool.poetry.group.lint.dependencies] dotenv-linter = "~0.5.0" -ruff = "~0.6.5" +ruff = "~0.6.8" diff --git a/api/tests/integration_tests/tools/__mock/http.py b/api/tests/integration_tests/tools/__mock/http.py index d3c1f3101c54bc..42cf87e317ab6a 100644 --- a/api/tests/integration_tests/tools/__mock/http.py +++ b/api/tests/integration_tests/tools/__mock/http.py @@ -17,7 +17,7 @@ def httpx_request( request = httpx.Request( method, url, params=kwargs.get("params"), headers=kwargs.get("headers"), cookies=kwargs.get("cookies") ) - data = kwargs.get("data", None) + data = kwargs.get("data") resp = json.dumps(data).encode("utf-8") if data else b"OK" response = httpx.Response( status_code=200, diff --git a/api/tests/integration_tests/workflow/nodes/__mock/http.py b/api/tests/integration_tests/workflow/nodes/__mock/http.py index f1ab23b0026aab..ec013183b7601d 100644 --- a/api/tests/integration_tests/workflow/nodes/__mock/http.py +++ b/api/tests/integration_tests/workflow/nodes/__mock/http.py @@ -22,8 +22,8 @@ def httpx_request( return response # get data, files - data = kwargs.get("data", None) - files = kwargs.get("files", None) + data = kwargs.get("data") + files = kwargs.get("files") if data is not None: resp = dumps(data).encode("utf-8") elif files is not None: From 4ec977eabafefc3859ef04425ec669ce87b5483e Mon Sep 17 00:00:00 2001 From: -LAN- Date: Sun, 29 Sep 2024 16:12:42 +0800 Subject: [PATCH 73/79] fix(workflow): update tagging logic in GitHub Actions (#8882) --- .github/workflows/build-push.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/build-push.yml b/.github/workflows/build-push.yml index 407bd47d9b0f8f..6daaaf5791dd24 100644 --- a/.github/workflows/build-push.yml +++ b/.github/workflows/build-push.yml @@ -125,7 +125,7 @@ jobs: with: images: ${{ env[matrix.image_name_env] }} tags: | - type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/') }} + type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/') && !contains(github.ref, '-') }} type=ref,event=branch type=sha,enable=true,priority=100,prefix=,suffix=,format=long type=raw,value=${{ github.ref_name }},enable=${{ startsWith(github.ref, 'refs/tags/') }} From e4ed916baadb10ffd22ec1772e56b4fcefa1b57f Mon Sep 17 00:00:00 2001 From: longzhihun <38651850@qq.com> Date: Sun, 29 Sep 2024 16:12:56 +0800 Subject: [PATCH 74/79] Add Jamba and Llama3.2 model support (#8878) --- .../bedrock/llm/_position.yaml | 6 ++++ .../llm/ai21.jamba-1-5-large-v1.0.yaml | 26 ++++++++++++++++ .../bedrock/llm/ai21.jamba-1-5-mini-v1.0.yaml | 26 ++++++++++++++++ .../model_providers/bedrock/llm/llm.py | 2 ++ .../us.meta.llama3-2-11b-instruct-v1.0.yaml | 29 +++++++++++++++++ .../us.meta.llama3-2-1b-instruct-v1.0.yaml | 26 ++++++++++++++++ .../us.meta.llama3-2-3b-instruct-v1.0.yaml | 26 ++++++++++++++++ .../us.meta.llama3-2-90b-instruct-v1.0.yaml | 31 +++++++++++++++++++ 8 files changed, 172 insertions(+) create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-large-v1.0.yaml create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-mini-v1.0.yaml create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-11b-instruct-v1.0.yaml create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-1b-instruct-v1.0.yaml create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-3b-instruct-v1.0.yaml create mode 100644 api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-90b-instruct-v1.0.yaml diff --git a/api/core/model_runtime/model_providers/bedrock/llm/_position.yaml b/api/core/model_runtime/model_providers/bedrock/llm/_position.yaml index 86c8061deefac8..47e2b020fd09a3 100644 --- a/api/core/model_runtime/model_providers/bedrock/llm/_position.yaml +++ b/api/core/model_runtime/model_providers/bedrock/llm/_position.yaml @@ -6,6 +6,8 @@ - anthropic.claude-v2:1 - anthropic.claude-3-sonnet-v1:0 - anthropic.claude-3-haiku-v1:0 +- ai21.jamba-1-5-large-v1:0 +- ai21.jamba-1-5-mini-v1:0 - cohere.command-light-text-v14 - cohere.command-text-v14 - cohere.command-r-plus-v1.0 @@ -15,6 +17,10 @@ - meta.llama3-1-405b-instruct-v1:0 - meta.llama3-8b-instruct-v1:0 - meta.llama3-70b-instruct-v1:0 +- us.meta.llama3-2-1b-instruct-v1:0 +- us.meta.llama3-2-3b-instruct-v1:0 +- us.meta.llama3-2-11b-instruct-v1:0 +- us.meta.llama3-2-90b-instruct-v1:0 - meta.llama2-13b-chat-v1 - meta.llama2-70b-chat-v1 - mistral.mistral-large-2407-v1:0 diff --git a/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-large-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-large-v1.0.yaml new file mode 100644 index 00000000000000..276c7312cee008 --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-large-v1.0.yaml @@ -0,0 +1,26 @@ +model: ai21.jamba-1-5-large-v1:0 +label: + en_US: Jamba 1.5 Large +model_type: llm +model_properties: + mode: completion + context_size: 256000 +parameter_rules: + - name: temperature + use_template: temperature + default: 1 + min: 0.0 + max: 2.0 + - name: top_p + use_template: top_p + - name: max_gen_len + use_template: max_tokens + required: true + default: 4096 + min: 1 + max: 4096 +pricing: + input: '0.002' + output: '0.008' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-mini-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-mini-v1.0.yaml new file mode 100644 index 00000000000000..3461d8ab71329d --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/ai21.jamba-1-5-mini-v1.0.yaml @@ -0,0 +1,26 @@ +model: ai21.jamba-1-5-mini-v1:0 +label: + en_US: Jamba 1.5 Mini +model_type: llm +model_properties: + mode: completion + context_size: 256000 +parameter_rules: + - name: temperature + use_template: temperature + default: 1 + min: 0.0 + max: 2.0 + - name: top_p + use_template: top_p + - name: max_gen_len + use_template: max_tokens + required: true + default: 4096 + min: 1 + max: 4096 +pricing: + input: '0.0002' + output: '0.0004' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/bedrock/llm/llm.py b/api/core/model_runtime/model_providers/bedrock/llm/llm.py index 77bab0c2945887..d1961784f2e841 100644 --- a/api/core/model_runtime/model_providers/bedrock/llm/llm.py +++ b/api/core/model_runtime/model_providers/bedrock/llm/llm.py @@ -63,6 +63,7 @@ class BedrockLargeLanguageModel(LargeLanguageModel): {"prefix": "us.anthropic.claude-3", "support_system_prompts": True, "support_tool_use": True}, {"prefix": "eu.anthropic.claude-3", "support_system_prompts": True, "support_tool_use": True}, {"prefix": "anthropic.claude-3", "support_system_prompts": True, "support_tool_use": True}, + {"prefix": "us.meta.llama3-2", "support_system_prompts": True, "support_tool_use": True}, {"prefix": "meta.llama", "support_system_prompts": True, "support_tool_use": False}, {"prefix": "mistral.mistral-7b-instruct", "support_system_prompts": False, "support_tool_use": False}, {"prefix": "mistral.mixtral-8x7b-instruct", "support_system_prompts": False, "support_tool_use": False}, @@ -70,6 +71,7 @@ class BedrockLargeLanguageModel(LargeLanguageModel): {"prefix": "mistral.mistral-small", "support_system_prompts": True, "support_tool_use": True}, {"prefix": "cohere.command-r", "support_system_prompts": True, "support_tool_use": True}, {"prefix": "amazon.titan", "support_system_prompts": False, "support_tool_use": False}, + {"prefix": "ai21.jamba-1-5", "support_system_prompts": True, "support_tool_use": False}, ] @staticmethod diff --git a/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-11b-instruct-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-11b-instruct-v1.0.yaml new file mode 100644 index 00000000000000..029f428776e0be --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-11b-instruct-v1.0.yaml @@ -0,0 +1,29 @@ +model: us.meta.llama3-2-11b-instruct-v1:0 +label: + en_US: US Meta Llama 3.2 11B Instruct +model_type: llm +features: + - vision + - tool-call +model_properties: + mode: completion + context_size: 128000 +parameter_rules: + - name: temperature + use_template: temperature + default: 0.5 + min: 0.0 + max: 1 + - name: top_p + use_template: top_p + - name: max_gen_len + use_template: max_tokens + required: true + default: 512 + min: 1 + max: 2048 +pricing: + input: '0.00035' + output: '0.00035' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-1b-instruct-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-1b-instruct-v1.0.yaml new file mode 100644 index 00000000000000..51c8474e54846d --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-1b-instruct-v1.0.yaml @@ -0,0 +1,26 @@ +model: us.meta.llama3-2-1b-instruct-v1:0 +label: + en_US: US Meta Llama 3.2 1B Instruct +model_type: llm +model_properties: + mode: completion + context_size: 128000 +parameter_rules: + - name: temperature + use_template: temperature + default: 0.5 + min: 0.0 + max: 1 + - name: top_p + use_template: top_p + - name: max_gen_len + use_template: max_tokens + required: true + default: 512 + min: 1 + max: 2048 +pricing: + input: '0.0001' + output: '0.0001' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-3b-instruct-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-3b-instruct-v1.0.yaml new file mode 100644 index 00000000000000..472cc7403e2d3e --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-3b-instruct-v1.0.yaml @@ -0,0 +1,26 @@ +model: us.meta.llama3-2-3b-instruct-v1:0 +label: + en_US: US Meta Llama 3.2 3B Instruct +model_type: llm +model_properties: + mode: completion + context_size: 128000 +parameter_rules: + - name: temperature + use_template: temperature + default: 0.5 + min: 0.0 + max: 1 + - name: top_p + use_template: top_p + - name: max_gen_len + use_template: max_tokens + required: true + default: 512 + min: 1 + max: 2048 +pricing: + input: '0.00015' + output: '0.00015' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-90b-instruct-v1.0.yaml b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-90b-instruct-v1.0.yaml new file mode 100644 index 00000000000000..cecd0236ca9e31 --- /dev/null +++ b/api/core/model_runtime/model_providers/bedrock/llm/us.meta.llama3-2-90b-instruct-v1.0.yaml @@ -0,0 +1,31 @@ +model: us.meta.llama3-2-90b-instruct-v1:0 +label: + en_US: US Meta Llama 3.2 90B Instruct +model_type: llm +features: + - tool-call +model_properties: + mode: completion + context_size: 128000 +parameter_rules: + - name: temperature + use_template: temperature + default: 0.5 + min: 0.0 + max: 1 + - name: top_p + use_template: top_p + default: 0.9 + min: 0 + max: 1 + - name: max_gen_len + use_template: max_tokens + required: true + default: 512 + min: 1 + max: 2048 +pricing: + input: '0.002' + output: '0.002' + unit: '0.001' + currency: USD From c531b4a911f34357e2beeff77bdc5e1ecb4333de Mon Sep 17 00:00:00 2001 From: chenxu9741 Date: Sun, 29 Sep 2024 16:13:20 +0800 Subject: [PATCH 75/79] =?UTF-8?q?fix:=20#8843=20event:=20tts=5Fmessage=5Fe?= =?UTF-8?q?nd=20always=20return=20in=20api=20streaming=20resp=E2=80=A6=20(?= =?UTF-8?q?#8846)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- api/core/app/apps/advanced_chat/generate_task_pipeline.py | 3 ++- api/core/app/apps/workflow/generate_task_pipeline.py | 3 ++- .../app/task_pipeline/easy_ui_based_generate_task_pipeline.py | 3 ++- .../model_providers/zhipuai/zhipuai_sdk/core/_base_models.py | 2 +- 4 files changed, 7 insertions(+), 4 deletions(-) diff --git a/api/core/app/apps/advanced_chat/generate_task_pipeline.py b/api/core/app/apps/advanced_chat/generate_task_pipeline.py index 94206a1b1cd678..897b6fd063c76b 100644 --- a/api/core/app/apps/advanced_chat/generate_task_pipeline.py +++ b/api/core/app/apps/advanced_chat/generate_task_pipeline.py @@ -231,7 +231,8 @@ def _wrapper_process_stream_response( except Exception as e: logger.error(e) break - yield MessageAudioEndStreamResponse(audio="", task_id=task_id) + if tts_publisher: + yield MessageAudioEndStreamResponse(audio="", task_id=task_id) def _process_stream_response( self, diff --git a/api/core/app/apps/workflow/generate_task_pipeline.py b/api/core/app/apps/workflow/generate_task_pipeline.py index 93edf8e0e882b5..798847a5070b8d 100644 --- a/api/core/app/apps/workflow/generate_task_pipeline.py +++ b/api/core/app/apps/workflow/generate_task_pipeline.py @@ -212,7 +212,8 @@ def _wrapper_process_stream_response( except Exception as e: logger.error(e) break - yield MessageAudioEndStreamResponse(audio="", task_id=task_id) + if tts_publisher: + yield MessageAudioEndStreamResponse(audio="", task_id=task_id) def _process_stream_response( self, diff --git a/api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py b/api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py index 8f834b6458ea4b..917649f34e769c 100644 --- a/api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py +++ b/api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py @@ -248,7 +248,8 @@ def _wrapper_process_stream_response( else: start_listener_time = time.time() yield MessageAudioStreamResponse(audio=audio.audio, task_id=task_id) - yield MessageAudioEndStreamResponse(audio="", task_id=task_id) + if publisher: + yield MessageAudioEndStreamResponse(audio="", task_id=task_id) def _process_stream_response( self, publisher: AppGeneratorTTSPublisher, trace_manager: Optional[TraceQueueManager] = None diff --git a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py index 6d8ba700b7b1dc..69b1d3a83dfef3 100644 --- a/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py +++ b/api/core/model_runtime/model_providers/zhipuai/zhipuai_sdk/core/_base_models.py @@ -48,7 +48,7 @@ ) if TYPE_CHECKING: - from pydantic_core.core_schema import LiteralSchema, ModelField, ModelFieldsSchema + from pydantic_core.core_schema import ModelField __all__ = ["BaseModel", "GenericModel"] _BaseModelT = TypeVar("_BaseModelT", bound="BaseModel") From 42dfde65468ec8b945eb086d88441baf1fc22e2d Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sun, 29 Sep 2024 16:16:56 +0800 Subject: [PATCH 76/79] docs: add english versions for the files customizable_model_scale_out and predefined_model_scale_out (#8871) --- .../en_US/customizable_model_scale_out.md | 310 ++++++++++++++++++ .../docs/en_US/images/index/image-1.png | Bin 0 -> 235102 bytes .../docs/en_US/images/index/image-2.png | Bin 0 -> 210087 bytes .../docs/en_US/images/index/image-3.png | Bin 0 -> 44778 bytes .../docs/en_US/images/index/image.png | Bin 0 -> 267979 bytes .../docs/en_US/predefined_model_scale_out.md | 173 ++++++++++ .../docs/en_US/provider_scale_out.md | 2 +- .../docs/zh_Hans/provider_scale_out.md | 2 +- 8 files changed, 485 insertions(+), 2 deletions(-) create mode 100644 api/core/model_runtime/docs/en_US/customizable_model_scale_out.md create mode 100644 api/core/model_runtime/docs/en_US/images/index/image-1.png create mode 100644 api/core/model_runtime/docs/en_US/images/index/image-2.png create mode 100644 api/core/model_runtime/docs/en_US/images/index/image-3.png create mode 100644 api/core/model_runtime/docs/en_US/images/index/image.png create mode 100644 api/core/model_runtime/docs/en_US/predefined_model_scale_out.md diff --git a/api/core/model_runtime/docs/en_US/customizable_model_scale_out.md b/api/core/model_runtime/docs/en_US/customizable_model_scale_out.md new file mode 100644 index 00000000000000..f5b806ade6f499 --- /dev/null +++ b/api/core/model_runtime/docs/en_US/customizable_model_scale_out.md @@ -0,0 +1,310 @@ +## Custom Integration of Pre-defined Models + +### Introduction + +After completing the vendors integration, the next step is to connect the vendor's models. To illustrate the entire connection process, we will use Xinference as an example to demonstrate a complete vendor integration. + +It is important to note that for custom models, each model connection requires a complete vendor credential. + +Unlike pre-defined models, a custom vendor integration always includes the following two parameters, which do not need to be defined in the vendor YAML file. + +![](images/index/image-3.png) + +As mentioned earlier, vendors do not need to implement validate_provider_credential. The runtime will automatically call the corresponding model layer's validate_credentials to validate the credentials based on the model type and name selected by the user. + +### Writing the Vendor YAML + +First, we need to identify the types of models supported by the vendor we are integrating. + +Currently supported model types are as follows: + +- `llm` Text Generation Models + +- `text_embedding` Text Embedding Models + +- `rerank` Rerank Models + +- `speech2text` Speech-to-Text + +- `tts` Text-to-Speech + +- `moderation` Moderation + +Xinference supports LLM, Text Embedding, and Rerank. So we will start by writing xinference.yaml. + +```yaml +provider: xinference #Define the vendor identifier +label: # Vendor display name, supports both en_US (English) and zh_Hans (Simplified Chinese). If zh_Hans is not set, it will use en_US by default. + en_US: Xorbits Inference +icon_small: # Small icon, refer to other vendors' icons stored in the _assets directory within the vendor implementation directory; follows the same language policy as the label + en_US: icon_s_en.svg +icon_large: # Large icon + en_US: icon_l_en.svg +help: # Help information + title: + en_US: How to deploy Xinference + zh_Hans: 如何部署 Xinference + url: + en_US: https://github.com/xorbitsai/inference +supported_model_types: # Supported model types. Xinference supports LLM, Text Embedding, and Rerank +- llm +- text-embedding +- rerank +configurate_methods: # Since Xinference is a locally deployed vendor with no predefined models, users need to deploy whatever models they need according to Xinference documentation. Thus, it only supports custom models. +- customizable-model +provider_credential_schema: + credential_form_schemas: +``` + + +Then, we need to determine what credentials are required to define a model in Xinference. + +- Since it supports three different types of models, we need to specify the model_type to denote the model type. Here is how we can define it: + +```yaml +provider_credential_schema: + credential_form_schemas: + - variable: model_type + type: select + label: + en_US: Model type + zh_Hans: 模型类型 + required: true + options: + - value: text-generation + label: + en_US: Language Model + zh_Hans: 语言模型 + - value: embeddings + label: + en_US: Text Embedding + - value: reranking + label: + en_US: Rerank +``` + +- Next, each model has its own model_name, so we need to define that here: + +```yaml + - variable: model_name + type: text-input + label: + en_US: Model name + zh_Hans: 模型名称 + required: true + placeholder: + zh_Hans: 填写模型名称 + en_US: Input model name +``` + +- Specify the Xinference local deployment address: + +```yaml + - variable: server_url + label: + zh_Hans: 服务器URL + en_US: Server url + type: text-input + required: true + placeholder: + zh_Hans: 在此输入Xinference的服务器地址，如 https://example.com/xxx + en_US: Enter the url of your Xinference, for example https://example.com/xxx +``` + +- Each model has a unique model_uid, so we also need to define that here: + +```yaml + - variable: model_uid + label: + zh_Hans: 模型UID + en_US: Model uid + type: text-input + required: true + placeholder: + zh_Hans: 在此输入您的Model UID + en_US: Enter the model uid +``` + +Now, we have completed the basic definition of the vendor. + +### Writing the Model Code + +Next, let's take the `llm` type as an example and write `xinference.llm.llm.py`. + +In `llm.py`, create a Xinference LLM class, we name it `XinferenceAILargeLanguageModel` (this can be arbitrary), inheriting from the `__base.large_language_model.LargeLanguageModel` base class, and implement the following methods: + +- LLM Invocation + +Implement the core method for LLM invocation, supporting both stream and synchronous responses. + +```python +def _invoke(self, model: str, credentials: dict, + prompt_messages: list[PromptMessage], model_parameters: dict, + tools: Optional[list[PromptMessageTool]] = None, stop: Optional[list[str]] = None, + stream: bool = True, user: Optional[str] = None) \ + -> Union[LLMResult, Generator]: + """ + Invoke large language model + + :param model: model name + :param credentials: model credentials + :param prompt_messages: prompt messages + :param model_parameters: model parameters + :param tools: tools for tool usage + :param stop: stop words + :param stream: is the response a stream + :param user: unique user id + :return: full response or stream response chunk generator result + """ +``` + +When implementing, ensure to use two functions to return data separately for synchronous and stream responses. This is important because Python treats functions containing the `yield` keyword as generator functions, mandating them to return `Generator` types. Here’s an example (note that the example uses simplified parameters; in real implementation, use the parameter list as defined above): + +```python +def _invoke(self, stream: bool, **kwargs) \ + -> Union[LLMResult, Generator]: + if stream: + return self._handle_stream_response(**kwargs) + return self._handle_sync_response(**kwargs) + +def _handle_stream_response(self, **kwargs) -> Generator: + for chunk in response: + yield chunk +def _handle_sync_response(self, **kwargs) -> LLMResult: + return LLMResult(**response) +``` + +- Pre-compute Input Tokens + +If the model does not provide an interface for pre-computing tokens, you can return 0 directly. + +```python +def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],tools: Optional[list[PromptMessageTool]] = None) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param prompt_messages: prompt messages + :param tools: tools for tool usage + :return: token count + """ +``` + + +Sometimes, you might not want to return 0 directly. In such cases, you can use `self._get_num_tokens_by_gpt2(text: str)` to get pre-computed tokens. This method is provided by the `AIModel` base class, and it uses GPT2's Tokenizer for calculation. However, it should be noted that this is only a substitute and may not be fully accurate. + +- Model Credentials Validation + +Similar to vendor credentials validation, this method validates individual model credentials. + +```python +def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: None + """ +``` + +- Model Parameter Schema + +Unlike custom types, since the YAML file does not define which parameters a model supports, we need to dynamically generate the model parameter schema. + +For instance, Xinference supports `max_tokens`, `temperature`, and `top_p` parameters. + +However, some vendors may support different parameters for different models. For example, the `OpenLLM` vendor supports `top_k`, but not all models provided by this vendor support `top_k`. Let's say model A supports `top_k` but model B does not. In such cases, we need to dynamically generate the model parameter schema, as illustrated below: + +```python + def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity | None: + """ + used to define customizable model schema + """ + rules = [ + ParameterRule( + name='temperature', type=ParameterType.FLOAT, + use_template='temperature', + label=I18nObject( + zh_Hans='温度', en_US='Temperature' + ) + ), + ParameterRule( + name='top_p', type=ParameterType.FLOAT, + use_template='top_p', + label=I18nObject( + zh_Hans='Top P', en_US='Top P' + ) + ), + ParameterRule( + name='max_tokens', type=ParameterType.INT, + use_template='max_tokens', + min=1, + default=512, + label=I18nObject( + zh_Hans='最大生成长度', en_US='Max Tokens' + ) + ) + ] + + # if model is A, add top_k to rules + if model == 'A': + rules.append( + ParameterRule( + name='top_k', type=ParameterType.INT, + use_template='top_k', + min=1, + default=50, + label=I18nObject( + zh_Hans='Top K', en_US='Top K' + ) + ) + ) + + """ + some NOT IMPORTANT code here + """ + + entity = AIModelEntity( + model=model, + label=I18nObject( + en_US=model + ), + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_type=model_type, + model_properties={ + ModelPropertyKey.MODE: ModelType.LLM, + }, + parameter_rules=rules + ) + + return entity +``` + +- Exception Error Mapping + +When a model invocation error occurs, it should be mapped to the runtime's specified `InvokeError` type, enabling Dify to handle different errors appropriately. + +Runtime Errors: + +- `InvokeConnectionError` Connection error during invocation +- `InvokeServerUnavailableError` Service provider unavailable +- `InvokeRateLimitError` Rate limit reached +- `InvokeAuthorizationError` Authorization failure +- `InvokeBadRequestError` Invalid request parameters + +```python + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + """ + Map model invoke error to unified error + The key is the error type thrown to the caller + The value is the error type thrown by the model, + which needs to be converted into a unified error type for the caller. + + :return: Invoke error mapping + """ +``` + +For interface method details, see: [Interfaces](./interfaces.md). For specific implementations, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py). \ No newline at end of file diff --git a/api/core/model_runtime/docs/en_US/images/index/image-1.png b/api/core/model_runtime/docs/en_US/images/index/image-1.png new file mode 100644 index 0000000000000000000000000000000000000000..b158d44b29dcc2a8fa6d6d349ef8d7fb9f7d4cdd GIT binary patch literal 235102 zcmeFZXH-+&);3I0P!Z6ZD&2~7REl&I=>pPwmkyzKLPS6iq&KAn1SwKN?+}WBfD~x~ zlF&OO^p+6vM$dV^an28)*Z<$k7(3Y`ti8utbI*CLd0lfR?_a4aQeI=aMnptJsjT!u zi-_n7;kf+$3K`*(ByK+wBBJZk4svp@l;z|&UU|6MIyl=95h=Y-(I+?1?xW8(*1CP^ z3J1si!v~S$L|V`7iSrpc6qHE9{;}dHn))e!8OBdvky~;niuHzsm7S*Z6axi!f0IE4 zkE@m}47d`whFzV-UwE#{*bQYFlM!LAX1_>`b|i`CSK*W4ib zoLT=-=V5~N)SFn^2Uqrj#D;Vhxy97=7p1O&!=6%o<0Fcn@RxLWK*W@Gdzg38?m1NdhC0IE@vwr6q{?o5KzGbM~v>Jb-!_YVl8?v6t-Q%X^Q>D?RwH-6em2?OtOG z7v4_nqJCwipUldjt04P6WpvW;tx1%vdT|r2DO7bZ zc?#zwRHe5@LZ=MdDWcv?U0J;04$#Su(ipxG?soef&!sG0?&SxvrPO;tme1T;`t9%7)`TrnoOyG`>di#Ce;iCVki2fDZlRp^N#b;I8-l@i(1FzUUXqnm3$ew{_xa z*mL)98D@6BXWC8ZUtB&je$K*48$rf>RQdV3gHmQclow)Q(Kgk?ds9>;{DjW-ei+wQ zXzi|l)cJx*#lZ7N67I5!27=-q7W^cZuPQvyQ%`cAJ~I1q@Jx%5v5+NN`+nmiZV|Ga(M;T{rYcKhMmbJ#OIf3Hio zqGs=ON3!f*T9mv&?6FY$@{x)pBi~iYcOTZ-fBIMKS++G+^kMN*GVQ^3Q}OS09#w@z zzQnxYeDsm=#r3=IZlK)c5Y#6fcNZQJi7;e_2U7l6YL;2Jnwzy7@0K;9+#vi+=MIi* zKr@GCGor{b;li5F&8cW3mVB}86)X2rK*MdfZ%7Lk5j6GPX1Ub7eq%GbRgD=?MhCU8 z+=`-mK~D1RWTGh4@Ll0;VFbDEa_vQ>bic_M`PSgA?K3gywTMlMd9g)14QJ^(c8Tb$ zZ$s9K6=dz-Vy$BgczkqX$@O8@-J492^o8F!$u}b}Zo1A7_463cn`|La=kn9~4>Nf| zilpjoucLI5Kk0sd-6ilgzK`x8>)j*MOj8r$c#wHA1R`1)wL|t9hRMn}ASY_;#bUA4 zpta68(h3-j^hHFQXC3e9#w*f!#O=11M)}KbXCXv$|9G9|-y)*1Aa-5|GI>M3OH9nc z7#3i^s_iNj^v*;X9X=Hl+oY89SDs!SdJ{eWxPc1Y zJ~4LL?2XwjRgl7su&|LEWL9rJ-qDLBU3^<{L*^di`xlGxwD&0cUVMwZ{^{M1XZ;B_ z-;G|=XoYUeQ{CoFI8c)Cpzwd=br+JL@$lW(Jn^DOjxkhthpV~Y4<4nqgtkliIRyTLQV3SGpF>O;{>I9 zq=}?`X|!sK3D=arzqukM#2d}D&tBXsDy~oGKo5w{R9UWI4%^WKo z@6;MPmP{NMWoZZI9BY~yBh9k%r1eY-+LVA9n5##|o38dk>B4+(QdM~J`94pKsQ*Nd zN!tqwWePPgk;HsrN?|%;(ibut{|u+g4RT;OT8R6iYrMgsW)&1&r-H($bE>Yh`18L z%cv!~+#i2}kvk$i49D1DnvWn`pvM+Wmzb9sd4=r-8|)&PXPA4K)E}F;`ps0>6%D-t z=bq=<$emEOx`Z)JGk4!o6P|G%cO18KwqkMQePm`G)ZftiV)$PBJ^1VJzDG#_?RUlC z;_qJJ8}4IkFHIk0d>SwrbhbQBzx65nQ{NAfPqm+bBK40ngajVVxdactb?m76USVMQ z)A^^9gQHl#N4ibfiVIH^-2nH1W6FeK$;qV1WZ&dLDRU`!dL8cO66NCPf}R<5#<_6L z=)ujZ@6J?idu&fmo7O3!kf`cy9+ZGs4iy{Kod|C#RM(ZRmUr(Xbg8%8bbMb+(2Grq zOS_GFMtO0&JGupkAw2vxcfWKEQ{h*)(V1H<-Dg8rP&XIuswfD>_@S(JI(>`18dp_3 zDc!R+X4WzNgB@z?b+f1fi7sF=R`SI51j3fhrWGe12hdO*78>Rrrb?yac~=|nJrq4p zw=O)k6$w*xrQUhI1Dpk&J;f}aZtwbwjN0bAEvF{L1q}lI+cJVAa36ymf|+sRxWRM( zi-`+rQeKiv#KxrARj8@y?X2yzkfD&#_Rx?WvDwUxOqyo~uk2qLzA~=xakJW;naAE< zz5VHS?d{fFbgs6XM%c~HJDsWTyxn3Ng8-EEqKGF!0e-4s0WX5S%|u2d94Cx^FVI{s z+SMFTM8;G_{i7)SBGevmxW69wNaPWXHC69F*49=$*5Ge3ks0*9BH1@(ZU-x#H;1){ zbzk-U$A!k`{h;z;exosQTXRK zI1FyD(_^lsT?N~o+VT$grEX+2N>)n@$dueEQT%qZjIOnxYG5`VC(S8tXRHnac%ade z9%e?fw=3(ODk2hH6J^Apri&b0DPHP_*$p065TWNywM1Nu`XdHKq@ZPEo4P9V*VIwx zQQXneFICcOKvk2k_m(HQjL}Bwu+p<=(TLz8I{Y8$? zCa~Ubz4a%_OCfNJmfWDO!kEu7qN3q|uaLU(@~M{ylbW->)Os_YlZkL?|Ao$Qnq5S)*L*3Vr2T`=v45-^H-ip8e<(@>Jjy_`U@JP0fq%hFv;K#~z0y-W&Fx)G-6L8yqud(n@qfDm6Y^F_eSHIo7tTivN z*-Y8|2x@lPGe1~Y4{*)ET4B)8fxwe9;52zCHC(1jIvbaKzOk{UY_4Rk2dc%+9P%x~ zkryo&!B;i3vROH^eMbt7&M#OiShztSN?0t9wx+a$#sCStJf0gGX^4bAW4%G#; z_qZLJK}B@z90C_8(qN~My-xY&lE9m$CaSuS)^sa#F3~#(a_5)l>Dg&i7cpEXnKH|` zgInKqk?q54AmZHFAUXa-UMm8jaJDf}wpCLj;wGG5A-Y7&M0A;OMoc)QiJAYua|Pmu zM5MnzCm|w=a3H$$?=k9x_g{bUgyYvW|9U6MA^D#XSG00S|L6Sj_g_P~SZwDBZ&%%v z3_Xd6=FoyUiS^&LjZb2*xGx2N;33-Le z&CN9{lJND3wFhah-BRFA{NKaxN4C_m?I4raSfPfmK-+0z;OZwkMD?%~`O9dM zB_EFNmV==!Ju`>0f2c;cwCLcz_p0 zMox~T5IBe9aFtN3Khvjh@ep^>c<KZNxR6()mpi~T zYx)!BC>UCd<}0IPTusLDEr+A3=wZ-BG!|Ap9P7askk{ZY9RFM+$G|2fB~UKb=+0l@ z#z-hhfJ*l}X25(l=)(U58^PxD#P-K}hn^RtABP-hI?61(FX2$FCTep-e(UwO|8;Wz zMoNnpB+ME;N7SA>(JP9-LHc{UJ)-=ZA_$96F*fT_lO)bK@wW<9yZ53r?Fwc8<_1jH z>;=2Uogh+*&{gYkuu|oGjRPR`25dYuzDMGw_y`*PoLnms}pZs-@Mm$J|aDCHeKOi@pK64~Ti0Uuy5n}eQ$Ksn+ z8`yfQ>npy9GXJApwm)LCq)fb2CGpB%+=KX%g1$w9JiWWCiPz7cd|oWcBHW~0@7}#b zsPL2vQl+cX{nx|$(;ahkLsIn3rZ=+jnL^&)9N~gTPv(NOkx@MSGMtU0dVljo0?aIC zYAz%0MDuR%;D)(Kxtm1d?~s|;r051*%;ymQP5A$KPck4-A2hS=$1BV9w1fVEcaxNj zJq2HbMPc-9nZKFD4aUpFQVDmmolRuT&COXPEmI{hFRpb(dDFin(>y%#t@-#gF!ko& z_<2q=Nzwx~k7)iTd(*2n)|WQ#99hUJDm=YXc+I}}e%kso`mZ1U%FiBDoLafobR$}R z`h}Q$V>7z-vqb7EnZ`F1Sj5GsrT_7NZbg3t-ouZw)q!}CV2e8~XzMABe8DhVtPh&Y zQ#88M=kM&&OLFkziH2&elQ-`()lV^+`IeOV_Js?iq;~_>yBRBA@&463KFUg*T)=N> zTb#ue`ol@N0$Mwku7aEXIPd@X#|XtQaCgzIPV1$w5stTdO3A98UU&HZ@2t3_o}=GO zQ6m}uS)chSI+m6@>@SSp&T?lY+M3dcP3cj@wp{xAzqo{m#Ir(Hm9Q}PuDUYtC z5a#!Vcy#mT(hqV09P*qzh`yEpYhP{E zzq#;1fh=b)k=J90y3-n&mRHs1tRE#~F%SMOdhN|~lB|ivKj~xniq#XS@);p?rRCzr ze+OY{mwT2B!o3I+w(PZ*lbDv?fBqMb!9tzx;OM>^C%M_S`(+GQH=$?RxFH(_lHSW7 zBZ^&6hqpR0i_VP~f9V6Olf7uM$vDY}qni&z0gX0k>aY@e{xdWk`h;3)Q99tX5OI|= zYqK1);EP}?fh^|D;oIdLu(YpOlyXDJCWBpMPUW!P&K%v_)741L-pqBg;1xO>R5^Bn z5)gNS+SW4&HiXWi@N8#YdFap4P%ynGN&w;;xKpXxG(_@eC$r|uc|-&R0!<-3_9@yE z@#U2H7B}SS#1{T#spZRuVeepOfl`z~n;W1n0|Mt*`IA#T?ujYrxN&W4<62A;!G zzA1eFUZW5lBTMGWbbiJmGXQgpD9~UDt{Bti6_;YYcyB$%DQt2+ zSp7~aevwlKC!Zd->DPV{*m$;<0U4VZT=B)So-|ZL_XbT2Ci>?~4I9-M%smh2Q6&Sq{7>{5O#9)yD+q)@0*7%M zBCU0u4_bY`6%2|7CK=m)`f|Rx7|oadqx$?gOBr0_N?$~;<#g51&9-24ebt-S8*gyO zDi?7FacOk5DMwZ;Uj_;a!Tvos3?`%cjY!&za`T?T*8N6fp!2M4$la*&$Gx8sH+9p$ z{UIXT8LxTah3TxP8p`RVQb}2xW-9I-#8Tyd)F#u+4D9GeS91pU?PO51dgPBu49XB9 zA7EjT0{}^#Eo-9>tA`*IkZT5RGc~DXT0uu2JWHz9l^SNcerF3Kz>tdz>H3+kJk4uh z&~3i$Mtj2$SA*9JsdZeV)cr;H4~Q`Li*U3wSSw=+JCu7#PYcG(U6NToa7EO|2OVpKFIE#?1Q z+b2)NmdJUM48(XB%<+-psBw zZ5qemW!*bYrS<;l-AQq{*`uQ)>``v+G@x$ zbmqfBczBlIzP?I54jmGa87zdVz)5(La;;_uqJ;)OyBgdxoqT?Nm?vnT6Ugb+_*SvO z1J3H(>;!3+9<&)46!qRJZ7pctScN0X15>H5y;`=>r-IInhJ}ag!e)6H_J^TT>jf|S z&OMl&c)$`D_M-TpgA|udzmH~Vx*|Z{TUTyEAHP_;Nsfm4n(DOnl$_Fj#hoC#bXSbi zi2uyDC@(EKZHp9juI|;58D6(eqKvfn7L=Eav|{W|c+E({dP_+YClrNdHR-Z$-_&U@tN=G#usmPV@%Vc++>Mz8iZ(50G-PsrIR> z2=5ahx{BRog+~y!+y?+l`5q0NAn}2PV81GtYuM9-1>Yh14A(^;;0~3$P@#TU5RQ3H zrE3x9_^`q;zW0+kv>oY#MM@({D1NNPKHa|Sz}96r!T}|fKHM|WqP#aE1XQ?qPt&Mm zU9*qbf;1E27{&)hJpQ#QvuW}P-B?`FH!B;k&Z?c;l{v3-1+6_kX*2aLjrJWV zt6wvV@E9RKQ6`f+X-vDwzyl#+6yw2#Fq`LpO#fslNPJx3ggp3SgEruwaLUgW{?#3{ z-Op(It(`f=K56PD0i?w~XZYLkeHV!FTa6w4D_+ z0gq<~MX-%Xj>`5$b9&=brPW@RKo6UJ1&vQro^uIE#uDCS%HNo3W|8e*(14Svu=~1K z?G(I0F8WXwvf_7O%Zzn#f;2Wk-SNnQ@|R0>%{hhIwB%a$L|w4qK{74cgVGtFHPaI@ z+|*%ZXbG6n(w%8API5J@21)PQO_y5T9Xo`};8TcS@@FvFZaYk=PU@hBG*Gi25}$+* z1Dk|IU0|}wgCf)4cY5#Bo0kgBKEVe}mQ_DHNHU`IFm$htQ8|bt6=+#+R+q6ElexgW z8GF?GM;a&iPl#X68?x(DM9~S$w?=1jt}?%{^rR3-eI*k4nH%uPrGB#d8D&+U#&lhf zpFlamc=7@wO-bI9F@mi|)?gf!dzxuLwy{&pgvYeGCkis=O<6NS{V!k3yqyD63|fii zJnc~jOLuJ)lZv?1uGpfJlj3T1eKCe>g#$DW&orh-XfnPQsLj>nV$U?%(g2e@{EHzZ zdm{2pEA_rj0$?8>CVc-uTM8B1h(+RQ?L&urGKv69m|2=#vv9E&18X9KtQSd2yyBi0R)~w}X2*wV|_)c6Hj)zi9&EPI_ z6@%IzFuTsT9j7E!#n7n&@8UdXqk^s~5OR&x*CFP()dG`)+Tb6W5ls$b^xwmsY<8r4 z+M zwIt5xFF!r@f!u5dwn7^yZA4CM><@%bvpL$jr#hz6bd8H z7+ue;(*MvF{+l}=HpxnU{9Nr)S*S8a!{rQESOVUV7b&{k7IE`Kl}nbeHURmf*#Lf% zGXJ41{p5kDfB({c@d|?S_fyyt>rN zX~ACP?<{Miipd|R{k$| z4Pue z1z5eK%r^kF)_PucY=1E=87oz`eN73K~Ntw!|5i4VELmw*z@0-%Lzz~;FBSPe9A{&qd8`2a$zDq8j5=Khc8 zFhWes{4zK2TUIMIAoPLrjqpL;4L|>8U*{EG06EimC9i_#E4f~7q=G^ho06t*PEJf! zApa)4E3#vQb=77J>x29WTq+gd1aV;nKuha&@SP!fY|wt}k+7cmeFj?|3DhZ|@EMiJ zG|j!5wuw~NIt!V|B}IMHq;hJzT;!p-3ZQ*@X(&RQ&EdCXkQw-kY95Ks8Y#M?ykX6r zL(Sh=1HOkex`1$7*6K1@nE>d)>|qLBzwcJP^Q_)6L0At9yf|YBO4l<3=gy@y{gh#Q zKzZkVVaW1TW>7Jy)I3-{aV~51#I?FX1?@mD%TXOP%0`Oua5X0m_8+ z58!9if6`|ygf1)i3Wghrn2~D=X@jqX@?tWCt6EW6tl?v3)5tg}>(w6x;k<(Qb2JnW z6k~9O0#Dr7#nwq1EBtICP;&{0k$$q5-{6ZjQhQc?_vzz84SogxjdfyTsV_y27{r=} zXFW23J*cqZKrL(Lu#kJ$sv$eTt;cX}?X4)n0*MHyPJsiOdT@KWa9o^J+2WV9edx}j z%Ji8rrz+U~QI5}1 zB|)~5mlvwpx4&@T41Y$uut0%4;%fRZNy`52=tb0i|B-In@#m?A_bt*pEe8z){9eF!5pC|nnka>_#pu?brv7L0=*r%UVd$jwWa;+1LE}1vklNP}S}1g4;4JXy&~s~lMD~4}xqo-WPU}twf$h11dw$Yz zt3O2-y*AA=7ES?&1)+X;BDu_^pS*C`#s36L_aOt|Sq|4*A3N?Z_iY65P+*X*qPs`U z07os5)cxDptLZ!go)$TC^&5#hW((!#9rEaKh6@@+)`=DU!ZmWOb!F87{_13qO@vwH zXRhGZ-X7+%ox`tuYBP#Vhc}?Hd8gyWgsm$f4&D!{wjK0X10DS1yS-Pq{8^xJc@e$I zBjC=%3w35{>&p=F`X!UVPu|3>WWU+Vm5V)uTnaoWU@mkw{3ElGwcx$vg_k|4tPW(# zr8hh^7pEr{vM4QWC`9yQV>dl(rJ1ZKOpAoLI$}{`Rb-VqXXy8-eK2hGzb}& z=?rd%osCbS*u>naYyu}jSxz8g;oM6OUC{!q%X&sQHOi%L%&_U4;CtFFrUM|{5)v>& zLP7g)hRj&d&d}x;AZ~GI=(z^->DCWIoQGPtuUr`5in!6Nsfcb3QR8Y|gI)1neqr-k#@g zTlRpW8rRI-FdN@$boV!l?#cvLU^j0q6f>ro1;rW%6yJvr8RaR@wGA30G=eeFqcsGg zQlBB@z)_DK^u6Hsfbp!&ad0u7AsJClra7PNa+v+R`)5 z_ph(A@b7Wg5MN+C8P^^@1wWNE3&zhjH_maIK|$*YLQl&TnZEx~ACqOcB;?nbeFxdN z9Ub~`n~q&1vU7C_C@6n?_`XGxepc+s%6M7Ol2M&h{2g_Kb7Rnme6@}rmXao^d@aYAEp#6zSUpsj!@4{X=22QRy6=$b_l%U|m&|!_Rr!hC4=-atcI#vASl{U{ zYt8ucHF;mT{cjNzZQjM$6^N^sDqmxm#A1arN2uA=&jCGSK?iEyOoDqIxZ|kM-idvg zi@g^{V0rqXY^BPL2h0TvMyDv0?^y`#WB96zF$M;FcwQcehW*eGoePNI!g=2i+#VI2{H5a5JH0_tgATQ!EogD4RWhw@0&mX>78f>bmPi5be|l_# zZrZ-shd~Q117^b14u(a+V{gLX=mt2s!;L<7@I{#UiV`D#f=Zj8adBMVOTf+6C#J476*~L*A5!#8+BB1!T(-)ZQv=a`GLzw3 z-kb-+GT^{5g)NwicDpyUsRMVIZr%h5+Pfs`z1i}`eNmQ@s?s5RC9PIoW^UERRhd9f z+rzT^E9c*iA>J{UrbVB8XJ_+PDXDUq@jS`G&0jNi^}Zl90dC|ua>z^SO6e{`KtSG@ zf3@8;#e8Yy`9TMSQ(Xm8a?wgDh&xC zuAj1s7RPz-6`Nz50?$_TB|~QU@O>IVLcM!Go(!*&X*!<_xg_?sLNcSr`uh`AdC{(t zwg12t*KeXUq*J6G&vt&g`wOZ+m{<@L4N6k3V|er9>XPz{#^&RuEuQ*bW;69{kG(-x zkO0+J<2<5=0knaSyW8`LPwWqABf$7YSla$f;>*QUvAAj5++F7_PNNzN#=V2345u4+F&-u6^nlw- zpjmQjL$_M@MNsuSY#XxI!drb$a7ak`S%~rLgm7eEVbCDHwyJx=ymf~1 zE`^hKIq}VdLDAZYb?Iy=uDr4rFMj9i_R0jYdk(lg-xE6FGwn}ae}Oly=ZE$t={pCV zRPk(e^3CtYu_gzt!Jz#2E^yZFY36`i`bRTYqOA|?vV%ofb#f~7#`s!xpQmP80@`ogbR<6sZUv4qhL?+$31=js92$%g7qgWQ49;0x{YP`QyPDq zao=57oY?W}ll3m&26;!(6d~v#e68VRhuv%=H^Vxx1>Xh8SZ3X@%ks4B&wX~o4!qax z1$;_z{R3kx(=TEeQb{xA9^pGd7Ysw5OdxgUHl*MyWV5CxCCo`Mqs#^Jo8vaGu8eSm zehncDFG3%fpQn$_Us7XAcJC4ZnS2;cIV#K!=wUthl;r6TT2bYkq|_Ucjr1HkPsFY- zfK`^*g4TZmQd0E2@UE@3V}iyFM|I84KOHQ3Y8)u?x9WBis8M#&;f+I{okb&Hvm_9z zhfoA1#)-@|*?HNfk^&yd0W z@`T2PesV>jTaN#rCA`e-Gm|;;PFKgWe34sqa{zda54_cr*>}4z*3JBMbVK@O95+lu z5{T`rMnv)OGLYBSo;MD&p+pdFt#zXa-?>n9SdxYd4`ssdhMA_jf=-jpCxg-aYG!B2 z(A?j0kw)-h9-!45<^jf@JGAX8;tv_>&nEFpZ9bdI>Kck4oyr@Vw2PnWGs|wU574*D8j!~H%IuGl?W!phAD1gulNIg~$n&KbU2PXZ+#_B-!;=?m;lX1g6003- zBC0+%GnufxHmPFIGGNS!)fvn0AJ*h4I8x|NEQxiJ$ew`@cOyz*N36NUxFzAU%m;Ii zlslILOCX0&xMU{GwK>}&f)rS>5Yer86#26)R-ELNVf1iU#>3fLqU-sI$R-QrPGF=Isxy~b5k zN!J|hMcIg|jHyqTebY76RrqTDq*F_V(!)o<@~2r?YT zp{X<+4uDFC#y_`YmpPPE2icrwFav%%0x;)m(x};{iPgpmAN}BydAYurtd6=wBv9ib zVa?_KgWo+(vWl0o{T8_jVMamN(E-yvVX8oMTQ&CE0y);>SDgk_eY~vQmY#}zpBLv_ z7<6{s_q(F>`YaSy{;P-XWj$_ddJS=ZZjg*QpP7mGe0#?Hi?MFnm^xm;Kp(;_#iOGB zCQ<$6bUvqDY0Ky}n+%mMGZ4=Eq$8Bd6Oa9#w{a|b57H)n{4uUjG{ZSmhV$Ua4V}Lf zw+Ge6AFe?$gK_sY0>$}nz40JqWnV*8QY!xu8wJ?R!~Di^6g|_0EnhZ06q2uR}52Ka)g-7++-FlS;Xlty56v z04wq=^j5RfKRWY9ZbPQLgPTX!3lGW6E4gYr>0}TsYW;k-UTa2Adg+w_#{5tUuLA7` zzZP?b*NDB#cW|F!5Q%)A(6zsvO0IpK>fJL5H42qG>pk`+dWKAJe78$5I~{36{!sTzEus+0NQ#VwCaWt?QP$;0dd1#<0qT zu}vtk6ZmoU%8jw!K5|3)iv#wI%~N!bj>wfS%t!T2@OfHe_8}Ag?f~r4Gli=MBO}{0 zzINVd45bAYhJRc%MXf7VB>4r_k~C*CDE&nDi0fUws9)a4bnpN2OFrTwuEi}1BR%%J zXlkE4{>|anFb27+U!$qZTuRE`#q{BUz$SHFsnZUOa(B?!bw>(EV|(8~p#KEtd9r!v zNSzg4szkeze!2-C`o+q+da&_JgUr_!7n$(mQT(D`&71vivWEb!c-e8tXE1~{{jB);{h)SvRb%2Xl}~iQ1>?F8t6MX3bKl^& zCO_UDT@OX^xgdl3>EcVzmkXiarH)%0vVN~7WRG(};Xd2&vH&KJmB1xb6gW072(>|6CxyYpRWRdX?(7#RBPzKdw9 z8Tsg@?6s^B_PxF(q*^^9yCmH#LfB=d#w3ED;kMx?yKFA>4@mH=%SpD>nA#k^BMw0* zK}EwYvA*{#foN>nnz?G#gz@Y)p}-g}bYai?Eo-!EL~I|GszJNu6Oz z56Wyp#dC7dL4rT}X8*B`pg@naZ>z}Dhsw6}si89bVJX{dw!K-S$rC^KtNin#|Di9O zHt~MMi4vXAw#2Y0#jYZDZVn7=LKF#?aG~B7RaVlAaCxu1*Buji|c%}h-M7H&xodtpk zm&Ro(`Y?e;`g6~<8)9z)Mk3l24PuutkuK+jcOry~q>2+Um6P=@daozbegveu2S#s{ z0lRng;9rB=R$RY!2uPZ0oiX{g;lkVe|LI$#+tMFr4)&)V0)xIng<=oRb!V~;6tG2gv80V zHj-=$l0gwILFZ#tPAD9SoP*H6Sjdq0T86{<8W`_e#qxtng)D_7@SH0+BYQiS_(~o@ z5Gk9w!Z7~K=!yT^@B%-gaGPda^W1_2^sbm8B`Ys;7KCPQl@mHCrX|(cRxD#C$6C!i zrnmOhznXctX@jL++uOcyX+j0D3JSkDW8GUI{?>=oJP$`#&d#^R#oAVA@Y6!IaPWBi zieY+LYFX=KP>3Ad|+wg)IR_~z-8l41+H(um9^JCji#pH8GyhoZ4LEe&(`YX z8HlW!PoF*-Hfgf&J`R1mD9Ibua+u7RBEWo5gW7AUT~bRa$83LV)aBcH5-%mzxtS*t zQF-%g$;Gw%%|-dU4O#9(Dh$jU28`(Z)Nr0BG16qnGU(SFoPo&n+DrKOs)9?58m~b3 z#uHxEVN*}tkCR&k4|%6v5Z8XUS+Um+;^m5^OfBKbqn{rTD0}fwZl~NCec9r$`t;hf zZe3Q@+OL9g|Iz+K=(K;B4cQ8a^QNo1P@L?ob!pT16o0DG-Oz!3qf1Kf3DL-Hsdc)3rVX)rL2pTFNXqI%X|x=QN)b9x__jusu6S|bxLSD!pKRydNR7)NAjfJS=9HlV zhbSsA%sk(ldpPJ;XR~sh(|busjo`ziE$$CI7bl$(6~QuR<715lJA{%L;#MJWc2HtR z5`8RzRt5lYKX9j>7#Ca?)_^`REJ#Z%pr{6XmDSXD9#L?AvCjog{(VZ|0}3-Q@_@5RuZkL zjUShj`cS1T^|^CUCS`iF=9k(K18wUE?ti4gy}3bo58;w_@}lp66V*MZjA`}zW=Y;$ z0%;slOBd=hJ&l2gN!{#vfHN}*suXP^l|FZ`U&HzwXOT5Z9T#E~W0e}7l^(FHSZgD6 z=2~XRuD)hv_a3Sm(hVx7^(CW~wn1Y;_;T-lLXOnLDo;>y>po+R)73ZZqy9L~@!c<= zx6akOm*wmc`)k%@zd(tBuc57Ea{Q}0DG}3jaGehVu&KHRDKl4Lv&^207VmBX(7q2!=Eu zxs<=rJFp{)k76vC@#RI6bS#6O!U?F>)LQfF0|~R`mt1vkE=3jXkV+3Vi} zc8D$#lyZNu_Zrrpb63?^gnA`~J;=S4&!q?|{M3Y~O1yV22_uPxk=!@j@|1GOX}JzI z-v@gn`+k>6p4^sb-_#LKE8NVHEZMIbj^1DQ!j+lXU^U^(@P6Nl6>VvlMO4T=u-i+6 z%+tWm&c((Q<_HJ>(fn=fJ!I0vD{JP?ReZ-LJ%&byQ6ILO(gs&v@9kLw4o>UEM zL7_dlPAFt%RIioj0F%QkP^J-+eqJwoTM^5#5?USIeCB2<(^sWEVT?GzKZx{ zvt$gZ-{uo5fn}aHAM4rkmRpyY!!#NHs%-CZ0;hpXEv~&8&28t!8f>ev;*0U+X>z}s zh~8XnIq&y^KK&W&TBNBVxtEO9*rz2^qY!AxF}B*(^=;@#_Qvvb6kKx=)W)p?Sm3#<0oiR$VJ&x%bKw>z*Goly$lJu?Lup*Jnez>$0zp z9u9DYNgY7!lFn^UG2C(T^k!3p*5qkdUZ0tBwcUI^dce^KxSn3(;&fl>g!Y!tn1D}( z4r@RmG;2vmx z?9nT-j+JEyJe(bFv7A9KdpbjSDxP)hT~t;`vqOSjD+FbhY{tZHnV5WH*wmf$LK?uN z`O$YN1XQ)@>FWphhYTdK{9^%48cxi$YBq3RJ0SKjYsP*D_Sn%n70rGt#YC2A3$vxn zdW_WB`Kci1q&Z=RVi%k7b!%=+2F88L+jrE{coGwrWb0iqtZ0~iY5?Je&lsh)8BRO@ zXKQK*`f8CtZd)$q?Vd1Oih&QX)A6&*CPfjx=-}MSBpQa6M8S*V4AvEUfz&C57mC$= zn8!`90l%Yd)rTK65GT^`^3Bcr*99yeS9U z_aFNk>}26<cfybbL!QlhA6K}VWCGtnrXw+U6UzhQwSFm36 zBs$n&d&Z=Fut0B7Xuw+171YH{?;hCV%6USdyxt7veg%X^_n08>qcn(3=tEk3Qs0{} z;PK1^d(fnaRO@2cQmn@gas!D@)*kkNUJR<$^$pwI*NA+TZ7tk-IERnb{h9o$t^VcH zkT0o5og0Nu%j(fa-8prW-c^k#5#!pFf-m3u?#`vF_o)@-*N+ILXG>R3q-N;$RJ(|S zMl=wHl{eRY0J$!8r8IjssfOI3QN4#YF%$ZGq6yxZ&otb)^%v%(4?2lP+(x$sO0yPU zbM6`%s?dnbqO$Hey-`H4@eT7=)@Jy3t4HnJn;2>pa=VjN%2X?k4RX?8608;iBwaqx z5qdIO6%LoA8mn4-5CH}+KD{=qu}Q?xRmG^{L@NB+94s9TpV6ZHQu|%GbhByW;Z-be z|C9G-+q(?W*HB6`HoZoP2yqzce(|khl*pUT{UQ_qx|zn8bEY_06}c+@uA1V(&v#xw zVYzzm#4ln9EfG4c35Mz3z4Y5pX#0ozC!4_MEi4aXl3sieaQ>E}%_2cY2Vl6r_;kG6?iVlo6+vJ(-@((KNK66keN(~ef15aMy!jj&iP%GqU2G-*-xaXV%$;!E#zr%h$X z&#j+2WLCXWj7B72jBU^SWA6h zdTIGeg>JX42!r!qLU^K!|!R^2NLMW z3GA&Rqu2H|N&U;oc<0WcCu4$lDl8yiv)mJ_ER^Qb&~ff$;`5Gr^Z z&^4|dQ+I^$?aK}RW}Wcl{?PIp-wzyYIX1znI0^uxIz~u6nw zHczLiv`JK(m$PfaQz)a26X+&+TU=qw%C~U(WVEMF>H2?!R}PJW$#-ZncAI! z{Pv=UTCegPOX_oN+rgX_iTmx-Z~<(uV_C^gP-gr693ZJ?w3|S|6bEYE!wMwxRrD>m7$Vp z?%KVAr{L+)nYfJO$=P>6^N(52>2diV?Q+;Qhuv#7wX*dvXT%vY(bOtjCl42m~8$4ZecEoN;^d#$kglZ7pB%n;Dj zdHi8|FTWFX(ZZ`O1mBrO7GF_mMq0L==4&i3t^(sgyNqu!h02z$K&dN}KjJY~y*MjM z7=b@_%C-xlCiYcaeX#bNKmKO3wxGQL-*qJ;`GtMzta{c7Rp&a34;2Sn6v>g;Z2Pc3 z*tJ!Py>bQ`IKp~iKz4SksxD!L53igtSF)Z<(aL8^cIZH>K;2ZVfbIyZWZjhX;d?6z zV43<7Lgcv!o^o)(IHfdPofT`-h{Z^6pi^5x|JLZTZ_n6hjzt5npVP{<+Lk=6qq>2(3rTZadzq>{`eJ!b_L(%h42IJ;q^x<8J zLp5|{F}2%UGyh@niMSRPe##k4&WX^F;C1brd}&x-->!BYmr z2zMNmd7Yw-HmVL=^kbRL<=L(4q(gyoa_+6kjzcuu)2BL178MRl&qb42;C!VPtjg^$ zDGoDx{90qr{rLTx4!r$K->EehX3CN|7OIN3xZC5ycO@&~q;ZyV2XD<#SAp}VIv`_& z6>5Dx-~X+19@Fl#WX#2VLH~AN1cpRKsOD9~a=pcLZ&lT3(@eEX0F3TH!2!oiL|IL$ z9xapkGkrA)#_#z{umZg90VgQP8hEmLGuGtB8UPW!!u4BVHio|IBW za-0Srws1#s+O{n`7+%k3fC76QIB z9C=#S?kkpIbA(KaH#hq95VtR+<4haKhwA}~XNC9A#|LuVwJ#Olv@8fHG_Fn<Org}ExwH9^*fAXey z%Yp1=-z6DK+~v+U1&S%tonw`GhmN=<#VMVT0+Yjeu~vHkf6JLCbn5gpyNOy5joP*> zj;~Yko}@bV^4u9~`viLc&FO7qaf#b{3UGm&Z7#sPj5qDY9@z%ogvK&jw>C;Pyp*Ok zjk5;hCX(aCf|$iCz7W&WbtTveb-JU-DyOqAK!8kCpoV9h!jObs7PkGj&P*(SI1~Am zJvZ3f(NLV#!>uVNU0F4&tR>m>h^!m7@b&$~>8DvxPL22N8!9oE}pzxsn>BX`ot;|RCu;%FwgiVACI~gS_@fu-uhOrSvv*%sxnuMOcdpgjrWT*D!4IaPf7q_KB(b`oChfI2{312Wp>xU5 zUwP<_dis@zlsRAaSD=2Uyq(^Pkkn+ezkj07ipeDW4bJPdPjkZ*mN2YNH8d6oWQSRu zax+Pl4+mX3r_&5opD^|_{bC&UQ~eYj2=VTivN$bIaVm(5+3{~;cgS++-W5^iS zPH@!Gsjn;@m93|bK9q6iQLVSYyk#nGI-8!PzQ9C(6zWx#WVZ_G^*n77TfAJ}?u3bp z6`zxSgkgeHF*25k7@ria!8@QJ4@YKcj7qpqANP=YXSUf?Dd%kQM2f&o63jK z&G=q;_p1})VPT^H9@I)sv&w@k#c)cG8r=!z(^tl`V*>+3 z78_s=OSRDo9JW}S1#aqGBUqb^fW-w+_B$XhSMrz2-;%uK3Ccbd^DvTPXG-KJQ)W+^ z4F-~?8SeUV*?p`++Sk+A88dCz6GH+;Xj>i5(Akx-(0c8J!_ zDNDSR%KaN6Oi}`3lupl8AFwr^LRd+A>zT{-S9b5H+D>9~?!bgBd+Q>go#$NY6BRN@ z{ZJ#pIU|GkJ(85^2r-rLrIoMX z&Le)+Mrn$tk`H^&7wz6N9%Q@Da(Utkn3PVbTc3$a4M`#jI7=(6A9X){X^oz7Z_XXf z6Ei+Fe59yWah?t#dU|TVm(i z0qNFGiZ?Okrj7AdK4bt;dV2Y7vG$?D1bKTMaz;07&a6J+`#*nxFTkGowNtd+30+j0 z=^=)scyInd42=j|RPUIU8c;F9_h1;Zc4fyETnHazA3I^wChOWLR8=|>!A{<~F2~VO z6%lkS_`#LNHYlca3X(3?-h#h?e1nQ#>boC}-5tN^&JyhQG_H7s%Ln=8N1~-i_4DTe zcA>*mNp;s^iDKq@cAq+OI!0%yZfbKC7W|$o|0?QpS^!B(l>=bCwz?c%-#KluKd+_U z*fP!b)0Y!{D|RY8F`PpmUWV0EpcQ=C6xEyR=PQeYibf^~Rik-#0Q23fo zKe<%%@>R+v^OExuiqdMl@GT|C3?XIzm1HaKJhj zc_N^``?>5GZMN)ffyMG&{3)=CrE4*Ca4%+WA!t)+CQu2ptwrutPMDFguX^^kvu+z(^P03J!WkGg*)JN?*tmd8u=iX?g3TF{q*QaYbNBe zFz8uT$KGt#adA9P9oJgirSXFwyKyWyoK zKaSm})&qo>54B*MWrbq=TGf?Tig2(&`{N(pYc4|1Uagt)J*j6%lyi`7CPi+>TpYL+ zaMlo9G*?}zmReZN-W{;Z0!cKYxhQL#3!bgGJlD!pUc*#1m0!%4SKlp{hpQh#1gN04KX5Xe^7Na@004WodO zrJ#=eHij;jx<%jb9X}2k32T7Bs5+|*O2I;A_}cJ@+sLJPw?PSaKm?Ef%p$oaIt^$A zBEYSH(Gc0n>CU%?yRFzw9j@XdRes&57DMy(OeX}!IIs$@7U3{k^nnYhjw3^+BmssW zUfi5#0sXG^pY*s=EAs=X7aLgesKDG#5N&eyc=$fyUp~H?ld6{)Jmu zQ_a6_q9SOrYf!v|2RWE5&LHNFZm-wRNzG}l{L;o!)G}y4tJP{@RJ4&FRt zb#BeGIC28*H9+;_^G3TNN$XR-C8m??GC(3Nqt`N|&d>d`)?`=(nNnWxWL;xvv+-#9 zW<}g(lQ&ZmYayJMX>GHOCUgBs+*W#L59W6e$Bb}} z&A=}M=xs}=`wAaZsTWd&DYRroxbk?Ce02dhCI$s9Kt{*2)|luoT{9&%E=8X<6&Uk; zFz#mRcA2eMuwa-*>3J?mRofqlEfc>U!@a1?j9^bIv%asg-=~y^fYTP8k|b8UHhu+K zx!sH3j;p1!Bd&HpTQ4D+lg0|&6=^KLQ`0D=HPhiN@>l@4)zGZ!MXUpXnup`+cZgK% zntRb|D45+v5~vX5Evus>zVz82;Ltm5k<&FNN^{IvHL#4?Mpy=BXjN6G?d)on07kkT z1MdnG7a59&hp9DgThDu)$}ENl_OLzSG2Ml+tOr>O!%1hgYzDTO*v-2~W$N}av~47k zr~9g>mCaY@vy_JDR+O=8vg(0a=lTo6o4u=f?CBss%6V=|T0362*M%YNtY*FT0F71L z-1ij?-k-AwZt&8!tc5O6tT1A>#H@r?xXC?EUt3waP-5L9a;>n!z&&ORkFJRn_1c|( zabOX3O${f+o=W4|HfHd(l4+@@qGU;<`I}jksyj8#OY6{C^fhw?0A0EUZ5Bd`p)7+E z++AT%It*~%8gr-O<&#-pZKdj}W=Ds3eSpfx@M^?JC2$uhFAxn7u(y!zXYWq~oV+4B zm}*8zB6FsY!Ezcmjz1mF5@sE9CnKZ~k2Jx=di%zjIWbvr3pv!DqA2>e6e&L ziQUAsw@?Voo;sG)PQC*l-adexJ#MO--EepS`Vkln`w?2B;>#Hf4v@{)__o*9Qtg#a zw7U}4xb}>9O=5CMb^asffQqvb&NiD+bLzq3i_ZroY|W=tVV5Woad&=7o($RthWWjU z>uc<-%u_s?8#~7dYVKI2W6B)!!|a#wI*c5ltZUACfmhR4pOjklJZE}jn1aon$-`?k zx|I%P?okyf$!`7HL@QJGx7~Mh7^~maE*1}zjuODgokPCv)cJ0hH7#-FzPqnuQzo6) z)@XW2CyBW!2eho&tAOFBl!$rIsGYgiWlQh4_UI#c!1ELV@MwesZ()>+CC_9|i(3uW zsx0Pd$@f&101jD$z`90>Y0?yRDTC9+q?C+_zUDNTkB6}Wn!oRfDn$R{A^!Bhvzk(C z$*zOIlU1Sjn0cqEfgp$bq=eLZFX!$qa$4#+nQ_5&7+DZPQpA%*rp2^BI$xWWi!MJz zSZDRfvR|MGM~HAacqiK+X+{+?Igsn5Y*l^p9aG4@)xDp2&KN)`U}VPAb*PXMlYFyp z|5OxQJFt^jxt_1~SfjOBhqPD)nt-y zWlK&`JriG1OnHCUgv<{a8|G{x$X9@;35-x|{e;Tysg`K(Qi{zLLU**mwC_|%-HmV? z5aF|@kZLDT?9|DyBud!>8wq2fY?U(-G%fbNWU0s52yn?vfxu95RyHiKiB{J#luw=D zgjWKOsVWDI3^jFkH-8-L0d&VUCKdS!SWOTwQ(LVTbN9~h8|vd+>4&VrO{`s=U5=MD zU0ExOaETc|ED;aWP8o`aM?kN}Z@Bj57rU?3?fY&QyxBJ_$d{uwMpoKfl%ZjBJk7nC z0_&#OGT05^s+cB{w$1>#)Ot-Zg5j|1oJ~N^p}W9UjH}H|vQ))`rjU@pN8Nia!RVC3 zjBnAK7@o)e$0rsVqG_sRD|)=TS50}@H;v!EWON)$mG}~psO@<^&f3X8GS|lBo96pN zUA;kpfj9nMkwzW%{`csWoeX}euZcT)TsQK};Hte`Df z?Zs7S+O<+7g>coz{w#$#ljK&)Ia?C8kAwPk@(~MZ_C1{hNbDg<(rR1W)OM7e!!os# zkVA_0%>uov1B}}P2SNZPfrxoj%RVqy{i1mf1)`xsEVuy<2cVzBIWT2PIaC3M=M zty!?A&8(t1513gXy(+&%f9_*) zKJ@+cTF<#UYKGu_Nj<`20Oa$olbE=8#+89woXYJtxKJbyN7dVI2ct@L)nXpQ2|=@x zfw8Ar0}XqHL|}N~QVlBLk13=ieyX*}qe}O9L&He#G=kE#f^De`rgim$@W&C{YcUNB z&CGY5_GR=p{Fj*GPpX4=Im%=yflmRTxr8DXQ;~&K*^~}$XOd!hT+UO%-qGu5vdi?i zi%E41$?-x>W+A2OR>hq9VcgnD8^Cst9ERv^VRG{!=@4M&rX(Va$_r(88=TJD6$^1)b)0YX?TxJzwazqSN);canyM`Y6_Z`&Ck#{~GRTGW$ed-1SLS=jxi8f) zr>FO@SVVfATVI;)Ic43Tqn-ktoNd0`-qYAq7xe+)r`h&_2^x$8CMAs!*7xEAu+@oQ zX`i)peeJ748PpyQe8M^~-Q$aCuT-@c0-xw4&0b!$#IDw16362h7}0!r2ss~ck2X}H zCeomz%aY?L)7(VD^H56%|*nTj}TG|b@0tK&pl^+ca? zNSOehsmV9%sth)0E0+UJ8^&X& zMy!>elmoU1(tt*nrm+Z%ze7pa32ZN6;eB&jLJ6I2!P)8imOH-F#eqK*bk5gv?4^g^ZWgcWs0>wXB1= zn>9?)Bh?d|M>}HQ)Cg@q^X82NhpL&y8C_TPV^)Jb!y2-1m4$O8F+z!nVjc0KK9<1! zS=!e3c*^T1ZiHCltcS8!RG-RNS9tW=&^f~6vo`hilc`Dh>`3JVO-hW3=q2 z7&Mk%rKj{grvmamkfh$tI~t zJe`D9t7X4Vz+9|_826>~u0D3C28+tl!qG{JrA>Ip`luv><;?@}=KzLoE8CA$jPT~b zrJ_65g*qd9SvU8U1JG_~RgN_Wvj)0b>1cbmrNEP@g*B5;{8}EoQgs>cav{cV9yRxV=Ug zzvD<)gm-&g?7H`w&u(OO_x#gkS@crOyDE?JTIvfW3mQ9XMUVMh8?D!}$!v5q35IFL zSqG>E)H?n}ktnmBYXulgs8Pp0m&xGL3R#3Vt)o)S~4U##{$45I8@PDDi3Q~v$0%C84x!RaqJaqh7 zjB%Ek)B^$hPgz6n$e3B-XU}pscCj~o&Cl#r&;*v?m4Jj;tCi(iZ^>us2*otMekZjm zv-E$gkmYcop8T@1BV5xBd;SH%PLFS2sPlmVw;=3V0()ho3a71sg%sn%NkY1&zJl{| z$BfNOa?e!5UMFtNLl~KEQ;!3U`NN^mdw$cD_x9CK^R2_wShZwbb@i2>{Wlc?Z(cpE zMO&;Fohp#5CF;!&%rvu6P(N|ydNT;r9v>7N2$;IDXQ;hB$T|*Rf~X*lxKdL~Qp4#u z{%XN>FzK*R@UAK}P#lletjD?P55C?Us!R{w9XyAErFtFA1X)4C#GJIo1MNipyvWo1Iah`+b7=_OloJR1E3v%SgLo5pXg6-|zhsi*^yj zZ{9Q3*FmUJS=ew0#vJ*RM~VquHqcbxF*z9v@zwAB^IvCS0qO0r0kZgO`I<@Iwf_sd zJt%?MfMjgf8w12eTf!yQ@Oz8@;Gy{a2K^B@htzl)XM{Un2*e92 zc}wwgH1wwQZY_rM`zu-df2Y|1h45?+JtXIPGsiw7O%VQg${U)lov4%afi~a!chUVJ({)8J^NYiy#Ir#9i#M0dp**(>;3w9G)3_L9cv(- zQ6Cr^sn&6LT2FV89MqQDG7b=Z^`O_nPvFsyK&*W(PpuC>PZPTQA2R#*PZw?cRsfz( z&hc%)xEBBdl%G$CI{x$aTMh3UK^ksvtVMOX|H9UdB7xruiKmm2p<0&99R8z(xxliY zatI^xzVVYH`*_;F#8ynI^q-lQGq=G*W|!B^&$wpaQgZTpixEl6eEeGxN-sHVtNBn_ zR7flq3zdWZGZ_Zrj-2-mW)riw`5`F-*(io61KF|r!e9CIW1u+fXR7pXXtS>M&Wbqy zNun=68$xsuy9;9?C&56EOFP0&|5RXR1Fsu@FMi0DcGo}F;$I>Dy2>O+)G?>CGH{nh z19hC##%)v9AKp_tILbFcYo9~eWZ(bTKS5@Ic#KDQQ~v+P7LN}y0ZxS7^@}8mf1PE( z3ns(;h>iFMIRBq_;UFeH?Uh(=z8h_xTnynECp;Gn2?mOoBK#gr`QJMzs4Q3wFEV-9 zlnB4SUb2kp)Y&n)1WI7sFzI_{o1GnA>%;nwQU%>e`~vn##_AXY=6r<1dRV^?L618@pmhE??JrxhmT*2yN~y) zzoO{T3*D1k8^3{+zu5A~nfW1jKlCxWoh*V!zn7>Un%#cZ%y9qpP!D!GP`cTea(?&C z_aanFqw&elpt?;qMGpK`HB3PQ(m7= z>FiuvUeo?9hyPtbIaG+N)fmgsz2!`h@?mLQB!Ai88ghEFCm*ZN0U!R`)4U4a4#%}M zSZ$^dF4P^VkDL*>i)xDhkyGWj%qmbCJyknz$kAOi$Nvdc7f~RW#N~sDoYMcr)FXnn z)#M%-%N+fxJiRsq=s=Dh^XWqL-`++8GqihrXo*oQCV2N+%4_alziJEx>;l0>pY>9Z z1f*G%reN`(rgE(H1v>vCD-py$rSKj@GjSYm@|(>S45zwSk181aR=+)n<|F77nfVHhi}#1wrfx?v&+QRlzb$fn@_F&-`(BUEu`V1occ+}L z;pccdX74>41tzBeESEY$b(uB_kMN$@1CUk_ojcZbN_%YAMbWMJ39QM z@2rlD3v|CsVl>}e|F?Z_Z?OivMbNw#T5<@N=pOKw4e)-?tGhKcyBwCpxO{<8(JyZ4P?O)j?TcZVkn&WJyeiqnlk{^J z$NJaMHj>cC!gDJ_KR&+dZ{P3l*WGV;Y}Uw&wzh_*kIJsQLa!D23ykeNwrg>_j7wp= zjE^Pk_`eT4-dErkdrd0ZX`9BxzR?AeNsCfxjv=1smG+elYRie zV{>|g-}DZ)$_gN|0)>C_JkNo^MiFxH*w(a>`Y8%lTwWJCI=nE*2V=y~W$U0QaGvoq10i9Ep{$RnH0Aqkqsg}PVzaqy~Gle$hb_$WzcG~sE=xIF&bjWg+&bf~+#B8q`@83^JR!2@YTe?I_v zk{lWAwL{UFv;6hrCcnYO1daBD3Q$rYxnD?)bE()6qwH7=!rzDa;od5o> z|Ngv(_zDb>a$OtidZ}MOv1fg*vJ+K~c<`kGhnN?*=-;2{fERuP^|L)(tYsI-nNisn zxUNUr;1XB$@+{ZVuj6ZpG#@8YiOa(rBm1?#-XLULTRw*{@K z7DfK9*Uy;1HK11n0^S6S#e~SJ9p6PO10q|3Mcp< z)&J1~9~3sXjkg8oq9+DjTc3aa^RM5r*kFlU>nGy|&`konyS+EcnjN?}EMdQuotX=2 zy|2X5wr_L~$LFg+{t!WSObeL5%6$P{50!~>t%^K6d0@uP`nxzq@hPCOIg-ndIwQ$6 z7srEBlKP#KQI6Wb@crI8lH?ahUm4#aAKzvQ3dne+k>dry)|ihDkk{ z40r|VhgbQRef?RiA<_2nw${JDI@ zVId-_?m8lJ{8K2=-d+u4tk*l;)0c|UzH}fxj z{L%{4algc@nk>-nkoG$Mm;8`ZUQCr3A1H85&i_Fj4ACO8(I{ zTf!))>#g9i0JiT&mA`|RI7$)R$vRyzutYXIFDhI1CNb$(MfGJ~Sf>~n8g{kn#*6+@ z7%?LA%cYdpocw}m%%m7MZipyh)PBFg!wX!WKeIjqLB9%xnwSs_RbJ39IQSPb@SniP z_{`WpFur^LZ2J&g4f%(Lpk&&YQWmfGzkfpG_SecV75FY;1pP;KdBOY%Kyo^~zwC(Z zZ4Q?3;t%QKVzI#ZiCGE~uieP%D)Lp{@;ERC}q5pxHMC;kHK|6?2sgyIXY<5+`++a3T$w8tC6DJ=134Qo&ps2MUTMO-pNsmwCT z%$CG06Vmb1=Byu(ASM#1G+2xY-^9*-;o~Y&&kF&_Tgcx#eQW0{(r&@b&VC2Q!D8{C zRQH&mQ`Z-k^l77puVC*TsdO6Yss}4w_L=-(?!Y?2^ZV=?$v-XHe^*Z>K1e;WsI%5{ z)D zw9yrDg#@vDoO%bJ$+Y7K;;53OMmN1YxQuvL)A?*qxNVoG6Z(7+cuG3SByZ)})2E~{ zrz)s=BXDD2BX`Ap-q2&+nlLjkX#W2|BCjTttSAzCsXv|FUD&)dfl2y`jBC1Az5Zsg zz4VQV&PeR;>PFrt!qPZQ=HmV`R?Q-TtAk~_MaP3GeSwOOdr0mqi_vuV^l|(Bsd$)C z9tl!PB#48!%IBEZM=eUyy9;t=Q-w!Kz2LO>rg7~t5dXtH`P<9-eUSU-IwbjpKdKz^ zaV{s-lH+xFYlCvdGC2yVXV&^C&$ZqAA_1(`np*Sei!ccp9@^Dm?aXz0sescPdV{T{ z%!{V$1Gt5&g}J(8FO@G=D%~Myoa-GBxZG}v%4KR9bZW}V@k`FfJa%kXk3*Ap{ml?O z12dK!x#DpH9~GT*KCG6L;8zl>VCftistd>{DAQAsa{b~F8+lw7fvt(`0 z`xfYk`fQbj{)u|ISH0)mm$TZm;g~ayXM#;W?5l1jsj*7>!EH;PDT)=!O6cCE0;E!D zq$txS8+o0%5(V(%0GgA=({7d3{_9VGfT2*k6Meqlz4(6|GIDmX{h6{DnSAN^_tkAG zlfmV&T1^gOz%w=s83`r2qNy?s7518Ae0S+Xs%qT{KfPE~>2*7N{IQd3CE6`SzJ zDUB`5Ip&MfXi^=UMRGE8F~?%RHID{*X>5_y;iN>h7mfr)b)46H|`u$=};M=@O`UEcinu5!BasPr4*WM+XCdcFG#pC)C8-E7YKMf17BDh1T zHu0ioI+jwp6Y&VAha#{dwV(^zJWp(vPrkvS(@C-Ooftl~_tY1P#7t2rCi4Z+S>Ga# zMhEWkqH{?SZgLftOveVXV4Q)iat>(zQbQA??NH0<>?9QB78hn_rC$HDZ#qXqWZ5>z zMRD)UjwV17AvQQ_<||e#IuM*LSDIAYe)DIgei=4>6fd$jk~yDu0$8%ke!3X{%pkn8 z3BDNZ)V>K6TvKAYTrm9gD_1UHBrBP<>4k-N11;For zCNz;uBoYZN2)Odp~mSO=|JZ-q~SYqN%bSVV0KeYp6|KO9G1T5p^ZAFh5f%0wbpV1IfzkJupq z2crxC{o4O zbneG)v&p*dgi6fMyOiA0GFbd)9vkh~PNqQCo15o(zqktGpoY7O_xF*S^F`n$7e$c} zShQ|iFINThSN(1%K-Z|=olPDwolSM5a4?|R!Io@XVgP8C0tUYca>On`r=AlCfnWrI zN9Rl42}Lo&S(cGHl@+e?F5stFVhO=c3U&REuB;5$5#Ht(d8xM3RPo~aMY=!IaA*{% z2dZ`qcfatWCn-a%X27ohXZjz+;|&^e=@dy$I2jCj!ZP1kAfWaz`XecJ$xC_z6-1Hz z(y{*46lg3%L;<_6gVZjK5YrQq?HhmO*aY=>`AU~G!oqN+C@&~~R8E+Xu=ZXQZWOm) zI_?*0mk?yQ3a{gAp{N%^M?}Iet#Qd$rT0uf{(6O?dSk2${%?0hjcOlIqJl|O=%UGf z%jxuGn+dl^1ELG{)oXazV(YD;B=$&}Rw!{q{TzBr1Uc#)eaUYiw!|3cb04nCWsiHy zi6%!U1dC7GBMBxM?%ah!EiQTeIF1hl&b+vb3*HVdP+QeyKHc3Yo2n%(+=E(btcHn>?4EsXmMqH0@iH zs)rhAy9}Vpv(34?H5V7_$Lskg-skf|5E1wlMu;2%MVMYMlR#8r{CWk{xo1CyZ=pcOj?4T}5+!(+jqAIH4USvnj^3ky?`I>k)csuWcZ8k=3f({NO zA`*F`EetZ~oygPge#57i#e&bGJ0x>-$&=A0sYZhgv+PgeQnRW=x6v~0< zEGhuBYL#6DBtGLSZ)~hQcOm)H-;!klU*dluxsEh~pKl{``8~Y{3=A%p z0n%1j7cV@Q0;zBQ)kB4 z?{n*HUDeQ#So7Q5ikBMN&;;$^{i0J6uLi!)$vdKG`Zefv`;J`J?}+k=2zPG?Ykh!R z6K6@ZEa1(MHRq6%Fv_nXpJ&T7l)en*;V0*=!C9JooAmxEui62sEl$DDJFMkS+HI$oo;QMrbs;<++_XpQPqHaLyp$ zaIb+s9(VT3H+&>9Sv23Q@pjJz`TN>((y9s4IH~q+aUi>2<#_DzKK(Pnpi1dXzr2FN zV>DyxU#@p16qFP6%dp;Hn`wbCT^L9UI@msQKlps?sZD@oA}9H4{s7r@dRb{=qqm>^ ziqIt_+w}H293pYg4KJ9T^#gg6y#`flAbf&E-{V&>#Js-O-|LLlrhp}&{l}Oh`+Fhr zBvP>BhtPvTPYUOH(|6w_^4hc=s6PfY$a}?P9=Bz{8V|HRk zD%#%)=Cy!!yO{3C1d}nFR%I`atjw6P{*1O~kirp;gFf#fZP=d}0j~6|8PoRxlJDql*;$|aK-kYaO-r6T`!4gK9hiCZ^n8b<*7q3O5}GC%b_#Rc~F?+f~C;q)EL z%GH(=dRS{P23DV;Tjbxq`@707pozg=F2eYmM$!cEzV$==4vtD%p_Z)~YjCEIg{OXo zLhn1J?Yf4`c`c4Y348nK2bOil=k}ef{ljvjyX=grt1c8}^9!uG)$0{kDjn${d3<~u z8T7U8x27!-p{_boNk=s&H{Yfq9dnj8lbndkD+h%ZE%o61P!YH;==%GiVFReOAAG6R zsuUX?4)sI&>n@N3z0mXDS3+=T?~cX$q)Oaketq%Jd_d|2A#`077?Nj;ULsMxSK-=M z71Gy8O&ejj(r&WgtfKfYp<|`mSx78maBo4g&_$pqrgED$j_GujGqD(vwC_6uJw&NF z{t!wf%E4XVyRPgH@jtyG&|JZk%N4CSSCo@3ahFaO2^_NL(<@*ld7ya?{%?;MOPnbe0aexW-MZf7b;?T~cT%xpFlg+vrESZ}+Y zR%p8pN<){gA3R4%%k=2_+7w{AnWEba~#sqwvJQmsGWG+^T0fjGw;xdAN_C$S|b{c>U%!`icxf~JoC`3ZdyGrY3 z5(ga9incs|#49E&M8P@!H&7%NpZ#1H;0`f}RO)pn8U|0y5}L554C;tdZ@x|$4FDIz z=v`e9EJO(3@v&q6JLY$Vz7T{HX4A{x`67hk&?1cJooCr+EpzQ{uMjU5=rV+gCwffV znY;e(J@mz1Tt1c+!iETyU;7KTp;&Jca4cgRZISpj&wx-oFgj~gt4O$`W?ih{0vN|u zn?>yoOpU%56{x@{uu0d4O*+r$v(3S6ss)+_KCpv?j4a$PgSN2webO+AzZLGi1Ko}V z!L7Xo5U=+~?OT^#SAy>p#Yck5CFLiY3(y6#T&#&jLWdpC5RapT6xp(>wHTM|MZj|a z{)pcIPln$1wogqdVmPc|+)b(n^JOyx(X*$|NqdrK2L)hH{^e3WK>4A-2x^@F7+Rzn z&dp>{X=UIAvE?O>_X{o2*gdH7Bev)ZrkYuZW)4rz-kw49&i(G?N= zr0d+)3Ia#OZ;P6U)?mQ;uQYBa#jYSPi4P14ityVo7~b0fddP9`fx6)!vZ%=TCGaI& zFnauB+$DI{lVhQFEyz7@UN1np7y7ZNAi34nck?$H9VIjlSHuoI;f z*2p;^wmhJOVsx9@&mDxrQH9wDE?z7W14E%gtuyl6)#F&pO_>-6pwvYO_)lJrthNN9 zg*00rKl>o$ebsXt>@~TO{0oye;Kl89s(X1Il&MD&>)s~rM-Up`#>oJiOWhE9s~iFD z-(FVi7|-4-5BiSGm7wqv)#6MZCmMSV4+21y;r95`f0Hk=AW$&rHWp(N4EEYMR?Kylt?RrSk zKP0Unvi}T|qsWEfjw1%)KRTw>+OJS)H#w&|?~Ki$b;)xn1iW$#{2M-_C`7dB2r}$A za; zKwb~@$@S{K-LCkDB4h(yoDpG*NSVCGhyT>Q+!1(=vdD~`86QvW$}bye>(zv%hO+vh zPY!wz1@bb++HOV}%CY=-h3;rT`8`8jbU`+wfZ4sm|m6H44;2qs7>7r8tm{k%v`PK8VZ z5AT^T8xYt{_rEZ?m;<)GatQ?TP&3sk*gk&{m<>H(>CB+}HJ1`WGiJ$eqVL8GoC!)+ znf>F{pM(v-=Mb^6Ur$S*R@+HE{}9Sj3~jn50GzNxinsJnyI(k>AcOHYM%i;0R4Ulz z_yj*VYBH2$uupG){4DI^MwO5s z&0X`~Dp4*M9CGR?34oZQH6C%p-J{zc7yl=EH~b3niA;3MOW6r*x$zb$+-bvfFhdeN zr1$jGyJrdC_06+CXVYBs+>MD18n1S8NaXn3oF!qB3j6?oNf**kEOL#`h;YXlvt&M( z=kF+1@uAQDIBSXD3sMXMlZ6*#bSvR?d~6KHwd3ePryb1xBu0g|JV6%j?L5u%zkknQ z4zk0-$}4Y_pYKF=!&`~!WMOF7_$As9k7Uc1lEv%is1h`5uj~P@O!(bK6)A|!K+QU( zCa=6z9)>pM8!s|j(0u40`cXGA`Qs>gGa0o`WmkjJKLpz`jnNareM^yjfvBqv%qR7k zadElQ_}B^!MYm!euZ1A)I*vFDAH#C84Y20Pp4zC2JN{(SvgWXjz@baQr4q)(mX+#U z7{c!iwhvhYDs7#+N%yLdPL57T7h`njMvV?AX|&>?ULlaBsqi-82F9wqq0?5B^Gt=%*R zAhfU-^SP1s^VHAGy5rMzWh;$3B(&>n8r2Gt8K@_GjkB<0PP@ISa6IR^i2j1*!1 zx>68X&TpY|+ZXciG5B>I)WLJxiJgi;bK$=s?|+L>Xei+8HPN|nav{T{g4wA{Rs!s_grg#^owWg!ChLdMfOjhO_1ub$_C zlJA<<3*oYXTs^%OSZFHNj^$N#zUkDyzk(T+ zoU&@O?IfH@rDy?TBwg2bKlHxLuM{@c5`dt}3K<%96l8rI@T@X`%C zQ4?5JomxLC^tx00gKKLUq>&p@1ptp+^}I+iznLQ!ES z;%n)T9S{aVKuK~;E3m$#CbZAu1w+VSJnz||3(c3h=A5gvL5Rg8*8W1P5^+M57k(4P zEtWV@>^%qHI%1wpOLvwxAfmwJjg@yeeD4rFetSNf8L=02IPpg9wvq4Kq0xT=4Hb5f zubREkRsVVgk_3bC6XtmXCKZ1{APq?MM~ny)v%;C|vijqs<37V7KKT@e&92@v%?ldW zk!l){J<43|941yufLL--S?bmtRLx;qEwHeM|5^*4W9kXG?jQ}!*)=EmS6~|8e~b;O zqDcDD*Y|T3t=T4wk)lCNC%=QI8}Di`g+tP>LW!vnaXt|C38SGIrD7p)`RAyF&!Ht# zZeP~EWmGJ55$lqJ!@+379K3p$WVCm2vz*LBwo;TiD>GYby@l_yJxb}|Rp5TqTA#bf zW*@e$=kjv%4+1ZS4PVcTE=A-J>_wQs;nsB_u(HtyY_#sGrdm|ks6RaLa%gQ3{(fAs zPCS1xIC6z_W8~YFry%ogIHcqJ?ABKd9EkngYPx6Jp9%*Q_hy)=jxavF80#G68D~|C z&T8mb|4Ar$clsujXO0T9-I+4tm#VLs^WnTSEaxA@RZIxY9gX6n8#*dI3<8kUNp)`M zL8*)q3L+V!TUq*9jC1sifp|gwUOHj2C_6*T?JU{Qi6^LRVL1kKOp@aLvYSGO_m`*g z5Ff5pc+tw>AxhP5WSIQ%59&wPw2LB9VTxn}&&ukk#19uwI%5Met;G}n#3%9`@L{wj zmVl8YTXqv*YmdrOeEJPn|2sUPq0A1(vg(h~O2>JJU!zC8%03|5AMR{;H+CH?^R`|N z@5@Jka~Ysni~HM~_#>_=!R3Z@1fp1X1mVaVp?wO%QEebL?!7K>*y4~tr*5pDpzq0*j{MwlFfoyk?dWy3eVBi$(~*6(;I>o} zG#(j{tExm}euOaA%bDfX-kNg0Gp-B$r6qo`v}xJT@g@ut=CKl-?2O&OqE#UqK8f&I z)lf$VYiEX&;yv$mOA?aVmU@!mbQn&T*}-hB_)jK{J7)R}XA>2zhWn=LO=9ZVqHKxA zKlp!aj6aL}j~9f-{$9J73g^aNC1M;fd2veVAhFc1KTz`(K67#R>J!iNBI`!D?hue9 zH7DMFG&$KWNPOI{9vMR`)hJm{&4B!GBnrqTov(@)nQhyNH{G}d1$C>co!)O9`tT&- z){6k>M8Z_q^{_9?tz|*=@SlwsaE_2^+z(^L2Gbx)e}AJ+R7M44Ji9tdZvsC1Y%x?w z&Gh%N6uGRpINoC0Ll?yqxj>;h)UJdTgNHhoBEjMx^?Q>gB`EW1YHBVE=7y`2*jEg!1&{@(2Y%~=LwlANmwN0Qqh};zWZ9@+}xrk$Y z_=Wy2yPv9Q&ZO(ahWI!HoeaBMfm#PsD_x;^FTVhZTjmbEV(M9ZVPqrH1F=c&0uQ7~;q*H_gwmU9}j0+F^r_j9>MDk?} z!k8trYp5br%s8U|F}2@b2~cAWCJ?(m#a0zlz&SLHq>Ei*NrRq|**O(C^g5*@{-t=3 zR{VVa&#L>k97ZevrveFh-{}zBmPrKh;4FU=;6Rvw7~L>qF!drThe}p6S+ml=w&lNW z#%F`XBu=5$KQq61s>@&;L~wZ0s0a#|%C@26ro)*H(3X2uH6hceKGOfZ@9o?#1={dY zsXEB-2L}Sj7Ynbh7#j8Xxs`1iU*!sjbeWri{+Wv3bS~tr0-OXIDk+em*EN8&T7s}p zIag8h#mT~0KF&y>SCjQQlhKL(E5-RM*XfajBV)9Qd@c2R54Z1R?fu%t6s(Uj)-0_3 z8)OAyazI3pEFgu~`^t$%xS9_W;V-!&U|;@l3i@L-YWj0y(%%?Lv&*iY4HSNGkhZZl zwWProhE$ffKl7p6{G%%XF2b1L!>HJ~Q4q0UWH|D$W!h{p&-$xQE`pV?(CY#Y%f>%U z$p75M;0v--4V-c&$#J6Nmfdi4R{o?V6Z-AXR{YmkbyL8Barqm|_-&Q(K9C|}qjF}_ zDx_rjcRtr0{LCK@isv0I&@yodz<)yu~U8scWny%W4CS0|IoRr+N47hpIe_1W$ ztz@?3EvD)dMmQ7@DjN>q`=M)yBCbUBKR{~1)5kw0WacdY$;tmR*h+epM%TQ2ytnVl z7T7N+w5ZUouo)61EmDR9Ga0;xQIUb{Ea4@n;rVKs?0>lohDOAj&$GORf+&}Gpu|6|G%TtVDKwel|!6RjrVBPRD~)YE6C*XdMtgYacu zIU=$i{O`lTgBs{oXa8USE){W@Sb%v`5Sarh2=_{i3xd_6yd22h-P;Lt*3znpH2V(` z6CnC4ep?^@u4J6S(lyUDGFon2gQw;_HiNpFmyvvoDIx6|)Pe2yyip_}+S`6rdCk`& z!<j#7PrdN!W&S(*)ZNx`jhzrPo97a0{Gz9ZwDC8 zo#O~*BY4LD%eYo$vaMRp-?r<=T&ZIR#;Ls>kX>U%Fj2vcq=0OrCJ<3dM~|uGTf9aI zE%}XXW^>^{xpAlD|J&I5sNqI1V}wq$#`Xzj<-z!JP7Tv4|9WQ#4c+K+x53*Yv#suV z1a8}9G~rO$0-|%=Y9AZ*a5@#MAbdh?+8aK{t-r>O4$=MFN3_?^|5K-mCTBEpRF-gi z_~;~2H)MxD8&)^=U#}Ap0w*NG#aHR}^WAe}mS?3?pMKrHmaRo*p&IOXPlK%sGBefH zTTM#(G&J=U?{d4s3^3eq3Am~LlejpBA!EV+`Z|Cs1x{d!6L8t>zrT$6wEwl+a}ROc z1B!{sgxuhp~OG%YSPDW&W)RL1RCs*~yfMqrZVSs=^?Om7YLO$I7zN-7r-WARAXPL51mPn_$wq{imp#iV98~ zsqhvW`X|Bp0|WvW*c7qy*#G5*{*NEwfovb9@Xv_~~iPz`LWB1pEWjPIgysItgZ9QA->tQ$XfH38{&{(REj@{>m>-zmJeu>7n zw24y9rJ0%ieVe|)!Px!%uN7GU6fY--cCYXI=y%lj^{73bE-GENc6ZW6_{ru-yc{0A zd~n3GeDS~#rG+=ite+gn{g@d`7{{~6p~rd7Dyt3o9nX_KCBVE`;$lP} zRIDby(7NnSxOAGff*ksj4SiX4eQy<&KK!Km)gz;~*bpNcNB*4cS8tu;MnBG$WGA>S zl~RpPIgV)3T6V$nT<-(C<;{>%lJ!l7zin--Ke43hAC5 zHXrqEW%*G?5phpkPGyGUGTVX&^>v<8b>DFG}qzj zv)FQdvJy&eN$0YMoXHQvxZJO1`w{VsX+1&hBvbjsELt48J6URWvCqGIU7F^b^ilki z>WfOROS0?BcJlZtvjO?j&7r)0zblxr$HVmtvze-mu~skMk(RkS+hiFs4+V{HIxLo! z!b0Qhv4}RHE6j(Rz49smD4NWQE4e|~o&Kpu4;VEBga^fb?I28!-|(}Mkn>e8F*}=K zTZ=_v=|pQNhY<0#FnMFw!W=^v(_n&rtQqqFlWr_BJJ!GajF>m-{zA2;@BPLLe#doL z6{PX@G&)1Y+gA*O;I&MzqbGn!a{OSfUNhUmZ?j(pkwvAk#i;H2g|GUq10s3%*Ok+N*>&(%5=O}qmE6JjOndBc&);nOq$9u-5z zAX!i4U4#&1_bfAiH7!!2yex4{eE9{T%0ean;%M##-(xg#mY-{`%igcuHO*Qbb-|S1 z6cHJs{lZMMMa=$8vyA<{PI;Rsli$OPoCj=Fyzy%F%|`N2dIo!YVn`H`Trm2xyIjKm z0X0=XXe0v8vJ@YKRhk%V4rX1d)E1T^K5uK226a7o&OGM8ZRNhS*UagA+{1N!vqyMn z)Ors^xU^n5-=A?Z8_Vy@w48?k+o;M!#sL94)K~ng0fBDis>RaMY9I<$|K?D~h=aEF zOqFq0NNiyJ?In5al%bzw_{M4GfJCjskJ##s{CFk&u<;nuH_1sXy4&ka8eiY#Zl{h( zpfwSfc!jK1ko(=KUX*MJg;7%A3)?UIgE@iCjezE3ZzGZ~xAiaBP_@~>^A5LMo0(M~ z@0V-{z({Ml-Vd%yteO1X>QM;90H@e6J^x4vvukSYi8b#v~KJI zF0Kc&jJFe#9s@}%1>PwTW>j3q1TlCUlk-9jMD&X1-*n_!n-()c%c@kk>4Fx~N+DHg zJhso}lETLC3KuW@xov+be0g_F_QCM_6g{Kxn}7wn_37p?j{9+iZ4A*@@2i}4L2+zb z=yyhC^ely;z3kT0GB4*|u2t7BYfmbrY0O*L@A89Qz#i;}8@KwhJBp;^xDPEgUV=#~ zW_K0S(d9WFcIef7_JWr49P++b+5ZIX%$G#gSbY=f3wC{%xB#hxJ9$@_`qcr0ZWyeR>^5Y+V_h zK^J#^e|uS4b%!Lyfbi0%-IVZ4G^t>nH5+@*Si{;f4lE!Oxw6K1pTY56DiJb-)3#+-Z3?jrGhJ!h`SX(QK^_~}*PV{~C_o0G^%rNd|iY4N^ z!)3v0r!7pM==`KU;%+9*vKHU9Uq>u4t0JNm?Wx+3rJ*nCr8f^`8-as>ZUF)U>i%9K zmvHz4i}cMBWE1p^Oqvek{o`rY;~4+LRj*(mvSl6vopIc7+Ij2A7ajww&LzLc*0)Lv zl09?3%JfT1VR$NCrbrlIm#Gpp-SW`e&@C#?*d3)OF-+{Jq_M~wo*p%lM5wsvR-4*i zB#@4HUp{%u8iWPtE5hDyJS>t7oxi;9Kq2p2?^ARBl$r|J!9T{huW#QF&N>6YrD2g< z<_u(s4_(}Gt|{=#MOnT|oL)CO^i+KsG?tsg8RdIBlch!8=abGG5c^Va>v_UuGr8}T zYNR-8!?JfG>tJA*ym7zv?8Mv~s_B-~VXxhktH`%EXFCNgRIX>Xn#HK|d4Lq|b)+ff z7w~jN4wu?U>-pZ)sSl=h+a0~f?Sb35$I}&++}XS9C-3JQ(E1U~OdHpt=OVtQdQC*q zq$g%y@or4wY;~j2f5wh0GVCJryJI}?=wxcM951{{aoZj>h#VvHV!m_R?0mu&D;|jK z{_F)0#%<8vGepuWAE|r5r9QnDx;z^n`Lfk+D{tiU$#W*hYg!i=nYq5|(`SPaG~oiz<`HJkv}s-xIWvll#Kj#PZj8 z!c5EB%xexSTzao!07Wh0`uHQ?G0;7~l}xHg|#Bt?sN^IIqaYliNJghdO1%7XeKpSz+B zP*62xDKHUzZ!Cj9lHP`}Ue7y){?N7vZqsD_fCpOL)E5veRI-nx;Ha=g-j{ru|GefQ zcQ~~E#+V?eS2%rX+2QLfQXAm0#(KoD*UVBxK(4Fdi{rlT8xRyEZyT@8)Ggs)9E%K> z5F_lx0SX;|qaq6PmiuR~w|C24m`H>%44ww_U&amxiBp&i1{n+LhC`xO^n77e_V0lmxZWKy%&(*Pxn7fi=H5A%W$@VpI_T^-E{ z3^>9=ylBwq5bS5EpQiz^@pI-v@RHhT@MCUf`U+tIg)3}XJeA!zj@f_&_OQc>Ao1yE zUir^)s>U@;GNJry8Kz~oWV|EW(-t$sqg8$Sfxkk?Rl3*)>iHhULe$;%t(p6(_qLhD zCH11B%$k@bao7P7a!X{1$@GYJpKqL5+5Fl?+I4IF+I&ZGi0?LLcVEF0(9>q!M9Gr< zTA$f3{HoK|_Ha;OF-*gzu~|AO$GsSsF1Pl8Ks#(ri^!6&+R+4VSoJE*5X8n-dp#SF1 z@~A!_1H)J&I)Cd&*sWVQ&KuKqI-$3{PLG}V&s5JoO7EPEUA3gIs?8?HS9@-@CK9Sr z{w`%y*+sRPZBTuZC8_Si8??mhpumMa1m z`JC1HOAt!XxRM}sp$Sm6!w$YC=c_QB3Wm2v#T>k%dZ6vse_0q0d*wcaY9V@8 zp+zo((O-A8d{D)!M`&ONohK-S=&wW$!PT2MJRH`uDa3I zDq%GStzLV)wK(CtlNU!zMdLY6pb%bvc4@T5&BMvkdQ8ct+oM()FVlvj2hlzDyRHO= z@wPh&)~5BhJG1BS>E{hNNaM`I`Ks-b8szMM9SovX`L6#kwtyouA;wsjP&sxw(E}2ykiZs#iBMFNOGv9 zyE7-;*r3vp+4_?(QR-f2rImc5>{`dbHYjb#?*wGd_z`fMnY=t+7bN_rgVT#l=~WyIx7&F!J43gJ z#B2NnW@AY~8#_f<&*yJ>z>9imT$UvD`Zw2Gi6;wg)d6}7?Z@ABm3-*4=G8qf6vpRX z&sPG;`4Ys~-G_kNDnDKnM+2QYAIhaqAXM{|Y{1CxX51$l<+;02iom&A<9^kfeBF9A zXYY_SF`vPd+nS=N$IC|`BXrI`AGs$Evn^+5deix>vl)?@p-K;pN zG3u<;*Y>Ed5PsLfbqJgU^*Nc$lByExL0jhHU@(rh6rumHZfL-&90wL+jEL5ZwO%_? zI+!O9h0W~ymzu7pwy!pL%JwOy`dzk0gvn`{Z9HAwbt*-bqNrFSWyysue|UVjPqdzC z*@Iuy@d2Qe-3}3q;t8;d;V|FYSKJN+@cU^n)HE$jF`eg2=;{yTSJV)=82Ir?fW_yu z7{vBwx9!k(TUwJcX8%TkV8sAO>Tgbjfq`!-A6VXY^Y+(96RaoQjvXU*4%~gT*!m`P zBtua5*H`UtQ**V}Z0-Y71f2)Gg&f<~-ITI{_3jkTSFaD^Ihrx?%##RcLS;c018nZi zSK4?$Amn{fvW(z0;W&gA13DD%7ufxl5HH|*;)3_}2_yXDbB1Gd3=_04hL5%y<3`Zk ziEpL>xjJP!wQtRR(x)KdgM4*{-VfR;{6$~t$82+u7m))>=Q@c5$_>eH)2MU#G(Eh; z^f&qwogxy$Q{9WHPmTQz(9;1#XVZt<|T>4pE*Wq+PRz65$$ zjZttaZf8x0fr%&Pt(OHLdJIH)fqG!qdf|zS5sdOD8;?BC4 z@;UI^*hpS)6BW*0f@2;&UE3=!5?rEISTQYheRZ%2@ZSKTV_$BAJf$kI2)9@)HBCT} z_bggWko)OGt2_+}xGQ@z;+ZIlx+$1fO~>HD_UHLWFw=zzd4`WOTo=?+ENm~Tq!AG6c_|I;bv4K z-gP;b^RY2s+Lg|f>bBI4+FslO(| z++Yp2s}{{3Ou|?S-*&bpiK(nX=z|V1ssM#hryGI+69B9+rZfd_D#HaN+0KD{u<92* zt8tgM!X35-Q%i*tp?Fx`{74xFUF{dsz00(Lz^{98R51uf4Tpu@WL4S`hbcePzji(? z5?x^0oF%$_UMck#Yz+X`06&q$mB%pPp1VC;*rMm5WVSO~_qpYK2L}>MrFt+&KEBZy zb0S1Iec*jM-gRPb#N36M#3H|M->U|})RPjEI^NB?6MlzihFQ&F_KV{{5!4rQUXIUo z|2pNXzzyeyOJ^7>w*MM77V))TGcC2U->P=7?mR3jJaoO)5`Kgzlh6@Cu%4Cvkyjvl zkW_G^;;!Pxy?8y-A!T(f6${hzdn1CfrDmoCIxoImyci<|tY<4r5FzCr8AD}}-~@SO z=#ckq;Qf@ISB9@H5mv}}+m(cpNMpKd>4!0QmSbpm$EAEC)2(ZY$A+ZLQ{)ZFR{iBf zEMo#a36w|{#N#)Su!1+q+uKCD-;#jc(FzBlP&uq17IfECgtT~~Q__OskA_UOH(Acc zZ9RcuwTRRubER~^anW8bU_cS@YQqkS5JdP+K()~{^xz%%BDI9aF@EP}s+6bUsfXf& zd}50q?@`049i#PA`}T|70j11DtB3?eV%}|t$!9jRfsOWO82~K8tkWR&_G>@o$L)!N zNKD8RZ;Vw*qktXS{zR4{0pq%U8FF8v4Q!9$E>dA+4`ZL(b4_r#i`y7@sZ9}0sO{Gr z%o$#{-3~as>hwAH0?K}9PvrZ^v^vnQtqHNvl0{j6ziq7t=Uw~-TrW5>a1}9RfN1i% z)%f1CsMdDYvzgBYcfIVl9AK*RC=oc9ZFR_5r|R794qV-Wt~ z+41!EbDNdU>!xk-T_S68Nkoa*S4N{0KDPQz>oNx`JqN?{3#?8)((biZpIZAiSgC*3 zQ=SE?m+AG?ml#Egw*iQ&l=p&+@v<8r)Gm2j@C7`RL&e^TgpAudc=&@YPjCKOF@7zP zcdP=49iT|#NR!}El}z6ghlQ^$)>EZq{psG=hnZE#uX?(EjNE za1U(_NG9cXe3z)r`elni4RaPT$070ML%T7ZZ3AsTzw>s5uO!*6wn`uc)VGw}V2Dl$ zV_tOa()nav+4(0>=ch74P4A^~dN$s%YC#jA83ODOFM%_C@Ja2%kG7tR7QI9;1q9*_p9|3v>L4tn?Tgp zWEA|6F9&ZfV0J-}x!k8c)3D}zgkMgq%i)9FO+!p~tfZY1J9iv0m5j~ zGG3UTQ8!aq3YTR>88I0Z>?`>j;0*0hPaI0v=P*mUvc}CX-s6fBnj`ye3ofBXO=*US z#l&P>o|lJ-BYCn7t}9*`9WlqNIIKR#s}5O_B{WdxS!?8z)kPwq+#&1Pk2~2$utnoY z25e-Uen5!?=!9Hmt4;DHBBF?SORND)KqeBo3n(s?_9HGVi|7RxG;>_(i?}lDXVEh~kUkwI?JGns!6%qZBlEwNE- zxE(&JSz;Py!vJBq!=YKsxrVBK2x9U-{!M>jhlsuY?qH#G`O9|phsRdNQF<8b!h_v`u=FW@+aS%;e)H;u9NBEloHc!%3bWzD(qG3{nM58(5$>aUZRS+ru zA-lUE6y;Y{qTB6(W2EDKwQ`ZbWMbvjnwOEyST~xvIvX1~|H}`$;9@VD`W3L!% z&kWx6Lt+mdlJMemo8D2i#_U8}WEfM~l)FSL7BGLvj_&u8BWy%L*eZ6E%*T^2YxPA$ zh!1)m0v*nD(w5t2oLkjXmt?jj#^GP=N!jo2XD!`mWH{R^3_U7QLt!h&y$|0an|3yJ z5tnT1#j!(_wyP)38{M*?=g5|1zS~Eh{!ar^zx4;A3SBRI8c4gWJ}eoMJHMa~O<+2+ z>>rOwRYE|cfOhh~4|;#7aFrq6Xq`x56m`?u=>pgO48RPMqca3{3It=8couWJoVg4( zF(2=r@_N!T*=wo#f+-MfN*7Z?RGz5u7CeiO1uccgh@hkHj%FV2@?uD!15yF8JG8+J z;s(v{P-97#auV#f%wq_xzMaAC_bm2`?Kitj=M&GCd@fsUrm+dcz6;YK?~VwaI=hsU zZX*N?oB_}*>xWnjgp06V|9u3k`J=`>f>FH+qfZa|;vkIdf`W5Y+$R(yhu4OljIeq8 z$9vQD>XEis+BZqN9TEzb9FFT7vV{i#QKANxd(2)^;awyYaQ;w^j4<~F9%^7^Us@&i zCE0c6!xliXAt7vCW%zp>lt-&PK4;a-55{QdZ30kPg^Fq^VvZkAc(1;*GlW~x0ywX2 z0F}Y2Zw};y`NLpXcx4p4&5`%><-dCY+=);R0n2TTQ64hgIlw+iaKWD-p5o7Y9n}#J ze*%r~Xxz#)4LBe3RT6_6Tj3SN?eZ`mu2tFTmENYbp9(w#ze!IPT44 zy>KSEDkeVEWF6onVYsJkBzr8Coc-N(OtM)6VUqVR^p+p>qPe6kVOD%THyHW_*TZQy z9O?*zL!@6%o-m$di872RpqWJLULG!FRDAr1aSiK=IrlK(@>`;doFsX>7^s4t6TCAWgp_Sk`6ZCJ!)#qT)wKanUQFPzLH(=s>7RGo?KpV zd))F`=(=1ce{|QfXxu3z&bJ(=b9J5hA{4HadLkE#LWBX!@_QJ~b~YfsdholLNsdlI zJ=mWp$umMvK3r(_FgwPO1cnz=*`mSDldzi)`|a`#VaW>7p$jRy3A@_xQ{KD9uitI> z!3I!Hnz9&xPWh32RIe`wG*6jyAZ5n(Te;i;Ai?dU>WuH!q@Z-=1dF6ZCQbf!(ksl4 z)A7-qEhGuKE9NSUs%fS*eb7k<~P$!w2uBKK#Uy$Cqb zE-_>awJ=&k6HPeZMc8*Vsl_Zum6>pUQDNG9YG;5&u6`OvKK#ohkkuB1RVf=-Lc{(P z>8p9|JIlrFj|NM%hKlgHLaGGSr_+03p++F`1hXLg-2&$8wI5ajw~}CzSnzWcvXC3+ zmu$RG%dH|fUvk2pd_rrF59~UvJ&EPeMA5AKPx^+pKhfvwE8! z_+E$lU1se5NE0<|x|EHFM?l$O$M`P({*^t6`c3UU%~WuRv8U_vc%7Jweze(8KH)XSQiB7*aT{EILlQZ?onF&ql}|7CWCYp(J75(-IsU z%|^mTVy!n0Ggkp9R!L8Xiiq%?my4ONrZ$+*HHwmR7G&2BQMniG%}Y*s=g(fJxUSOD z?sjqa*4ge}Zbel-Chmb-x4m{DaI1?v(=Y#ud~tJ5 zzT7Q24FU_Y3zE!Tv3?(mxbZS=`Ps1bGG;?0nCVGxFY3Z;J!gkl0yaH^?|}XX|1^U%}nk0=D-LbRkg4gLxoj#6bV` zHT9A>(#sZ=hFxDjnbl*3E~}M=@`>6x)_FRlz!^yco}y+?Is2&AZHp`&*>&1B!S4#&G0Qc&2qm13}Ns!+lw)gLZ%B zrH|LIYHrS0WR*yxg()jO{aA2SAbmBQ(VS2pzjZ>&dlzl8i>G%4Zrq?dDCtQ+DEH*U z8s%Z@0%!9p&mUudYz2DEWgC%LqjGAL8{{FGihufrJRDrVr#6(l(Y^L?LFZ_k?q)n$oq9}QV#lj$E{xHnV= zAlbHGDL1=}uK5-3xc|KB!n;j?qpC2h*Vj`=&ct=ENN;1;C-d0zS6cc_IILczxjOW; zMHA0X6>ToaigB@f8_{xe_Ze%qu~vw-YfE6G%6%8AAF;f|9Oyz6l9_hU%v6(}H^HIL z(`mF9qYcd>H5he4)%v(#Z~Uw_Lb<}&D)AXH=vKYpZA1%SVbZP^hElbsQ{nv0l>xo| z#EkkX&t$k7nkJy?iVJ`a>qDZ^xay4aLU3vuQ$4-lB+xOmayW)8J(S_r!s=S~>k5I~9~Cb8 zLyABcVb6f>tBEK6%jT9BBds+{L6W_vUPrBY@zB7~#{d3%bhqk zEn>aV2aTQYlJf-ojLNcJlGZs}X4=Ddrk5u?acilyv*~E@sN%S^eT`?%_EUVvZ2Kcu z@09cG@^o5-RU?R0A;ze)rJpgH-8o4SI6j;W!PFfOl93}M%VjqvRX#HGv{bjeEbR-; zNxe5a_mi)+)^ImdhP-<)0T&53**z)Mhw4@%DaWUsZu#az0ZU$KH_J2m#Zve=vCBm0 zKc~*h+uuq=h;1G7_vgK#LFRzRxD62BR2EGonz){k3TI@`1}i(^$u;(5AXB014Iq~? zssLo%s)s(8M8NojM+~HOchZ`nzRlXKe?9!>a`hXpL9*b)5yoQdtBZ|fLq00>u&O3E zwan;ucbh}$6S_%;Xl4B1_Un*&$361(pyw#pKW9&MQ>@@oi~iJ(o8QM94t73w7{vyBN#<~5hcm{$Ql^l4`_IF3U91=!_tQFy z5we$tJ{H;5Bbluyze#|m<691!X%*Uj>u)GuG`?z;2+1gnffr1%(H7BLkeexUp}M=j zNmJ+?JTkq-(2YTRRP+Up>ZKI|$^}-)dSx(#l{(iS1>OA!N<0rsf;hrbdBJV2sWCo` zUkmc{t$xB?t)-Z#4856Oq0fiHv0~Utc`byj-2oc_5|Sy=Q-T(5XkfT4pXPpje-|1H zyR40Fo#!5^fgsQ!0L1>mavxw<6GR`cAF|<{K9d`a@Y2O3hk+ajoO@9aa@kQ0T9Ju| z+!W~LWBsn$mI*v&Ot?wuz5)>ISUngzK!B)GL8nNxXO{m6#oQtVdiobC@t~F&H6ZuQ z+KUjg;kah_o-ga*M`a^)lAvY+KkqC>$XuH$5y4!03a8d0wRX@^ZcD)Tc;sj3hCIU7 z<{or`>kqjD!cMCkoHAkqR}~Kruj>u}dKA!c3DdFxLuWYIkU?a-p>C7)7KP{u3q0c0p!4J-9X5>#i$ z%U-;8^N6IOmMk(+M1+=~+IB`?oGd-I<*ZTK5sTbO_zZpLCQ)PU81Ej=@bC$86$cO+ zdntTAEG5xprsyM2{TAv{sG58Mm9;t`H3c!)UiS!^+h~3eF@aP z2@>Ljh3R}HW)v-UQSa#Q2`#M*R;xU^sEF&$`$vcoA{=xZ;2Z)bYHZ=#SkPi75)HmNBww7%K)n)pF#{mU0Dq|d+mjAdze^fTIx7f zZY^fN`V1?EVk3cIa&tYAHLap_^4e+X2_#e_Ou|>oN3W@tjr<*=6wUX3hME^oqwhcG z7kdBhlM@#Sr)KM?m(`{f}ylKg;xoTexot66KMe!v0;CS}vE zl4{I*>Rh*+Lyq9j?v6_EislHu(J7g=|C`sd>ybBX{hNyE@3uBP&$Fk78K=IU?N93N zcl17&NAg>N;k>%2)pQBPSk&PO9HhfSE?!wBMa~JXj8{G^48~^y1pI+0m|-4)3*Eg| zjc>xW*oRP%4_6k+#)+2kN>)~gz}XP;ZajNGzaj3tjY>b!Rc8rl9SyD>079cR8(3d3 zSsi+*B|M39-^+3Nr54}griA`%PN%*7;k!4uZ$oz3w{zk8V>C18Pz0;WeR|paMIL99 z_cau3t-bN~*Q3d`l9hZ-xoC-_`$4cL3YhB*953kv8;Pdz%WeR)t z2W$>0$6lXq3iPnN&E$glbc!|cw;yM4fo*~AOlcchHYtTLXpTBqFppWs%8?)7k}0=c3$RZwl=<9#sP5&u4<&CY7vT0wQaHco z&iMWTuaLUaA-a1|_ch-l-ULd>+WQt^3O%ncs)qci-Elqa!o5#<7x(%7O^+m(eD5c9 z%OyWkC}h1WawYZH+@a?lzdqf1`C@Qu$6dknycXBI1KN3EGhU|eH952NZo^9=Yj26c zcxiGN+5{c%MWh*kXe6;2ZlazoCw}3xaV$%w3k$rt zaZh;}?OLgX%di7*K|*eJ%La`WBS0{0XN4{b#F0jxVhc=jU{G_a5eyGO^7SYN8@xNh zEd=F3nZ5`52GBFse83D;z5qH#&*}3Z5&>f766y~hW^2NUJ~T<7y&@ZzSuP zq2M}s1k|q+;mywBHtaWr>1CP-U%uc?RzjT;s0dnpsO(}GC5&OLM&!IePJD&=eRerxSS#=KPc%`5HWfe!j4%!nLtHG?T;#?N zd}wHk2J}H~o6CZP!-4X}p4cd#qH#`^Z+`Wr-Jnbf2}w$lteXKxfVi|}@h-Zza-v3C zlay6990h6Z074$dnpn`_ach8h&Chx;c{yG1)7$mw}mbPdqiJiLXg+ zYIpZzYbW37JrJ&3$=F5+QgSMg^f?4tH>C6!h{N(4m#0qKzLknZkAI)-MD5D*kpTBJKi>71dvYk;AV z9D0BOhWKCib3D)SzF+1m%r*Pkd#`hy=WhknHJ<(+yef-;KdOZp^Z{V+Q#mUZ90q@! zzr=l$f-!8vC;V&nvCmvK8ZAzx+OfZU>GLv##d_I%EzaI62)}Z7B_`mD6u13^ zLsq2P>hM!<`)c7oOLdrsaH&X0m49UhVgJl)m74<;F6~Tw>!+=*62Kgn8QGhLa3Mf0 zzahvxqnPn@C;hZ=_4iPy4u75T5w+-rK*iwOrDnx{ZqK$w@S0!h&liv8)-~x^q;)(n zGyHwO7WBnQN&Ns}5I~={F&!1Q#w9`OV{b|S>W9cj53Fz|L&BLSg5_nZo{IZJ@!j5axX|s_Toj>o{%#ZY z?xTdysdzEzH+U49_#$CT5&Vy$>Baopo9=LcDG4^1>1jbaJw!%8HTUlDGXS0)`fgcNsqWUpp%zR{NOZFYnSkqO)tbU*(CSa{-h zTCCuG08=b1oZPVduy5FXWPoLRiM_!&)emyhMXDkbX!r@vhz=pIYN9^w%vfkiHYf_G z#WP((llbQ|M4UG=%kHnf0Tx~n{0LRF7*qnjy%u>z(Bc|_N1QY!d9R&%n-R!-@>Kh` z_b8d7oYm<%*b;E*E==Jqd|nCiA4lmQ&RuQ;+|j16ndbB4E~1a$*gHxe}@ zX0Rjur|2=!ZiHJTT zG?7B^x&WmL|7K93ng(g>c+C5l9BbD(s-V#Q)0a4WcviYrpDB({A4FUZKGRAQ87Cq7 zv~lB`@x)TfL#x3pGsn@?{Th8#3;@95j$$#*bV)C}@#44X+*=FBSYAgQk31WR?5Cxu z3p}SaYVkGGq^^2%@>C}c>2ID7Tf?3b9a)JA7w<+`sZqF2D=OwMqY;{5Bxq88)eyQ@z(8 z*->kJe-vX)h=vxF35$eHrb;0e#?oq8c8XIHH>|o=LD82$I%EDc|u`8*IlYb z=--IU^C)6EVWfQzxSu})%}g@Xv*!w)1I%`1qr;#29=dE;K~2yvcpWa#)lvZ!F`vI% zcSW67`wTo4XLRYN+sUCTiDdG}Fmz%!tVz>Ez?AS`enISqyb#z^*4Wx9`7G5z-Oco8 z=A*~(iwC_@ltc%S{vC!vZ-T?#cM8S`4jOug{Myx0{l%tNHaKIDkm8UXplly1sUn^u z^!W5LjD*6FMn&o2%XYw}cn+2^#CnN4I+80J{TMGyEyF$pQ^}#T{!B7r8%MHBie=P{ zXs)G)9vnb+mqPosmtPRbG}G-L`&S0lQDvrXnRbHjF%zK!{aP>`?B1~YEC7TeXG6{X zS-^38eD_h0u=N~T6-lY)`J`B2Sz;dfrF(9-kQY{x28 zoHx0lcT2)}=1%S{x9WNtd=OTF;;YSOJ`SA_hW{Lt5YonFso=*z9xu_AuW1@Ft^&G?87KX zI}gN$v*K9<1`K|c8kqf-ot+KLiJ&tN4QX`F1ENtMRy0l!s}fj`D879du1sUKSbM_Zh@?H^scz_qrc<73?GvgjwdpgWlL zJ_aWGuLDsA2@7ys)dNyrCHC(< zq~fcHz1hn0tAV1AaBeCMv6Cd?qDDd5_qn6Bf2zIgUORWFaJBvXV@UI1FqJEn8&+ap zf^GhgEzpI)O^fj>)-#0!FoKxwyE@*Nw>lD@kaWOSWt$c36@w6HtD?Ng*?)egjK&k$X_D|eEk{b-)ykAW`*Jkr{ zmeAvr)ms3W+1UPT)k_WOtt-D$xS(I{9bWbH9+<@!rp+wHD9RjJ7Bqybl3!&#az#wew^fJ{JPZsGUzHzj7FSc3jRpQmgE5F8_k-TV@70K~Q+D0qK=)#a(@h+7W|A;B)T^QJ#YGnT=?Ct>)|Hiz`D{ zY>j0GRF+m0`ATWoOGhB}g#4B0EgPi$hqaof`^W-(6w2pZWFl`*D{5%43OidLcjOy1 z{F%KjO6jwS<`eW*v_SB1S=`Rl(USx&_Gn#HWCkVYt8?`F5%?gNyL{e@3B}kHw`t#< z^u6$G`FV&jXIVItSb%-d@skm!j?{gg@U~&(pLAKr%Lv<*)~knZ4EGMvm;%ca zq2_-IgNg)DDm@wYp{Pi7F}tmYJEYe?xl&iYqlfwIRy!sREX3*eZ+BZo(|?TC660)6 z34gOS7Wy*Kf;`YIY1E>lX;AKXDX|noHwj=gTwG&0sMoO7`($bVrtNm z$GSViEd@qkK|e4km)AZ`Dsjm@Mu7@EJg1%sDA^~3vRdw(O=;9+uD}-deGpHpxxuIr z7lZQ$8CtR+y(DRwH+MYcOc#6QklJ|^l#gEY#o8Xuy_0BlAbJUH+DnKVQNr&s*G>#t zLzZRw9*}_gV!_K*je60L3xf_~hpd1oVljrjsOuBvcP90Jj&9FdUi0$tnF+a_$w8FQ6V8mSU-|*Qk{| zxi>nbJIh?V1K(5v3d#0wt(!n5MJUZke_s3ZHic7@vn$jMimQ37P)A*OzAbw?YT1uq zPWu_XT~XsnFA|0D8N3E!h;f{7PrpQFjN%bhPvpXL-hcBu07|ZwW>x_V-M;yo=E1xa2~1Jwvob1~TjsP>5srylk;heF zor;X|(4?mvLA{YU?p!g*;F^2Mcq0M@Yb`_Xo&K$kHg^RtROIT6D^H_pX*^W$y(!D2g23-R;xzfj`=YHq+}55RnkQI~gxW z${>g{N-b0TysL#H-*F8DFu5{*4y|AS{~aKDUs^^?^d*FyjV`8P0$OR@!o>Wj;I`w( zgCt!{h}ZVuRxGqhax07a?=lfXOUHD{FQ*qxoQQoL7B9x)m$Jd_aSiA&T0cn`vsKA= zU4$(#|M=z(Cmg>+8A;!oy+6Te%DwV#a;CvbepfM((oRWJtFM9!I;DcfxA1MsrU%UO zA(=U4x;$uGMb*#2%jtHaY!N*c^9;YYi;LB=jd!GCpS?3)Nc0A&Aq8EBi?f5*dClhd zT)^aq7nuz!Ui9vvH&gHAeZ5r2wg>9-+j0uIGq!n11{Ny%IErE7$*lJ^{M(`d>RCr;i`yJ{d=ukXJWdkx9dgt^W8QiQhUl zPo@HI&-RvBfPE2?OL2w<7IxYf^4^wh1aR9rmI*Da1njCaZ7j9!UXzda;Vvw#u~BfX zi`1K>u7vlK;)l6hV%u3lcD%=#w~F*LZza_>%d4W8PBsQ$?I;-2iB8!T@Aa)vZ?ZMteS(}dQ!P7-($iuk-=#Fg0nH-$x)lVrdszqps{haMf6 z&Cg-8n11Vx%632ZBX`@-xMR^(@>U3rjld3p<>3xCCfrW zOE=qTXk+jNizH|&WTo*4k8Z9gZ>&};|7ir)Ez&qWHECX+KBzbotN`|nZhpl0U2YW# zu?c1G^Q*@-$PfX2oRo^SY?~CT%n@bS$*GVVh-HYCvss9yq(|b2E(OzvG$1wJKvrW` zHEDmS4=MG&UeW)sIG7iDk1{K+ujXZKYmIt!Y$UPHr|NUGEMV1;6ef+t!#3ep?JfB5 zW91rJDeMj93%y0syM;({<^Fxmmvp*p@v_S5aququ$j2MldVZmo={BPDm3+_Z_SjLe zSs6A72S3OjsoECM_6uc545G>e8PtL;-!jf4mvL1NYt@@+|pGo!66Nt{k_^%(<@K2L2}CrxEy= zMBi)qlSbt1P3nolymsyWcUh8A4ezTmWP}^I+Q=}rSq!~MIKhri){{+;mGx&WfoT`U zZUMdaAvE*(0wFsAj`r%*VDiA&M)W%rEP19V{KrCrBu7O6dKr#^GCth3>1}oAicQ;D zH~`>9LKSj53a4qMIdAR8G!-5ZtL3)8wlHpRvlH<^}gUd&q?3spKA z&ws@MXKk^VNM_d8fvpBh5`>s5?xxlcwxCq-x!W?@8V8a zhaoI) zsheeD(+^znOSuRNVz^B92Sl?AnM6fJ?f*_rN#4!n^Pu=m73Mrhd6M_(LG>UaA6?|@ zOF%-h<5!I!{i#{JrAW>)S6;vM3xy$7`o$>;J@7fq_S!4;!;_=-k; zjfMSGV=*9=DEQp=3KmduGM+nz06i5g550Ob$s^`8smQ~k6^;R2?}n(={jOlAZQkM& z^^qy7o--%T%X4`ElqMgH_2`ShX^{oVu>ltB0sB-ryuean6R>6KaP9#5%_QP{(GF7*#^ zts*JS)(Z0Av0Q;;xEfJcSMr`FRi^f}>BMy25&ijm2XZ%>!oYP~z5JVPm)6w<+I{l% zwZ!oXU-is(j{0C9Rp^^+b?c%HKCkXaLN`CnVT0Th7Cwb%0o$b9q6d%NR|1|{6z8Ca z1~I)DNB?YF*_)F1v8;S2thWVH(!gac8%%puSur-`T=n?A@}`&WhkE6QMCObvaYD37 z({Ye#=HAG=B$te1D=k2R%VthmG8`F2#^L1B7D!WO3lrf8#1T??a^ zol)rVBR{i)s-v@Y5nt_N>7!+F*X;`v`^u5_g1nV8wt*8AzAA=*=&oMGi@!KM4{z9`cp>nyHe&{MAHZWE>72&Z#=H`UbVaI}c@!gTB9`cN;aaDII(>Temg`Me6ncRtV_e?|cJ31~GFM0qDmDjz?@)%4Gr%Jg4q z&&uAdaL`4ATfIxX}{Gr!w+FEiBsKqaVxn|AUDm%P`G zkh2WgINMGle0-;O7J%Y-^U808NjO)t>9ada;V4Q0EU&rrC62)5r4+k{t-o4-L zuSjYU`8e6@$P!{GPi%G4*QejZ@ev``#$GKlS3!h1GFh%zX`Irnc}Ph>mS$BGe5pOE z@c!CMGoL^KQ@&3tlx48AD9Z2keLIy3eo*x8Z`%{SYTn@|(YmpC+8jnrT>!>$4V3JQmQ{T2=$$ zkdB@SCp*8(MLf3J2KttcymQL59PQAR>k0Xw`HII-)pXu@5Ece=r;ma!#JdvHc*Y#~ zaH(qb5}9?BPqvH=OuS@SM~EFZZYR36UmF{k8Az*Ui4A~kVS>Sp!pMDZne6-41rihULS;Rw zlDwp=yMpiPk4kk}voS%d)S~~Ss`s%h#?qe5`uLdt!G0&g=a4rxz@q5j)WWplFCHj4 z)4d&lWBPn%cc&nXf_A93QQZ+AZNwrIE{hBnuxZvJ`@Edd(Vn>~;wFOL*6DmTGXCWW z?k?;1&QvNFZk13vTb;U(*4s4@opBS8VD@xPL%=+VN_BU96J`znD8nXRobycJa!P}6 z#3Wy_No}`=Q8J7OcL|nf47tdvTiT9Uz%1j5^SgtkWTs#X?D4@l#y4KL7ds$Iv2^27 z?UbPxdHns+3tt~q6Fmb31}CD@5w%V=t9jNLvq9z%=@7TKCQ>)P0+Cu;l*H);aOFRa z{S4VN%+&i7H(4~mzwkJWRAj#LLU+Ap1qVT@ZZOca;0g@x{h;yvx;*}T@tNIQ!Xvtn z-b~w?z&X>C54+5&F+_wjf3$2ch}G`}>5~5>Lwfl|gJ<#3qB1az^%brULY-JRgE{S| zo9r%(F8>-%0k6t_RSt=b=5|gM$*cnui!AAH&YvFWdA@x1u9Y;-1cAKR#5f6dQQye` zOZKn)1PQh^PiN@SzJ9Tt;Y2bV0Cjt;dALkcDAC-jgV8}ch9RWk6uX2;YWTiA;|?_K zI6Y|a&dWx2T>Y9QZIa*#Q0N+_lFQ06lP@`$bUYIB9$TNB(nGu?ziOTh(eQvZ{so3{=G-BM~EG zRvXp=%ceg|f8*xw z58{uJc}Ow9tCiHv1n>EchBgXX>42*F^1GC8+)Z8TV@CwJU+N98+w_QYRean1C%G2G zL3RASY-F`M+QQ5^sqJM&ZM~61@I;4iUkat#)b*xXg0InI;MEtg`FxAATU%G6Xgg$Q zUr2y$EoW1u&7Q9EF1c$$THi7}zTua6FOAY;GZ^&2DDX8==p z0MuD!-S9anBo=SNaV4p)ll0_?ON0|3M+RX5jtrqG7iDvh(tj!&VbPA~FpnE6fzwXe z3IzJgN}=_I@T1vQiTZqj{aEvNIspeWR#h9OG@MCT+ALTW0atGtR9XIzJT_gc!uBNA zIQ~MM;(wIcuoMV*ok1H@5DVDoS31#IJJs>QTYU4?1=e9nVu7`|3ofJizO3W9Df4$+ zsiitFj#;T#VeeN4<&G8#2vur_Y;QN8hF^2#UYv~{nrBdEUS4ZbW1DBpO{gV5Z&R|* z&$8|L1TOPNjX*dOXREjN*R`dGd$BJB|GYE;Gy&3B%}iZ`$Ib!VxNV2-3lvS068?W6 z>L<3FvxqzE8xO{^eE0^8A#|A4QW5edvlULc7D$v_t{d-EWIwW4t~LEI1n6A!a3sJ` z0BOsVyVL3bfgrC|Tc=VyVrjwJJZMnNc)Zb@j&~ts5Cda3f!YLfDgs}c(|_cLb#s&UKgXR7 zA-hc%ktcD2mOZJf$T3oz{ARd$Xv9bg+3u$wi0E0Bv57>DJ2=o=4mkAIaovgB2Bv)@ zQ+B~9Y?5)cFBG`1Sh1tAo&m5rmn33YmW%Dt+m6uM^~iKtmQ%>o%77> zhZ`^&q7)BOg$W|Zev@)}IiaJeXLcB8vi zWjzG;wOsqo$S;Qtn>rfeuA}esA;>>dNAw`Fw$s6KH!2E>KR_ll@-UtwEA&p@4+PyS zN!$m{2;YY+#?l6b#VEsg;i8ONx9n|~h=bC^!R=b|NDxcEZ#dO7nk1_=9uZ$|t z2lWALqs6=pp!5J@KcqCcu3}gJ$WYOerWDPRf$Cj@0E{tIzF)O-RdcXur|6{Po&0x(JWbg}Q;R!-(CHbkobyHtBy^HuEZDtFWgS7~QX z0doUb@8xSsyCKE7lEGS^`f#a;JklqV-bx{z8h`zn9JWtB&h8tNz_glZWG)2~5vWdV zD^*V}zrgHyD@%*r-$KOI|7M2-x;rg3d58LBhg@^$D26t7kgb44`*4GuGDO^`b7e&{ z%&b1_aPBub9HgNLt&-0SDO%xwH=bC;CrWn8GL?(HisyB9R#*OiX?uTdU|XSxIH2k;J-pDG0Mj={~f!3o}nfq3#~A;i4fb* z<{Iq&q5e4NcLPzyVYukleF`1*I^tUAo_32Qqs@A^Y|(W$AeRxSO=3A=JL{5jqf^C; zlxtHiE82juz0+JF6YGEksa)#hJ0{u%JkUH zwh@SnSnt~_a#Cn~Rkh^hs_e}Y))Je@3hB}Kps&mpNTA26p6Rg7Z^iQ=Uh(fh@&Gtgym@yh|Bm=+2)mSVb8uH>a$5Xx|MN|p*~p-VbC3 z{;L+~^dU&A$U+KM(V4+qUZM{+7qx9frMb^T+(3yCg|C-m4_(XMVSF}NA_OhBKaDLq z81}~XJZ6urAdQm3_>wl~{#@5xt7*gJQ~M6#u~`h6kIv`8XTrJc7v}+`mZ>d=tzE2_ z^5@~M<{oSEZ}w&?O-Rv3RkxSujfCFd6y!;-g0`6Jn<6o9du7@MyJx2lUs!<-$!p_G{ zAb~c5SFgu8Evsw!8@6{2uGZ1L%n#u1y`y?4p|Y z%7#W@cjq~Tv{z%HS+z#9qiUk(uqqGqeBpFfcu+7|&u1!<+WCdDX84yV>WQFY#RRtx z*mksYx$n+XD_F^7!YG3Ff4GWfCNem)m{`o~`V1H6`HC!j!lPGw<~I?fOAloWlvvL` zT0LeCIEY)&9*>L_8A*1P>#7Mp?l45HMcbHJPSQC1K1@MUUyp@CXVj=JRJ z)@|bpOO}@!TN(?P^&7c{SO@Erl)3!Ho+kLH4gC!(;IJ8)mA2ANKs4_C%%V0+WM?JV zJE`UR{sM9I)kq_A89bM`+AI>)2t>e!+tABL@wiz5TTQFhLI;xB>r||kYSIeV!#a4W z(JumDLt_!qKnz!O7byzb{^UZu7~UFw0eYz$Jn={9nrL==2w)!X z-mN29FL!~0DlU8Ee}v;{LNX;SB8`c!M-9K|UKrWMGIe#M+}*jkT3ev~2>tD=aX)YD z$ANxjE($o+G^T-p0;m~NOrr~TryZIZkFwiuft`%Wa4>o3&yNI}4kNeb0PGCnwG{DT zv$0Bn1~WsZUjrQm`;+xQhq7{BDoSM5^}O(};KN_A5^1KV{jxaA%{W+1#xeM3ydlW0 zW-RN6B_3?p!S5KYvhKS(c)*h}LH3%B-b5eJQ*Mq?f>cU*!wj^(T@HUQ7XsP8%Or>z zF3{`p_zUfxX;reN13m>ggiep7M7@a9?Io^jhCyvTjeTcre*&Kj+xq9m1gZI_;~kTB zsTNKUxim*+wXbp|$kh>s_7C|+sI4OBhQr6D3mb(=GPm92K?@Q0PbOaS>nhR*hqx%} zcPV^ZH}GT$7!^kcXE1w&UyYLF=jL+U)Yq9Oyy&AgGpmSaviPZn*=#=yC`o^3Tl#iik01hUtkrfJE68?_uXQo_uE|FN@9kz;f_k!jj&G(4vpG zqXKRl-r#%=j>QeKXGH@fI^}CwMs8$BD!QoHED!k%f!z(ZI*FvOt5MsC zLT_7_$xqI^4yoknP7cb*(SzkxJV&jS<#F~+=z`GSUM>KhWp}bnuSJW#D<%LF<;=AMY5%pM@t)_P{nOxu5n4iT@u9KoSujQehQyc2%mX4DCO8O}Ag{!f+yw z%nBJ5zWBItvLzaXHBAbrJDyYaBTH$6m9ByX>l_P`c9yYR1m;8%ekVhTOI-UHg=`yW zffpKd+I&8yQ3h8ucm3zkv~o{`S&X6#xj>34(ijEnXI;>XG)fVG47JFl7`0Y`paCc| z+1P8&>Mtya+?g&z0{V@xV7z+1#(||GalfW#4BgOCk5BY|g(Qr7`I|iC>MrQDHp}f; zr6rl}5n72%@MVrsfX-80t3juyRz^EbPVpu->=_=HkG`VTR}N0WO?0V|K%Nxz)($ zNA;&SbZm^qZyjX1EP*AlyfjoEaooqSEgwODW>yGF&Xvi!I|3LGzOJP@r?V#aPwQrf z2NEU!c8UAO-RX~ejsbk%JR@w1abAzr2Y#Pn&aZnUar;d}&XcRW4c{KjAPycI z25jn02b9C3xtgub4XTF>j?w}7nt5bz4gPk&vRV=7k^yj!ruZdYSGPrK@bkPQ=QfAs zfNiseW{DsQxm>g>#9Z8!Tg4T%SKf6Hbo?o8YXf`d)+V7esh8fh*zYKdGX{B!mh>^= zt~EXmDJEG5qCu_zmf+X8^)c#AqoyX_GLukqz0otn&tf)L(b5{&(j4R@0Uw@nFb;BfRAMpcaQhiJUjlzHQkwrT)+ z=tjVZxkk=CkUVg2a5LNYv)xWw`7sd%6s3S#lrGg9^CO@9TJaqFO}%Yn|NLBlFxm?` zbR%*o_y>9~Q>I_%p|zch80{OTq+4guk4AC(hyF4!&~zR@gKKA|rtVEXG8^7?kY z>TX%d!6L3Y{{F?xha7rQ-(Q7829$Ov`Gpj1)LYlhkD|=`v#us;=O1TAUR-?40#;_x zVtf>K8tR}yL6-2hk2aTq0ZSM~T~)-pLTwG-pD&?SmBstm#)|kAEkRn!K_Ig7ex=QI z6;1(@1{Yt05@b6{qhWFDMY+181Tga9gXoMdoS}jgdghhSf8BA~M+foqbH@xF9CnYO z#>AbI7ECvr>D+4PIWN3`uT9}|(te9>a%J~Wn3uir_U$_e^~}r;z3&s}$vjVxyeLt7 z(=xs2h|RU0%}zXX4A?vx`4dGZRz(RYzk05J5byHsP_6C|+j}lvc|=pir<*z295|OM zD414wMB~tHyF2de+ZpL;Kiu*o(Y>s^J*i4?riIry_-d1CYsSF*LXBdeG7I3Lx3cHH z6Nd$Iv3keVTO`SIp6>e1&e!MXWd>a4;+rbg)${b*Cp#^RUHr@2Lh?B@hL8-VLM*4; z+`f?R$w%=!O?tBWD!bPpDSdTpIpHS-Hy-V~b8%=ldY*0Gc-%Y!$n zr(n=FhD&u9d?*9wqQF6sM7Ulz@Q$Ci_B&jC#85lG4wj&psXtqorrx4tUtqMtT%WPL zMSTp76aP@kziA zP0z5uuH%dw!rD`vC6aY6nq%5@dLblrWj(J@rf`0<^Iie=&4(kgg7x4Yn^f%IvlS3L z^Ds8}_HMHI`cD7&4>X^^xR${gUR4t|L$`lo*a6j$1eh(UzGhBW$s)^CW+P&NZPxp2 zLh-ZtLFJk%g5FK7`PUpECaZwL?ff-CdY9YGXZaNV)!D4}Bmo*vdkxw!-7LzHq6P{*CF=C(@Gl|61-;zC&n0qCmDTkz}-`Tso z9DWqIB!0QtHM)0)Ir;`pq4YH|wt(@Ry{D+i-L{Zsg68Yb>*JB^Gh9pP<%|xe*DRt6 zNnDJ|YOa2&aK;~|M144ZC?2Rf*sBc<-VZW=_V>%B_$)dMy z33@Mrxlhu3ZZ5yrcgOq?bX>I!YF!q#Z}U0Y!F-TS-l3K|Y5^#)jc^C6iq@73>2#+_ z*P&2@UHuiI-Kyf-S)_`z>4G`f++snIX2(#Fy}UXolM?OD?3x%Z^?|U2Fux#yGRqWG)6{eoO^{v+@6SssvoqPo)Da~aY2Y0E z`FN+`^Q*aYK){G^4X@2VSC(b5v|@l+Sl|~*e3C)Y&adKs+r4WWf#yz@`(-<_59{z6xpqAdSb z3^iP1CUQ5JT;3RA6Z*6~g5KD!hWI8%k{h;v%3k}?aiAXmv}2fPFlM6K&UvayC3XDz zgs)Dr#$LWk@W|uxi9%OPWMTJ=#Kn$hMVrWxhY&w7HYeocS+OMnWm^g@zYFzBaPDCr z{A3je}y8pjp*LKGA` zGFr&-mY_h1^1)io%S+1?<>aRkiN>|>yKxP>VWYS8Q*AT#Hy4v;ERaBga5g2a2lCRf zVDdc(qkOh-oim)RzpT)n-flw4S9PJnNC)Uot&F1xb|oH!2<>`Hhsx^_?3s@rv9&pz z&zp?DiU@xmLFp4z5e;a%6vYo{-o{D`rTuu7$w9t*7|y^ecOECg^m*YqYMz;I5D#C> zE2N(0q9ZEpD@6lIg;k(AOm&7pk3YCa#jo35_t;2~V8)|=@JLLPg+D^UP$8BCzwfRc zykdMag%??(Tk+lMj<+OAi{5VYJ?1_|F@hxN<@&A(npQc0%x>_w7fb0QV=D{Sv z)I%A16t5pJSBPhdP0d6$TffXq6mBlFDy6s21K=KtvZFi7ZlKd`mOl!z zjXUD09F2@Fl7ylfqki`r>O!qDW9DtHf0W>zg6V(mS_34T%g+GjxwVGzw09WY9YJB_ zU>FTB`|B@=3HITqoupw~U|mbNc30tlZ-gb5J0nY*DmgYF#`2PfpXuae7J0qA4!N7XK4#I!kk|_v|9# zfFp_XQdXLYK-t1NVqpW~o1jmUKGzJ&z>GRAHS2u=!!I9D;2r}NA|+2>#>`=qN06vC-o>SdUUpO7>CST8HL{L<$($Hebw zqd)rnd0;;=I<`%Fz5iLVDR6oi`+TI|P14Nykh#A|66+^1Pcfs-`~}#zCt?Fs9ZM3L zVE^ndHP3NjJ@C$#HBkGOoKg zC6bTJFaeanZO6RB;Jal^8+x7vGD30X1&Z2lLrZK`)Qk#X5K__Wtl#L4+|RaEnhh}Q z8*=7vAh|)O#@U;X7%gjX=3nZu7HUXkhb=XOxI#pGU%UkVc%i9O-;==yIk48)sz$CI zJ#Fk%A7ehD*!VGm^)1qd4dLK(4O7X)R94$wAyK&8}*q@vP=123W*WS}w z^kyDDTJ9F~Gap3_wS}((jkz+3p5SQt^SDh>GSbO3YfsAprz2Uv3gcDbK|tleX#v)) zMr^@q$lB6*)#z7-dTLym%c~Cts@1De(FSyNYjM?D6pFyJL5eigZhzA`%4|3BG<1mqdX{{piEF?*|CIMH1#fPJBo}s~=qEyO zDTwydgKs{-)SKoe*l+g5+>b&7GUuxW{6g`5>(1=cP!rf3JI(B){?6FNYuZRINu`88Xxl;DrdA1*atrAt@OosQw|1V*MhY{T$f{+HQAH$(JDX{*tNs#x~c@;?^T# zf&urxG?P=9j=J~=iR9k_-E#A`0-=!$vbjyH=DovkLav2POb5w-Eh8Yd0DL6<{#5k` zwibn4#A7$${Dl&E(|~r_MZY3r_;(x>U@ArexPSk9Zh!?CSEjixE-2A1KLl5eb#Cpm zpI|qRbfkD7u5#^1DynkzjJHc!Qe`_xCtU z+8vxPd+lu+&$8J^T&ZTcT`T0P^FD>Q4r;_TC}O|;h{%Z|PvFRk7=!n^@iT-{esHY} z(T<7%_jSjsRX+IE!ddBrxA^xJE$hh|RmRAZcY+yJ(OVKJkEgs5S|))nB-5LENJxLu z735wG4E}?7M>y!R4B4a7_?ha}aLZZ$6f_PN&6%-Qjf`^wU;Rm#ah-LvFB=nSM)a5e zdm#Tk6+2xigGD$5=1voGCg$>gPXKrfzV1K=G4|w2hfyABKM;MEc5ZIQi%j&BBlE5m3Lnaovr{ubi^XtII-g6@C}oRF8{8*SDGU> z!=30WKP6Iu0uFMk^mj?Q>s%2Z(@gGaM4H=LTrEGuG1ukKvny!C65Zz#=(&<9`_>V8 zBgmW1^Vy4pPlk7tq7p4Zb|@tlvO@Vp0!}ifMz^`z$#{5mg6n@CmZ_9X`)hR54~os9 zZM@0I$|14;-`D@&rSL!N;s5(sMKek<_HYJ1yY-;2%-AA9ll?#TzA`S#bbVO5M35Aa zFhGzFNhuMK5|EOX8oFU<5Rg<#nxR2JX&AahI)?7<4(Vq854)?2d(NKSPw%JqgZpEr zp8LMKuTHI<%LaRJZ=#PS_+*WIDQ7&&kEy!ALCN;8utJqAZ$%xZSR~?!>=)9*LVR>f z)0Bd|w(IGUM@c8eAUiCxxU^B<1Ks?oeunQPIw%PJGhTvHjii{GC&4t+2t%iJ)sO2#PiuGk77i^JeX`|r}bczsYgG6JgyDY58`hyvu+2>$nC3_eJ6RBzc3>?@h93-t^~1nr_*% z9Hg6?Q=M>g$T*>av7FV{6V8!0b(I0+8^lQ(gLfSqqnu_UOF%;*6{8QG$(2S?%*CH^g7ENgV^w|sSiGS$dHPO0ZLYKziAUt&S9(+Mivx}?eR&9OJQb(9nL ze=_%{s9r}bKf(OGS_N!;@E)Eg~FE^s+2ioDTi9@FR`n@oZk8KaI=4gY;l zJVSDY(Wje& z$^#@}YW^|93LUn~i&N+RYGHjrv58{C_Cw4nIT8+{e1BwNj-pExI=A31HvA0HBD08& zPL{JV9n~(uLdnlJ-_o_t>j7)P|2Hm$s3L!h5W#HF18fh?j8f39ob?TCqj(vW#0&Lss1pdML7PBRtiAN~@5WK%( z?)><@?&XxH_vqYCD|HC;_|xha{ zd-xsiZ)jFX(zEG(bW8%DLkzHlHas#ivWN-m_Vt5$PAa_>3Y&%cLGp7+0IL22dqLD&9KhCh4bJXI9Z%W^j}-lpiYYO$DIV$zj~@3aT&1$D)5 zcLH&oi49FyGc4g<97If+hInkuOI$*B85~}S96x3EFdpqqdL2v}_m`0%7f9=}S08|l=Dt=LDyR!3-U zl<_1mm^F@7Z;FfEu#00@%cbPAT#6S~@ONQ?S_BWlm?quTK7*wcKZ*cU;p0kRq7B&d z4rXonMo~Q|$?fp)JwSgD%?n@*d0b|qb8{zzi@8!7e{Y$YEQPzjU|qIA^oCPT_BkZqxF45`b~T&Gl!;-Lzsi=n&C2@H}m99z&%jMrp508LICt zx8;b?4BYQ<7$qA6gKS1C?XyX^EWCgz--S#oZ~D?hWPr-wG#r%<0&e*#uz6J%3RAfb zFliWcx=*qcsiO6y?(!5Uxt^{~-~lEJ$ke&J!Fp1dWAp10jnzucW{bqLSW(oDcY%eP zVn}zH%HO;-?$h@u2K=SyooA{K2{u?PCb(^Z;kd_2sb%e2DRn^YLdRXZ9?z?)XbbhR zjP{jYOGv#pd?urRl2Rqs;|HZ*p`}0_BO+EP5HpqpRD6%R$>Ly5-w^Pt8d3IDC9VQq z*9kY71@S5uISO#QCVo3hZkD^R}w0 z>XQG1TMi=5v{}MOO2AYB_}-ECD5D1TS-w_HH;s8Jydx|2foxRLq$S*kkWr0-@(e)koR{gfIT z0*D4nWAm7c;G{=B{G!15!I3wTBC8}5H-FXJbvT~QL1g0Iv|b@AUy*R^9HJRi@e&fk z!e-gzjeHRwE9YsDXl7ubu-x@Hw`SUV-s$yp3+SDM;tY86VxCkg^Mbq!$!~z`Klait^+_e(2br~PvFd68f z$RyejQW9~ZJ!`B6LI7~zndEy@LTPQ_^Stk#f4)Mlnm+uAwzm7DX69{7f?m?d}5?m!>mb~}Z9p{^*GSu?P4 z7KS6(Y42f@4(twIdWRp(4lJI=0z$-{qVI`f?};Qn0Rj~=%G9gp`XaXaps3i!S_$*l zKXCZFQ@elG58YR3sjdNVKVk^sd={Co>ZZUe)q{@wuyFd$1wU85;P)fHL z=FiXn@y#kpArktlR@O_Pnzrf$n2Xv8lqVTSd;${60#baTz(|&5QOa+k#F|FH0Z41| z1AqWh%f#}nUw#OeMs5j_?wwqTuCHpbQ)Und=Yh_ZEh7g;<`~Ekz^D6h${**AfPMvb zCcr+BuuF@HW$>JTKgK?#-INxE?V2a=!*q8cL;aJvbV&$Y>j4!1(9xD ze)GYC5d&q5{g_$kQGngy-~dE>ev^d2tYkLA$BG3AB<&`>+-e624pZ!p;;rI8E9b^_ zn6WzVtAh?VEYmvPo7^9~JT}jHUD!8a+g$?!RI+;%2_HEAq(AA>sHoUfZ24C$^+_aI z8ugXG*5?h=kz0)i>UWmX?5YVyu>fCQ0>KUSF+Peg0rlsu(qCvE0oe-?(`uzRt?EjS z|E7H57t@s&Nis4O;U<6^R>5-?u8f**QxYXW8!9k*vD}S&J;yx*5a>&L-E}NL zJoyW(oOLSRroF9}^lAXY52RDcNg|u6sEsnN3FU-E>J&YW&_tr(0eo$K1h=Kt;t;C9 zqN9mDATD+d6P~Gfa)*9z+G;5Iria@!x0at<0HceIfee@Bm52Zt83hAN(yAEnu8M!T zChgL$#Jip{?@BUK5aOe6Y&=GM(VE}@o;m-9AVgZl2$X02nF-)8tD}33F{S4Tot@t= z4m-SYxGoFR$0+MA`#wocrc)?X*eq(W_<(@9I*ls7&4BrNB+0Pfd<2K)?T}`y4~!%< z)>vU$A^TC$PKbf_RT)7{~{>YC<*u zl=8UIRCz!)`G9G2noM*b`W&(|oyG^=U;+LCX_ovzCOyRKLJXWMlrT0bM!p#v8)o`( zj<4^%^Y*u39LbZWIg2W6JB2J7BP$#ah?>^-*Ra4xO3~z$(UxI%a^Kv{0{xg zh9gHMZMB=UKG)b+T@)w0#>^YBNa_&McyVqrAr2r|^3&VA2)azcvr*;x#EopZt`X*< zR~6YC1V11C2id$#e&&Oj0km?R^kvBjXL2>jTCY7c0d~vm4r=MpKN?R0-YTDcO7JvA zkkDXYWvoO>hpX-M7qYvFI4e%lRFE?HNy45(8Ty#h*Ue$n;1K^hi2vjs$wHwZ(kGos zaM$8kUi&XVM?@@HUb<#ggV|wF&a3L-4bp2zG*AItjH0uNz1CmI_UA7rzeE0__}L8L zs4wqIn?Zj*1)v5N#3n(OH#&#G_W%96*Vy2<-}A&HpbAYzE*$+27+t^ZPj-v206Y}Q zqrHZy>w4lRGyeSC9}FslfeBP&dX1Q2{yh9&4{-er$uhtV`E+)ovwi(efBTJOWEAvt zT46IzpT9l8^*27L17PjDI4^McuUES!Q+~UKoEIYUs|YGi$g{saz_m9#`xt=`$Ev&g z2;Z-U!u9)IyA`eQ6(DgTGXD;at3&{n>;wfj$gk1K-(Mpl82}`;KP=HZxYC$F%ywd_~(FJcNnFVX)e_rJlx{}XcOEC2kt-w<_s>#!e4`m+KF=IKpJG@NUE z^mhc9Y<5MXUfu9eLv2TSk2p1&$CHWio z7{2b!q2c1%C&b!2;(P0s@1^=JikSZe0irL^$q96mD$?R1pordda(`{49w%tt%wRkF zmK~O2t`=j=8LGSXp*1Or-x)qVao!Ut1ymh`&mQWomzr;tTBo6Ah?STu&2_+yC@PQg zGmOK=1bf`_*m+J$C)cMIzAFyrNVeAb?xex06xY7ckg(C&%)w2 z#}SB`(Z{r$wsWW07_Adw2oty~p(p?(A&h^zS5H*mNl<^{UNcv%X~Zt#%5VCzmXq}) zoEi4|M^FQygRzkxvCeLzqdJJEdH5Fa1%E0-oz z%jqP}Z(gINMNS-h>66Sf!rWj18w8)Ww`eLbOXf3=rv_`NaH^0s8NB zDMyM7$j&80;RgFGTB`d>#@RP8AVJTl@37pCPcpffdp(q*NoG5&Z7G?(TDyFU(=eg8 zWuJdB9lDnJ1qy*6btoM%RDFDgtsJplI%!fALb<7XeE{Sm_Or`prr) zw7paxo%c|w5yfw>M@5bEn{Mj_-j7KEMhk&Ir6&s@(CHM^a7(FN_wxmpRqi$WTm1M~4xNTzv4P``Ob;8;l!epE|0 z_xwPjmv`O6@p@B)2Q2BKhg7fXk-pTp*Gz$A28{DcBh}B2 z;WCJr+i_ zzT0xRO@zafM3d|=6(KQWrmJot_hW1KZ%n}e(eM;Cq-0nmjA3+mvAdt1aP3Xs6Y(y_ ze9C{Ex+i_%cEt4kG{^s+b*dcrN&$N=YYF3;o4?spW>a#sS)g9i;T?K&NAr|k=jrM$W+ zg4}2`-OWmS(8jwqP$;Ywn66);KlrL^x==6EX>BwbMv{4OKlJ`5f?osZ|DCfw0@5T* zI>UdNzQ1vy&@$#;cdc#)EQZj$#(Z650*aH`Lr&?K@-}ehd_|I)qc{0dV5wb-`Nhas zKRD|nFiWYq({doWcC5C#Kd596Q0@q-|UKcM6h^ewFMW%Ov zH%4X=={_pILHHY8_zSH%p#w)llaptD$;I-_llXL8Ni>%FaJnvqXRWo7%AZTD#uS(* zs0YqX!A_Tpt($p7{37K*+szHS>Na;Zg#nY*9?N;TOjN4V%cm~p0WXx%2>-WAqJNTZ zKr&)v>U)He>gy1SN5~eeaet*^8 z%!mvn_TrpcJa4`mZDz6et-9BT|^9mnmtmOIH9J5CzU$UhJ;l@Bgw`U$5G` zYCc{D2_(eR=JhKvH>UA8{lv4mfDda35d>PSOEQfotm;;reG-m!wh;RZbRO=B>Wu4N zHDmpwUjK%S=hDz%!kRCHD3a*bbfs)-hxwuQSXa7M4@lIQ`!IGKq4vF=H6(jseO{>f z%+z9{yiKwFOGH7=p1dD~%<`nlrE`7Al4bRl-p3xfk0iy*u9nVr>k;~N#s9_wqOAtR zgb$arQ9T;}LxvZ&%>yLtuz__DDT>T*>PxL3)M-=ZKAL=KNaU34HC?mGIp3IH$H?h2 zn7wY;G?AxOJfxVdMh!~bx(hav;<8-n$cK>Slbe|+=10k8CbI+W#70pJnijWslZ}NT zZ}|TWSat!Q?7nR3xl#`SbNNN5xa+^&W9GG?nVgGV!9kjVmycG8sH)#ZMfmIt?N`!A=trrj;4 zTs*ug4%^?K=RhqSHYSV{@9ci;5K9rl{*R#Ie>-dwb*QQac61Y%aAAi15iMYa+{aY= zi1MOb=zTl`v57hYv8blNnXa#~oChV47vf2~5l}WozK%#G7EGaMQMWGKZwD94keK)F zg;`8hH%+*%3wD-G!=7{tp1kF$c2nOQ@kn9^!YBSiy^T4e)3G6fC-*^GR}r2M$Eu)j zjdV+Xl*73Xz$)#6&>Rf8v zzc#aSG1KUY#_-8&d#aI-NwDSdlm!xZeJYOX;W75>T+h8}6=g~2PZMoEo}39%vYw5061C%{ zynT<;;$db{vD#jcA)SHU*?JRMp0=auOiz#vM}V_&pjhO?g!MRfhDL>r#iZb*RmGK= z32M(&fVoggTZtrWglYEg^VLN2lyqZch#i=ujK87zXFj-r=yau@qI;M$rb-X)^`<%g0=P1YkE1JSBc88~lAZV22x->E;6L#Buy&6udC|5kGx2A1B>H}(O zS6ZrsY_q70F76SZZ$vo=d=ZhZnIRucI)FKdk#JABp1BeCq`0kmSZFKcnnb=)QsANQ zFjpVft^r25xwCdvfA3pgsuR(&JM(BZ=9fv}H*Lvj_~#k~Kn(BV0TP8d}+J^~g z;-(5d%nz@N>?(LJBJjg4K)rLHGPWzcVFE70>^PTu%o!q8oGJDl(aeUI&K>qkVeqGV zTbIN0d0hs?NtRz)4Bj8FPgAmB5nYi@qHx$o@Q4W7y9s$e4NZ6t~GlgM}9JX zuq=e9$zo_(m(c7)T-fqlWOz0;IEiU@C+gYX3gfw6m08 zXbOkr!uVzV45Az>tal4UvI?n}OuM)m1xQf9?d6-NJnc%@DQ1(NtBx zJAO&)MZ|y2e;*YKw6w^8K)x3pc4SN|qExF}Pg>AcCFODcgstaC!V$t2G|>a};e73+ z{ERQkxuj|y3t7FrC%l}J+yd5}n~ak)1k7FQz2A`dm?TWSZeqIm$yd#PzRlHz zlLxRu!%_zXMK`Gaja)&Gh!8s2UBuI)!z-3tV1Cf_5crC=de^oPOg@;Av+B>o+)-#Fk>h3W#8cO&Be>I<~DuVTLMVuSw4?)nYN zU$?HR9JjtR6o&i_!TyR2x%*d(e?>W-lfqYZOtY%Glt0X5PfaRdGkCCO!QiBA z`0LHYZZemkpN|Nwy;0{>D5?E6?#E%iZZXTbA9x%XXr=zqCQc&Qf55Ff==9c_EC5z;HSa=m}~0)CL^ zC!Nb*1Pdb{i)$D!lQ7bQ#g%~iml!ej<_#6|l?N5`;t2&IXFu*L6iYaBOjVPaIPbk5 zzx$B{#IUrEA*!9w&U@Rp-ewbP^UzT17ok-rDHqsQsl5$l{wS- zs*3Ga>?25Lr}W;YgJ9gFBq!?Wv%b%Ie6~xgHP42}&GE6~E zeh}~Ia+kW%;YeDBsl|w?Wo+(vX>5~JRO}ZCucv*ZTk+c-c7xD_v690VuML34b|(L4 z_)z5M zCFI>*8>8c{4js_2ZabkaIk`XC#X^QVr?8v|-o@CmzW6?mxCbkn7B6F8s8C7kZv$eV zD~4(%+$=hU+Yu@(z7uory|~T~aRjoA&*$z})akF<6uV!b7LS3{jfy-((#$YFw-+ta zcNQ7Y!)zdtX_o>n6;veJbs*TRIHEE$IA?D zx}cwCGBwjo2)13vj&S`6kH8D){4rJUL-$)PXgl}e-BWH~CFBjT?h1tCi*%_9$stGt z24d9vBUmf3q$cMf*w;f*2^!V}@GFZzxbj3S5AMi6N01Loza!`MJS^U~jeR?jPvOZR zKYv(aPe1(3MMMQ?b;w@JR@GKHinV?HQxi)v1Il(a_s);NUEt?_y?ESCnWasEGI4gIs%4*Zc7XhYhk|ps#Cmx_wV;c^XLycqo>gN_PuLl1t97h z1W4UaRZUJ%aC0~)*_h!cTvbV+)E*n0)z#JAwROw0PNg`UCL3E@x5HjS7-x#WSmj-! z@`r7^Ny;Oer`;(DUX(Aw42n$7*L&ea$fhxU_nj#NO}p#P1t9@xS*)Z4sh)gD-Mp%# zoLSrK(%y23F2+Th>thw+8tfbQlQyYyjf6zxJxNJ#zS01P-nGSna)>8q9DVnr^swhL zUNeR-VFS7-Bzg}mVOH=tjVxqKjcDuz+7HEyn}PV0lAw0yDdj@!cUQS%2XBfWqU;8O zHo()a{ICDWDtY1(_b^vk)|@;h6DqWr6stI&6UH+yQiKX-m2^jLjuv<4nfktM=!`!8 zWb9!vb)W8Rul*^vCwk>{hG5x?B5lrhoVn^FmqOYfTte*oFvQQ0U#$>tF}twI+~hq| zJVQC@@2-6u*QEkMMYLNR>Ea9O-eucsI^U)r9ue5K?@p+bn{r%6N-CUY(s+D24}C;Y zp{}c^r@tCf$E~?LP>!)x1ow&(=*6QoFn11~62gT+!=j>?c6NM_)Fk=JP&tg0{k4^I zBF4opHA%TlwyAayk0cTI+UZum)jweW_&2VK!aZe`pLw zg>O#PlaG~!J#eP5s7iDPr!tDpPUdw`cAxfLI-YIU?9pt0VT9c#Av8W?V=s2~LXY=6 zZ93PJty!?WiM{XWG{*ta6lpQW_>i4<$E-t*LF+WUnwN)Js>vK)uBKLM1j_7ti(#It z#(9L%SCBw`9u9%%2^d|70@YHQ-&(6-#xx3IPGi+5)5K%uy&owTNAx6#BdibKURp?e z92nrhhwetZ@@(_1e?dSA33d~vps#I`X~*bywa z*7_|LYXfyp79JiPncj$8bqZ)3p=uAv^=TdcDfu|y<}mnOS? zNxZhDy6*)F?%Aat+mDM+4syaGisTJV6))@R%Sy343062Ie{{sZ9|8NGTu+caoD(j> zTT<3Sw=IY>$=ANcYq(Xm`_jFqL6kV&96L2F%`70)FwtcYKhgJZP%l93T6D%t|BT)6Z7yT4yz6 z?s*!&6L3U~R~{EZRW0BRd}4j={E`Cx4xpjj_dt5<+4P{^%-TSZj2MX$hn+q7#Aj2Oj-X(N){!^^fG9AsJ?L$1G}10o#)FccH6W zmBl^v`c%DCSeeE|?>}3@?3A5;r_n1yrG55x8L4nP-$`KRB-tMx(DSFoDd#0rUxrD2 zQ=k2Zyp~Hq0jdgVNqfHcW@?WyV8dArxS`0+y)`Z)l>82`SKmEk1(?;#HVdL{8lFUt zGLH@CPU}mWR2!X<4&*;73#k^XMj!xtpJYx7=O&1rMQe7o-yK4< z=rdgw;cE|QA*Y^ptCQ|}I{z>!2iMx#dN+hDFJfEfh8Nk_H*CUClUSAZhFF&|b;?%C zbYJU+};u`jNJxA&}^cZx#&VVu4QD#1RK8qc!i1 zsQgf08J#UveL55vi?liKZ#<{qmf1FC87sqzuB@kSIiS;UPVdkmUo;IrXlyG=0dt*S zu=URPe-b=5lbsEKS2H+FI@dHpsP35G4!hl#ZK&cRh(9+!iFXg2ww$y{Q{~SU`mH!X z^oR7l;i-%4D1bYAA~a$1_B8w3o-ixiTL=FKVBZSe^DJ`vrfuZF`-j;^br;2*%YY>C zT=m3HYPU+2$wtH77|p>7mxzP?>y?g3SX~V0UvN+;eWo7O_Izi+L)+S$b|1_2k-G4@ zGHmfTFPa$6VaQAq4#P4u z;qNTlFeeSVfqvkzh@*|EIKtV^{xib3)sOd_hLg!?q@5cW&*Y8Xu#yueI%G^bPkrs* zh`x+-`&PH;h)CB?Dqd)QcI%`=AE$Tdi-D~KCz$`1U?=}O)Q5Lp?4cGz53xAtx~tVr z4bl{|z;!pR!>te^ur_~Yt!NvU}w=iuS zhN#D~1?7oDh2wXgzEmd#Yqrc4YdcL3wGyK>+v>5Ba9|nwr0Pn`=&&IR{SPrBM2$gk zl&GDA5#PM$`9Qu>B@w3)l_B~i{dd|#9K@=gKcnw*ML<8fB&RO&AINY-7C6@rZaZaZ z=~viK4L)1xPc1WD4T9v=rLG_>j`U@Vs|2u&QfWDD%B;Rw8fL&dUA&OWmkyWlxHK>{ z>;K}T;W%|eUF=!i+aV!zn8r>^We}T3{ZVQ?fkTLs25O>x;6DE)EhM2HuK4QJVU^~~ zfDamUtpm9C3_{X-*zib!4j~!GO-3#j-H}D&k&o1)qlsU!ixmK0CVe`frFDauc;s!E zYEpOcCniJ~m&2t49x<6ISMD)C^T>0RTPEYL@~&*fz`&2aI8Tic<%$eC`MIQYpLy}g zSiG#aUyHspFvhCYoVx{R3=6PovB%wk(ny$Oq3IDtxR+xwbRUb#%Hc6c2Y#t|vQV5q zaXXG~K?njuu$HVdY7QUEaP@x_+z7(34z>oYs19IQuo=?*E35GQe0{rb5;*|MwaBBK zt3_`*Y#Bj9>CUtq!}_uxSR9fz&(WjnLnr=b?>Y+5((1MD3Hw|a`UU)Sl_hTj^;oZBc+h{x+7t1w zblSKd==+5H0*&~q9HPcbOeMyL81;kcly;^*co_vdAFQ?dn=(hkYrn@adZ(z~hDP{t zV(0mDP!lh8X&ZSHO}WAGPh7i(ZPG~cg)yf}O^>u3p4_`P#pJ^NYJ|CT@-|v&!A6Vr z2JsGW$8CO`XuyB$wLLp18{BFvEMx_%vMafowtKIjRj>;tnzxJFVZ_Wn6a3Efg&f>k zznjO}Z+3Hx1AP%B`&H_th42k=RjC+;V{oC%YZ&Gn2c!0E#L@BTxagABf<7 z?z-e?n0E66M*c}uZP&{tSFROdB>n17ug{N|H%}!V-F}Y^#+40CfRee}w zvFJUjgoZ|rzDtMILejg-g9vv|k9Tb*DF(tEp+a>%LpIo#n;|7+X3?O~9%;O>YHe`z z0dwI7vgmpLVnHxx|2vTF9h=Ko8mw<2oa^t{H3YM45&1ci=jO1wS?h5WJ~`xD;87AU z-`sN=6heA5@_?Pa0|uVGNorI?WAybcXr~9KMexEkFMbT}ST*XB5O*&HDuYBFKXWrP z$FJBj#KNy_fXi@Xa9f(H6~!C1Jg@7{_Bv$SeMR#?b;z{JZD87b=vET8fky9FU&`qG zP)lN_l3V$*F#6m%!64KZMsG!vhf>_gAh7wdY^EqTz{=6LzQa^!qYLi#IVAj{E%o>*d>5V$6h)=3D%KsZk*U8- z(ev{*!xc@|GXdYgc8@tbc&0F=7iPz$y|J#XASy0iq3z0dX1o+2$YoiWjaFj3Y6lSt zG;^q6proqgTVx){R#%VXGSwZPo1Zs}4H90C2jvS(SzNhI7X2HfFbHQ@xa`_IyN}jU zgPujCKECuTMVFZ=`MHRP@<|RE9vke6jL^mwBpi4P$f@TeftvfX<-@jWdEY(?ZI>eX zoaq7yi)w=p-K+_b`};RfnvkvUf?uFzW-I2CQuU{3s*~7teeQcGs4faDO4997-*zV` zpxL{ZF!nx*tC#I=(-VF_0r8_E5e~lQgfcCaiSN2)o1<0T7Aj=0YT=(XieJKoB(_hS z1aS8fYxfaaFOQ;@krHvvJsvpa#@rOLh)R@9Fe*B(N_?z`hSMAug;gC!v)6K&nS3@noYGDGP)4X126 zed^$LPs7CfHO>SPEj`SmnYB;)^>dDeuk65h>$Ly_erWe`rM*c9wFqi^P&cb8G1U0;Eyjx(b6{&tp)l?qkt?%G zOdpou#V$m17>`u^Xls_-Bsv;?4jCVw(kHc21A0)3X<1ooTdcSu3PPizwosW?PH+Hk zpKDKem?^=Gs;OB9vdKVtqc6w!0hf!9JFv>zFKM$BNAToXr9FS+r1CeN_Y)Hx5j?SQ znC^+&A}Rj6K1HaRq_VQ`Qcz4x%>APS3psYi$XO0-f*nh>Qr)2qr_!^QB;^-kV!m|U zw}E7Ko`cb}GHGn`$3U|<`MKtGU&)BJO4hx3=I87tP?OP4#x&sJdh^q-6J4B&F5G%w zJ$`g|czih`3bKAY6V6k*9j7Cbt~JTKb*Pr1{{eJjHCCNpSLR8eme@32XFF_R=#L%T zOGN?)uWRC*A*&x-8vaSTDhP=one+)xeFY>f z$CzL@dn%P|9c)1$;bfuY9E7c}x#qPoE=_lb z=FB{)beSdZ3yfR&IfBRgu9_3F(cd>*QWulDoDpyalSvI^_v$xu-fwYW`Y zOI)9=r_&KoD>W_!vHNm-sfH?Cdi9C?E(YVJGIXY3B{iLvlIhMH2!+_om#Y9>?mYc}#E59}cQ938&l)X&eH!P^^V zkN3YUvd1r#Se<@bRqAlcQLmGzJ+b)gGHR|~W)-daXZa7UD&XZM;z(XHVg%g=B16q? za}0klTk;Zm1oN^TtinAH`46dQ68iRXu?`i-sNd`s>B5!K(@5vj3uxVsX60K5*;lIi?N0IRzzCO}E;R56amGV`kZw=LC@sj9dz@)qQ0>9+_>kpse;&=lMwhD1 z-FexXhlSQo29ICyd!kQyR>HpwnwqOXlo=n=(e;AbteTCYP%|}o^^-Al{ckgd$@5i2 zVQ2K1$Fa`#-CJ(wKN8gOtL4&*x@-U3K|;Ed-g<7MZtiHXOE*@VEA|F{CiMwyecwf! zm{D^pZ_VCV;(e;xQ3FmSon^HpmuXZ`4Gr1zqKaB8qd58JfX8IE+nZ^tKeW7;UW%b> z?Vd?CkguXYBl!|szb(f7+B)c`6=V$fZ2$#v)IA!szuAp8c+h2~^~amIOh8EOT}Q$a z@9(fgM*@@svt9n~h@SK4VwrRXYG#7V;peVYY;G%+&SxPO3FcIyAmykWPVJd8%i(6! zC$5H)K1Oty*lNcc?@vmfdv6Vw+DqZ%hSJ8B+~?&Ti%2rA)lp9ia1)2u79iiGAPo z$?>ydQb%^l6%82ViFL8D*=VKAem*47#5w3p+BhO6=FW*}2j|KDUd2`qul>i##Z;n! z9^;yjjU@(DR~hLSR-5*3qmHcF3mw_rVlW>Vn1r2xD#INT*<^H4W#i9FA2yz$wsRve zo$?jv^n;S<5~$FVx^B>S$k2ohnhuRJs`)N8$S#c-Y+s7+udQvxCM1||yJ(o@q#No- z99J!FwIcrTgK@=vkNw~t6DI_FygxpQEIc!FXukMC;jFz~$-|#X zMY*3js!n^~n?D|8v+@S<9@LmKfcTMgSe_rW zcg1=6q^)ut@zJHZ>JF`A2V|Q?M@bF0R;l9XQm>3GVZ1RQSZ>L#V8nA9Ti4Nkmv2fF zlaVq`duV=N;UyH|9`Q|H5g_`hY%~!xjBVd-SYP&j(c?sw>N%!3EfA4CqEYbkST`tg zG@22SFbAAsYV4iSOCb8Ouj-(@Rk`VGSlF4r2JghzYp+g%-{N!Fib^ONuN%)ZP5$6P z7xNYg$!?3aF65<bMxT-A%K=n+hE)f2(a3V z;D|)GFLhDcJO-If%W^Y)R)<<_^f_i13iB{$RI>6x-Ew`pE@kW?XUbFC9oE-XZPLME zAW(a3;(6C&BW#Y51fQ7r5n~*eeaFzs>x+x0U*3J7S(9S3KO-5TyuN z_6{5{u=vAX6iPltfA0EyR=naeGfkWAxD;I-e`0(&c;_rlbkm_Mmi+W+-O?dDQgK(C z-g()OYlzdkqjwGEyWO+lX&1E^LXSza*CbR+W;5@v=zqPGWeUdPY zV`uzBRJwY~K*}W2Q0-wZ73fD7@kjjH-$Q7(3{9U>-YsY9Mnkh$5k%@nyS)EN8}Kvt zl|TG$ngCe}#il*s0)|(wb_KGTHeDqoBoO>wCb9N!4jZOD>lQyrLA|_iTCcT6zto!u z=%~Q$JOOI0mu22?U17j85C{`^2?Yk8z;v&gxPOEf+8@agQbynzY1X>dFnm3y3g+7*cDCET&kw`F{-5{w0sY6afYF6n z3PyZHQ_TybFy1KwP^62)52)&368Doge0Il#B7u{iQaP6v7GAu3nfF{=oY2KTC&z5G z#HEbGOs;Ok+Uzf#-T!f~tCIlK0{~9gK@5qO!!dQh6Td0G<@h`q(`_(0Fz{y9NHmZa zG#_Ii$FW#W?EQcv7=9!eLp`75(z$uIlc_C=^a!Y9-Z;$FC=v73bbwT|FNzwErH74c zkNJ4;pMDs4nMci~#Bu1efF}VNr$y;J8J7vJRozG4;-X?vkMr?I z#kCQi1TQzHHRqhue6jU`AY&AFUKYYC2}!CTTAxL&{Re%Bsq;d;N4v1O+KF_D&U_M^ z)0m5Zn}H^e%hxuWk$Y&-_FTRi6t^qEXFYyY}3`#CE>J#kno+{U|;u z_13J}a3|0mbg<#DdG}=i&ZK2yROB{!2`voxNOVWC<4nWg!Hn40%B`Jxs1^o zPT`%GyQWDBBNe7)(n%+u8=D(nq_(NuUJ=jN8``KqC|f;VP%Cj=(*NspiT^^!Ghcxs zed{;wK8N5Ui51^BQL|QF^id3FZH&aJEq#}7%ER-tD+?O+f#R#Pa+-<*(&#fzCWE-D*W>sNB7Cg`&~p9?73PJP$EDZy zX5s1G!H!yYPM5^1z;uyY%gbH^bHfDdqRva*TFou+aY|p74$^3$ZM50x9`ktFNq+93 z5`47SA>Rm?x)=qc=Hyfcth6OmG9x%(V3zA~Sx!BsJ3H(co@OX=%T<3;Hb03V{JO{1 z&2p^cQ{3vb>{HX0i@m6Iq~LBCEYjcK|Lr)U{Vg8&-GGqy;h!yQnWJwCl=XxvWe>M& ztCv<>m=5KYb(|e#R?>7egJ7f`0JHIASKogtgiX)`uU&|NcRi1qtz&-@>M3=-=}8i0 zpWjI}@q&I$-$@{gR(0vm2#ZkiKH}gapgavbTaf_0Hz%~;UXYKfU}UT10;Z>J>rGQP z?A;)|%uPjlbRH62WLqhF#7wVTxqdlr=Z(r|Wc7`?ZYc<|-}hQVTH5MJb+3iD zLVpCd+m0ps;+dBvUFf1dpUTLkki`y?kdK@z8VB_}r;vX*psjd`@xUWlzg zQ_uyFXLZ_3h+Vp)LaeHmmUZC)^iNfs9I~?YYDA;@MU(=bk+DseY9{KruH9*;>|%p# zT_{ZuX1ljhK_h*B%`E$cwtCs94W%A@8!+*d0ka5-vhOXA-D9JssGq*D6myW(hi1le zwi|d||J4J-low8?CdMB>DDYZF%IEI5CY}^=Jij;#WKpP4x*Gz!3sI)_09i^aC4J*9 zsswjELf35d`jmlT6+9bN>n0iR2sq@Vo>5-ML}|4q!M2u>?bdvjUoEgmuBz{z z4}?JEicwNi?^DzBsqaVrVmT;1kkgzB-*c6Fa+H+zGZunQHA9#Mm||15x)pOk+hZ2D zI$v>FgkRKJ&sR+Ishb2ZseHJke1feJA31}WTeW5*&U!WpM~jMz7VA9(8JDr2irnJ# zM0E}Z>I5R`^C+m(wyRZnDaU&Gt1}8u9EYNdfPj^RYeyISycQ2WNvQ61 zN+G09-flanaZ=(+#QAc1%j$Y_W}`bun>-PxoH+8TI(m_NJk3bx9T#mX9*$;)$fr-4 z9Y839s_JMXlI+sG!u51ir7uGX%W1dwOuwD1e{qd$E0mtON-nQ_-!&j0AWYl+WD-?= zMHs3fo=>4UId48*s}RR&ypX3-=F`E%D?8-`*K>mDJC$#(bx7Jf=f8CC0F82gg6}j1 zZ`;~2JIzg}8SEp4bO7O!#a@p|M(njS5bS%>8|`-Y(NZ@*4}fMsC&~BT3c(0_2x^z$ z56ARVdvMHunBKot*~$U)&k~q9$TlC(|?jvNKDDqXt6u9=_TiL=(0k>GneHM&34P_(A`Zd34Cx!PKuOBG-3_%lfwqi@R{uJjNbwc{W}vc6 zDUOR?z{MB!dFP3V=Y(rR!CUH==Qxhi+ljmFd$O&qd|T0Em)2z;$j*|DFd|dOT8fuq zHYjWZ$2JZr?e%dSVo{KBvtki@dwYwz&yPZWCG$t2iD5bdWD$OeP!3cMi+m||JbqI;zEkW{-OICf7A?+$w zRBS8F(nR63H;JP^5sm`{03ZvKtCSrjDw=Em$KF@|Rkf{cOG=2eY`R-ex=UKR1qr3Q zyO)57DBTUx-5t{1AS}8|deQmLwcTgytBt8)_O)WUaHs0)hQLd5S^Ys8AYwUGL zSB!Ah4!j(UJ%A_V1jTfK2pBN%@!zT-<(L>MFMN5wV8D#}_v5Cyqi%WF$;}oV3>MTM zc`8o2_^$r+;rttfo!kYH&OMYgrxeJT!Joi3bs+Ww`rBUb0c=k&xv-$iwrkXM#j>(wEm+Mo};`{laCx9D)GoxE# zO$;Q-_65q9X2tjGT|_FBb{8rn1+Pv}YdrSqKRbiBW#r^4z+6}70_uW^0#n4BI^4|) zHhF4{y|YHKlcE($jb|+`RDyy-XIWWUvr|%$&5LF`B&4x5^V<-g&n!M!@}dtw2~25O z0_*`coZIARr=r?15#HZ(`TztA+5vuO zEAdpV^1)({y6@p(|HE0|z|Q`d;5(rI>6sa$jA8(yyy>{_BP-t6@ubV7(>IOgU;Xb~ zXKxS|`@Qbnm5m#DT4(o|X{l*u>`AZE-^Fx}SfV8muapN5MWzC#!#I*VV{>``BIC)} z*tnH*u!6A{y-hO&l;k=$YX9el`=8m~7#=N!G(Bh}l_^2w(^P3ZjgT$w>E za3biKRP|_})lQrB#c*4JLxp5A=$Xbcl|)A$;Q_|FLTJeOz@nx&5aChg2n&HM%W+r1 z+_F}=*ydy*Av@re2_u}eWX;&iNsw5Tm5r8rZIa|ACz;-%6q9ZBS+9})>bK=S2PGvD zfo<;6GDMo*eTr~{!HqtDGz^sQ=XGiRF7E@OnJnz&@@1MV?Z8qRw2X`+u_&GN2HEBe z$pq}SXU59ST?By8fGV28Q2zA}T9Pf_;5$Sm?(V}gjgI36Ho+JJ_wy)!{|n!7y^ER( z&#Q+?3kuoIm(ALq%fj(BZ_-aVThSH@F&0`AL{<_HVtSJ5Pz}#w^=9Rg(ZAEV%yYIO zKIJx;vbh;I%rRXqjJd7?AW25YjQ8|*_q6wrh(gVV1tv%t(1DE*K7C#OapiXGm!XH` z1Y|2rN9ka?>?Q%x1^0t?Ec7qy8qky00pVCMZ4k9k>ba)nSypDkn~!Z^YxCj6(Ge)!j-{=irF9_Tcr8jeM;q?M3laG zDDFDd)GwYK@~3n1>YdI|yGNs*&ui-T;-<;jb!Hy`5Z18!bj}^ZVZZQ|ElKU#{meIV z{W4ZBwnPs3GuSwlk-6!sNrTJ!a`!!>L$lVK3I%@KVW#RriT;>D&v$g}d>!7Ab?HXI z#tS-(mVoM^PYvi1fi25k5&jwXedD@&-C_Cntb-#-nMLdn`t!1?iQc7tn^G{FydZ#p zYp_Y3$V=WG;+mUVbQt0X>r4fl8+)7gwck*kFRl`HgolMKv4!9Q#@Y>5{%XPh;uW~e zqJ%2`5vNj$vM|TN$2y8%mQjA7u(NXrnsKonS6yGfJ2wSCDK{429RGcQE7rkI9$lu{ zBeb2H0)g{k(>VZQR3?@=?ssx!Q{)EWEIh?^*COCfFQV|H-Gw>6Hyf`HkaY!4lJEB^ zuHUjFTCCQptW%jS7aGO)2_U(Ea^cq;Jq`A2e?*nJ`$HAZ2P?G(Ax{Bt-ycWk>Go4a zT(w!xH+-&}Dtf)Q*8uFSNZE`sB6mK}HB^#opT(iV~V7fKI^ z0ej?p;fh3iF&F@?xGolGeCu@!?S+UKU!tCEI~{E{X~kb08?!Me+DEMovyHc`4`+eP=;ACB3UJfo3@$Vwp;Euu}zmh z(Qi8)Pq-L;HKPTylJh#R?+9g{V>VOs3SL||uuKZA6|w;qK)%Auk-s+igs)8D>z*SJ zN)2M!3RZp&4UQ4GR;KPW+#3Cvko@5lu>fAt>wCp7&MyiH#gtei8LkWYNu%aJ3H%a& za$|sYZai(traGImu%RMQ@^87!=OEa7v8O<=z2dO9m^gr*I&h_&;%)iqR=Cx~;O?cP z>O#pRQ1BD)*Ax>MM<7toZzheoXV^nO9_&hMCBEBS7Xrs`>MSp+t<}e%j{>Ynw4Gi` z^f~e)J3E-}1<%{(5WMAULx9nEJb;ZZ#n%Essa5gbEGPAe`GF$W*?j!E&bU|lGLECVa zDJAPz+0oVIzaeh-?bi^=0QBjTEZAznVX7$Ea3PQnOqba+EW`j&20TVttB_xWzI>&e zyaku;R6~Rf*NNVL@sIf0Cf+Z`W<*6z-4|0dJFjxTy}7uS{rUiVaLCrM#M-*WMFftP zC52;Yr?l|R@%Hb#+R07jmo6^J~py7U@6{sORSl1MT@2~sg+jaeaWi@my0vaDp#Zj zXP3`z_U2$B!Z4oMw48_E(|iLIWD$|qyrdFbMC6&)wzlY`)s7 zWY3}U2s@|mgVL9IL7i{S`PMqTE)fkjElTReSE572p3!sO=k2VW@UYqfE04Dq&tE}P zIpQf9fS2%~(}$#GF=SjBrzIYXeBG=-whNJI6p&qL>aXx9y9N;7Jr*hpELpt!I86(Z z`z@j?EBe9nl3Y7_??Uf3E8ErSxIf*qNU&6P{ z>}`e_A7bDmIZUpH=HXQ)=3Wo~lIT_~_w|;_KPWqGs&3x=cD@j>7^-nBeMww=czNi# zB-z36OD0ML6DeTRFHM^+G72g)79@ttCJaZ@rI4U*YQHybiJ=C@T3Ox0y8_~E*%On6 zcLB!(z*eJo5^Ahqn!?<=VK)HWsX4)8Ex>aO7}OpG(3k*YotBD&gK31h^K9_-VH4Ba z(cHIw`rXk^U#=dv>~p#~9lz#m@1Uf?7MGRX9$)(&hzI0iG*T|kO)r-s!S>)opB|u| z@x+b5rV#>SxA`Up%^eZ5Z5G(9P2ADeZi@(AY&j76vR>R&%(I9y5sf$GRI)UyEG8&T zU|jvVthRg{C@4=QRZt}p(n?H&LO(}a@i~Jh^f|Va>wXo>VbBpCKi%=L@Cq$Ds5%v> zX+HmUzI3JASPvB3@@I*cJ%8FA^9bQw0n!Po>m64n0%HYKV2+hqGX38m0g&%nTGGnZ z@ zJzm?G?V8@j1Ol(g{V$~aa}M2d_PwQvmK$hj{JDTo#eoRFMIBGwe!^Yg&Y+8N-`jbdqYsuTBy+t&^{_dvZkXL7`!>pcLZyK5b)UewX7}*{y zz9}8(Y-Obf2q41GuGd-Rzi44W&M$8E%8>r^*cx!O8 z{=F)N2l0@qHvJNOB=hf3kTa2v1xA3ebs-Tz8ons_WH;r!0`No2UBUL}IhkG_YX=%* zwdIym_L$HL+kp?a410UWcFsw(odKz%AKZ|Ti*MuULf~;BveZZ+L8zc(>`j$Z;+(_1 z!PvRKf#HA-rPrPQ7oh5SygMCexJHcLQ^mDFuOPrF0b~VKx1uU1x{gydrr*HNb6cEg zOAq%P_O2}LIsz#qo?s()ueBo+qs<|mV?75!AjV=IZx%=UmKG#QQrn&&zcYq1M9kTp|mhxZh!9dz=eKk-|rSCWp1NU&yAF^n%4md3&r&&VTizOMV} zQ-Dw~MpT;szeGyixq5;F)O#g3LnxMh(3+fz!hHiH*9&QdIV$^`NC2F^0q`#_FVjXk zM=(_ZqSw1V`m_F{E0k9Mm_TUZ{$AE_Of3*mhoA-jEuDk+6cM0_RIT z1M5h`$X<$(qyWU&vM+$|`zP{uYwO+jsX_u^QYG(<%Oy;o2UPFPy7bAabgcr}{yKaH zc3=Gwo*YeVV#vZL1;f?s!HDp9dQwmRSzgnYnr*$TmYYS%Ecs+Z8!qTXiz0lyDpkFY zOiaSDfgLWFQ=EOA7ajG+C+QJ4N0eXxui0O1*egk*Yh~rVP-7>VkkDTAxE{to*Y*RH z#q*_gUhROrH4kfq)^9*)@dpp~&-i{q7?c16s&F=PRdtyk&;@UD(^6jCnemu)pbnLj zRn?9IC22;yyxVIG`#^QFG%!(;G6_PHn+(ARrBNfko6x-PS}6wS`%TsS zlG@sqY2*P{?5~zqH67NVX$*8i6tqJ}OD+z{P0uRfJLi4;YOd~^V$YLaJZGMv35va$ ztVnjVq}U>Tz^-#zh<;aR;%EEtiTJ!6P4@8M9N$pgaqfYNduAl^3Z2ckgOr?HB)~t? zE2b|W)T_>wJD{wW0V-CzL{}}J0c1IimKK%bNYLY}rPbl}nrMsNi*HP?d+2XNxqQrY z`->syu*+m=q?YEZ_cl1qKxgvMK@VJ&jrNJlu*UnAM1wxctSpn4=34bsG#xT8UL{|%A@YL#uWWb>ZR;pVJ$dtAd8$VYY?==k~rq+Dx8$%2{l*Y_=;CXSHGvxm#O@ z=zJrql8sEKpk$O`Pm_V0)Ys40Q>eVeu0n!<>@sV%9F773=>Ai!bJ!;lFa z9v^SyQ)!udM=okv>83zLXtapB={1tG&VT%lo;~+PVg(5+o~8VR8_9^n#wA(QKGEjS z1iRuW?DB=DVgx;)ZHjzpol_ZDBAHce_AQ8HzM&@~%Z*yVu`cH_{YMF0G0K-KVNCUs zg*0g=)Hq};l_=;rakw33wxx3AE!pHH2G5{j{s_8yh1?U=+tObX)WUK=7}KY;Ayg9NP@>s^?FO&mcoal#0&K_EN+yoG zXoi+V=9=-*1~w2{34b9gy!gQUxj!U5k0y=|Dckm?<@T92WZBNpdr3_8u0T(&7^A^i z{@3iz3PR&SBr2<1To|$fSpf4Uz;H_8bzGzcGUqoo@UMSfxlNPoCI!9i()!``3~+!{ z(?R~417z5ths_9qeIBB(!qQ(%L6mPdyri_l0^R{l{LG*J7z7PPCWv4b&wCcgP-r1y z3kAAmY~S3_oC4aWS|K92iz-8;f2W*)Dx~KkW#AN~71Uc+1Ei+|C|mz})#P+(Hi%7N z`V*{USh{0o2{twELFBX^)IwYLUkG2qX~xK|LeHHRu6ezNfGfGwWg~5I00}{reVVDC zjWjSG!5m!Q*?I^fz1QO&2C608i(bWlVoTE`kX#p$c=1#ZVe-enfay_UG=DWw*(et+ zLTAMuvF+96pN?{r*X8(Y9r`01x zoxC5RKULBhiiP}w3`paLrEWT48cyP{ON~eZBAOtVWQ)J=txX{<^So)9whcEHs{ef4 zS4R3{50Nu}CG5j!0VN6kvZq9FmM+OY#L4~=5j}tKf+$ zOA7!H2$Bku_4{c_{==tqD3g3^QMCA9n2Mi{<3D|RM+FQIBa27;dma6!zk8tvJpM>a zp!R%t}i+d7#gx@dwk8cEn3hH{m_p<+YqpzTF&pUxNl>g6N{J1GHDPTA}a)jA` zH~Ivqx**+p^x*#Eo&K4w_~nam7*L=?!lD}g??!I{+P#=p=V6S$&DTGF3nGJZnW?0* z|2rX*5j4E+J?-a)oC@3g) zsU;FLbcU|;wlxM%5ujv|H_|coDbNJph4i?m#P=(p;3wmRn%grp#Hm*s1RxbJlWfFG zw(L$(TjI7>d;}cq?WtyOqS%WwhGQ>*n*i-lEy2qhZmuCHs^&PXYX5ErewyUmr!br7 zs1*V}Z*p64l)~i|#dXufjgry3k(pp#dZ@14qTq-v6Y2WVC&}W<>#4Ogb1(a_UV*y( z>Wpso4pO7g*EB9+cpU4|p@Dtv$YwLLU2=XV|1C7SbGEJI&2g?}Gj( zCXxGnp5_j!qen`DaDc9lLY0D}tKkV}*Tng(WBeUGu<9hH>oe=?Y91j+6MKpmN> zgDEC{tV28y!&HM$UK>ULIdL0a`69Oz$Ph0Bol}M)yeQ^;uxq{lowoxzMgbBCcjRuB z$k`)fr0npjz#>Dr6R?G)i+Rd~6t(r1Rmn9GQQed*;P&X#_$S=SqCcTiYnb!YoHSbc z`uYP3cNUJ7o>!OJ@q!SQ_><#l`{ZhGOz!p1+x9>SFw48CexHbt^N#*?wRLQJ1JTY| z8PwDQC4c#lY|{S6E{MfHue7ByIXLEq4uHGG0_1>Wu+cWU z(CjH7BXYiN0!3XSX@H2Zi3>R9e&KBWFLw-sALbUW%QT3+6({NnDaG7$NDT024-yfT zE*^?rkhTep`wuFzm?hjlZxCEKvDgj;QgM%J1>%w!HEzirHkfU5ZtVBD&PtN?Qn)+M zoW8rq5rGR58w93NZ5jnspm0;-nX1$Hs|InIEXZ48YI4DaoK&D#DGbOkAY~wRHw9rOI`ez&0UW zdX_M|ZyWt3i}#9Z-qrw)&s~+c`hD4gfZpfYmvV9r*E7}D)BvBfz;fm!manVfu>Dxi z@H!;S5}U-r)5&M5Q^` z`RItKxTOlKJFdmx7&v^-JFwvG(916pXi4I_P~#bfH*#f3uq5#}kN1C@PtZrak>NE1 ziVP|Wii{QXF3kL7Xb90rogm(3){5dj^65#ijV{0;(m&#d~L}2-lL>6zD5I zLmOtNWh0xOF#Zf$f8Rzh@C*vVLW!Y4A^DAP_L=6C7@Af?neUT7APT@Yn59XG4-oD` zp7whZcphCQUE239IiT)wn(T9x8}GT}J6{F_RF6VqbN@);YPT=h`1Jp2vnaa3UV$6E zB@|GFfbfQ4D+RzYYFz=)x@vaGuy+iVje%?j0M49oYd}WGs#@BlVyU6?xI{8Fyl581 zB#O=~8&Ln9@fs$vEq}zj>6p6Admg+>+wYd$0zd-B`flgDwZee^nAepDc$nklL} zGUfpfUR6?fI6WY#a6A=b1+et=dJ+#j$p9KkjIKa^-%xUcm-96sOtCf_Ame!j#S?9N z+>BN~V`Ed<_t=t#;=}9Kvn>o~&xi#LI3Qj1Xo+j_w4{fBSsN}GCI_+qw0iQfM}~L4 zWb`O0I=Q4Bhsu(PV+?Wt(hD?6*;(94zJbu9iw0T1&?x|2pN}9};Bm4>TcBf@ZsQ4s z^O`R`QgGY24B}+#s542X@q5;y3qU(5H&eV4lI+a&xC}ZD3hr#tqqP~O6WFb{ihYjO zr<{|prOI6NJ{EF8+ZCEdZdu}CRk#@ z+J%RPcVX7+2&r!t&4s~+WNJSE_L~NXZu=zkl4m={Il2|jOv_$5lCX)eJjeg_E5r|x| zT0dGwZkXQ_mhI|2zO^LO_jqNKHP`+s!{>X4g~Toc;m2Do`^P7xP}+z8dtaI_WR&8j zbBse+f6vHo)*)>Jb<5=J3M?=wV1C#1jS&!&?K#)1kT;D%0^lBly%H=p$F1vv5o|f~ zOHn9xG=$!r7$!%)dTbTtYip4qkq)xRp@4Q+kHfPj8bFC3h5O*{Hc#+22pL)MJJime ze0o&#OrYrvc<-rky8NDD{Nc@N=O>K#mN)6LZV)g0^EE&55R2(F-SF9I=o?ivfq6%l z51C({`?qhkP09C|WZ7W~NOrC?eUrsprIll3HyF0o$!vr{ClF)NpV)fgVBQUaH-S$m zT@3OP^LaUrsOuU%T&iFqd!|XsXWhtzwz9{csq*l^ysV#vg~eW~`>ZoWv&y}Qy-=^D z!>&`l4eE`iB|s@WayE~0{zsziS6hQo4eORVFX_K`8aN&RAoxHoui@wb=8pD0*+=1= zYzK2f;1Gr{P?btsjhf)v&ME^y0SAv|WKRaASOC!Ude`@2+nW39Ad`8cCx1Th%O^df zH*cW8x8GjW-{0nP0w&f=wTg#OK9d3pmnEAoF==_dGLT@E@w62~t1aW{sti8}?vIf| zO#>(k$o{>Sbn~pf+g@$sLkk&2%lWC?t1 z{Y5$odR+exBcMk1)#O~wS64jnQ(?yH`@Oe-_kp8-RKCs%yp>Xp<+dN8(7%1SKYx+S z3~QK~+rgKf#SawcjYGE#(_gJ@6SCM>LZP_B_}k7oYO&=c?t(&3@71udSCUmKN}!TI zw|RNxyWOc?S58KN`(dw`uKO=VtG_4jpDvuM3r&_Cd7o@8h*JqEzVM^Dfg^1c!Ib~% zAEH^lOPd%?@gWB*D(`NlTrML6mX@h*J+&EzbOBD+1v0;&=2B;JUcAqW`rAzWwmQZ@jB`L_pg#{ksQs{w$Ej&phA15{bZYqZb>n5k76li$vI{2fE!$8 z5t{7SQKcaTfH1F!(ttf$cU+DN`1f5oEt%2G$$ZiPO+J`N9iYSm=%&yPvDiC}@`8AW zk+0*yNs(y(_?-W9Q6NRBk*)Zl3-z5Y(sNCiDrJ+RZiO zo7XDFP4~rsN=WpZA-n{aIwI5Py5Jbd2q{`F(!@=LjHz^(a2B6qQjNA$THV%vxa?1Z z3tEDvH2M2qu$esjp1S)$_CQ?R5&qGD*11CBiY#@2= zSi=qVh9z*#fT6?Z(TsQt3+RsJ0OCuZsA#A#6ft+W zxcrAxg+F36yCw1dZcKiEY}I06Ev@>@%*^_DT438#2{L~(RQa$Rd~d&if+pKLFc4LF z@(4RE!WvGMRW|IUvhur*?(S#@b!XAE8(@n^ErtmE|I82BCwO4fS$4rU8KR|06-SF5 ziec=x#2kP4$O=PGWBNkwd|##v2KSo%4K#d34A;~LrgOG7M|kvo8~Vh ztA97^he)Q|$N)*@i9Rpf$@9X7Ov z{A6*WT~H^r1}Ak7V9-E)SZS4#M$-R6akif@b=~*=JNDMgwj}1wLLBo8ejWZqqtLtqgBEa`ADq}wkIk8oib81@BOeg%>IGNVztWPrrikMezIEXV>D zg8lu|WN~_)u8tYgfZcL&%ay_ZQnB*R`yp_uc zF=Ynxy-x7`cwi|kJ``hwHEUzZ;!%$L&zHu?K{`4H7BRi2zF#=*1N1n_V2X<7> zGlomIdJqy$(y6VubB6U?Jp>z=enYj{#+dCD`Zw43)1wg%gBHYC>BfIpX&_%oFtUWG zL^Pq6u~S@>)<47nL(o8dwS>KUGsMEU^+xumqT~Nu{D0m(5(wu8>#ye!jz?hU1C%Ai zFx6}f;?JLbf2`EoPhBp;ml)p01E}S`;r~xYP>}$%HTFpV(Z9wlA(={}Ta7+EkIr2q z-L*Hm!Jf2DU9-d^lugHHp&)cA`P6O?phJfFxmwT81TzAOoD4%BP;UKQb^c$AiOdoD z;PW2c@jp^dircURV#`reAMzWts2Yyt$yw=T^dAc;v--TEONgfM?!k+8u?<;6(dN9N zxjeh3zWiR&1IRklbLnyo%n!X7#Owk1r~?C!Ad&Rp;o(?>gp5i`{DKoneA)sfYg3Er zy_r*sF=u5l)jY-FyjCmoD+I{N(woj^S=F*OoJjLy;4M^1p-mCGa%Z1S$(}*?Qw7uc(S`d zy~NFni6jXlh>Z3#zh>(TJNnRxLj*@W_T<&o>ktqBePS66=>c#~?NRt>0af{(NJNFf zXXLFox>22C+fWruRQU}All>Le=bjC=L7tn4f2{2(VZ%K_c}I6#d=$Sa_gowaLHS}) zWpm-lsn(ppGxOAhk<{d5rq?xh)QOzMk(HI~K?+h~01Z6r?P@l_uhI$n@F5pDwwL1q zd3Q0D0g4{hvjJ;xY#AS3nnZg^R!EwgCjinlJqu0r>oa;TV`OOu&3eY?&*;PS0U>!P z$q`&{C0Q5R9rFA7hF4jt07KRm0#iy=?wWVd>+W{?cS%uiF2+j$9>1Fix)IMpbgT}Y zhMt?!R~`0O2L0-hh|}y|Bsl=P4+Lbf5T5)CGa}3at0mt)D4|rcMfOp{#IUb`jcq>O zxFu|h8L#U3PLQMK)oWe%;B=GDo{dON+%WcCx7GQm%c z073zQ>XKDD&YLBHgOwN#xPWusP}X&U7kO^OP3)Uqt5qA#fp0G(PEJ$-KCq24ZC!Oo zm)J}5mme)>s^mCr7RZu$ouVaZ(lRpi=>ayVAw+L>chBrwW!Ys~nq%v$@r4BX#bXAWHHfhQjfbF zH39nFr#cVSplg-~pnHIAu#}l!{4gXn8QeFP=!+wg$)!~G;0Z>VVe(+6u#^CTor5;6%A~!eYjHR-42>P z4jK-wzm@VG4Adb}sRI_V!=3f7_~2As7>>R&)0wZwQ_s4F(#zqY(v}@s;RoD5(z}z% zP-Yk4-#*&ijy8>_=f@}P7{7XOWTW|%msVj`Bs^P0Cj#E`JtDSZMm+tiZBExw9M;3yLjZPp3^+?rtHvxbDt<>9HS~THf$!||)GhP= zLvR95`@#9_Cq(J^S4>2zv8ArtlQn(t54Kb};+8-5GlyC0@O2A7o?`pkEr4^n<^>y< z_vZo7ZZBz|#=~!6-WW%eXnH^`_c8C@jVu{ zm15C@z~O#K6Z$}KXMQyJ*8Io1n4p27B@}`E4`UN24s$*_?_#-bm<{zq@Kc zQmT857%c%q0}<6FF~Tzyrv%KDLX)>*6^qNiW~N`R9EV=`J!wo5a9jNI-lGi;jbN~` ztHN0o@rWzE1q~z|Wfkhy3qN6cO|1!zA1hGs*-nTd-#gnT6Q?og$mndDPD9W+7G;yd zNnj5X?-XBwT<^kpmUJ_2Z{tUVhm&h+ex_n#a*v7Ea#(h0o5OkBm~Y^TmgTO-zbf4tT2VaFt=G9_Ad5Uy02^By808D^f1 zYcgV|>2q;6>S1_YfCOc(q`k`~6zCBfdu7y$#U~bcW|gH`kDiJi-*{cd&t^;$>+X$;t5qyM0!UTjQRB}*f8q~)k)EpDPI z`-DF&@eu*Wm5WuSTZv|auDzVI3{%2h8$HlsnnR|Z)5&?y!X@(59@y^38JA>|h4#nD zl*pMM5IV|u=>cF;a-)^(51k#*4z6vv(w3$`{H!1C`2k&&KDJbPma9=p=BXt^J$s;W zmh0A4`RH(KNHN+Vf4bGFTI|IirnJra#c|%ohtN=I;gW|W_3LoU)5m}Z1GLexKMtbwBX)`p zrA_f3MO5*;=#8~|sJ!)n;E#%##`Lis_+Bb^&ERo)VkUwMjDn}^B50&OXVGZxwa-H= zF_?5x^WBc)WBVT64nRs~>Q^O+slP}(j$IAqyB&5Yo0t85QoE3Wt;|`C!P7*dH2tU8^F?@Me03at^{GoC9q=; zK2HbJ+)xW9>SP!}p3vyNfTMXyC4O%b^NNn8729|UUOn#91DePuPd7Rnw0+6b)l9D<#IW z&`owPTi_Sy0FH~16h;8a)H?rr67Miijy;d-JWMY(Y5{>}w{^=|@=yU5k@7)SQkT0^ z_f1D+H>>gbtSybga)4CfQEL|2RjwEN(a3*mHYe~F%*Y~>@>IKVAY|2W)pW7gs&L%? z(Qg3-NQ?4t{071PzWqL&@6JBikwd0~bE6_9L# zG|M%LFGB2MY44nB9QP|rimr^ip4Qzd_&AUn6N22Ko#>kClePbHmZhH@?mglGA#!D& zThn-B?=D$YaUr!Ke?l_T4kB*QmZB^{%a!Uq^#8PEHsAjApiGKAu5G;fEz^Ng_|iL! z%CY49vru;9!urD{)SMzjZPVu;_%PaevO#zT&zicxMMqPv1Nu6l4ltunPihh4Jzcgf z?cOP@Yy%2-H;cG`Atbdqeb7 z5qYwvpJJ1olTOmawb{I*M;TNm72kmI&NgeXTSbiLvdONrp4Xb$$G$LV=)*4h`heiU z(FV|mU?q#ua5x2amW@(Chu7KinMnwb(oD4(F6+$1v+RD!)VpD0!;8THlivQ4CQ3&2 z;xc>^{;t)n<-lI#FQL_~6e@yO+Sy)g<0RuE?4s|oD~P6BTzcZ_RYu*S@V}~+8TE5k z?3=xRM5K7eKO-BPH6o0{!~-1Sb-F2dpdozFa`FQre0UY83R^#ez*>BOg zI@^C6qacSjh`5Zi>4!AbOcJcRN#qjKIdKb++C~ja=_&)CFzGU*3u5)~ruPS)>`%@& z>TyYSDIgiTI$l9O3X4MJ$jBF*ivLu%djRS~x#-U}ROmhaIOR!QPz%TYVrcmp3fPK{ zuZ*Ol+WY|-4H6;$n|CyGntx;A8p6E=RWjDT|Y#MX)s@XC;(;W_#k~wyx*R7qAdHposxpv*{PcMl zwzD)kw#Hjs4!!^d)?1qCHp9%G=kXXQD9-&eRg6ZHJ4TZvQe|&fjomIrkDQr#<#KT1 zxE`G>U;5prn8tfBT!sV8Us6TUAW% z<1hn~iI&4LpY^fep(-0eZa;Nn68S}`5)C~TW@V>R>;9Op1bJGc*-oZuQl(00R$DX0 zqsl(T?BC1G!9WXdWW~$$fk!I*^!A5KFm@(D*--J7Va!JcfD1pxHya*ewt@S4+MAhfn@yrdD04tPoN5hX)cgrW1rv~cKRzY zg@L)x9$V*sxq_~HVHd8)5D}k|u_wNwAff@1`3zPijR<_!gjpeq_4xfmtlC&-B=m?6 zm8$-bk@{@oVu=w#ocFkO{`@19+ZfP3}B z^#xg$xiks1;(Sp{Eayr`t4oG#B6B>SJj&qS9Ei-)7~XgGS-25DUU--~vxhEp_ky`) zZOSM++T`2i02lt#A)#~j(JCYl)jprbD?QDv7^s4gZ?dZ+-!Lo`u02@?%kl&wXO7p*$1P8h889wc7 zi-%3^skO(|;-YIH|LwF_Mk{80)J{>&E#)6#cHiP6-)XWqK2*MXs=2hDu1M6bnA7;Hn=;c08XCs{R-pb6W;K00w)ArIva zWV|6j3PFNr@ppNwr!RWW(L*`Mul&_di|?#{Z(xcikdIsJds6j~otG0VS5p1q<(Br` zszks4@=3{7fuQH`Ku?s_Yc45u^=fwOF~FrAmlgKEq7-mI(Wi{s%w+a-e+!jXKJJW4(;S#G{2^DF{emVrZV8qD);O(Z!jFp zR|5^8RHt96Pr7lhMxE!MVC$t&CD@4uBZjJ00|$3uxsJgZp}>w5xZbJ3^Jub$Q&cQ; zU+1&cb)1(Yv&EEEsTBd&Pzd;&mu*!q!SP+7ict$Q@%n|5WF2}mIOT3n7TN~N@DOyn zcHhu!`T9X`I)J*%ZqW&Jf1)04xp;(`)R)F9xR4NObY1+YS5} zHl98*j4bVD;M6Xy@|h)Q*fZm!do?w}1sji&g38)~B89|hZ>3t*Vv!p*c=#0#?a=n& zgW8%ZgA2R)P_;P`)Up}1(>S!1=OCheEPC!vFYTIkHcQj_$>PE%^`yFO-IHH~58t3V zp`lkJk>)U56FeBrG-HeBQDGy@?76#ksp+RTeew~@C+K3}n3tLa66lnF8QI3)Skq4t z)f^KiR12g_s!tQ|aq+GJsE7)_=SWz-Y@V{00`FU~32#5dvnmiyJ-oDtNV<*^hZLA6@JlkfzUpHeqJvZlhy2#>jZ zghG8*ytkY0y^mWZRn`yC@AzDs7id;Dh<%#nty?zVUYgH5;#n#6TLXFAS%!YM|8nKs zS@EpVI`(1R=3V#N>Aif6$3bJB?ECH_p4oQ>#$0{nv-?@kC=vAF`AR{;0svzmBW3Z; zU)mKQe-u~3PJtvuxf@efWp44By=FvpF?w806fxzoj874B=3Z&UMg9-r#Vj!o(hx2k?_;VS%LmPgrFqVPPCLgRaz z-?Nj1{@3a>V{RFrNo~9;vYBK#gJ_KDgNd-8g)`7QyM^K$;o{Cx3n?pq>@;|X!9Ms! zXQ3TlIgSn$+a+Az5AB5is@o++lbO5|UJXzR>h{?2uO82mRxi}1432#o3srI7zY&PA zX?ENjc_pr*C8701w8D1FiNiwu4CF2GiJF{&FZ7jzBkA}f^q~mKhv+0^-SI$O=n0y= zZArWq*iQeH?<%+5C|~oP;q805Le3e6h%7_`{WjWWV?m8S(cFD_ZU#B+IKNC z*>a&U50F3Mp!%RcK>ua^zWC;|PK@%R1=6#t=FTj%pC$sy>e9ol@2U{V-Rh16`{#N}i7nDasFgCxT_2U@G%@ij4-VT3cv&gW4oQYF1 zjHj@wdQ?4(r4ZQ!gLXHUv2<6fGPY|^ysvU5Clh`tewA5eIr<5Ae}zQmyVpjNR*v}a z_G0<-Qk~EE^lh;QrPy@0Pt4pNo*#)>eP_kre+m!v_?ewBQ#wV4@wVT3iC@v058)Ei z@{hkm#N4tcbN{?=QLDg+)__LG;C^=SzQ?FXN%dRFn%CZL&`av|o9c>5> zd!*yKkignI7M)kG$O1YMXjc7^;|>$=od*k;H`7}Y2o&E=4YQXJsD#&e!~(7M=DlHG zB&xY+Ryh0dbKFt0h2Ef~@^)(5WV}?{z#JxG$5?JU7pjs&e=cK` z!L_SZZbVSaz-c{4@w$5B!%%_8`R#7_{(OC{R_VW*i2loq7d}P4F^!>Yx~gyK`FiKI zb{y=(q#$rzKIb@(dv-88Ku5X;==gaFtwt)`)2ayGX@{_o>i@BgPh)15ht;f_-5mOb z^yLU^siUUK{I`K*`JfXR?9D$)kS1J!m%DRKym9qo9Z9A6`g~W8%oI|*Y53UL!mXh! z_dEbm)iliy@s+9FPFHdxSKLa-MqUO8r^f0TX9)=p-z;Ux=H>e7JScY?i1$mq=2uNi zCA31b$qg(h1!`j6*8(kfWZ(0p3B#eyX%b9Gz|l}de>@4r`G9FQU^pS?xq&%KhM6N) z-}MT_B;RtZ%kiO&B?!oVVFVZy%4F1T5(vt>hRIwsq1cdGwKP8X01#pi%LJ(suE84+NspA0qT(Ak8h(0Rvx(d?onuNammJHv zT9J=l1Y&AU;W!CdOUsg}WwF`lOQ-3si-H1~b(+QLO6u8#+EwZ_3YSl3vSPvZ9bo$1&yVfy6X-5MML?muCsCw-rYJ51V7`Sg+JYm z%Ia6|U4pxgj-R@3P9Z;#y?b?awnx61>Yx?Bs`_iK4q8W3Ook;uN;eu(5bB3JUtln~ z*@3@22{yA|&ho*~Ia@vOxmM|MB=O$GzkVCQ$V|+LiuWuUpKY_JQSV7_RE=I3yG{Z} zA-9aSxzpl;z&lAVny+p2hMBfhu2vsljqVG~KA>p-t03EfGvbS&0NDrnFk;OzH`zE~ zGlf$GI8=j|kZFC}@H@byE-&BcG~MKR1Ihs19BFOzfYew^&$C!PyVIL0nUaje#(n)zRo)G+zEzH#U_54=|J=uL zbA4S}e#p6kNFu+|d|pCXAkP^2vnPc9o@uO1rN%tz`_{r<7Dm& z)+hDroC`eZNur>(9|7IY2ZDc5PWuypkoFB}cP84zkHYcA7a;uy{tPR_+`_Z%#n+u= zH7ger)D-`&${rF!bPW^Rt0EW;G_M=;b~UW&9l1QM#0LcHs?hc$go^#w+UJ@#MB?;}*4VGrGpn{r6oippEq3K9vK=$3B*Wr(yI zx%ZjQebxOo+@Qx=B{H5*O*bN=Y2q^a7Xa3=)RTS2<1HP~8`OCKLaoLA=0)s*El~WZ zQ*PW)IXF|gVzfCbmyYJyFdyf(6|u8ZD?NSM6iQEG&JC$ zP1F?a8a(pc=$oJoCawNqq2X{H8pSp}<);qbNtBwO%d?Ci8-SV}Cp~*S9Xqk%$6m2? zy$0vK7&&A50Ey4#vc@nimK-3Qhv+Yhz@eG=>0&? z<7)~V*#ygmoolZ7dS|=$ESgeq{frjkmDHSmpTWXuBmi7kDb$8J_J@$Buo}WmqW2?2 zCD&Oe%*FT7&QkZ)`h;6)AT)xLrN>Dl50ID=G$}d_qepIaB@+}RzsJ&}-fg1pChoOc z$t@)se?ejfEAhTZG`Zl!n|9c$tHdc;AjI4=EV6|a*m?0>e90Vckt6p7X^6#W8j4Gp zc$zr)V1EhKgg)>iJf3Qm4;Y`inTFr??R&eAxQ<~L_)$U`*b`|b*5gtFl=Gx$5hLCS z0%F0sU%u6`>1}mvNSAaY-QA6Zz|h^@-97JO>;2q&ySMv!etthbf6d6uthKHGx1}hyl4=v(W&S9&%P-F`$=Q5~ID;C8)7&2dJ!i5sTI}`tWR0kyR$1ewYOGS(wiB zE&J8J2_qLwmrnykgluQ;0~EAC44A+cttM zKYC(l$i0tT4H-HF{Y$*%> z)J$cC`If-Z9xaMqu+{3So9ymz+f@$eM1n!V>lsQ(zvTT;f<+GV+{rm?@_|C2YE4_b zsv4m$IlcC|9RGWNj6$BpF6tBBi&eLG$0878I!K=c#dATma09$-TTq{+>|C($xOB`5|F4*S5$e>3QF6w4*xZ}7p& z(Et8Md{cs6&r-VbTiY>?0^Fy`=Tt?%f2|mxO?nA6x^Co}9ru8Q;I)d=X zT2Xlf2U|U8JQ5bC`;!_aAcHSzzWFp1ml=Jh-4-5G<7Whn=tV+HLt?R5{ zJNYKZ+{woivgCR$${@pNK3LWB{PKkth6o`EUHxgrh599f)5V2|r#YEBt71W&S>DVT zTC1e1ll*>k&>F43c+{lXOv$I(Nek@{JNbiYA1Y>M!!R!3@b%f{z`Xc(EG9v-tQkM_ zg?a7nJ5HivLc~+R53mJ%;`+k+qRI%Q%qx%~{0r30p@S5Pt_E{MoHGW=RbG({H-f;PaM(&0e=sW6~r1JN`i5!b|k%$>a$^1?YiZSkv_Mi8QTcXS< zZKm}?C;+&5@U7gohQHRB?&{i$!E%nB6GeA}_|5@@95JX zM2MB>EXd>Xrl4Jv?_@o0O_PWUIvhTag2S+aK)8h0xB1VA%l$Q!tDU?OX>93~GJ^Ue z$qwH!&PMTQaLLmGUJips)WOZEacIdr4@cPVJE+8(AK-5eqJz#hih zV(sYR7s-z45X6*jlPaJemKsyq>Qm{S)*{IuBO`AMa+Q?_nYb%ybFrORuqZ&E$so5p z{*Gz>D=oZ&V3%Yv+5%G(AE6f#arajwc_Un0FhOMbi?IArHa_fLfm6TJF{JbFz4qwF zy9bE)tRSW-0s5#)t7|-&wK{egu8cDp&RVabcBc{hSW-bak}@wOk#!X{fNjiB{j3A1 zg~yy$6kn|NG*8HkZV8u`0nU>$yZwaBYEKL$E=xg_JwAKv{C#D3{*CX?XkVjsvJ`|T z)?daZ`+o{+bcxbq9`u>jpYraj+;U4;Zf=CK)zwVj6gP5f9#hpBA$m%*BM`(|&OSMZ z`j3*SG$Gf0wxtC{7itf5fgElqCtnxE64-!$?dRVr$`w?jWIqv;o zmb++o3V+4U19<`9xqa@UK*$J7Dg86GJXI=u#fpf}rjZLpgZG{7ejm zlfo|Yx?ktvP1p13lAfOUd&peL;Ac3D3z+0Yt@v*j^lGd;DC8mJ;Svw%zg0Se6HHAS z{o{9$of24YL#sLle3pA#PgN3*Q>u=d=EORl-Of+55g6m4m{<`salky6=0GXxU&KQz z9J>L0;J%jxK8KU*cr!2$r9bkWpLl6!a)f&hF<&!9nmbPIgA8XQsx5m}oC{~pI0jXu z)z5O5cFV?ecRx&Z#XwQ}bOxhKrJI>ZphCVB+xO#)^?1svC#seU=ainqwi7Rls!`GG3b3*hWAsc4>)S_W|ImInnjS=HT=_}QKaCQ*iaJT z$;t8^P8+O zFPiUAogGUT3FfHIlDPT?rwVcG0PTj%*kO~DMHO0kFdz;{39*Ed=pSwvN*+em&O-wP zZv8~=#Rc}z2;xzucJ>Uk0j47lg;W~D#{{3}sr%$cZhM`=Z#H`y=6kpX*<2nyN@KPS z5<0R9I#P3jxw&b`$`a)yd(H=y-@kGVOANJuCE?UZL@A>w@B*} z{fJkn9;?3zNptk8WfDPHVx&^~*6<3qEkp^DxPY18&7>WBp_f#V-xDWN%)^rGeoCud z+_gcAy=qt^aAm)`#;DSU_#XuAf8e@GI^Cl?r{w~*^@KR0WP!IR`7xyN+h_}GmwQtR zdNO?Aw3tRvFzYtaMwR7+K5NSr>(#}Di`i%=gEU8Tc8Myk!|JN|*F5Tt$?{SVS|?l9 z9`ep~w!5YOx(|7LazIN<>R@B5z*d6WcSp_2cmDVlc2q~;1{~@Q1MDv~Du+EjvajSz zN=X_Gp}j8Qo~`RX-d(QaIgG^IBwI z9c|QW<9&eP2=q0xa8Jf8?0Un86$g9r#@nt$bo~&2m-him(f-~8_<(O%g2ip7y{gQg z`AZc(yPCpLOxpwjoUSfL!sFHiHH4H@;P9B1YFSNYL5aKEHtCU~CszxDlpuQ&qQ6&^ z=+g!B@xu6zw=iLei4B!D0Y3BK3KyR5Aq1z{-u0eEBmMBUl&Cl2f`C-j2LL=0lx=s3 zWjXNC*J2g&;%^1xa;fCH#-tfObgTpyC8Z3nv%NvK5Pjl772{P{s7d0NIc}dLbqK~CcHkQ zn}j7_R}?hw{;h#WSZb$;USH~a_g8qHu#va_7Bp?7a_1N1T|q{E_yw|*;9=%C*cu#8Gb7k9p%~U< zirTCXLe4`^f1Yxi`X}5UOsMU9gF!Zp5Zk!tP9Uo@zic)zagEYN1)4c8?L-1!Uk+MaN zrBpWUPw@VNrZQZ=Zma*^*1?0cI%xBku>E#Y(gj7zCtA;@S+=@r{O?Q^9EnIknGddE zN7UsM*V7-+5EL0567o}V?n6u*g7HF8@B^e= z@`EKW)5SP2?iUxH+ekj}#Vq`?ycZHSquyrPUbfWEmVsz@{*ZoqkYG~ClVVVW6VGzb zd1SYC&k(7Q+*LJT9{e0vafeWM-3veGp^fBt8$J3zJO*z*#yC*>Tpt=j-BBB>sb)$; z&Yvn0Iv%jiRf3ovmkbS^8fX;^_`cm=c5-sU+6+xdz%Rbx*posX%}h)C?Dpx)pseTq zWs{o`9=~H;$cIqGW4WoMc%}OOJDw8>7g66`iPo{*Mo6RuB}= z?U?~sI$MRgm|Vk`Woi~a%#bM=|&sL-JObs=DP2~U=2FE@VJ%~n2xRUs`cH}=M(E^RFA~7ix+CxA9 zZ=wTn6Z;popK{rDDnubW??O=0tdQQWlU_)^N2W<^&kh?*bth{Y({uSjYkz&7@5W=wY&gDYl}0_?sWc-j zGE&JVe7*-*bG$T|RiEi@L}~OTCWJAdE7qe zmRJE-F6PipTK(8rR#6p9A{-EUhf;gAJ(oHC!j(HI-eCax z9bqjf)Ss3$clwm^?&)!IK5}I7qnPb`J|vMZ^4hYBfd~dGufqQ=tNx!OD*{tgMkcBb z;M#C-0|V5ikVAq9+dMNC%~65B4%j$bI}bD`ze z>4%e}SXaw62UlZe27M#|0Kq zay!B+99FqmFAOd%r?n;4f7mZ~xf*n-rmYh|k31F=6H{!rfHQoePOL0rn4FInQohdB z|6}}s24>WiBAY5t3PiMRawODkT(bqdTA+|LoTKA+&E=9do<`lpEN7qi<;%S5SM&25 z9PA1a##RAm=Lu5!5_*QOb||A;&k;U-T0C(i33_Z6Js+qb!>*L*bc=Q2eEEp&va@xG z0MyzJIUFg&w%!n0YnvfgYwhEPs<~Y{!GJLbK4$%^R%MP{vYb*RW-OpwACTNFp~upd z(_YFlt+2AwTe*5^E58Q=Y!sl+Dzzp}qVMYjMf?+UY&7y{{;1n56je$$)ha>S>P1A3 z9T$tnHwE(^cL8YQ?GE3Xe0N=i(QJ>_R1*wnIqRK|c-&}ySJ{X5JHX@}NJc-1JU7)Y zH`02#lVk-_$@6(jZRThlV6g>MMkaQ1#LWw}Y769V*1dja-~PK`|NJV%@vi$WA%V2^ z@KC>_m)9HKgYwDVHi4Wcg~w=nq8X+Ient45-bcx&)^-Pvm+?e?q!P-VD3!A$`Re5% zaRs&fwAJexQvy@f42oVT3i-8!Zr4KluHih_)D82nW8yf0EyRcRanPYDKn}|Fw2krV zWcp!YcX;qLRRVhjT4IMqRr`Iekm&kP{!!n*%5XmWP;`g;?mhA|wG2FW?+WI0G}MS; zBSw`b{mm-+i491==I zQ%v|A1pww@GMsp_R7Y=k<^O%M2oO`bIvJ}a_!5=mGK{Ez$C{;`-zhAoqVjUjoso%k zZ)2|a3yr#s+up+MC^w0YO z;%XBLvPMyd{qJnq&>*n<=t87&8%m@=1exh0^}ZLjYeM*S^y^Oq9(emp?qnE{z-D;yP^uFHP)s|}_ny@+(RMwWp=NVFUsxp^ErCuL8$5#e zy)oiV(wV!dv3+1s`de+XRN5jRH&_jS+q95J$@Czta{i1js+ox$Be4{rS?{wOY92?EP zqB!7lqt@QIA3GeW&sAYDLk6T0>J!F_J03+CKw>A&AvHxh3r{viN^55dH!E*JE(kv3 zRZbES00{xyNWZX+u5Au7J2D``SWw&@J?*lVl=i{?>war$N9FRe;I?|k8>~M+z~3LL zoSzr&m2EXzI42nB-#8I%QV1?cm)7WNTk1vbRq!NZ!~jQ1&`zrtPP{S^0BXO5JB+C8 zMUC4-6@jW#lmd6fB6XIPaJnbVZ5dJ}p1N4ul#l+mVf<+i?Cz>Xkj0uhpfSU$%Zxl^ zG1bQNgjrX}xPGXBOlvOT+LvqR;A;?SqT6w?$i&VWB#iLWC!yxc9J4&FQtt73{qEJ~ zotlaT>ESX{PYUz=0@!RE?>V5xWH zK%>nJ#>rj^w%BNyMiqXlV+cXUN)J$Kva(Bb2IL1wIm_^vw)^K95mDAj0M*=Zprhb&i zHxZl+4?FZ%ZFj43yN9c(q}jIKpA}5oAMNQMXYuZfcf$Ow|KcCI_QQw)+b^&36^;G& zLfI3^jt$MB$jibsLG!Q`UfoFDA^B6SxmCzZbAU%Z%6@rCn1C`6cTdx~b?Klpn>;WK znXaC4PbNk!o!2jQ@x9XyGNWK0Xnd`(EWHvMOH4+0j^X@?R zs+R1M!ifp$EG*a-)PFrF?xrS04={sakG2WiB2^0nJYU9J`>OVm)EUhH>1d%+rSXb_ z(@7;b8sd|Ocd$DlApltWh1bdxQQj*`j>LP4*RQbgLaSocEK2E3bb0Ti;Bc$|~ z<4sIEJ=D?h}AyO9Xy^D_M|$dih;qJb7hfWtmXeI ztV+K;&$pm(*JdYiW^wgBOrL7G6tMiIeXP}zxDIV|gx8fWEPCpFKs)AAFAqanLkWi) z2IoOHo&QR=@!a>0Bsa=9X3KH`O{WPJw!LlbPkMFZ4Z4Q>QBMZ+1pws;jfw1`KWRbz z&r$XZVL|d!pid`sJ}#LLQ(61$8OJ8ziPAimP3g8}ouKG?=wU^7@WxSxhs{c7Zk{^f z&9P$C&1cM_$7UMe1J8J3OAWnuk1FWi_)2F z2m6eZzk7|P*>s>cu|v8h!WkLR1#C1@5fq*tgB}XYYd~mPJa{*BO;c-bf2k&*bK2>lnYaHMSpxxQ zo+=%qjsP4-js3>U!iIwat_*-!rmUAzt5A6wlyKRbz%Ck0(IAa`$6xomN%_wop4{n; zzQ^_?(%iS;?0OTCg)VrzgLIbum}HKaIbjH|^!T z!s>gDWz2BWZCv5xb9&W@OUM0=o;po%=hi1~_sd>G5)C2fP=A=$FDoJ3Hd&$L^h2PA z#x_Ph5&#JDs+x=%GXlUOQvjS3iudH(UGD?iO`^OxGxZcf-5gsM$qPO~9C_Wh#SRtMoqxXuBX4+XQKhpTk4vmo|0eX+)ZbY%cj`Z_lR;vUx7lTE*$0T%{)f+V#!0G zaT0Qx|8OpnR@*ppBGVz*OH#RGXd54ZHTas=h08pryhBD2fDWCiZ$^2%jt5qFEb5os z*k<6JfNoCyR#aFNabCCy#wnD4)BZN`;mt!!84JRu_ZLKVe8Pmj^o6cJ!;Wgo2yk>F z=UDmuw}k_Vhsg>|LRC{+Rjn7tb-g1v_KRB4ip_0ff-MbuXe&YHCi@VF(rtP?U@Q|t zj6tUD9NQsqZ5Sk;K39SS!6F5Xr`8werO@wG>D{1nS8k`H9W+=z%uE76(#s~M_Tk?e zb$(5O14BGQ#4GFvPh86&uzrZEw%d61bRZ8?RTbNMEvb;srlN!3Qay&<;Ztdu&*@k? zSBw8u{6fd88fc1YszCQ0LL};xq7d+SC-X1~U5~3JZ}%Cnxot(AG)>w*pr|jw(boXqc`611G;D%vA~zzH}&%3cQeyPZ`&xS(x0 z0EMV6754%-!6X3MR-lnp0VNYyycOnij~CSOxRZwg~CaN2&Hz#SM&JO4y3FAtzd0jd5F$s zKuC)C#OPzocXD`-*POFWId8N|y@d27h}$WUuNW&40{}1?75(9Q%3zIRcb7Cx+8lr^ zrSn0j^QZcHxf>>N9N)wEo$9;^kLT;J*lluiIOTxgM)EoPN5rQ_8Gsf?ixY}DX@!&% z5Oj0johOD2=Q$D{-==uRDcjqDCgslKY|A#I#}^@ETM=jET-vMiKnXZ&*l`#9N+v5yp_p-P%ES9(RnFQ}vA&s?|$3#NO7s zUW`=I+nmY_9(}Z+FLF=PA%w64XDj%yuOl&nXVy9j&$$z9KY@wTZ?#Hb)&rCo81KWB zmiTBcl5-N}@o8{A`xQC+--Csh1_O|O0_`OZz6SX9MO#3lE1bvO^#MLMIGViz407La z;TQD2W|_HXY700g><`8rLL3-&Nc$2hdGDgOtGI16Ku=GKp&uHOp1oy`SRKh0+Mk2k@SjF{^-iAUT_L3hy~z5OH!j z?698FM4Mo~6U1(uwnc%#(*KqPEX9Y8T$J+LRN2sHU0A>6G$fMk287~W#9JSWIzc@t z^%MKQ+#H+wd5s~QwYq7QaXa>~2H`zc*-ZIhuhc4sQY%mXoOpKLw`kX1xg$#)Dylm= zR+F?!7oy*-Jkf1BTbB*l4T%YTQIpqUYS}=(mSC4-L>4Icg>bExMIv?Ew+7RPV6K=! z8E~noAhFEZ{gg~`ZFkA5bSkCq5mABpvf4N#vQY^3D+rO;CY?|V)-xF>7tO0vw%UJl zvv^UE)6$*I$u@AQO`Ql2={m=CcpvEg`GJkOY5RnA?YNXIz(2iuGn9Iysi!2i!CZS(8BJ;+!`EzTSh zps13Bk^ke9S#mIe&Qd4Nrnac@cz^!$jt4 zCqRTQ*sOb8^Pw_~MZyg>%&@!jr<*QBgf7`kPbk)`FZoi~09(w=kcrH?rk6vwvcK!R*BIC(dR$ZyM+5xQ~SXMUw+O4vWK^ z((M?nhlvT_1-SWg-d;bOyfza*-ljggpCuX^7WM=~R_?v_)@;OU-QNuc415?#a!g{q z!G;-Vgzf?C0ALM2lL`E+E!oL7U989`U%fuS1jrBIsIF^t{VWvzvK0O@!&!T<(#c+# zj7G%!w@#o&lL9(9U_-tX5Bu^sLN z`|&0VmCFQRj3NWSxB1Qdeb|qyetX(K2Y2Jvc;55u-{PRI2S!#x#dIOn6I%zLJ6t2k z3Mi(cV-fGF7M=|^8h+Q_=`y{jVm-{h*^BLlJtb9yr)6t-)ppFVSZY?ZwT08r_`#&h9U->_n5t4I?j!qEk_4ikBrL zc>R{E=wJOfAdsK+e|dlT5bxjSB#a#Weuhp=@G|n)Z76bD*3H>x>>myxQ4Ke4x6{7o zFX@p#z2!XtPTIrESpc$TKl~fn4){(%xeklxL%*fG%sPUtW6@r-e3woK zB)`Om1|I+M)Zop>`!Oq=pPlmwOnZNytNjQ_5kj_Hj=y-51M}WmyQcBx?umKvpY-Rw zcdPhar6Pael_+}(Z4^QxiIeQrN5?3qeKw~o*L06xB3rH0$f9u z3JCz#pN!V4GK1ctjxemUE-A2W;vY!+bB?md(G?Zcvr#@^eTAP=liLIrw@ z^Xq7ScuvVF4~^&_q^jdy%D{~MHY+Bp4i5wXPsGL|YeTPrlt-wmZtsr;!21c%w8(mm z4FtA--JAb?E1Gj}Z-zXgQVB&GxNr@=*!bibLR;VCr;Y>6fQ z+28E$@gHmct$dS=TakFz!0#0jz{~f60NPr^bSVp7ez?81CI0uf_um`kO%?`RldQ+~ z(m#5Yc+Fr2J=;zfD1PT>0htn-CYQ4j8ZDbOl91HZZMOfkul++(#g3CKD*{ z*#GpOzyAsl0{~AE&t>>)xvPyU37G_VF^OV{pj7Qf8EKfDE; zQ(%bnS-dv>FP{b|i~*KbIh)h?A9OAM+YOn-0I0C?ZSQXk2_7E&-@B*qWfUg=Zx=#< zm#6BlXl8-*&5=~s9=>M>5EL}3sxgO+#th4imhKcT#u#oGc;x)bVegVDCX1E!j_ZF4LGxBPcCX@ibLI?m%_XUJ_@c+#y|8X}t+D&Um z67SxnU}9lO6gwfREYgLBZ6*mPaO?WTfc(gXBqSn^fdsX)s!EFV+_3X&Ct!We zlQmqNVyYIK1l|lL`?aFs%Kqm%(^Gkhs~#}Zg;36 z0^pGTlAQ_rzs=Mux?~v4+M^>WvWwI%G=3SXWH33-NBMQSNPdivFvG-f1HH6~04V(1 zeF1reJDd;ytZw91)oPbNzjMR&Bu9zy(P&7GD2L4&O-NxTR0mC8FHHj4z}3y@qC1Fr z;nciHf3kqZ(-TVtP(*3iU8rn~*g%bZvi}v@0gM;;!GrAy1tKS>i{>7Tz!LhGL!4(4 z5)vfg{0KA&t2oi}Z{NOjdGJ8$qc|8vmXeY0)x?BqSj=~%4t#R0+`9sdxSGRMVR#+a z1qLOvoE*>(X-tyZg=;dhX?J-2jl~mEP_+$fJk82c)%Q2d>8~4TSPOi8&Bd(i(JKo_ z5gTm?(v!-%ypb0u9D`4bPDW;e%tRQ%y|OoM5*PCY(xpl_F9SaEBHZn-dy%!-b9+(Q z+LD!u0%$y)Ad1BbgP`dB!L<3}J21@Y-XvQUp)m;!lOioA^jOpOfmz6(dqcO z{4vu1wv3bc?|7?K?fE52GbX}jD02C4jPa5_A}Ruc_22Z)MI-PzjsW6naq$N+l#5GW z6+M8P5=Ga`dpCPE&v0SizO56!Ab=JM)YcHrzM7}S^7`??T?%sG6>Gt90>PcAz`tWZ z?!iVp=}#vIisv-&a&P^LS`y-Uv2Z2dZL6}Uw|LoRc>q>GvB|$OcWeon8Q*30IkxqF z(>yosuLJz6Wst2!uwRQKCF{Wi>Yi&b?mLMeZlV*^o&j78`rO9^>m`H$ltciSWD{BL zcgF)^jhVbbxNMTAJN*SXy`7Z*+o%yehK&tOO-}AaZS%Yq{H$d5h-ofCTW62&D|O@g zAn@B^cP{J`ue*FB&|wcRqE+vlfX8($OG!-~wnC+Zv}Grt&~H1Yo8dafAoCcYOYCQd?-; z=e1>g0xpvbwUH+r_Ke_1fVAm+0YFm1QzfG#7~+9)=Z}EZUcNraA6ss#JmvP1X&>gk z-SnR?f^&}Zxz0?Wfpb24I9e>n1ZI1<+C8nkii9+<115`xT_QsM=Q&e>(NyD?Fc7W* zt*Dbwz?k`kMHyp2!$`wCLC;4#_hr%=Z zH=~@*J_S06w~k*xQ5vVoo&~VQDl11SSTs|5>S3DOF8m(8XuwrXr=m1(gZ?Ej~4UKRs(B=l2i-PWKFR8NXR#&_wT z*Kdd@7#MB*AbqYU*=ohbe~^9WvE)C_s`{9@sJWg3J0(#nyl$^s%IQ;kyjk4{lt`IX zMwUKWSYKz62r)<+O~lknTpqJGx!G-wbB{~2F!Hvl_I?$(0`UhG|4;ztl|ml&CVH?o zN#2k!O)3|bvwi)_;~~U7_f-5YXBh$q6>4c1?F^6MQCCo z?akF`nGesEt8`Oi2sHo= zH&K5(~a4DiI|G+$ifGs9R{Xe%uEJl_TaB*oc!;iGw zKYC8Q0*?Xx283N(K$~9w8=&>q$jzl^Y%J_6jX%|$aK=eIck@~#h;7sAjvPl$qz5uX z3NX3o1Vig|ihC~Kpa(42Qqa;y_BCF&@0v6TaY}fzf_{o`6Jf)aGkxZdI!pf52L9zv zzN6pWfFg)HiSu^&n@-`*l!G{6dG4A8FZQMgyJB>u?-STC+wV@3hG$AfST@9wj%mE< zZMc*&@JFV*QLeGwCCXQ;D3SuoToQhguRle~bv!C9bp4Y|Ml$n*<+Z)g+Fh4n>7Csv z%dBHeU}40XO~jJi@rXv!#ERgHxT8Tk1fXvw7V4~93EgnZ@sUiF=JOqRO5AZZ0Rn?T zV0VXn&r>NE)^fd|!5(j{1b|+}&AN0AhmmqrOwn2@>{GmDhvu(=Hl}^L2~lHxuj|Vs zRkAzM2Iv*=k3mdx(F$nkE@3iM?|cJV65cu?5TOdby>gl;W?-3I2)T4K5NjWj#-_nx zO#!j*Olg&r(ChC^qR_YMJsCc_D~7lK0_OqBV9`fIeb#N|m6NJ7*B!abUog(xh?-+I z^_6Gzpo#XUt=deV7a>un++QE@xb=Qg_EuIjalag{d^Hj*eSg^g&e^UL0B|-dyY8G8 zo`hxoZt|Oh9}MGUzNL3TW2iA4rjO_bSP->AHYLYyRHnr7yY*GZ0nU;4>?uA6>6l=IF?lNI!6kgt#-r|8`P!r9(z2Q3WH?fy~QVLu!E+B_xEzRXFsf6w*bahEQzn zUkq1(G{yTaZdBHk-ePu6XTabsG@5Shja@lxJsBfiw|2&5w(uQSD%C7AADVa4^-T~sNY@&&$B&z@Nnl8gvL!d zJIEVzNBu)B6*toFfJu5h{P5+zzI)Dsh0O-uW zkk&e^b=L!8i#%Vf1`rGtjV!XYEH!qR=2Jnn6%{)@6Fd{jTn>AQGdri7Oj{m=^bXw7 z$%L=R0-j^SZjxr7VQ zGCqzd1=MF9Qg4D7jQxD@-9pkb_fUe)Aqa+@<_@AjjVWH3I`yj8@+Y}NdDqQ z{1f{qr4vdE3tx0-@@PA611vk?7h3P#X?f^lu8%D=PQiP{ozu&xJMyV%QdF7 z7&7B?^>^jshkYC*Mh87!$}2#JO-Il5b2BsMULJ>lfHyJty)D5(;D0US`tJLdEX zi$$vNW#3Hl$tFJ$!n&c+`jc4^fB|@*RC8R|RUkR&4&w>Nbv)f57aya=HxCg!kF$Z3 zeCo%eDL@+nR1cN!Kg(xaZtQk=N? zhd(`vaz5TDtmq>f9p6uruL%=8xjCivvb*5S&o~*;ztT59@?~$tz(bcI0=vvEOerG2 z8Ys3Uz|!s;ZnnSBP?H8Az|+B0Ju<|JhS=oh50Qgfep*2`$G0M#UboD4ox#hHh#I{GH7!;mP{6v9{AXRmPWYE0h$jnHzcmS0wk~;Bxrp9RXzD2^)d)WA6TQND5J*{_l9|i3E3W#$M zEbIqF^2s?%z-G_YmG;ora+HD;pleQH32ZjKF$|5q(c5)biJG;`?!v;% zG|F1`1s6Em>Yj)vbv|<-sKP*5$m!t%xps7|xmx8HDmorz=d(?V?X9TH^+Plq(G-c* z1na>MN_LWkEE+D^jKpOYI6CQlJ-_1JOn_+L`dBI%@USk*MyQ{}(lY`mw#>x8RXlTQKi<;ZER#5V-?DVm z(R+`#01-YospN4sc3$>^a>rMO7Y|g)!Ce(pn}f#Hqg2XanT{>(b7wbob_@*<7Oc9x z#&YiR>m#+98jpVjU>`7^yLB035tDr4)(B!|Z9*Z5N1C*VLGd_2Icv5rdSlz`3~_er zJDuh6J_a`!h+4U1St34>Onf<@m$C1zL)eXHn>cFyd{wB*b!dfi+E3jZZEJG+(Fwb) zMr9Iq!q_nK9(??$9p^JkXfhK1YZY0~sq@Lc5_^V+qsnk{i}-qJe9f0wc`DeXWm32# zrs^-#T`Hy9pE$-CL<#vv#T4eNF|%wjTBX#9dlzGBX9n#oLkn1CcHQR1Y_fe39zv~9 z(kN)KCaP^RtCzK%o%Q(c#lF*>I^8j-@E#s+;^`I-Y_7DNduLEGG#@JJhv>DAOE4OT z6M7t&VGf9i2ljySdkS(#uZ?RqCUas#Q>Ee}4`L7rYnXbo$d^nV7&*>#M-) zpdx?Xm8bVzI|{6|(Mo^j*}$(}qHiu(Q%7w+{fyqrcEZ=-w$bVgorgSj#aGf>jgxBlnIzibkB;))nq<6i;0&lKxfN6avm;`*Kk|~Z4a`nx+asCym5OL zuptbIL!g5sT~;2F(dbdrBHqtdg-qjIXdP3r5<-qUF2gCQ+G+%@=h&)^@4_FkU$72r zBd9Uzwgn^^#7>P?0>A?s1#OWT+a$#ukojz$uZGPE!t9L-v&mQiX)Nixjgp8GM_=;( zR_<9P67{w#LN_v-0G0=NscU0BSR#aWpRsv$Mz*%1w4p=_Jka(Yw-2M(p38xi8J#91 z^Ss^3;0|mFq+QJ%z16)Vy-mZpXbH_|lmN@v?*$HhyVmi=uF3Az8x=lj#i#9iEQd3-EL%BqC38%=5xsPicN(*}x!VaJ!4F!Fn|(kd{lsl?A=9YwQEFtwv=) z>rM)MAY_K|_9dZCsoB*x2g8LuGR5nz^rm{zcG7c)4m-@Auk-Zz^O$RF185(<2h<6w zueY`(dd`F2?R7K%T9m+2)Byq*>po8z(T&72zC5Qy#`))L8G8$dWyQw6*JI zY*#%mUQ7)DWa}>hGoW9i-t$AZ3{3xYpY!$J;$;zs+>q}opWREF0iQ`}p}wuiwx;a* z3cCjx_fODy64xADcnP~bPjvV&?U8OS`ibsk>iEebvIi;jdyM*Es3h-{8Cii5C0S`k z9tVPNADtsfJN7PL5EKN~2V0mH`6vt=48NI9<*~UDiO1>Z#M}EQ{q8NbcCp&%P5o?f zx&c>_A8qx*TlSqX>1_#T*76TIQ{_`fyfYioqpEyZr%SQTFA_=Z`r^<(2?=qxZca1W zv;YABX!AsY;IiE4yY-R+@mj1zhzd%XOz4~542b&j#u_<}Dz$aK0C%Bf-3j>ma0)Mp zeP+U}&NmhO+?kG1-LqchY{7+O%hO#4%x=uvRb#gWS&`qqnXVwOuXjN8l-VVu|J=ge zo5!U}Qnm|Ngs%Q(@^Q58q7?G)U}4_R>UZ?X5VzZoD_oeDN{!Fj8_9442cY9$0igS- z!I1J?G&SJyM-F+%6Vb%F(5tehlb;YqfqpFwzKU(4)ukLsazpM}bkr%tv&q?TDUn+5 zwIZ=v80Le6rCr!>b?y#PKzvU4_Trn1WT}-}y3kPuH{oW{=1w6{z9@AVyM;lwOrP@( z65~tGxhvL;>{5*V%o@cbOcPAEUofs(`JyHlt!|&*n^O~SHF?&Xd+K}qqG9Ty78jd< zn8ttzd=N4VU2vyH8_kzQTj(L3)WNOP2@{2TXdA?iw6nH{;}0Gx7Xg?W$2Q&FZLf_w zY#wU)l39J*09e9!aNMh)%r*t}T_4mLL|G#+kd99b60O*G-)0`5L%DA-kGeOi6f0Zr z3yX@2AJVSqjC3q1dSU=vxw@$O?nH%P6z|@FI^^-$aM6U zv=A5Ou^|dB=IaPe;dX@3nf5rW`gHXuX+WFHd;l|E?eXw$v7M6!Pr9$;BQC4Lk{i3VUT&@7fd%$7VzELz3s zLpoy#53OX+-Dsv!ukb}AQi&lzL&yljnb;pR5E}-sVAOE#v7HLkDmV2ZL$?du#^+e&B)Sp>?$@c z&T!O}>=C3who&v1`Mwkm%NG}d{j^1m!;S|XybsE&76%`{2ek-#($I24C99WaJ9=z@ z3>7m#k3YGEldE=J2+4bNU>;XI?F%wSF~2>`-Rp#~1RbB|y`$br(Y6qZLOgp4x! zxF*;ctJ)rvIxHT5AlrnHGuKxcX~U(Q0hUy!zvi#9SYgWn2!V?Q;Yvb(3JA%i z(oTTqt$fG1PWucR=dJ>w*;-Is;Tu>wPHdX%shjg7>ZQH2%A4H!EADhVg8Sfr$hq-NK|0X|`5A#$; zDUaUcF~G*=>Vp1hq+xErCkBjJQwZ~HXR5py)k%AXnvLY2|CzU;#t@`zQ+I7Y9XhgTMCk1C2IC-;@M zC;Bh69cQrziv`q^1pXg;Zy8qAx_%D}QX*2KQUVepE!`y{h*HuW($d`^-QA5M-Q9xH zEIOn?x}}@-K1=sGXCL==pWpR<{(oD_y4IZYnNQsJxW^bbjUv=&XO|^59hZOMRhsTK zcYbiywI@zu>M&8w!rb!{SiH?q923(XXJ%3-sotlv_z@xrX|%=n_$FsOUSNbI8;!$l zpnK}Dp*U-H6~@VwM-n({aveFhO8b;=W@tvq>CJn`rYo*3niR!~IGDuB(z4VMU7c%b z<|H}Ja}Y`4KIX3XTY@^Z!25T$I5)m~nLeLpAKl)|ZbxQWYoIJ4o4@fFajYCXoO&R0 zzeDr<6-IVK`6vY&dgq>NvC(-of1-r7p_fG&mvOzgxzykT07qWbR z$0I@xcMJbXdJjLs?Np_cAahI~N#mK+1kgMBK%-I@JkbyN=ZE?5u@*P=F0*&;vt&+iiO#quxgVf5x_5tCTq-SHd^5In zwt}C0EeKy{_wfk%vWt?4>}aEwVRNrt^))FxVEFU}3k4nIOp5!}8J=0f^^BBNeCd{$oq~RidhiAL{&~`F(stOk zv-e9E*Rtx}TrZ?fn&MZne2vyie#8b&pd31`JX%1O7+Mh*yzOE z2}=IW5X_u!@O`^g@_Rum`+scBBy4c5Id6HD#?n#gu6sBTdU!InSuea_fA=OjVBm&M zQ%D3CUrCw~T0ixHeyt-mvAdDYNR}(Hbau5XQQ}BpMRa^n2ca4hK)c|iKDWy(8ilYC z5!JnIFTDy4?{7LC>xE}*?c~Vl5jRZWG)ZmXTIaP^B~~MLGbFa2+Me21q~m1q;;^+GtIKz6dY$tw&T;O>ii66&>-<)0p{XEU*2C8?Q)b z2BhA?z=f&8uMQ>rdzJPOk}xAp7qU7YpT-gPcl0a-BBD-p!MHogmhW?NzIJj&kWR8i z?(b!5Is%`yn)D>BmqYq#v5|TRtZ&A09@V`Gbe)J0Yeyedh6$CXNalBnj| zb{3hu^`LN`$-l{{9P}7muh8T+8c0u-qEU}r>h9@h-8>v-I^JIzX6Wi3HCdis5X!R6 zUpW>EowB_XWgU`sJKIXDe@Fi)q?$(y|KUOL=UxHSLLQJcCD%#+rL9EP16;Qi%t1Z2 z`_NLgkGRBr9LpaTdrh1s9WYu{&KDcXrP%qxF=yvaF=lzWXrsy*E}Kgx2T?#S4a;Ae zy?Tj%)G0l!?w7z*x2A}a!{J=((w}KHvTUk9Pkgv`Ep89xyJs{<$1TXSvKGBfMzpBl z!$_*U^UO>9(u>d+BlKF(Rr!K~3wPyIH!Ko|EWb57?N!{Z;wG}h!HfYVgV!gOgjUj0 zcwDF6;c>i6(Tk0y)7SCD7rj0EtD5df65wPweBH2pw>CqNF|yHZ*Rb{7vTjYwdG<-&WZCKkT!@3qw?bMMiA zWDz?Zb&TJ=-S)Sj8ZmSFeKA$0DtJ8shrza;CFdq2ox}Z7EK^O;1p_|3JI6*dLpe}! z&=8g$nAYEe7-}=rhtLuvDrbJE5q2VuzpL;OF&{csI0b=>6{;%9&yE8z(_>l!k$OFH z59FkgM7J0O;dU!4;*hNZ58XjI8^h*eLfbEb`gWf7vlnv5CUXiK<5HKOckN71BC0N` z*lwb_rIibu6S>9%)#8^11oft8aWew~fqM+iRG)oQ$uI0TND>kfznIGuPnIQ%6z(&mXr+h8jPIUZmYcj~iM2eNlQ55eb$EvuLYlEznQFHYutFf9 zpb2r&VZi^+XqZkZlG`}jbmfWl_5I=zbL?@BJoA)Wr-AC~U-UHTOt%^iyH#h``_xOm zf~oi5(cQTeKG0_$QjlO@5Wb~ zUs;It`BLaAQZ(13o9%MFY|pfqov`xRI3*v6+we?x9hG@rFiviu7pvoldm4@7DMtJE zSLNii8f9wN*9KlL7oYoQEBSPBnF@-s(eX)iiH=3=JEhi1bMh!)_i5@zNhFMES85vyi5XyJIe z_n_DLviJ7g<&QFoVs>aPjCFDW@`tNSS&Fl4_Ml~)yyO~`;cR!YKWY4GjGMSiWHpV- zcuNkSC5Q4ku0$aM4KcnlplyEpE`)xl97&->9QvyUPg({sxs1bDRQk$Ur3ZbnF;Q_g zp7yS$JNrELj_aB@p5r>_5H$Y13ttW9z<`1_-7TGbox7gKF!omEzXS+3;4=_Wn_VbD ziq86MkGMSRd-Zzc*yfY`ZkN$44s4k!+MXDv?P-3?Et|f5H!;oGReL6?FfMg^-I3O* zp2d+z{hXHjWWs96Oo*%J$D%E+`1R5;Rw%Hw1#%DL4-`;GFh->a??ygTuP9}el z>g?N_(%k9b?-pLIF{~`A8p(3hnyNEbwQ*Q!6eYSceN;DIa3v9opVkSwbcRd_$NAK3 z1RnIJM^d=cK7b5p-@AhkG^$XvJ7%k;=j$e+3pOpBpM{wP*W&Rp(}4-wV=EXg8Z%Bw zpv*8<-*mQM7fSwFItXe>v*=|z{q~K^u*vwrE0=9`W(%R4#6>l4ylpx1#vw`7-G{qw zk-KiEZCi67{Kw;3LKs};E~-B#1z#+;oVJB7W=eDrI$1CbQV12uzJ0UuRtiR|CZy#Y5jHZup}FFwy$sh4OZo>^*{??g*L99?i_qCYa?>& z&DAH~-h9~O&`dFLoEl^=&4!ut1zw!k^|EUg`s`TE|B4@9Ncl}%b1(Y2XHRv5<%tYz8tb;e>A1KMVG>;+9x^}l*xL?1U}8z{9#+$-D#2b;3k^%B>hxeTX!uIAJq#FxMy--<2VCY97JwX zX(U{J?8nndiHou=&{eq`ysT2ln5@Ci%p`TmW*z$WX{>h1gFERQmI9g+)vsurEA5>p z_T~3MWvDx|t>^gqktt!Jdj3Z|=?n?xwVPMF2A9T223xc0BxBh$J9S4Y)dB6*$r(50x;&=I@-W7s|wX9r)-{WHV(f|BtqkX%j0dV~GAJqkmRh51;(xVOw zd1#K5rsh2A0~ooqyG_vCf-CEk6YS7+=NdGQ9mCxwvz!e54@wnp5|oSaI70dMwK&HN z^QRaZPV3H?mxrm^A3{D_t@{|XaiGrpB6`RR`UKh?K_pCPE3^~T%iKzqB59sW#O}?0 za^J02#%If6AS-T^2eh6gX5=f(_T>j*Pp@P7zM$z3oandnoEHEsP-C#`68$=`@a0_j z;C_Li^>s?PzA%B4-{~aFu-~g=j<#CKz!*h8&2{SV@;bBndy(U?_ zc+Gbh`hNH5%-!Spu`u!Jxbv~A(A~=^2joyIveZTSzIs9m!kK?%7fha+KFuOyEr#l4 zIvM1j3s(c2TTV1_9Q`jx7G0*zDj|qbGd7u&jXdF z-w|vnlVm1TzxELQ0C#&&Nx#i~hYJ~MGVGdCoa6(=mAsH_p)vDhmb~&BHJxo}BM=@e z6JcJr)(TC?cr=2Iu|2s%b}kQi(FgZwBQ)JoO!)}Md+#Szzdbs8-<`SuAhn_*K^!hA z^)a20y{8%-m4WoGmLpis<#I|#dRKXP)Z}IIJkKO$yeO$vZ}<$0tP>aE3um-oBA={3 z@&m)dp+_krq&}wG*3HVazD>2#Kf7E00M`@edHiR#98IZ>@vvP2a>)aWI;6iyiJvfd znyGVj8F@)XmqU5Kpk)qVlvPjJ6PuHWNLF8k^u~8|GeWsqPp7!sxFPSuIF!y0yCbO5 zTr*Jz>eQ_3w6UZWw9)$&=#`e*hKXtDpq3~@q_l0>>_7byZ1p@uK z;{=hbk;e!5zS?zbrN=GYCtPs9Rw@%vQoF{(DLCIrq1GqySzcr@zk_bkt)s2}xJ+$dxAJRYxogO~j`M_$mT53H&IJx=3Mn@}DMJ3d#7q8_KUxa&uaARsAAD^2#2 zN8cJ%pWGTAnjZB7{7#mNh0%5^{KPYWfIT0~6R}QXztwxwg~r~~lk)KHo#a8ZhY1O+ zOW$r^(``>xF){~uCnoE2)$Y~xguJCVfOrsYzDj?TO-P(a)A7qnz|4t+A6IkCIH^0l zdHdY{z(CWD#x;y#j~Qc2(EQogTjY+o!f6 z#_XxeH!>4%q+p790P4;y0X?ywC5XI00VA^|Jc<|97xakXPl9@#QZ+B%t2CTt$y_*!x*HwHkd7z5mGz#^ zqg`XS%CvzJ=lj98u~?^HYa@9{9{Nf=CSTJ_7`W-^UP3l0%2~L+MCUh;Xl_rWfk&xJ%~I7%1*|7({@~D#nvjKtz1OZ zJ>IJ=b$n{O#&(0qi%uGl9L$L`b89rw0l^t29VP1mlwvb$IX^8p(&Jj^9j|hsoC_`D zj>)k!IOsRI)w#zbO;i2JNKxL-g-+C>Bf6M5ZW0;msuD8(#W87OJSNsZ@_Tl!!i~It)Y{wgA5$ zlw%9tNaRULjP|3;^DNIB#|N9WmEK(LEN}=bpY$$YrL5217L@ z+Kqu8rkpFA(n^?UYL1Q;Dz%_H@m~gAt$v2aR~DNdHgP}{HxGNDZmON6%rHzfTKxJWz%#=wBda+yD{STE&v~Z!Ntyc-< zMXYj|A}svY*VZ~ep^5*N-;mS@$bSD!NGAlhBZ8-+c<>qvIOVzDsB!r@C7#`SgT})S zFl2`p!wQjZUp{$MCB~3ywBT@sf+P-m$7@?_747IdpTk@clWi-vHX?RSh;9mhBni56 z`^{=g-NJD_&Q`j;dJOwoQ07ugj$J{q4>j9 zI%>p|l4h&EfM(c?v4Fo;b9{lh!2^A(Mr<*Yaqi7kG{HhzEBTOtXdHi?7sK`;;onJ; zEbc#HnXMi>dCNTX8Z-(xPa0Tzvl=RyM{B(Cx!zclyN@2+MZ z&FZ#)f(4;|#R)0v?)b+)4*(R*uL()!WAZ0Zb0s(mIDus~nF_nw=pq$Tip}vCk$Vzy zvO#u{V||@m()ngznR`?39=*bNS;)K*i5>xgVe1y__u`YSi}pvS)qCiK z%c49TWZ|UUNdoM!MdJAdI9M!L7Up>31k`Q@|4#iNrv|64paVJwtRN(@yPsw6ai`;@cviG<}X{IfEoge1t}RBgwUj;uaj?- zIPR9*fgr7QBQip%+ySvKQ-cdURX1kR>t7R{e;)vU{G`D0-OldbXmZ`jx&~phRD*zx zZu@&N=IzFXE5^kq9`6K{VdZ;;q2%97A^zA_sIxrupLO;f^K{Nt@y2NdGSviuh;2+| z1o#K*dGt{=z~|1=vHcOA{;kOLj~^mLG#=>;9~zP+^Vf+x^Jp$95i=g`U4d(UuW%UC z)<%)p<7oI8m2N_rRd|~UB{~bl5PG?g+W+;-=|iY*^vA+MPSFK9qUFrMxF~00{vaw^ z_aQ3_OAc!}WpqBEwe{IVW&J16^XSJl-aW|_&Qg5|5R@-po*Voc3;cS)97*_dElsb- z`}+raCp*s$TAE{cGq^a(_f?pghJtV0i#-tgz^IE1It!E5U_AddsQQy)_>aLM_Kc2x zW33zpHg>8fPo9=h;fTvvmPW@$zsQx6uzPmPa3$cw61n&*aYlGKS&AP%uo(j!sTP3R zjxE`H>(|@+^#bYCsD=VQ^e5WnQb6w^a1UFo=-GJd?@!zRo`Z~FT43EJPWj)D0fsG| z8a9byb{>!aUUvUy4)OnwABWpn>782*aFbt=RZ+o@TYqr>zAvlfb&@%`aFAOfDKF|vNCJMe$d50lt=j(qjg@-_tPP3N*F>dngAxf_A0oa$N6B_Gtz zPcvuYu`Rf=Xz82h z_Xm?jXKacqX6%V!47_*8^>3C8*PRcwt#-a=2u`&v&RZ{PM(On^bbrQ}#t&N{?1(=- zuirmc-R<$)ZFs`U#6&nH{dYWcT&v?LwJ4n2~+%$K*%-!2RQ=IGKtqzHN z974dQk?+2|GlRH)U|znw6cu{Fzf|6=z4%zU&LvCD{-}-bXkqJ8Cz$`}U|lfDZsV&4 z;7U9)B!GIsI=8xynMZ@1>MK-Q=>jA*ah43tkJ2PBRBV#|4^u;$4=ly#(t3AHAN_O= zXtFkwN&gU$f``wM%2*2yjY1(~8ACy2O#XN$vFLr@Hi_5-&%50t6DUt{S-v}P)||kV zY5G+~CO6pY#RucB%f6xY%%I4gy54W^(OvdU+KnpjYkomgxB3|5=!yqQCFZy9UuhGK z1rxS(_c!;kVeHP;3YF-Gc8p!t_MxHRIkYl{c02&ZqNdZ03x9Yj8}ZQ1=W}H0^)|s= z;hBP0bykz}cb8+>ANPRL>89=`hYa=qVYOhxymp@{r65-#%P7I&D0p}t<$2yEE3O^w zeUZ_E!y9}9VKe?9XRpw8Dd53UwX${Ad#opeMcaH#BMvpq#z!jFoMD#z!Ir!U^v|{R zJ`9M$9NPE#uN7gK6`tdE;r>7)d)I>Bq1|>dvOXl$JhfV5hD&hj%8-`E>Wj83IUQ>q z@to^aMWh4+;f_P-$;X$hObRZG4!yZ$B{^bdwbsF~ZB*5)W+*p6SrOnY%xG{BA z6WO-69ZmM1w__6%i*f5ZMrnz7u^Hwgs$xyMOb7}NDbdOBN-rabHiT_XRzN&fj5mkvNUY;=Bc1nZW8wtLV!kvi{xTNLfK zlLAB_1COSnag@@~kArBW7j-ZisP4OZ=I#x!Qc#G9hT&5kZH~sg^+Q9&$JzT15bUdK zm6I@9^`;jeT^EwDbcVdMg*yhB3}%71K~J2es7F_H7(e^wbW;k~8Um-00Dtvo{l?Hz zmp0?ti>ghpsssKU40cV{*o6U-DCC{Nl=>?Af*tUKv7u7^;^DOG~}@oNeONtfYM;VOADrdg$(4?WixT^_n&v#RSQN3+pJ zD>dkTIz^MEy4ln0D^!3C5N0(a|DY$rj~M459u0N}r?=OdQ`(N*Z(I@lqXDlTwO0Rz z=10nvH+rywZ9-Jg;Zr6_Z|ZVVv<4Jb4UP_pU}QUrHVU{`w{Hgl7iM7_xI^Qq0TKx%Wd&rUdzudq?BS z)*;qy{FQx~Z4%4;)x0Fe`_oY;64mr6g#A)589Qda0O|FeAX~}9zYFw#9@XhwFvHHH z+6%Df4K`L8MB|#%N@&kvu=fTx3+>T-uzu&bgIORa_yp(b8$bPvjs4o`jXw!GJ{)1fHlK|)t z5uGZ_?afOT*l>M&6dsvOF4C5@ha1TF5AhcX5}a#)s;nzO4xfB>E|joQd#V#}2s(?t zv^1P+S!t5Y@Uwh*^GU@l_2lt8>;4iunnRQKzDE?GY-_hYO9vu=2J6wQo}xPOqJ}Ls zUCt=8*>f>=vzirIpe!yg&^X9$INNz{)z(Hv&Rp)h`7||)SLdSR?Pk?BD1eYoRquws z1A%UXUjF7;!A+pt=NmU4RV_Y&o=}NPN>TSYJQxRF-` zMzR2|B@2ZIb6-s8;-N4)ge$d``ToeZ>4a4_Pg{?H#f%fX~Y>t^HW&g zhO?MVa?BHg9ytP|cLy}PWrOGQb;xpp{wd;E1};#ecg1A*j+bo}Kq@6Ti~2F1)wxLu z2{q+Ke{xZ#goojAV4w<%#OF+ko;U4M%*!4-7ZSxl85&S}=ohO;qu_Xxip^}2wm-yK z9@Y!&^B!TlqAM z!h_GGc%8xw1={nYW}_%XpLV8HtC`r1E-M~R--VvUf5@S>6eH)T^yR%FAh#q!?Bwl!y;UxSHbX zB#;DJl#IW7 z85zK%sxq50LlRfIYITHkcSnE5Sf|r;XC}K82;AoWfCaiWejfrBR)h#PxtZ8xK9WWy zEL`YUIlte8VK`YYMNP=b8r3MwVuq=1RbebZ*||(g*hd;gj!bI8E=*jo6UA2`Z!btiTZZxeU&DugzbR=^c$nFn6RNk{kQ%T0H>c4> z4JV=YSMfMU@slU;WVeMgZ&Cy3e)%ccaH}JkWE}eTB(8)E3rHFn3pg3$IEB_n&l*NSX6zsBxcVK%tJKWU?2P=h0VBoEf7R;Hy zc82{Axa#R}>ksETC*WKcF@NC(Hg!u*FQnIN<6oce#0t|PTI~mPw6lzZWC-U@dnbt6 zoOf~Wl~lBAZe9cX(?|~rK^@LqBr~4lfo<9MzPtO6&d%z|$;f!M*7RQyNCd5<_F{6| zZAEAGCo`r`569&EIyMwR(tJVX!PH`8bJh>yFN{nQn^ha?;`&`Yw{R1|FCzRdNqxiHjf0w zSu|*mK`PJ~&4ehx>B@9Wt#)th0s?9e9UH9fFRq+yT_sOO*_oF%%?&zl>k=R4hS_H8 z1?j^HtwlaeU^PYvCKspyb@0(#K@?q5G~Y4@uUi3=YNL+{*1@xN#<{5YKdbY zCmPGQNeR3&tL8lwj9bRNFg#2kw-Lv9DAdzCaR8HLVY6(m`MQE6#%kHNIshr?rK@{p zZ9xYI+5G2b+H2i}tvSZ0G@nC3e#SyFnMr*iqK^YbwQf2E+%3|(Uq5s|JJ zjP5PijthWPw>qcafk#xh;m?}K_4Q-(*?P^Bx54+1i1Sp|ltZ%nQU{H`Nz4mo0h5ply;5365tD)O^ zTV(clJt2S^zCK$SMQEu+Tzgam?LIZ=5Q}OL5xxMKM~^3Oz4MeZ0qkfBSR^X9Yb?I~ zfG$RceQt9G!+tU4gpV<)>h0i2X%C>`h1BxWx;ib#<55)%26UXn069oEnN$N1OR*o( zLl^iz0Tq|t5cg|9U73pkN(#N>=AFEty0l)FgZJ3GeM6F#Gn~yMp}be&Qh-{tQwXh90n8o}MNN0V zm(ZQ(jvKXckWW6ocQ#aCH56AG>q~&ud9!I_tPtjAH*1%{24iS7T(4YYN*Ea`1mO5p zoLH|iMBXm05d=W&GCCQ7Pb1_zvtQ?42azd~$7Q*j*NF|jSmyY}mtltP`41!o|MCL( zl(E{y(0omAKVy}GN1GXfbcaV0owp`#+Ty^q%&?2eoF5{*H<+#2-6_!}x*(zrCIdEU zdh`7HzTkNe(LW8(GvFM)vEsamAmiPh=S$ZqCUeWY?%s!poZLBw&8z*Xf?8Q-ZVAA^ zc$2!)#BKIeQ?>H#V+OvH5*SC299i7-y(_Wnce~ov49Yi$rWlE?l$X1M9#keJKKjKL z)aqQjlxgm#iNlY~Q~HjkcHNJ6Hrp+)LUGhJzBD&nk0I|bF*77eI8HkrRKj##(8nwL zgn<4X4bX=6c&KT6c0&;46B^{>V1+mt*+u(|=)AvqEY_X&flq-F#dYF4&V~hRw#LnV zM2_vfN`rtw4&Nt)2ulUc;jSo^L4&L=4E2roH6KP?LQWiUe`THj>3Hd;2FBN~gHHyJ zV)S^ZH`QsA94TE7_bB0R=OY(Y?l?ZtO^d7!nPXy9+2z%@SsSlTn<&(`J54~`TFot( z0!WfP%lyJStj|jzweauJGxYU}7@x%-XH0oqWpnmIHzCM$<&o=?YE3^oeQBT9Qx+R6 zxMaM%5Jrnf$=5BDmuKU1nG&{!FD~TB8ePjjhKR(WT2ceOt0l)Gvhy?<-t}JTq zabys`C@2C}Bs+2?QGM zKEu6Y`QF0sEQmYvj!pbayr5L4APm9|25#CWgYa&(Pv=_NrQ0`s@n1$nzD1OGfT!zr(Mc|m zg*6VKGP@h~2kpPc9Ho(S%3j zi72Pg$)dddR`4&TM}Jz#2zLjx0+~8abIhYP3g%2R15*{K!8qBFE&3O~zG~Q&STONnHPjUM!`E zhmH2G7^yIm%uBSN4-OD(9zmQ-Y{dXGyw|@4fO3T4McUeBi4k7QWpc;H=VFUR;3|s# z{5r`F2>o8Vd%pkeNCHUa8v;)Y4(&k%$3h@K6KG5XC1Mo%`J}?*2D+RWZkYesA%Fe! zgaFP}N@Ff5TTTij_j#+Xt9d^#LF$AX2~Vge?&#M2X|VtMCSo)^PF85H=z`DTfw^d6 zXz0&xMgu3u(qv?rVm)1m@~@ZsPYJaIfccKDVX3lLF!rC@Nixk38)YyUiL3=R-kL+y zga2f#E-xl9_ujxKjM_%%a{~@MRjYR<-cP7~`H?he((3q+CgBahBz!k3hYgs7%O(BG zBwUUSj#&>7AY-Y)$cfHX55dKk)05)Kd!j)$~2Q0}0Nfwq<{q6(-XpuvI z)8($9avxaDWg{zUKgUa$L|W|`i3oxk%Wr3h0Esq$2-BWUqH;b3?{~lIDHZJgJU8gm zyPo))74GmOs{NrL`Rm&f5WyNa*%BvLm)Q~hwkk*>@-2k(=Fg$Rx{WD}#rTgV;W;pq z@aXDiFq3eff18BEh5av+@HUu9c%UsmHq0a(``;$v!c@}F+)AD-ul=S8IJkWQ9R!D6 zKE|SZmT~}X`lhZe`PXOqbI9<=U#00`mCf7Slz%Lonj7ma?=CEGe)l?Vc$VDYx$pPAi2i(6_g;c& zjcO+p|BKP!pTF~`HyND<#&Ox8KjHT!h(Fy>`a2l^p~Vbk_$NmgRCanblvGj}(oB9| z{sFGjBw&0HYi_;#zx8mT286=4c+sQeHo%1FqFwBv=0#sKHang^FdjYl!+hQrt8+bB zPMtnx^XJ+5x9i;#6=ubSH-~#C6&Jcsp1|pBe#6He_3mRwb_QI=-svVlduXs-t}J!7 z_XM?9?ie)V4T1rXsK_3^up)8jczp6#b?(#J(f@r%$x9U-(W*MMfDE-M>JhVRvm@N;@`wCG?e7tNE*zs6=vB%60Kla{i5zp`0@6n((uwxXPp094S$+lH#}=@8{QDn zaBBiO3fz;@{3#E7hfKI9h!i5SS_mJOR<@bSmb1kb^dH0B8{PKHfLCV#tm_6m$sPwF zNH9yAzjd)*n2zc?ntsdqhA-0V-b7waSoytC;#u{0fjXsm!<8rJ9IJcLZ4SlktH~~4 zp+TPJev|_AbS(2VjhSwZbweL4g`DL_rpBy@L`2$Fa9PXCmtcSd@QsiWRkI0!2_JO( zW-^J=ejc->x4`_Tw!R+|@0HBUu_y5~=IBgkID`G?#{Q>u_T$W!v-4E1w*yj$CiVki z1%2NUz$Y)G`mi_8-<_22N#*0?Q)XiCiIrY9g(gk)^qpK;Idhh1Nbr;ASp@Kdx5^X{N1spegeh$*VI;?}Q5Lt(d{xehU9DXTlly^&UL;#^_ez5oQ(ndsyf=52 z{^)+98(-+fJ{){M3~z$;o6O)=k)Yn4j|zR$mHv@wcIUI8GJa99{7;k8sdTa@DX>uFQ8@kn}NXO z1$!)UZt#z^*T}%YfV;}wz1wg0`~$fz9o%A1_J462@hAgR96}X2cdc*?^=3r8hzqP* z{I&IcZ<+89DVm65KwdCfAM@ZjLJDP&^}L;T0AvV3%d9aP5#u!Spu}y@-4!7-kA~|Y zq7#zDo44WEssr{3(O~iU&1ev85RL;m%y9O0Q{i%O24xSZC9!VS@65>nVQg#m9qu3R zIpE&Fr&|#042mHV)b01K1`M>3Mt6s7*xv-KxI9A6g8Ta^K{8LNrJwmUHf2CI+pcQQ zqxsx5d^(K$Q9Tt^S3B}Y6FfXVS-q)aHD(JD;_#>h(nr&18-Q5jbNeOj=rJ00DUyuR z!FP<01A0OQSpo%Ydrs;lYZ$fHADw#_Y|cW}H*l7ujL_*wn>l73=Y79HrybWJj>A^XLZua^9wS+_C9Y=g>0Ea>9ehFA7yy)&0~< ztwLU&omwX_n@5-=wTIw$0%ei~ui3*j>v=dUK#@_Z7u57iwarbooa|z<8Mq}lWr+#f zY0|e4Sxz5iTfaz1sVV4J)PK+(@b~*?dMnD(@@mk^!ss$j0rKTNeJgnfQe$;PTA!4- z=Bd+e_Bh$x*0`Q5c@hvyd~;#Gj61X-m`$y0ozj=O;j3Q?WXxMF_a@B%nA_87V$K?O z>Tj=rs{VL!L#A3H|2|}?`DBLlrWE&F=gI)6Ko#t!WJ23_tyf@1Fg9BQW6*H>W+3(z zKr?@j!VYv;l&aNcK_w3$KjYfGPw|7fOdHpvyN7*5$%G1QsOHGOBcXaCfOi=IiQR?_DTT48vbloP(eT7hZ#RMOwI}Kvpqq+m2HWZS;7izo+@6H%YKdfKClf~ zfe>&a7-TsB@I4r%j@I#Q@54yt_r9F(@18>?CaXjo;)MN;jK*vZVeo;M5;I^kYDp$Qi^V@y@&a!}GQVY<-`< zk4xlV0W$cf2J#>7{AJDc#zn{YFEdU^nTHPu%eM?#@$#d<+_G$%Qo!`qkFlUWCPBU9 z4H7hg3yPLScCG}c65uxkEG;g%9is#lo;+( zdS7)+m+2)HTsSwP5jTzB4~$QKCQzLSZ2u{Hw-VGU|4m^gq41bBEkQ-fo7q8v?(C+@^>CQsxORR_7=N0?ES;4~=c&3%nxJz7;${Y58~q@pK?aG?*+wn) z_!Jm&S@nuT?HUT))A+>th%M_*l!8`XEge-Wh#MK@l2cG590AJ_rlaz8*iv?wD*e&u zTKXCuA7QWCxh=E!nj%>%Q0C+T;$aF$tuR;tZ(V$dgBRR)9CoLQi%IEUWggsSE0*q9 z+-`ID+t(p?-*X7N-#LkPQ^w7{x2$?cteF2L-QJ) z_f#MsulUEsEinfKd>6cA-|=~4*d=lXy?}>*RMChe)T=B!Z*_c zg2VmgL*h5vY3}u`dt(@UXAVZAWeN6kO&JLkllT%Igmz+Dm|vEt(q33ul{MCG=_1G) zD%To^;E82i&AFTi;c?94ic5vf2{EQcmBA7vx!0arp6D531;~eH+lP;f^ot8=RINmY3pUe6)aCm9_IZ*H>c+-_$&Qmj*hHW=9GxZ%DNDrzri<6VmBcL{R<-bkcmh%@p|5YJwVRv ztY4nbb>c^gBbMyJXEQE<1fh=br$yp&%gJ1|qn=cZyIcAWKRNbBDKaic2$p!tJ^yr9 zo2jN<{?HgiA!Z%#*9FItriGj1o1^Kd*_$J7i2<#NQSRKfWiOPGWoq!&c#?O*?s4cq z*p+SsjQT46K5rxi?JP-UOF1fG#8D4{+a_y(1Fdb^4a&R?Q)9kYHS!O36(KhkDtdjd zc!rxVWZ2B6Nr5YYR&l8DMSt^k?9R-w{g$en8tuA~Cx2d}lT8${+xaWsWF~=B2d4%Y zM`FT0O(9hGClMRd9nPc{<>3UXlZA6QF|^!Sv;v-DYjKLWO7$R6?-u68SIH@-rhV$V zrS2`CR0DD?)n(!}18k2pfp9U6VBg{;s+h}&zW3$Og1={ha+%)q0oI)su{aKNI zE{(rkoWeAir#4)Z5o@bt^Egj3@Tlo=8D9d*#2X|rLxC@dtEmgy*~-)vY3|>>K{3R% zGUT{Au(4TQQK2{}jy4WZl-el5bqcY>U>=c#*{NyG5=_Q#5D2;VOT1uYm_#DLeRQF{ ziU0b?wt}e$mpu=6@!4_n4bVA!;U$ZxvhL;qj;Ezs7}J&N93nRnthhR$<3!E_zI~&p zThg~pW{;h+Yq=g)pV)8cjv8q?SKYc}Hi&i~-(M?b&e+7=mU~6Pe3ziuH08aj<+-wz ziNs$VV*c~od-wgJcYPfCORz(-J+XPM>k>UNuUjz3^prd?@X@qgl{>bI@hvidDwacG z-uuY7a+qz87VM5z(D>|Dq9GN#rC9C`?-_%`92#1&aM<`caA+?uw%7*101^^1~agG(Qf4cZkXG7yN5sHY&%#x367! zbuZ|zKA3|4RE0yo*8O6wxteD^lnHk~)~|7P-9=KFcNoBAo@m0U54o-OsjUO@lO^f-+rpZ!SD#;i7mn=NnPRBfaFhlQDKHxVFDNohV$>}GX@(R?<#|s@$Gbl@=C8NdG;O7K@^8h z^hwSB1x>cM({B}6f%6kIytWuhmZP4p@TnB!$wq+fRWMf~F3GnL@39DYc3@W{m;8i^ zTawnYE!a>w?MF|wpqboWT0OW3Ms19-keGS$Y#cXx{z(Sl2DC$5ReJ$ zKgt}RZhRe}U3mg6=S(><+pJ~rZ&g7O=C#o{=5n5M=!2NjW_vuyFTD3%d;1SyEE@EDsVy_aY?~_P+jMgI>ffUtCSo=3$3Xdf zD7gv+aKu(AF=O=4@0n*F7vMF0Qj>n>EyHO%F${Cgjr)@Lt-AN$p7p^ z{zXP0E_|T(*0}(W+!&&(Sh41$M$l_%pQ(|8rm@dQ(w?`Rcp__SQ`o1Ssy#w2E;`<) zXq3O6_oh$7dgbA5o0hpwuydqzGF;e5gkf7eb1wV0+&KZg`vWiU0&YNLTpR<87&=km z7?jL?S>aDlHM;N_B6-vBlSui;)PlWcNTLbbptb$Z*V$=}@?(YrAxPbDz8|IPWnJ@l z(!>e`30U+8!iZ2F1XE<$LH3Se5#!U;y2OoBN3{C$K0;%t&VCk?U#BQ_oY(hbtj1l# zBm@OLdG_YGv$Va=XNT8hAF9~*sO!EtMO5s8ne`(dZ|(3B87~6!Yl+QR2xz<`!Xof3 zsJQ9DRfl@@oXe^Z%qXe?DEfDw*4sbmbT_NjypQKea96`YF4f8H9naBpkCjCl5%3lh zmoExgWl$rXgPziO;%IQfG6T=k+rsDRGqW0AREOHNd==H>^$?~Zzm0|9b-N;^O)ouF zf7>mK^Zd=6i=c*AWTgWa+K6>Z2{a8iDGY_mL%+u1i4Pdz-v2wa#`DFa^nRocer8d# znlnCSQv@a+{%ck;q09TV$9X~IoL%J&CAN=%Io)&7WXHTp8mRNxoyD7-DrT5{#jwrhX|Xf*#y!D_zCO<`@r29W*ECQS*glZ zF{cO%Q^~i_(%vB{PP<$PG4-QFIvnp^d~mI2vudLtd-KI-yps}e67PoXn%p6Sxj){l z+ufZp_}K&g8_JrO>t;*s(SU}5q5{UwB1MKEvZUbM@bA_7dMMa-^c& zOU>dQVV}-=#p6krN}*K1NGOG=n>F*?Z|KZ+_d`$E)}8BtxA@n}Y&tx2zq)z+4BB&8 z;9=?*gVdV-PE_|OyqBe;Rl$CI&E>I2o$F)Dbz9A8Cv_X6g^M%j20}MU4{!FU9M-1X zzmx26Hd+la&6m7M8_Kyp$7j2OK8Kl;2_wm^wHwcW0jUVDCyuCqPpqLARlzcU<~A{e zf*yEF&oXCr&ARRCu7Axa>+o{S5xiV*^e-+Sp+yWp=kS;R zjEZ!NE!2HLprmS#NB?9r-P;1(JEtL3_+Cm_qv3#WbJS~(z($Oz|D4Ew`XpxycPG$c zKcYSugU6jDPLroKDdgLxx%;7pQ&@7*?6w#ipd(=pH0ayKDk~}}E!`ObExZbx^1UjH zl{cwSlCpT{ELB!J8GQXLw)j5R1zX;3 zr#B`y&7BO`sn!BkUM-%0<#^#Za5BSf<$onz!U@a?#YFKEX!>}raPWXLXzkWUk}x!b z|GMV=^5G*&@u>DYkR1S2?gvJo8`eZH}b z-S+IP%&|wI_L$$kOTK+CCVwo^FO+2nPv*yfc1wP8!X@fsBD}^JzB0&d`Tw|k%dn`n z{sHuepa?3Ugo5-@Vn``zB$bd3X+h}$k#0vpN*GW|kd&?g=@>vl5TvC`=^+IMh8nng zz~c#rbKd)Z?x)MAc{Y2mz1FXLlkIYOeV+gBZ4fv;dN3E{3^*&lGfI3Ah)nQ#?03d@ zP$Biyg{_(`@S^*&le5cMX-D2EmV3*~UQQ$9&0b&Hz^O!@zhE&xTJejbU$Z@0_BGGHsvaX`Jlj1{_tDaUZC{6Q zns&{;UsH$oV+2v}Ra)VU?_!b{QJq>Yob0BiMTr?7QPvdui)Fp{P=nk`ioi67twe(+ zQLUow_K;cyJGUv8w6b==n%&{VG$QZm*M1JoG+qyd9=Cft896go9rqtg- zEQ17QSh}P_J(qY8EXuAn?&vJ0j6HLZDqH&KVuDAi0Q^AH4=h|0pLfx|yNu(moY>Ue zrW=e`*7oA;?sj_jzYd;jCGy^@itBXa|I;$V5kn-b1?Y9!d zb92v|@C{lq{adJY8>g3jIog`JtH4M7aPf<^=xS0qC+btkK=RYe{RHZbTa+#^3+_Gr zy!k4UWO=T*&D8O*(C@&*J^#tRWAR&~-9S`Mvx; zWf2F%b&vIdeNV;@(CU0`V0~gbiKo9|Ui9M!n(lEMt(}6M$H3X1&&W|SgT$@rhTd=f zqO4xGMxHPpyn142xHUP|)6y~;`DjYGX`}X_T{2RqT~yw)(|NqYoHUfPdVaFLpW090 z!v0jQ@>{(!ty~?S(T&htJ!L!``0sRgx*Tf{vNW?sU#)kaA*cDDI8lB^GGvbW!qCX6hCO&8C_C4%un!2HzTD$M? zO4ahK@r^Bkho^%0ENyIZky~!fBG*0BR^WD~P zvg<(0Yo9EoQKlSiWX)GtpG&beBEbq>Jn3m!PXFu__ttjcW0{i2%{KAiy!o0g*0Op7`F0mz>a#l=Ua)XlcX z_4*ukl5M>3{yxm-=P>1x$4_E6=}{){bALw_>S-ag{zmzp(ZPaWlXFxElmfb1F}H%~ z5&~5kts99m_rtUWSY_h~Al{dWQ|G&0%EmBwn>Yon_uXK0!aZmMj#LkZmivMrH#9($Z9sHRoYlXC^ub>Ci2KssQ?H;AKT@$VLdYD+JvQ@BB9T> z3A;zkeN=afT@HHRZj_Agqp1vE<8hj47c;|9Xj=VnF|QT3nT+67JNWVz5<=wje!B+w z)z>tE@s29NWcezt?qQzn1WCy%gn^5mw_D7J{W5Ss2^ai`{99qr%pQWvJ0?5da zOpNR)jYT`C4MR<_90$$(cn=Vi__ugBzIaGa3gi|zTETW$*(rkOIsl8)YR=`sHuh~_K=;Iw+prJ z+kIBsQ!c-1hujv0C8C{JwRAIQ!pgM4)jON}C1g_`^=lz%?HghWMjLrj@|`79<^9+A ztolOjEnj2VxB$VmC*M!8cpspm1yX6z+rzxT?1g)VE|QcTrRae752ijMVH5`Xw`x|$ zQTK@4zMm#+-aN4HwevB81uf#9k(UTFNlG7SIHvM`!fB)JXn*`k!gZ^h`;8^rmdN1s zYUp}*XQ?w>t`8A_uC@Vq6s&>4qLd%<**`)cjs7Y-B4WK0z&H-qE zMIH;FY62sX332762zQ+6-0F-A8-gRRXGhMPw%5FCTE3|}=2GIl7!TECC12kUzJEJq zJjP0J78T|HV0|OclbxuuQHyD3i>5GJ&Mk14u#nBJ*#iq*!r2PIBN{@BuM?FP4Uzl2 zH*db8St}^9Y(F59*EV!=av#_ed*wXjmcgcB@3q^ybExI|*=xlm5g8-ye;Mc=8RTAxKR-59s!~$O7;Bb@w7Kftxkn*_U60TqzO0qe_#4 zKQ$yacb0kg(?rn^&H9t`fgd_A>pq6W^-NSFkjSxBqo|8*^&~`|bz+vTuD9S*>^_l( z8K3U;@HlxPf0M6>SxQcG(p~}v<1Tmxl^up#< zUhg9ibiKow$suZtMiaRi?|;sh=)UtbMUU2NYbYC*+MPGG^3gtigHzp4Q?An9=A-oA z^#3tdF5@H$%>_5v=2z}kN`^7qTPknZV!#K^2XOzAQ2pl+%GtzF&%cW+BtQGogxQthb#c%AuQae3Y+BO#8=rLMFm@DEOBG6yF zvfxQmH38YZX}xC}<3qUvpJX%#vSupuN&?E9=^0&`-b=RrhsmZ6w0|*> zery|(Mp{0vH}R$yZB&EALuAg0h@z)s6VvTH7hfSagAYWTqG;WQR&S$}N9@Wf3((DZ zWn&5VJm${!WONF0Qlm%BQ=7$Cx{9AP1xfZH5LIPE@o(OIvQs|XY3C$^T2qTGd^e}y zXofeHzIf5|cMuUcz#S)Ds(lOoCbV! zm7z(Ev0ihBR;rZc%(Nyu6pJ$h?-JKJQB(Rfq?t$=&)l$Zqt1~W{?dK-TDTTLrUqe! z>gBgM4^LdJBk?AlTCE1%EKq=|skQ2$jQuhzC+ff}#7LQCDy zI~Yb?)7`n9YhwSUC@~?_^W*CtnIV2Uk1>zVVw)k?Ze2T|8S!J!9TA zxeDbHVwo=>*hs5AZ;zJSVf*L1xa$2&9xmB6;Rj!im-`#jciG`ot-Mzkn$0xl+t)EQ zHAQ483)$R5udNjduEOahlHLPbc|>7;Bz1p-P~sGtNO)S(>Nzx0Oc$+EY{|GBT=yw< zS}Kx#&z~^|rcyIy-r$IY=INFjkB6{}keKS+VIiXWV$JB&k8jJ?vD;dk@==Q;vebcPxtaS{7C*Y8lPfW!oJBE`z4v1mH53-pgv;_dQX0C7!kgTmzML#vJuUncnxt5R?*{0}z@mN>ffHZkm z(CFRM(7T_a8^GWbQBDK#m)A37a|Q+b7k5BVXibrzfj8u|ilM5~D!@%7+15G6;U!_E9|r z9gC!>P-|?kn8uN6kg_y$zoDe0lo1~OZNGTnB7v`0L=2&y1QW)nPtGMeIsn&K9Jd$- zA8-`oirzfDeJf4umm_1gERQEEC)?AH00VBsO+qy`5PZ5^3g-ds^;t>!u5+itv+gl3 z9*Y@bi#tg@0R{zUVi!c_u<`?0yX7wD|(1;+mNENFtjtK9@&ZVU&2 zMeD`nKl@20UB=N$o%ek&cFg*aSe(S2)};eyw=`Efry%_X5Vl<^uyEbH`8JB33UvNk z$0k+O6Xw%HF19od#16^C^S)2TE#p=C&mKK^g$rUPN!AT>m6fh&)bPG^YYcn~hUH zh;!Sn3f;lW$LBIKDUT8ycqC3Jne-axUHp=w(J?tn91=?7Gt^&L8=lX; z)J)C&s*jchmy;6C6s>>9X9sSGmA$%N)-^V+^YZdW@Y}?UJz*$~*WoT?V>jYY`1tW- zZfw_iv%6#X%8K{y0Ue&DDI;>-Vx-6dRwMFO~2tq)iSCiQ`XW*SkSzv@e$=X&GPl{*mc@i6&>7yvu!RMZm)C zL0QV!pz{Z|>m&$YzC54Nm!oB7w56Ppr&nf{!(yW_lAaF4ffQq}TD+O171K(%LP;b4 zj*;=c#3SAiC{%Cgotvynv5rJoN>frSBR2Z}0}sYrxC|;JBVG@KnV5l)NO;|2TG3!R zEL!28i?Ct*7*$W(8^{09WqCZqM-I16%P7U73Xg1U$SgoGu>uaq^i9`w143&cJeD5C zr*I+wEC6Yh!p&6jgT*f5Uvp`^bJo2Z?+@_&ag3iUBvk^+$%#t)q)q&M5ONBL*<2|& zhbqU0zP~vK5*P=0fj|Fm7k_y(W(%ip03M0)NKDHv32zmL4W+Hoq!BSsD!5(?0hJ|0{(NmQ;EC42R2$0Q`=x~FQ{oO988(} zBtPD5oPx?rk`+k;v)J#Jjk|jco1zRr;ev9&>1#*}7-Wp@<@jS@^(!?%Xh^@m@gll6 z%<#g-8Y=UZ&%qj(nZ|yQghldFi6m(nSaz&lbO;{LMkQb?zY}ydeyL2$DDl?bVd|9O z{z9>3@2qa0J|K-GjLtQ3K4NS*&>Yi#z3>1#p(f4YP8PY}4oZkCmAL?5I(X80X;!;> zc6HX5>7wq&*ZPn?9J0boGKg$4lFR@Ut^1*6h*NosZuI^DQIFoe!)C{&Zn;z^CYsl> zva&(!hKIpMX`NncIZdecYiv}s$}q1Ph1uePd32}u?CT!E#JwlGyR~Q!_pPFY(RuI! z5FbsN!+TL=)RCT^;c*}+f(0H&5SrJ7Z08zYWjn<){&6Q%liEZku%}Mi_jd+j`>(O+ zAW@3a?=2m3*NK!GcdDhS{7T>Rj~FCuZ_8zrRj-vG-gf$2k6#zaNZX|1C$GBJSx+py z>Iiic_t|csC8YP4A6R$D(T2&Zs}t_sd|YPWa<|iS_;QcA7%D9{Bx}<38X%IACh*?N zu;j{jufAT|IRTc|cxKw~=7+ys)rrG>!LhyT$1?-&{2Y-;DL*q8v|7aJDzEUC)3A+0 zWTl{)j02`CsHzL53uW96i`Ma*lwrCSc~QvNnJ;KxmtK19**n5MdF}k|uDyGp{t)nO z6V0G%Pt=IfC!&#EAf+YCeP-m3Spee7uv;@qh#FLdi~C{K_*<<706R4AruauJP<-O&SeD7w2kzIIG-cdF@XROA(Q0_I2_NE5w8GSb<`SJ_c?`} zga(~z37a`?e&BoP=wLg0gzjL$11a@*n10236ErO0uw-)Q`Z|ui(p9kI4!ENOztMaHt;X?vZ&+Nl|cnmgQTW^$EPJJADv^}-&G^j^T>uzcB}b}w~X6$7*+ zZpgQ9Y>MG8iny8_O$9P#un8{=v+}YNgH$nF)Bwyr%8P2!W4AfcMdWv*KR>ioN%3q7 zT&JJSQHn_D)J_XMWd?kdKV7;-;bFNVEi0+13I-Nba8|ZtkZuDt9gw9*5?<@yQ|j7a2vmn?JRBC zqS|W{ci`^J$lc8f1zhBAuj7a9jXq|R;81yx?DT-bX_CGd@peRN?xMq=%eDgsw4z3A zYZ>#5Y9HKty+#`W3^IDuQ6WX^b|)yY8e_`F+3lFR_kG$ZLu_|LVM-J}a~^Jeu`o<< zvX|3t+o@e>lVjz5+dV`(%*P3Cl@II_h=CA12;zMh2Te$jd48M^_^XaT(vh~81C~X0 z!PkS-rqAjJrzWxa03j&ke0#dyj!|;87@dTIrp1>YiE^Cb2e1o(N?24~+5bBcKFeW!VOf?Oz8p~Jxiv_$omLtt%Y z1;dD&@klAyj~XvM!@TSgbfb_!X8*d!Bdw>>gonTL7WI9E4DnMFvZvh+_A6Bd1Xser zM?t;E`s-tdcjvqhK#|$xo#65J*U|eFwThhLi0#P=KbmXagYU@QbM^pI^+H#|P2^na1OlPh!RMs8uAMs~_is*s?IPZytI2adZT9S-T*kb5GhWcS8Ku?CT`Oc4JkpZz7a-$=wY zTS10IC6ehl6X9HWzNQxa7Kw0*jE>O+{S-%*;#Y{Y&s}+_=%RH@=fT9J!!xP~%x` z6_N1kMhp&;0;?Lbb+DnpEtQHFn zv`$Qev^1=kN;jBuT~)o+<0hb?p{e4+({-hr>+hpA43|^D)K)P79=te+b2L65q{Lje z+1tM_O7r8Dze=nmO~66Od-Zvk1(#`MxPvj2h?P)HT01)0rdJi=o_voyqD3Linc+_I ze|sf)Ko-IRZkgPqL#A)H z@XP-6bZPcLTm=wQu~L3#R`=yUJOxM!V04C?tequ!d&{EzdGLr!M41j{t_81!s=M;| zIGHv-jd9DW<>uny!TO70OzvmD5Sl%Lj1;dCLVxX{?ay^-IIQ9pUB&BrPR?h?QMFy6 z?8Uqc9}sQpa6j2W{9(o~08Yvm0junf5bG@vE(=HPdNTrJ{<0czr)YICtj3)KsP^;nDhXk(HB{iIafXeir|yL&VFPi8&WnjbgWlJ ztPdQx^4qL6&~{dkJ8^V&^s%~)sAhF*&c=F;?REBN4Bk(`z?4rC*)U(tuZ zF9UfVGFYFm+akDZAZJqOM56hzgG?4K{f+Ek!b{8sEDWXJGqqR^s#1Oj@?Q}Fea!1> z!*bsF2C!Zp9>i(dgNH&JQ*-HFwB^cGmE9cwwX)6=;AH%Yw{huA2%*e71X0xBgwK1}>-nQj=vpPkc7_R=6K`igXvlf;1f?XSXv` zE4xoMUz79IiM`df-XU5KpW!)g)SKSrcAst2-Ay~&Scm+cIPr0Jy|iAkPW1M|Vvnfv zX!2xc*P+qO(ys%*O#qM_h} z{@9Dj7XCX(fFcopfW?G0R;_rnVz*byYF`4~ zU?L2?mnR(#otbuO&H(hQM{JBvLSpR3ZZVTMM?5lRu2WDO(JoNqba%Htx&!8DIbH$t zJa}i>OkJAjQWJi;(ETHcTy1K2_QHS37f28Q7P9-g9yf7_({dE1L8=;yX^vFT%-1jz z)jaR2j%d3X)nR2{X<0crP`-w$E@voe=k;Eyzqk!}2O+rXGWekCD6I=vR}4V#s=cCr zKCT?qB}b#ix7}?ajB&;O4KHPWmFwy;6C~Q z3ewOid17MQbjnMcGBZ<#7^KpeO;&gz0Em-7ZeP-=eW3o$9_?WoW;RxIVh{pO_q}hU zR_oeqSC>jqqToz^=Ix!GA(Vpy{*2RXanVzOJe$rSQo@+0O{14S1c+QCNe)_nVqDrI z6``|s`@e%cLLeahet%UEJ_897WCD^b|8`;w^F%2cyMA)-KPf-*7my{j;pX1yestO7 z){)o(^a{_AmMPDW_@p^T3x<-N5I^Z!#gUSd@{_yPVL0Wwl`)Yxyio$g{8BVu4IgjM zfG+GT6QIIRlNxaSF7=l;{%!Bn6fK7)-!wiMalPbfYI#hP9dkj~(bHIH=^P+Q5@K#V z|NZE!C3uslIs^j28;J(ukl1Oi&J%*KZ4(!3P~rB__0|P*aTj)d?%MY}r`dt1TUYPu zH%xp6$ke2#<1+6)p!t!j`DJ5{_vj8xO!lRdOddWgcc5O(!sIR7O=8a73`s^B!AperAK;R`0B{6mnG(z&q$Z;`;LQL78Md~EabIyEZBDo zcz1iaW(}mJI~IwLQDtm*y3q~~eUI*?YdWwg-S~bMa)MBIl%*4U4egwb{6GpsTtB;^ zzF;2fGR4tKnJ+xYqmP~b=G{&uwxjk*Ikr#XvXCpz$sXO-brq6jBTSSv(4n(FSBg;R^|T1ZlPlHLwF zldN$ZJXfISVwWx?UD`b+)?c5NIhEK)!E$C62-O;=4?o9<5ee8i(x~&7Q|C)8c??fD zBM)RqljZ&O)JO;k0GgPrt)wH+SmThSo*WKHqxk*?-P3uSNe+-1cfnpq@)2?aAbUKP z*sp5GwI086P_a^-`4;%ya_b8`mdYo(=YuflI%7ab{w~ax8YzyKkcb_Bl`?c^Z9 zc~i^@_x*W5$aMhVI+vW|j|4ts7$@p%&L834Uym|h3>-z6y~YuJPK_-do`9XX{v!4u z9GHU8_vg5;VMqL5USW&?q)$|bVdn`T`Eyl?=yUjLjzhjzF8-ILKv&=hhy?|6nE8Lc z9?%ZBER{Br=l%Z#&I!xw2mum1t!RMfg#Z6}sIEuA>katnUr7DA`xBODiwBsrXK3-w z3IG3-j^}Sex;dzz-~PPv=L`Q-mq|Z^B6P2X91dWwX+)FDdpFtgy2W4OZ|v$2x(ub9CbVtC zV{pB6Dk86Uq^F-P3@H4Y1!^~Rbh)|dRl*tFmdgQit@)D^5aT-;U=lC_oAK)F;K72d z^7Y1`z()YAS*ck;%#QAob#CgT+TU4Ubmgf4GvxnqSk4Iu%_E-7NL?x^7{Gk}eCe zo^qozjjAuRE{h&K4?o-2DWU)>sDkf^F-FvTc7LeoARKXFr+6+I)rM2p78>Iyuvaf_ zlICw2>}ZfDjgQMKxVV9BomqT*_0QkFzrcbt2nSg}Br$+|wwHiYI7D`-(lOIZFn3gB zwl160OR&jcM+G2zLH0Ap0mw5_boHc#eg0U}aS&Uw!Ad9V2Ns1l>sta~+QOpB4`{Yj z*DL0cSC9g(I0vuATpd;dkLXdN8P~Sll(k7Y?@6zD$EEa4ylG&T{jZo&VC535$m=>> ze>9~a6nc9_1So(o31b)WFb8;CXeyWD=ozlZI3!88P_asuuw~f=u_`VZ0|qi|0#k

@X9M|S}m8FsNi-XPvc zYL{-#M{MtVKiTu-m>qchLc$tIRq8Hg9hYEpea9g@r#fv>Pg>!Nt@{DDQpT8HwwwLk z)g}9w;13Ch_r&0IGE60#$AI3|Pf`nD=!lHcwC<0piFaXB0r-v*B2d}W7I>eSj8!^Wjw#GR*@uIjsaX1kTUPV< zYl_6&wS>06?j%P%zbL|7TapIr)<)hpY~8bV)bCq}9O>I&p@p6lx-c#Tfbz3M=87>y zR+3be$$M+qHkG-g*`6sa(LhlAsY z8!ynWULOjLxsNPwbj&d^NZSMLHq&-CoMz2VDT)fsYlC0uM$rA6@X7w!Sx&w)$j>VN^&ympeQV<+Q`XW;U z)>F~O5nm-FXS0YcRCt9KqGM^&6cku&Kb7vd>v1SJw%)AQOBm{9tA73+qYI2!fAJ%} z{!-yyz8`C}fxYH8zTI)usXfxr!wHz3Vr}hCRQ=lUX%uJmIC5knLx)6)zN-_b0loh? z;YY5{0$O*yc6`{ep^#O^Ef5eCGv*oj3K*f^P_|S$Vw*Lr+z=9>p^xh6_bD`@d1CP> zAm!maSeT90$FaDJOC}F|Z+PoB${%$me-%`KnQ@l-`&VQjw3=zz2_{O&D7<(%QE3&O zQPhL{{JA=uX17JEM@}4>KiSzp<8GYXQJR(I<^+nrH9ODPbN29Co?oJfh$t zmR{Hj4RV)Pq-AyEmuD=ddZUBM`&6V4)Rw-y#>-dN`XapPz7CD65b84(+g6b`aVyT0Z7*=mp&cIi1RJIudLSeeeL54t| ziwGMrh@-IrUsZ!O_HGT5sWHM<{h~W8?c`v%|Aw z3r$w!wH0eF2zEP>M;%^icKJ11=7{ICI}g8oDU_pqb5Xy@2r(nr3^Y0@4SSWv_u5vi z+?=Zx0MdqlfEY0E+CCN#XYt63C}qSx4^A@TT=gHf77^#096wV%Jk7)!!jq8Gvt!lXr7F(=aBo8ajxv|k{XMr5vM`*x#(lL7``YHCgk53k5Mq>0E= zu|F87y`QFKDy>gDGDs7`f^xb36Z!#1Vr7O(*YI9#PJ%$;w@|+%@>=EU2)c4C0 z2yor+0_Tf1YtuP8U)?1F-uIME?^Xm8wmM3e1K35ZB1OVf*%iYl;`~`By>LMo7Bfjo z3D;zqY}=R55!Lr@nsK90*s8Q0TXMe9?t6j!=HPlH4SA0&BdXCsgCLLhiQN~L+A=bb zDerI@wpND9+r}HS?TFgLxga+wLD%Kj`iiwGt+)VzX0s1|=HiQ=HWtJd!ikjR#5O)5 z+Kk`@uHd2VF`umE1do-3i(b?8P1cbYId)*Qq{;(1RAldz09k>d%#oMwG)H7q(1_#l z=ynjk#`J?gtA@yF*z;^uCjRH;7^l*pola zg0zgX$tXHRc=XPh9++GqYc7v@UAQh|d1W$}r_;(+Ca7m`ox51r_!Vjji-RvwhHq2m zu6a~j4#f5S=na4nbNkid|D}2F{bP4WB?$0o}`9e5D&1On9 zM>U7Y=`_*#WWI)%6kwB9~f&Av;3Lm^t+8{Fj4s@$)~LSH)CB6V@b zXg)gMtZ-11t#!l@)l=0#XP?*8N>a!wRRnqF-l@0Jn~$35)XPa+H@Dk*V_mY0#R`TP zlIX$4$c3elHrh{MFgzCF4y2NS19ws5kc8es*A5I&EBdmV zjD(5!hEEWG2!#5qeXJjG4N2^gXiBU>eRiCSj84bb(g3abdG1TorSwzmmE>z$SH0OF z>1Sr2xealp1>O2gCFeYHpBC)-$bX)3VGd2E^^O^FYvSJT&luCZWDczd`iasbpjD}e zXAa}uh#lEbMy~f9d$rYr{fyDyu?Sq_{JQ0`iExImT^7Jsf1EXg+&3rCTeW%GX^+vc z@$aE87P6$11^`{)?sX>QIN;+LST{ZtMKg1~o^;NzC~F^CjVin;;}*ui+EhvteK_4= zf)>#sp%2Dakq%3tP)O^N#HC}!6N?>UU@Qyh<5=gCSsSfn%YHn-dL<16Zrr@0?9C3V z4I8imM?kL+hjLHncuGsW)4F3yZ)$&Bj{G;j9pX*uIw(b?NL92fTC-acOaP_Ov$2NRUx0 zG{C+|hmxH{pLvx>E4k$~X z@xSIN5tMj{p1FXK6R+V{QdZ7*JrE!Eq6jRWVl!CIja~!7_busw~ZJ|&)8J@IZoppW_Mb(U0Im1q@Md%KApazWRlfbBPmmk6va>qwG`u*MnEQCPoT9d}{(c5NG7*St9JXgM z9Ty2w1I|oX%h}g<8{K!t_7ikrw0o0Eq9p0%OFeC~BC>EYYF$9$MS}r}(713p?L*Lf zQI9j@4EPI`hZ}lble-nxyjLB{C3Bq`Tj$Sb0-F*3$d27<^WI7dQSp$MFH_}2Y%mp8 z&wDV&`1ocGG}XX)oP|(RKe6V?K`oi>AdvRE3*Fi97x>JCtxZq>|C)Q}n&S#}9P7oI zE{~8-VV@o4mA5V&;`U8__IUzYX)VpyU?FG7v-1>H?eP3%an1pG&Yf}-#gNhH0lVpv zYG^qV4o$bRKIp;|APKPi!8&f7KQ-Jw6Psr3+C(i|54Nr4Vgq4S( zXdw_;`P>8I56{>4Hc)-fh|*5^$?gUU&44sCs$kt@EDTGkVxc0^B;aOp`k4VneUfkC zvU{2QvUAw{pt!n>jvbS=@c&~QVzKJiR}(f$%{Bt6CI0xj`{r61)zaH3JyxVMo9=5~ zZaQ=c$#0S*e6MHY-N<&+&^?!;JGSrnvZzOAV{ylt@wPv{>huHXevKsa?;cUrjeGJS z+qX{XbTYQnziLL~0z7YrDZ9B-*Z!}qivV-BSfaX)nbQq|FCu*CCWI(O&~NKrZsAcX6ZKXQziE^<7w|Cip^C!A{`RLX%5S!dRxgr1ezHc85?>%pFbn4IdPdFQHnwEISrwt-)**$dIi?pgOHI zodb>-rzpVYH4$CZH0-85Hds01tGvBXnv`YSqJ9X#Od>_i}>$(%au;$8~7J zS|$cOME8n}oRVimZUco*a(Q9W#j3FQ zHlHWvyW13YpW||O-6tKkGTrmW{RSl&w1O>~qBq^>l6#`4cS?HJG=#L>;oVd6=q5O! zr8o|4$f$N+=X_aD6az_Drjzn0g@ekt4^6H`)4nt<0TyH%ULN$RJ)1s9v-ZVtbOXxf z%WQ%dAYQRL-bB|Fi>8f8*ePz#XrT45proc|5gRwXh%Rz$8RkRm6V|@z5JM^WNyzxJ zZ;hZ%H3{X34Ze3ei0a5&k@CP5!o=q+?a{~bfrdPQSUL|VzU%IQo2h4)Jd(?6=JaJ= zutAWZnoDVdhlUv?Pp|a`aV{`Q0JF1+b66vyB_)gx5bIV>6WR-$wOMLO^tD>ce5g!k zQE}QF1q}k(#-gaR+_Cp(CR3dpv|9oW>a73~2-=Eb)8V{4yBjhuEnHnk;^yl|SAgu7Ly0DHw|qgyd>S>&E~n=;jK&%2>M=Kg#G zVLRqXoKqXDe6A^Kbc~X}Ewe`MT5B6^4u=rS)I5_m6tIcAYWf9MjCW^;{CBrKY`sJ0 zQ|j`omLft07w+>(k*0C{XoqBLS|E43N=pm!Wo6Lm&Q7#Yrw7_5fqt;EtD7Q!P*&lv zH173<$V`XccD-J)DIPv;XuxkxWZDx$Do+`9^;@h+*V4Rksen?AY7GJ*>8~lgj;bs+ zBp~*PDEZ~aP05&~Tw8<}p=f@;(YWExHqc-1%8qK$PG(*$dIQ9{@}1pgmOZc-BuCkh zkUy#-K+V7y$F{EO>xMy2ty)%r8%@^t8KJ^U_I~MFc{WRS1s%%+!WcUW2nAtku4&x5 z4&jAPPxQ2vy=*2_?Si2_XUnJNT+#ve;OlcO3Si(%6BQw>{a`9_!t*fp!%MRfb?m9b z65NBwZD7APkChsSJgfYY`(Hm!93*!&bOP(LT9bKh?!(9+CVJ6sdR3!j ze9YqFd)~GeldJC)e#4_Feh>eUFoSpR>2d0>E{X73Y@WRR3UyI(v@Bcch`CcxBEC8s zuLKGD?e?gsA2a1C?GrOuIt4(k=^0}5I3sgL+ZYId_i*psOVxRk^5)WmAoI8DPmxSh zSq1UUO-6aOtW6y#!|;Ul`S>1n+9LHJZ|s#4j%sw|^ND+?dEud|8hN}{eg$Uubc}=o zCcxdA#eHOlBQ|dz-Mp*u6)fOsNL!js>}oLsjN>iz3LusNX_jCsx0Wcv0j5Ox2zYHo&1%nqOF1_|gwtC)`WpE& zY>@tQKy=M1O83}+Jk72`>8jJ>Oe<)J%e3NQ-D>?peAPlP!SdOMY!2h)8d znN>BpXu?5?6Hg2%jXzpF>zAy=QfJEUQ$iW@vwo@Fc+A$tGUrcD({y%x3Bn=4&IsKFzV z^7lmIt`wb?q)x>XhP;UamND<CjHhgUNc<4&{G}h5 z0gYu_e-N-Jh>N>Wy3_}7#$U@Qf2ZZYx}+V`dxKWcl(t-^ih%(8d?qqdLl zp9U?93@_)R|ECtaIuQua##~=Y1G;gPHZJNTEd~0l)GQ-?G$4vSI$7k?I0E#nQ`IiE3vWW(~pzwf0k1xPNEH_e&-8!;TMl9UMCxogQU zO8BpYgSofhe<6XRRsJ!5dk**S9jnVES12j1lpA!hiJBu$Bmezt(H>sm=Cl%! zWD*3QQNmnBbI{3lukxJOX*GbUM6u>6{k}$^#x~j!e|&NPsjZ z`(F#3Q$tP{Up*RcgsggM^l{hgg(#LWCzy4RvA{zf;E&{DhPQ+KMk;&eTk_|}BEOKq?>fuA;6JGI=Ne;lUwQUyWBfWoZil5f^t&HFq=ac2yp*pLo^{*+ z`YWn+IfxVc2-S+;mgG>8rp_-QxNl*I@`@wm!HCECUD{KfC$w5CuzfAsy`J#&`n+M7 z$+<44yGQt^Bnuc51=1jyQP?8MnAazj;1mIZwp1z4sTK%|pgO}zFgftLr$EnkH;Qsi_i<8diu3cxo0T*@&rmmoCGQk@= zJ`vFI5z%CbuJ1o(VeofO{z-8VMXXhL&v6txocY{dZ$Iu?}Sp$5#!n4X>Ewg2|RxDIcUXKOoGRLnsee;ViA z4&+BwLXrR;F!GsDAp{fQ%UU|QPu_}9`YA){T3BM{GH<3f9?Kc0Sy6!_u8ow~PZ6NX zydn?iTfm3gO2N7V3<0{FBNbEN^ksUHEAdTPMhQ@gu_7GD^=jjn0@9i%nbe9&QUg?E z)|fmZZfX7P1V;pXbOp#|l8SM5(%d{8_1r4KQpJ0N_|x!q6W5-3bjat&CoBT&W(>qN z3YX%!`JZC`sf0OTbqFoy3H;}+Nx{IKm99_DpV-#<6Fc*{0E3ysH?fYA_(|lLAC85< zofOKU$wwaW-~RPaP{)%2(k!=Z{Nz*rkVy~@a^TL%Pu#r!eJ0FQhxpBY;QNX{1se7r zT}{B91|otl{)aOG`D%Has-m6)Lq!A#9v|9z$oAc*eRuV;sScYk3E2@0US z#&K=>3YL7!KTg?oA0X2X_r?DKp(enc90}K?{`*WqAgvIm#QG-AH@0MB9{@ zJN0j5e;%+slb7fTVfOml2vo+FXC7eGSVX2eAM*tT{&aF+76QS>+l_bnChiH?D!|3@ z>uh^B@rtOeh)ss_Dh=7L$HRGli zIojk(V=Fc}=H|T10iEMhJg$D2r%oR~<&SG7Md}^6n9RbXfgKa`^guP*kQQw{sVJ}M z^3`lQ84vxEW9u{dbweB=kraFS|4 zR3qjeT~auHMm_&ox~@KCjy!m8N~>h?9LjK~n=ZoYXjNkt+;l(YsAmR=KVI?=mX%0i z2CN^3rn+lDPVKXXhYAM%^L&suY?G9CwNIpGV2B^4cHK)X;^f+OZmr>+KV|{^f>p># zb?mva-MCRFT9hFei;~fnf}j0Px8G}%J4c5Cir5Q&lHi}Z4`WDR$TvwC*VA@_dZQ~# z72p?nUGd|q=T0g3@-Eu_i{lg7aVS0VQ}nkM?K>UA}-VA6s`~ZN8%fFsp=GV&xSFpbsB1oIjp;6+!Sh zqQ?FtOAw!}0@z=)CxHx29r5{`27mbKMCj0dwu{KNR2-u9pfgFJq|`;K>y1HOi!`IDMvewkY@^sR18)!3-%=0NuEO%gPE79KdM1P_9MaKR5~-T7ls@gVk;d!vzo2?PmhXF1Pgn zg~&rz1g16HMqhE|xE>uxd-w=5ZCG$MO3mmaBSB0(>oaZSgV`haZ2CZj?KkI@lR9jU z=+yWrE-;H)Ux%BXMi7TIM>~%+!Q#Vyvlb07`iky4p^`5meo=rFhis9t;3m)O8#kiL zk|>mjEr6v{>_L1dDa0eXz5;~Xr>l-4VeTUc* z<|lmy0(#=K(k(cv`rB|{w2gb5)SvYS?O(?!+?2MNBasP6!A%eGC($^!$ozoTA&GcC z0$_lQ)J-f&<3E8bp-f;i8Uf}G4-qNdD7Yjum(G>jAhKk75eP|MDkzcr`TQ9Tg2L*CA7&ePt?ccrF^EGNs~FIN((4Zr6Wj)f=Gb|`k%Ku6TbZ_y0x)C|Fi zg7pX3De3NRkVcRWK|s0=9n#&>c>w9| zJnu&I`Q7JUc)s`j=YB57wfCMiJ+o%k6sN?=zb)bWY&uzz{tOPtrpXyyq@d}S0o$zd zF1?OZA1J*$9f5{{FV3s6&PKY0K9io+#Q;rzSnhr`L=m#$p<1hYxM?v5u5BJ@%IjLR z*>OMoN9;7)YjjTo417U-lSDk#m<|k%Wc{z_?R*@xdp|}hrYXq z&5xk;WB?t5gVZb5i{idGOqyxOSEHna80ePm*Yz$V(_-f0q%T#Lz$Qg96VFuI}jutV6nhz`)8Zy-NOMox2bY`U$i9zbC^~KItnsx;N90dbjO7J3I4>gP7VWh*o>bp`vDO#d6g+-FSE8A(u-=>} zb3I4;Z&Q?kKRHOzklD(M0}$^eyo{5p6eKYrcEy?Mr*`R}1=%O!mJ zgL{yynJhV(Q7){vRz3()^7@(l_t4pjh29pm=k}+deZfd|Cu5RY?QX{DV z87?jd{A^KO4CE@WL522WJ->y#7&Bv}k1^yn!Yg@jYnzBFQmzaK>M; z<^Md#2L~&#wq?mIW5*(fUYP88F|3(inEz7{yDKyVWqXMNCt8z)d)0%c(a|(T5M2) zSg{4nN~6d&31FG{z@zYc(M`RJHsnD=Cl+IQgpMSnpYBbrs%Dw-L9u_^3vAwcP&MJV z(aO1a+a%^?FTQ!sPYEw$jxiY0oJJumiRD|U!WC*#l%UbXLk=kudtUuZx10b^?2tJc-nq)W0~ zavMu25?PfPLqXl@KPdbM5!gxUKVq?Ohdv^^CQk*$yW-bEA8eaEFoXCXyjYBdB_o?t zFcsCt#rJ?;v*j#%t`YAi&Z59NBePzvLp}M`85BQxPE(;oko5$&H(67Ss1k{M=#6b= z{_L8J_*Bgcf72nQo!}TcZ#~r?i{@4(o^5%#vIoVz0j{U*7%xv`38CA%Rq)1dpB{uv zQ6;7_?L~wn zvy)(4rFNm*fOL=z-qs5jX$b29hshnB@q(CV4dSgB8=STM1s@89K78YF(W35o^7&=) zBI-Hz*TZROL|-L#p5Dx8yR6fq42L;(*enPd3Eb8!+ABXXU^?0FkSUO_R%hmJNIYIG zU5)>>&93k{T|hAV@hKTXZJ}5|{;~gYi_gJTX;q8sTWjeKKl#fZZ5v~ZsU`Y@{B38qe|p~CsHu#Vzu0H?-?5=I z3Ug|oB|yS-e(P*E!j_N#!EKkqFAF?fF~U}?ED&k@7EQ3GGPx@@x1v~z*8|Vs@~RON z3A5%{FJUpKA#NlfLrE`DU(+U_2|=77rL;5sas?J+x>8-D*4(jk?o*oO=&h7%GrjA9 zb#mcU!FkQ*B2H`p#|K7W-%ugF7Qc~Z4da!thCyCx&4dcF=TC*s-&y8Im1SiaZHlG#1Zlpu z5<$)`F}(j%1LA>w5Hk6#SSnJY{nHCD-d8`kC*iF+#g?N?q*ynOhICMQcMyA&R#YSr z9K(}7mSQwrBqfj%75)0~q#BrfxYs%N))$ygY5<-IRv@ygZwpQoaJ+9J09Y_-E~*Fq zza)+CA%c>6gNu6^9t}iJM(K$_U4w#Yv7CNpgKn}PF}SL@PpGM<-@&%8MrcCWF(eVU z(8`BUhBmBbkwdTn9J!b3EJQG74M`tMCx*b;8>{h=8g5n5=&%J0%Pe1h!AZ@$yl|z8 zQUqkCqR;E1TFI}0zA6=zKg4jHg5^4=Cz|Wt2vdmufZl&g2mPJ zXmNNi`1GXbAKx!h2lhS8OZLF;&%T%8#-pqYxV$K-x%^`&==V z(Bmo^x_MHFom3PVkkC12#;Kxb18qTOCuAt5`p~H2f|1deLS@2hMtq6xjcNNTwiHR@ za~r|&8hdA^>0J%4PG?+uS9Hk9uqP<0b<~C=i6!RjHL#&2aug*tF+=Bb6M56RzE3WG z5@zL{(uD__u2Ud%RZ(O~+`~Tt6ZZKHtq;029R%HlCls7b3&doW!BX3;X?iITH!MCy zUba$1rbwwqU5V$k^PS3?%>=rC0!g~?r)tgQ`==2_<0Bb_VH*?@7cYFHIMNp4dTBUT zkuAXEymCT_FXG-il6Qzx{)k=cv*FokTz`z-#<=K5-liYGQLb`F*_PGBU`XtOdYKVD zZXCL&2OrOs>@Q_>)?CYe9}*0Xn^{uA8r&3qkY5l9et3a0<)k^%i?mO6W(~sBIrm0F zycy{tG-tH=1MQ*<_vVuZt8gl{^YhAW`u8T?FyJ644u`Dh9DSX~B0=3NXx`%XhXkaQ z8s*L1@N>Lx1>x{h3ox>booOD9JXwiTJQI(>qqw{iF`j1X!yDgttdjyrwfJ}m9o{F`J?!)ta^vYCqu$ZRJ507OB#!x6lWrR zUOjm|GTuUon@F#hWFKp2(abu7rXstr3+jpJWZ-E1aj}?qwpdmN4Mb9y+;!zOvqt4Q zXFM5`qVKioA3u2M!Nq0{v>W@+b79&k&&@wtFtHM_B3lYjm0|sLBvu8Z`Eq(REMhZx{9cH;qXY?gD z_KvqsH-Wa(;nMk9SMi2+ zE5mlc*ApH8j@`aNpslIA?)>m4!9pyfbE`Lv4K@9Iq4rW^6ER)pXIki_YW0Zg{MEC)lKYsCBy=+nP5@u{ml}Kg)~NEL62&O;OBnW8V0p)m4CZHVJR0J5jL|Iag)c z!Hi$ju$ce8hQ?1ylR(bEy>Rv@?PqB}Nc2t#OfB zWyKzZztWvHs-j|`0B*IZ`};Z1(Q zu@>*Os>lf%eIA+&ROgg5{ebmE@lW=l-$Tz2NkW@pr1uAhjO+#Q@gM*!l#$V1bB+B_ z6c`*Ovy}%(uk~LF@K)Q*At1@?JI`YfWqAU(=f6y&ZuP*)L&D7~xXkHP>LbH%7vmnw z!%M1Y2yV(Fc9t?8S94?|7jOjWTMqc1z<>RL$$qVP|K{KE zp&}s@GAY$sZRlYrDxZ$4>{jtkE*{b^n2nNZW3(DIZiTpBgWJ{yPT$gsu0dv&374?c8c~fDzCulW zW*}qST$0U)ROx_#^vkfB7n`=apnMIxmLD5DgS=;$9BWPRJ2G*MfM+7xW2jW2ghDb&WDPhmeKybOQ7|M7M6q*$Xsh`am(d$Mp?P`7m@=87aU$NcN9;>EAsrW}Jc%zA9>G(9kBC_X zes(2>jhdRwdq+o6eVh1Th>|OV2y;ROpk*fU$1b<&sO^SxdtpYA|&x9 zo>#WS*{z-r+qFyYDVW{`R1GJif?kBGQX3h!7BF}FU*@73>M@V~A9GK%}}3`6f<7PZPFQ>-5QKaZgd~*(0k~Bt(3@+=9!E&W1u^>jLH$_q>8_WqAfE$fOilb|m=x2VPKM)bb zG?^X*u-Oz3}8*0C1mX(+>_xIMr@I`(L}^8e`OOFMkcF5 z13UQh1$$iu;?|{hT()%Xpi8HT9bKvPk$_?J%FDW`wCOVFOPWhae(ccoRP?_ihA$6c zeXLlKZnke)R)aXt5B1l5D~tv^HxXVyd{jkH>o*p(|6&?xK@| zSoF;V8c`S;ljkk*9%L5>*(H-+;uF{DknTUl2lshaZZx3dbig-zpsg&8R^-V!`uAFC zDwg^bwrr+XU~VitX0Q(G)V4zeIqr*8lehYtuBSz&vl{U}+rlVJ=P&{XEIPe~dp zn?xO>Np^zb;wOd6j9VMA+19)=<=W+We<)M?u&&f;{?MZNjFvD^(9TbLqm;B^xNC>7 zOMJ7o4}l}L9|&rT8vV?nqt&sT=&0$h65B}I*)e`JJvX!I{4yQY25or_@IX#aW$qkZ zkDBJQMO0G%-hg&W97ot-GgU-{W8v3KtLjX59%`{NYJ`D~q@|SE0D}d7?+Z&?GHTy% z=6w+R#%QEg>wDnP9q-{P7T>cdE7>Cxw(9+4mg>aH1>qz06a!P7cYe4s2Fgywdg|6T zA@$XP>0k)ej!e>4115jXQmnO!2xgu4jY^IWOpv>P{_o z#*u5@e*2ca=ue$wJ)!-O=M@#PeffmR+C)_DqA}~-OBiPfAh1eun4F}$ram`a@?8X6 zFR@BV7U>jrxHuC5-aEE+JufiR%;wV7maCpuk#>#AJhOi2rgC(*O z6MGiELK9&~W$DocnR$r6#MR8}e`eF#1l&Q4z@-5(L99@9YE<6S$aw3PER)E(ii2MN zbY33IG3qa_%frDtGslfzK|k4S$90mne7J|PW&OAPfG-mOpf!=m< z@h$}PIQvusG69t}xJd>f%?h{+ixarGF)}b?ioSD%%anPQhjCK?NU~~k{^&H(4P2-UHk#;qv_h#`NUw7ewm*M3kM1-t zK0+0Vm29@LE}CBZwgEq2Cm4TX;myt5ye#GJ;}bf0m{GO3+N(~-lO&@{$f*W+OeCEO zIpGAZJ1Te_?+birhdb_}LOW6RLJPkvHeu?59oz0xj7teO@&KwFxLsS{1oIab-9N&Y z^LZ|k8qtKbD0#-IHLMzCJ%4 z&V?>8x|(5+<+1tkv0&5MK?e%-OzNcTheq`72IpesmTHn2Ee%Z+K|U0Wg>{T!ZEUQ1 z^eKA+&}Y!|=rZt4`fQ8}&4T6u;io{E@_u0!JoKsJP(ixoP#zOb87P5Thal>zBo+f& z3ZvHDj|&$Imbtu~U!_!MHh+M%gN$YfKW-e=2bx06AxSn884f19-(5Hz#FM1HSGh7i zpc$&TOo!v(A{VwB0ozO(AW5yqcGEFpjE%Vi+}-d(w+L&w3+K{6O0P_RuHy~)-?j6V zClYGz&bfIi3WgiWtbbRa5!dxZ(BhC5II2nd-s$71dT5Tso`R&paV z+ogmMaMC>A^pOv2ejz{)hZ1smP7!>gYwO1M6PC?*dVC>u?`KbVNP)F6S+3GzV})3 z-)Z{l>yn;M!C?7cM4AbXKcah=vQ$ZR*bNLBOy{8!Ls#1*S4gt>P;wNxEff*yllbmp zkv~hLoMe%q`U^qh9U8b@XX?YzxJg~ZUgchzlZJdNyr>d9uKlI%29nY>g>p0DFa`n$ z47L0Ez5!w@M1j5@v-a%i6Cy=@> zh2~vMjaD8Z_FjCmgIkq+xWRy)<+^@^F8mt z5FeNq4v4Ls57UPQvx>uSGVT795}t3Uv4AV=Mv zMO?>1>4v^VM-=L}Rd#T?RJDb9R-^^`$mIYfG>J2XUjPQ}1~}P=Ps06nSIB0~-E}Y0 zCcZ}aEoU2lXd%V7^z(kA8mIdpXAZ$|xv=H%!$uCnwCK9sYRvGPkO(#3|g`@6L)#KPmfWoHF`n-V))c`*RAlDf;Me ztW|v`ZDCmzw(AJ=Z^o+4t3J)&xAR0Vi6L6eDH!SXC^Fi7%j0->S;-g{{vM8btQgmU z$23p2DSfOU@$>In_Gxyv=aCkBkv1*e5s_3RxwM>8rEE=yevH`KM(_Lb&tL;8c-oK{x7%7_5&gQZZ76W$}8`V)&h0AQN zKO8hIqFF8y+~na+iJ^%&HSGGR{Ejyzg@8>_?RcOwW%M!Q#ma&IJ8w2=-Y6%Rhei`0 z5u|j#e706l=n~O%LdnMWgs=Io4Bdx00D_O75^w7pTgD-F2YM}>Zv}B~&fOQidN!@>uI40VH+TD#;Cy5X^_q$3~?=cMhj#GeRMf@nn@NJVg9g^60ei^Y- z-|!jQ(2xZ5zkPQ-Z&mJ?fnD-axC9Fh_fc#sNHd@a{E!i7W5;JiKs{{ zZ_|Cl1EKPAHuHJHYDM!MK3P+gwJ6YgTl>_Os8aYYlX+svRavX)5&T{nX{0=g;Qe}$ zQ|1qw7*dIfmO;fI^)g+Bs-AKP4_OK{hKR8mWH?R@bYyo$>B{?%v;G%sC zw!L6U$SU7R1Vv8xaq2-jm{zJ|C1`c>~=*kUh>wB7@bY zu2QD)Fr|x3wX7S?8++n3Pxr3RVF&r~rpJSRg#h_@@PfozMwO_nQ?pD{)r=qbEgR8@ zo?O+eFUELLUSS@$$|0SaxKTMbl7!uZhY@Xw7HV2)lGWrbAN&@w3xWJ_?XE~+`fGvg zsAUY2EKq@+l|QPCecbONMQEn_3WCeprc1+Eq`#X@?s zZ&MhYmh6AFz_haGrw^f3q@zWff$*7!0gXc;}`{Zgtrg8O~ix!=(|(q5RS&t9Dm?Z(UE9cUsm8& zrE|aC7`)^Tf9xdoP-N79$4pp)I`*^Q$b_K-F^%!qtpyWQuVw>#a`BNLaN3aNR!LnK zObkCKB<=G61Q6$4=Cylt#~lp2iaV9wX|8Yg-DC7n2V^;#-Tgq5S+;4F>jDr$10etQ zv{?YtM!lK^H&YB%p&ZM^a-ehH4q^6De z-qfnI4Bq<#u&K~&nS34MKaLV`0leO<*%gx>M8#Te2gOV}5OCvvwvG~7{g%os(d3EN z9luD(%9cL-V&~-!GJcfol1=y?_IOf8W{Cm8ha2NO00OrAFz#+|wUqbn@YHW37nNclXF`H#^nI0cMi5#FVm?*og|MI}i#*!M=h{|hVU>WYj z@lXN%Yg@fs37WUE%5z8EFKu@?qQ&3~N<=+8Q2sAy_>%42Q&R=}g~;$zGp&iY;XF|W z@^d0lkK>kBt&E>w*?=-fC9xDFKmkqbH1O|b0TV@4T7h$$UijnZE#&Z-n}^wP5oLV~ zU|`oIK8(h~m3XbmckaTQ5+Kc^j3jg4lRqmr+ZJT6^a333+bzWJml}~7KTHWoj2dMA z{r?5`G6?4Id#yLsH^UEUc?`=u$05CmebJ8{cGZ!PF7RJzvWFzHeBBxb>M=?-8(;rl zZ{s5ou3y@#)Zvy0Xq45^$Sv&+?^|iM`??+m8$9VNoyf<3$DIpf`)7ZR+0%b&)j!z#YhIf#8gTKg z{6ivc5xzfy^_#7M0%_TW!NCRfTY>rOKO}+x^dXvWAJiq-)=f4GzkB$D;(T&h`j^TL2BTOj+Hf*9f_)Pw#|E*}4UO zz5S~IF1E&w9vLgxbWyV?pC~=Tx(1ky99B+t+w!LT$n|f)k;HbQGaHw8SCBi2@XwqI zjOzypF$y#~#L8{_KDONcRjVqu8bz7V)CuL{NsutCNqR(F5 zDw7V%Zbz$*CwKQ77J+tH7>?-c%A@2O7;q(Av22NApw0jIay}nW+PAq8Fj1dKt>PCCxj)8wvVCGK;x|FCZm12S-jNqYpunr`n3a$ zuWjF#TK>nwf+YB1c{LX% zMgiomP6htgYd*UXbrIg)GiRZzZ@3mJ<}*QXcLObKX;o(?dU zyPYN4Z0{bqR7}4=civkztes&!8d1A4IG}Ks>u1`S?t+HL8GTNPx@SkA*1=qiwU-z!7cr)jP;=fx>hXa7p zjOv?mW9`k~`G8Lsa1-AT$7TP!)i_t>BoN=mzJG#w0}+6%7Hmw-<3E7AOCku7k)dBi z9Sz7?*=?}^+Wzo(D$2&|G^H) z7nsA^N@wNdhNV5-{c%;gvnv8XnYFyGIgI7$yTNc)FM;;raMp=niT@#H`s(!!SaV`Z zh)z6Z3!;0unHz>MLWZZ~m`g6d^#85^$>U(L0(UfFbuLMh8`JeyP$etE8nvT3_D)SZ zW{|wQ+7J7m{oHSe4K9s&^5lQ;exEbU;dE6EoD9Dk2`ulu>y-FnihNZk%kFi~E?8ivbh^&sg$=r8gsQv$tQZPtiXe< zmhZ4`-`8LGI288EO_(KpcK`0xosaLoIsM|{}J;91PJ#|2MK#)d?6)yIYEG)h% zaGJ0{lyPgvA+3H1v#t)v-?$w+b;0#j$#`dDrp@a8!RYO<15qmMlj}^=sC%*gc)S~| z44s=Su}P={O4?P{Q8PtDA|Mdq4GWxzn#k4WykY|XFU;}rzz+=%i_6PPY5nvMhX~Nn zA}<>uy6_xtE!+dYlX&(ha=g+3C3C34O`^uE0S%%oZ3D&ra?`TxkrN^`)ttplRU$#*9PvN#DuOo{N`w_ z-1)9damh?T)Nk7MCBUGNBFL|YjsIAqC^5*k+s9q>xF;$$ThI?7V0qOvagJ{{H-$Y= zU=WyC{JP{S@L)q+huoz+e4)1l5mUe=&Zr)-{byz5{I zgY)gt%jBeCvM;FnBkBR25`t$Dnc)rX^k<}UXr=LQ&8LhXRNiaV2 z%?b%Qxx*lQBqkSbmHIR=ujhw#EZgKk>_oyULmzqiBU|{xH>;`Yf=;PpoC{EZ47B`y zFpo%ex{eo(?9k}VF_C*7;^+V-kfg+ZZencm=Pm#NEVUm-GbcOA0Z4M0+@Oq--XMi3 z^1kyV2OF80eUCpZ$%5TyFe;z|2K}(i7vJb8#pr(}{JoD~_L;t$kkv)i=?`dBC8g== znq#UZn6d|TYl8YmFBfB}DSdQb9-_aUGlZZz$?t18fsIjgC2mqgU)_v(@f!+ht$Vi! z*YX4RReUuqfwRIzE(r0qtzel|zmJ=8T*cS~b23RN&0XEJwn&zdKRcyH8(uKGIHi|` zZlO}9(riLjL?wV*{V8(l?vK=#zgZg*H*2Iy$kM7gMkRhqRL3xqlCC**hViya-Bl$I z2g(m8zGKeZ#Hr;a)KMnB=25Kt0<6#ohEzHW;bCvcQCo-NFuS%o8he1!& zEpyy7Z(uY>vn_fyuoYBlKv)p?hg<(nC4pp8EDOC8?M^h?R3zz@)a#?IQNj%JzOTS% z+5GiyHwwSs=`2rz5yu{_!|kv|{R`@v7Z`byihz>b|3uyG{6A8vzg_@bGKP;!vbmiV z-`5Sp6=?Dxy$a__Ly0xsA^?B0Z_9>&&HjJ!z%BUw&%2U^-92zZ{&Qf5w(RGYhdEkM z_BM_52~maAG{gLy0t}FD&a5G(@g^h`rUS-74x3ax6q~;bxN~RW70sEL_}`XarHH8zpI9ko|g6| zB%@~m-CrmT3|9D-ZP*0PMaE}&1%15`@ql%vhLb}ie0C})M#B-`zT7)sQT=1W6*Wddy;p1ecEX9 z;`9Jy@)Nwax;ps?0!1rvq@kfP7zjqZ(Hq@;{cbEk-l99YZp~z4V#Ha!=X~^>s?2_I zsg4r6>WC#PL{Z5PG7Vc7xW#5)Gb=a3mbE6?49Or~?lOq?%WGm!GMdQ}dnbwwec_a8dBp`!cXw29i z9Sy5=F0N!GMWGO|bpTCe?}yiwEq=!6m22D1T)a>&)(ukfd)GDEZ?Q^~(KeKoWnjs9 zSkFrwKBDd{--uqC7sBX#63C?01V8q**m%->Nv&wE5yj+V1hq_fAY*NtH;uZ3c_G)~ z@jQdubFC&b_QO%_g@p*2;Xc87HO@==#}7Lp(KwyqicuYBTdv8Z0c0`H=O1k>HP}T& zDk4QsI;ig z^T{PcI@%-5lcK;f-wx@bXb&|+tXx7*)6VqE{ot2M)f_7PRDrF!X~*w!tKQGOE>3^t zi8W11P9dvEK7DyF&1TTSe@GEzcc!}otk%oWcX4t&+>FRbd`qEHW@7DJ1GMfG$*l1} zy$x0ENL*yt<{LDihqm^|DZXk-^|}d?+<`XDr+w<&N_XGou0gyKAvy-h;4%sw*djh%0y8oIdyc0YTga#Ro_vzJyFAU zT<~7f2*rLnZBxvpva5B1TSbUo}G7bQlv5UAtPh$|gSNm`Sj zPSIQ!UBw)nB7NOLuXC7X#h8KnFGMG4HRaL_yEovmIyWT-$q1KG=f; zhx1yW-?tD3;>|oRnwH-!G18Q#nspN8H?kWvrYqat3u6QhmqwoI9iFce|K-+jcB5$wPW4En%a(Y-ngu2yQy(8jZxg}J53cmFKhA_ z42yo{$Nui}@YNy!#fX=?6LF@YhPC^IxL*s#Ear85YwK_{hP6A=$lpV}8X>-ubYTAk z8zUtW$Vl1r%G+^b8Rf#+rUax-sjOnR=5uBS=`2f5CqWV!z`a>m@mvI7`zmdp`M5|@ zNWBXfg3tqL^)2`#5O(gk_3-qq3AbYO@38))3~w3v{(Z~{Wh+;W+IIQ`7civv-naFU z7O8sPm0?iLB=tCIiYi=|K*-MWM&5Ltb%}A=XJKFO6Ra9g>!<)*wbVL)+6C@VY~v-i z6Jw}ejxa$$Rk3JlAm9W?uf0dl`kism{SnEHMx0iH2jeg6xFttUqdP?6-R6M!>W)Wu*l(nGuC{k{h1oa;;xL^ zPuBK)CXVQpLfjNtw8!gXL zb$3dDRMno1hQfasWg^}=Jx18M7zo!tv*XT(|HqO4d!c%d`&G7EQR$%QN=oJ)UZ`rt z*{0*d6zkTiByvCAvr3N8*x0xBM^mny&PP*>(34zKo4p(?(yq=tY zg-m;++N>polIqI40RYkx9WSZi9^1JQ^cLxKmV|sL)B#CzZoG+MvbN~hr7;c)n*Wnp z`Rzqt|G5xR-$&}X2pVyrG6#V)96>HezEzX8-j5%0zL$i2HXgNTQ~)L|Q-$#Clps%= zhGMPBR?AUInH{jjHGl1%Sa7X7vM!sMER)|lS1~!+NSaImF>;9p#}1UUm^z&4JXPrQ zn-jF6nNL~q9Tg?XxMIZpdF}1%kG~fok)28Iv>Xa7D-nxcTRxY37=vfrtKKYdk2d+~ zCn2k0Z{yu#OqsUK*W^=@h@7??R?IY`@E#~AYWRMzScl72&b?e0IjR;r&71gYtIupK z`1$uoctM}?)ob43Fci5;Zcsm+QBN2J{}Q6THIgscLdYZ?oWu&6?q5@UvmO;~g%hv>j1H%D+uFK*Rx+(*A>ey*U`}d$=JOy@5-l+v_4s?l2^B*>z~$b=hz~;)Ol{-+G%Rl z9#pki!K-Q3y~JSYx)7P!7i7R8*xVRGZza!rx3m?fdS_Wc-He|fKYB0n=) zWZ*2Co|ZP8ETJfmPfu~Q*13w#{scQKH!C8D;6;V&EHWM&&`1eeWIos;jXK6 zU0ZXukYGF9253n-T=i{n8amo_tE3&wvx7BiZ*O6aHNo^Z^75;@M5rG|zc1;IKA}Sv zgI?ZarND^VHJ*TEu-NBjX7+6+_qU)Bq z|KemID49kC(qA1<^J~WzE`t+!nt#A*6-$lwq~oi;Ji;V49ymN6Yi(oKbAGewy1mm> zD4=HXLjCZ%PZxd0`0(JWPj~gUK3!D!b;nJ@F?(n>)4R<%rsa5TiauXYtaP1pCWC=6 zruWzKfPZIQJBT+kE~u50y$43l;X zGW&j?VCA_LD%b2C_TyR4bSK9OI%l1%2YjjSsBt>l=f^uH{=wZUVDO_f_y?YV0~u6g zu}G`-fMUIc*>)4b;&=}8eL;^TdY^0L#zdLSw5w}EHT1i0_XC=oZl#(!z{y2_f#hAP zQDu^_mi^`)MFcf5O-5G|_Z{YSofnk}Lb-aDMr@Kp~}j zNDPmBa9sa6G3lpOVt2Kp#q5QrNp%A2HC#^*ec$Y8T}(uO0tCK&m36Bbl~R2q@2s94 zg-L(b4JZ2GrPevA|Dpf}b&)$C(bXLTUwqqhuTukkaxf^4h7_`xk+Z?h7~itCvN{(e z>u{`I8f>a6gN?j<9_<7Zlqu?Ee&%)rKXzPy?qAt6-M2m0`2H{_hU1H%clE(A6AEi& z=gj4y?a96#!NkkAOx?ufD(^%VhPn}W2GH$5C{^OZiS8i(9R0;$fY zEhcIK^PZ@@k=AQ7GLMVKp7$RU%XXW&_Lnh+tuwebBd-LVgQR|V(j6nZHFQPAt=VZWVuR6s*iDIr=p z)Ua;VPEeAC*uCFh?%(xhR3+Cq>9CenR;CHzKK%UHenDvL3)F7q!wT;^x5wGcs`0MV z?7K^KbY{$;c&h<_Cevy0T@wQBnR0`ObCn68l054Dj4SlQ+9%L+G8B+AHNDSvf^=e< zo#E9dKbb1T^wtX&UkwDp%D~`{7q|j5x?z%fyN$xsdvCr8%31AQ$mQJ#*vI24t7~6? zN~Rx|8N8>`DpHd>rY=W5Hl<(CVJaAaMD0it$^3OyJP9Vu2E5FEshcqL@?yaY2!W33 zpxnweAJXF;RG}I z?j1(=Ivp|($Y8EJvhJJp3*!cKd-cB!B}+xMzc{ttovu9C91sR#`u4EnmtQ4ga@I(L zBY$Ko8YVgFoCIJO2}dK~G3#1Yi$hACCfMxHcc@wHcHXA;uo?M;EG4?~eA=ayYjV*589fjE+@aNVjIEk7>ETu&5Xo-Jybx^%BBn_LN1H92q@Ll|3JO$J)>Q&^ z>dL1dB1pV*yN~C5f_6tSp%QfTOuAiGHZBEJA|1&)>?c(aUh1y)e-8vMt{Pau0@Q*fN`w~C*p*E)sRIjd*cJLdu=(@iRGDctY*0*SI2 z@9c%?EElyF?d{p`o7vPNrr$rB&$qPMFU8X%;Y`w>9m?BG=m5+PGWm$l`fbYDWMGVK z7;uw0Z8xo!hJO5LPxTIbKUTD8G6O@jHd3<{#Zg!P(6Fz#G2y*%-StF zPLI}EoOn9LyYGWa%#c)j&SbyP-}?SjkM3)Abp}urUH=$EaFm>MO#VrHEq0&yyTTh! z<@k_bqSdW}N0|r=FKeWUT=r?$*FMUrb2{Qx+XXx=)iaxIZpP^fs;@)V?juF~7zVctsk)KVsXP4~-a z9%%q7&m;+fRSR%+bsb%6*GrK@0+!5l(5stGeiE9#K76~eYo$uw<$AM;?wE3`bTmx_ zRU2^K#6pY4xq{C_yquJP(%t$-s0ESlM?!k)#An4*Y&SP4=fPtUalCmDZ&$v)=a}vS z@*c_=OU#fBhw~Y5fg+^^WU_(@ zu>TBRn%^;QK5}rDs*h^C=!KpZ(y8xOb`6>}&CR2l7ne_O*Ud!ss$N*r%+$)|o%I|b zM1_ZIRqULag82E@KqpkveL-ur7S_8affKyY#aw!#`8zNtR~NY?B~l$-=Y#oB5ss}^ zg57b6sND-2(?0CrXU%|Bc)L(Rg*^hVEzaj;Ip$yW3m-cyuR7;X6pXWZ;bXkm+&ftS zaw}2TrgijHv@C6@P_quJv6?CYgCoE*AEj`hFU=ie@f;{DC?HDZ4iwZacUNU}=JK)&z) zWA81)s@l3gP(ctBl@4h^x|ME}ZcsV}kp^kllpBx4rX{7Nr8lw34XA{4cPSwa(hYY3 zdOUjIecyY(-v0-mXUntJTyu_|W6a+i+P^`cpEjvjTIK5;c*I(o0LOJtk4;4-aPFK3 z{KEK1Rj)nhVjt)>2l0tM+!6^bcuoVkQ8zU^;Ky75b2bw-c9di`YzPSLkrzs+$ zttz74Qxu1BSq$bBJ+*iPq=Rn>>>BNU{%R$48%-z`A8m-yWd_$}rF?m%^J4~H)BB9p zD7jd!mZA;;mX=7Cfzdz-z--(-v1#iwDAH1gvevf_Juu2hPK#we>X4KA8lOdGsXcKSmqo+1= zMqA?I^E{L8q_>dWH{@-^$v41)bhR{O?(0yyp7X)&MQT^E6Fs&>{pM`%0#O^2Hl|K> zvQK0|C;A$2-|3{OyVM)N88_u<6xnn7>J%n28`%uL^N}LQIQlNsC1GJ$`x4>4$DuQi z>9~|t?#CksDSu}kVPhL^&8rBHd~Wr4Y==EU?|I8_V6;9hwL%8mI#UuuD+YrK1w6WebsNo zo>fm&RhsVl1cY5hADoEq87+Ov3zG_?mUb@~-+FNDOQ~Jtez4sfDj{z_wV>9okggyk zeOquh>03nzlyJY`4to`En74ogk?SMc)UUoyX`Uwt_wD*+?_}l|v9q%|G*5>MHjp_+ z$GJ4z!7?)m>4VmO7T2Mk&P++iN|g*XK>52q^sA^to`{^Fz3JAxd?Ctv6FrpBI4_?2 zE9;%knnbPR6oVboy{`&Zk$kHaYRvp6Q?|`bcB#dwbz85r*_Fj(YrvJxRYNe~CV7AD=9&Gl)sqZ7r(qu?6_AVm%9O7K(R@ z>W<69^V|G{6|_uDHe?H0!sKvv;GkT;x7qvbY`YrlJvZdS36EFWMb~bjJW*AxpkmHR zE-T}PKC53Ds~mIq_%+0v9Z~toliqvW`Zb0_0T`T*Us<|tlF?k{kbP=S;?16^-?-G$ zU}8wbZjxkf=-{Zs09@y!H^>pgz!ccEp9xfZ^mOw(wuBTv5pjBN8W=+J;jTO4jqmeb z`P_7zRh{p6L<_U<8%tpt3LWN88K%QH?BJqcCQB)k=ns^{;1VNC|4so&T>wt|%+ zesg|ahH76Ax=vEQ??2JM(*T*7<#JQ=zB#T?{k>#WW^a4$Abia`#s1r%c|LG&iStKU z{O8Wwuc_9HVm)Z^(j}Q|_UqA~4fjgjKMC-Ej7ZRA$y;gZIlVP^6lpRvUq#I z@$(n(vS0=c;kEr;#x``pLrJf`@ew&s&w-<~RG4Ib@Jfa_cK66V-j!)_r|rf7oOn0X zxSz*&$5_5Bt5K)XkYE92dpuHzMJM*^x@r#R+S?k-O|pC7 zDTOG>NAb=O#=tRl4EOVo5X{_MTks&G&@gs zOx__LNV5pWAANXueLF&zvEZrRG96POl+t}reaYOnJt$P$;(GAzebT0_`H^MIC=b$$ zaSzk}e*);qR0E0ZUQkG~8FeF_##qmerR+w@H0xIn>UbKP0_Go2e{IBXPQ)LbMa0E} zMLUc3P=Tgq?YBypw^5XImVGubKU1Hc-r8g*_yHMm*Aae%TK)7rS#_yT9amBfyUFLI zr%tc`a$LpIQ%OrFtHK&C7NLl6)xod6e~ZU&9?5B6iT)+PF;+9S(#F&u& z6HYCP3M7f-dmJ}bGeffQBc;pSe2{_&XsBp9FctXvW6@;!mCf`#bRQB4hIr59%gn+1 zu+|ppcb39^C5+1n?4oyX6Hq36Os+E;`lHhS`+sAwzjLZdk^DU3(TR%?>`lPGujQRP zVodeqG5)gr&L@Wtz^&@>S0o)HULy*^+hjXrIX_}OMExryZqoOkg z{OjA&$$hW2ld#2CDyKEOYS-P_YrpNnDHBEo8Teqh;2qmjSDGj3%2L5D=Qi6MUKl~A z$zxFBx7H{i7&tnWVCrbk*w#Rgu=P}oQB z-Tl2`%kr>2{`*xgSoW^JYdHMZZg+aBkgbsIs<^A4)?&;wC zHbe&Rp59xfsc+*eyFV+%_U!k_%RiE^u+riaU$K{kM4h)2`Lh?qGL|n%K-9Le-g1@( zj|%u&rIcFOo~(X7A|jotZ0408ty2b5H&A)^#jQIH;#HMEyt#V4-wp1C#4kQW@ z|I3}24evJM7zT!9*DWK5vpIrAa6Wq^lE7P*dzN(%H%vQV`LkM}n{x7LT_a?n4f6cV zmSrbWLs_^^tsWapCzGzUd3?uYVJ?jphskj|{?jvz8(4d1H2cQ{&PU$95%ApeCMVm2 z2x=OdB4NlQE!@&lyfhAM)B|nUf~P5 zJyO-QP1+)&Es`+UPq{SZ3w?nx=M9LZ6>atpp)!%upWc>8bJ(eJjTx_|lyf_gFg-D@ ziKb>ba*P=-VQ3qN2?Rb9!84?Zj}~764MaAFoj0oaQydv^3GUpHUn~+j=6#`me|Z>A zL*i2Rlz~b4mE^aX>5~M}Kh}HB3KO5&hP5^JP-Nx0W>hG=Eb9X%fH>$miFu7Q1YTu)C$u7RfYFjyywk9FSohvl|m;DzAL(ftW37yf%tDMRK zKhw>!%G9qG^HuDO3uRQYen0zx{PX)J{=zwb6`1xo55tZ1DlBJ5Z(R4H8-6%a5Q3VV zsbRdWnqne3BaVWEq-=jhCG`Bw+GZs0q6j+gU?I-H=PW3{>DW(-f<*K`d2g9II&v-# zRSG43NsNn=_A&C@HS(%~r#ui42(x~NmF-yvHJ$Rt*#>kE2hcq^ONam1FR!x{acYXF zFxS_oJHf&T5Nat74)T+mZ$vTGFIGHBdnC z8aP*eF7qOik_+pDg@JY}BOnXnL@8`+Y@q+lV^cE_Z(Dn7`VjJa9TRE{28Elq(9^m? zLm7pQ*mJc?+nY(-YDX6#!036b*?(Q2oqtmqTtbb3fD0XyfH`};x&r37$6DpRpUm)2 zre56g2#k9jHB}umvGu<`TC0hLAqkBXf2?=$`u<^8|8LNL&@|}*BG8&Ao=*M!WhAaG zL~?VgVT4t;-U~@)@sUe8b@BSSXs9XbisrRmxOv`zPt&)L=Eg^Yn`@9QkI*%+B!F< z3ruj%Ch3by$0+sn(uTfp@-KV`x&*IRN<@GzxjB8J=O4R=Q4{X>LI>;ZS(m#D9Cp$& z;+hUz7y>KWwg)59fk4jeL8tS&rlUW&P5SB>bx!DS)%@*8PT|Yd1vQr|s_#Hxoa~o* z(&t>u5?)$)ssd3hNxiJ;5Bj zJn@Y1_9FVUXs%mmTKGvS=T86M%)bxbl^3jzb-I~;8YdDv!A}h=SP+MRG)Vu($^6Qv zcRkN87MeLq@Gi4pb(j+nVh%8N@|j9+x>(pc$q?;CaMHUD42yXOq!ol` zEh5gtS6&KMe4wFkXd&)@Zs@!Eg#?KoIpLgHoj!=r)WJcMrH+1|z=2EVdfNSPO~=)3 z)`~BflrUVy3L*?qA-&Rf)*%x-bK|B_%}`)qP&#H8w1p}>_+qiWrcGXL7`!?>#pi?q zeyCr}y(WKdM>jKSfmWK;y@Rc-vN7$tw=bp_&_S?V7)h~?lXQ|MhVRyb4jAVZ@B02( zLvV9xa*wW9G@zkxQBjc#(9rkWADZ@ylDBV6Po&L$z<#8f^!*7tF5q>;8|BU=Qx6fk zsa{0^;Y6gP$V}dDuE3hTY%T@;$m^>uMe=wPu^sUIE!{=EN+3cgmQm(8`x4NAlcJE* zgRS*=85(JN`!}`tr!@esSN3)-HPsaVheNsK;kna|$OxjKJr8P$YVOzWQrsfBXf!asxGWF0XX>kH7z$17K1-;PanerI=ni*w2@QodBC< zKcHp#-=9QB13v%tLl^sH0{^#EBvk_Y!HUIP&j0?T&#B4${>9{fXcMdh?pjgJ_%UXZ z7GhBmUEA}MUR{lPo0ZELVy>faVnTb%D86-Kir+K$-;$4l1WEFnfBTlvBpHP;_4*}L z(8omr^e1J!J8DMbtfOxpvRcl=q9;Q%z(V(*C9#K{*${SY--)ZngFpyaZtjtr3I62L?-t5lg}hoV zSPkh(K)_9>603-$tao5`ye_X8)y^)O^VmEo;mcCOaAupf2^+;-&Y^QsT2>xaIpqf?cD zkp%YRKmEBoHEWnyV}mr#Wfn;|$sKWR%+(ELm7F}}y)P?YzM`46W{s!5UiJ#hD9W<#Tbh+NUjjd1*!usL)d9 zj@Y^%o=IrP^P*eB7PDVPE)%0MOyAY%sY;hdeLw{XT5SdZlnXH40|XHqcTWA7OBy=5 zvwW%Ct0wzs9~>JK`R z;CzpFN3n4&FabOUH*L&$OlZJ%15Mfa_9fpEwnp6@ujXoar-^V8lDGi= zZ#Hn7Apgzg%YJlT1q>cYAA|j(wASx0&D{c`1324l&C4R%Uu@redFt~CkmN4Y;$Le6 zJpw$rb-*jtOBC=Qa}WURQ!>gm5$s7W7N;2r=D+}TbbSX2vE{#g$z_%wL){CA zOqa$=SxC59%m31T(~s?28-^wt%)O&7htVYUd(6#Rn5*FSUGadpi4I#*L+YY@zHUA+yNYXz#`TCGU zuw+pbC>5!4rce(EuA+dgfSSQC6M=N?q5jBg{B*QZooJZlNaPNeU?X%U;6ZmJlUDZ~ zt-_t9(Tn6?_yl!nEDC7#%fSUZZS}j;z?4Hnld*0oXmHbD1ReOwsrGso@}9}Aa3Y%a z#-vY42}&b;O2)-y)_q^el4(c9*rPgkhAFK8HJW>N)J_?bTd9p`V9LRYmj?cPvA@`1 z(L7tq?V?0n5(=^b)n8Xps5c_teqKGk@_KLF9c7(ukXta!U=jGbJeqSeFH?@&c@(JM zh8AbnPL@&B4$}Ta4xz6tpN$La&nJ=z?YMAl)@ZV_veUN$`ewM-7&?Rs_cyPc*8 zTCVn*>U-zYs*S;=MI=prg}K>D>zjNN;{!gFSBya-+!`%JH8u6zhy1pHP9E^k+o7r# zBOv>BXXqQW0|Q1pnN>k>#0`KhXprOZjI&})d+sxu-^JC2<06$6z;eoQwBRa8+Fa}>kf z3hxVsX$Q|v9CEhe^oU#4bMxvki3g?036pRQowjD1OE%W1K*wFcpqR@m|J)rZS2H!j zEH}I;3^*X5M(L}x%?Q|H7-Ch40MW%>u`HA{avsq?%P z>dgY-6iXh$P|;d3%tOUgQwL-Sh5H&Q0`7?P2O?-iJu*t^Rl08H3u_vudOb|TI$E6^ zeja21`7u{SOdOkjjbQUZsQTSxw_dblUlj)CvM-Mv)ZX}#2Rd``HM?j@P2@1{0L<@{ z{lajvvGu%k?1FtID+Z!!DuM6QeyJa-!GqYBPL zZR8Dw|EBmJY8Eg&HelNGw zlRiKvmK1XBPccs;|N6~jzzd5IJGlR_7|; zOM@DI>IwMt#!Rq7htnwR0sL#54%@r?s9r`dL|k0FJ1t5lL3fMKP6S=X z^u%B9d(LJzzEG)EBx--zTDsP{D+)FBLb^?8ZPJawVM9d;-J0`w4V;dMZ^c(UCSF}w z*FV51rw7-t#)Y7^`&vGPTnUuK7QtSlC$kaqVrrHsuM*FG8s1Ynb(lY~-LXt7euoOo z8u%q#Tm(m~8#SGKuUH~Yg!GlTd+_yRc`_x5-O~H=E(K!v!@cu@iN{|KIzwsIqV{8unNeB*=x8az z9EvE-Zw?&`=u-q4Jt_WuuXau8%Xmgy`1nT9M)%^xZogao=ab`gvH9Tf?@{w|o`(nX zi6<)^v!PWRdV(R$haaOh%smhK?SLqP=-XH@y2QlFXXJrM+ppe4*77IE?}9nn*1cv0 zbH8rim3vSjmGnohp@0{0rjI-;Ta2+ptk064!r1`8MxuGq!k#ne)@8?36DvY{E7W4D z`9n#to9ab91#K*bPN>%U0ll}pni_PU#;`fO+6L+8r5Fc@R z(rC|Q@fs6B*(;TI>lt0t?{Pvg^NVRrpA4>=ubvqLHD8(L7o)Tzb&t$z;u^#ydgy2p zx${F}t>GLeZuwy=p(~y>2d$L@bDuZ{j_;k;IBSlkvh&_kRW2-;Kkn5MTyY7mOv}zM zXFh6iJxOtM`LKuWxgrY&?snIna{?^lk)YS`j|_CNj{XSDl7YEsupxz(7Lj&Ra}BZx zIDCN^0|1znSoW#xlRYCgt~V{KVPPu z(8`HRu5xBs@;6^S&e&LIQi60SttdA$te31;>*o|k5H66gsm>JV9?ObOL>g+XSg#%8 z8EQKQ!6JK9N5&Z~_jHTyC$~y!1w8HwGSg`jRNIgku3xZFO>q6)A?W8`7yVMVsJEeI z+GttXou^aTGf{ISuu;Ly#}~_duf=8?* zaJ+EBIQnhp)yX#c#+&*`N<%}to)+lX=W+PKx3L}YpT#yX46PT<(IoB7X<(C(QxPZu zofLIR^8-1A=Zg)mL%UV(Hyd&A2CHB;OUM$csY%mLdXjcTxY6gm)!sy?p$1Zv< zc4$uP?Rv^+cGN_d9{f(~qzPnZ_|p2OvGv};=6-HXgnZeE&PFazwPP36eEjO3@Ta_^ z?~Q61&K$f6d&P&n;!{MlEkj?(`MB)j2rm9X5 zW6lozRnWTT+t01`Ff&HBc+RHp6$l#h?vrE4#-#5RzpEO5Ib+_$6v0{?2%OOZ;g~{p zB#H7VEN$Q<49{_lI6y331d|@J9*vk`2lbdchRLOgLHpRXKA?iZBR3;oz!6f6Ne3Bl zcs)0&s{^@U)!?qGmn8GQ)eWDJt|XdaYWB6Ilw**1&`4_xolRdL%6?Y|eG;rXtjQu* zl3?cA?6PjdTSg*Yr0)xwQVz!k-)dW+?XGv_K}~jp@5|ych=upmx%5>ZKG2cVTD2>J zwzUNJJ&xJT_)OOE44x8UT%s#D@Vnuz)x<1Yq2Wr^bD1GaW$^r-ve~V70Jk4<`lcGE zUgMzy=_7}zmDJe4fyGW86|}{gl}>z@IAH*Qvhd|^;?I} ziB8_jFR0R(`e)&3;U@9l)tROn&&GZq2rMvvQ(%E%K*Lgt?Socc1}gdc<1IH?h?-g< z?QC{}n%Yh#*H_&(?8NMB<+^F-l&Tp{@yeD&|H729ui8wMi1)HMF{?04TNqx6#=J1= z-?4=U4r5|32deVX4r_X_EM$Yidx|%zDgxyIGMdO_Z>d}*gBa7VN996zV3#yfI^S1wATb|1QNf?#Jfa6@ zi+7xO+=K4ntO7b_IT80*7Tgi!fzK%DMFGdvqdl2V>;N;ZbWmnKkTASwx>M-{aUGa= z@pyUo`}ad!Ix{YS3s(R;e3kWc#gA+(qZ~Qn9mo|Q=lpA&jJd@P?@KL)-V8kM#Rp@} zrrWLK(gN;fjmL`-!t{v35}xGWP*UQN!8@cO3u|zdjl&CzuXlyNiIqcb{L9Q`1QIRF zw2-N$%9qbEyq+{VJ2zE1A?VrSZDQv_;R=1}WgFWf-*^4HZAC5y60{`{F`m*pA+PG#@NY&bSCJ9aAcBXA%`m7j9Xy%fNyEq=&w5g_4SZr4*2#gR) zHH>Eo`2rTdCf3!5%bx{vanaD1?wWAc%V`A};cV5zlXJ@O47E74nu^q!`I_^O86y(I z%Jkg(vqXsd4ns4Ddmp!v?U4L#L-a9_61FyJ(kqEwW%tK~+#ZdOsh#8xWOyFN3dvy- zv+;059DjGP&O=t>-*_*Sh@vF=(KYUc8^{ngTz7Ju!j;ideKfWiTeGvj8D%{jbDPh? z?hy>=-`lrB;=wxBx+%M?>0ll_e2WI@ns9AD=`S{@(!Utxs}3LDqaWVLdLWbt{Akvh zKj-8gN&=i4;2t~7Qd zDW*COHtv$jf^x4eb<-fMajRLQy#$0^sw5Be5hsU8Vq&9X8uRleXVsT%y>tQOjP^ffVZqKi8#;w^rH2yXVC{gchHDmnAnnK3>{CtH4 zpHuscf2>$-ee2RYVMxVW?R%``K-TZtvm}S}VhM(Nf-v;H7>`rV?X{R(yOUtqoST^0 zqIIf73{LKR+4!ZGfIu!NN9mYL}+3URPEo>Zn;*fnAr zGn^4L3;qZ72L^az74mSkS2vosL@SB5^7+e2T$;lsXPTC6UHUd`$J!8i3HO)=@(76) z-F2G{)hsupg(SwH$Wndz+EytqA%!-n`LIfDampgw7}bszBbCPYW_@!m@72^etW(w) zTR-Mjxyx*w1UaG@Q3Um}Zu4%}95S0+7yZ+I5CUIwFENXn$*fUJ` zkfA0l55BUm88bN@sa1-fVcOHMcIfu}n<9Dr=g*SM)!2xysP)Z*MTl&U&UbMw7|zYE z)*9k@4H|-*aVZL3KXToJ81zfU%Rp%yWo1+Dhij!t#7DluTHyGcR$2?f++zt5G{}2 z+p>5KL1i}S?O3O0bP;LU%^+^=rgcrFYshfUkAfvw!tDd|I2MdkHv1NHj{}CrQ|M^ikHf%3ND%j7f##k1rc#y@T@TQqO=G$NQSf|sKPe*J4Z77=wEAMd z?_O#14>3MqZK=l5s)o?!%R@~*2*f2t-(%)tqvp@fUqG^Cz&=)Te zxbYIAHps@V&K_zsQDAE#p3RlN&$Vb*bo^JN?K{HFP%r z@v`3<>dxd4-FQlM`IiW%-mIY7cj5?4C8CzReih9^)=CyUpD7#4dADWWSB!S3ZbYax z`|?8x}WU4&cjo8V`81_BwTRcuOl@o@yN@*ey2Lo`s#Y_svDR4;iFY)!xXfPzdYTxN=goQ zc($g8_lk1`%<$x|Y3YAeXw!F{6ZpBSrhd^7`JlUeK5lpx^^LC@l6o&%2yLF_2#jfB zAq6$d*A}wO)%*dk#4E}fxRoR_$bw2a-p#T^qm5dX(j=)WS44KQ`d~R%E*Wo&i(aVN zs!=InI+EQSv{83%XY<(T ztXP-p_{VcC9vHY}?_{qmk9s`ku=rH0aYer5Bv3HQQ=CGsNIz-4!>CA*t8yTdB=3({ zNd$8c?y5kFgL&kBs9~HcN5u6_=cE=zD5hqj!d-%ph^VE*-_xpMufN!7wc7r;Q7FP= zwNx(@?#X*{_?pFoL+kD2SFyt3gL=Ia)#bzRFom3WGHzBSK0USfDUP46gAp(aNL#Ml zEM)$X_iw>2dol9b;eq%|(H!nBlgi-JV4u_i^DX7FhOq#%0+9>k3j{TX){9|5wR_q? zNj&VHOtp#yg5tA)8nfW0WAXe2z3?a+1^@e#@WZI@Yvp}?qz02PwE_^ZJWX>fyX;}J zo=U^dECio)75c1}R5qay|Hkw9i>F7jq_S&b8|=jS1f`=x$a<7JVa4X-nAr|cAu^x+ z^eO#~If48&`Pv?#7mxBjoqKrTaCY+tOtE8BlT3VBU<>2`xjJ;xql6@b^8tO0rx~J_ z!wBuAgdx+BiS5mpFLUS1yD_Uicj+#c)QYbSY+6f`WyPPVbFF^;(6o?0(IOWKHN1yq z#_dC$9p`mzYXiZ3b?Y}}5fxcs18h(d!TMC2a7>491BIYwi87TD^=iF133Abru|Avk z*T`{P{C6^IK%#Gzwrvk|nexYH7qag7ZULRXjSF~5xAw#MB0Q*D3~?MHN1m{FZk8-4xhF&| z-X;6ZUtr8}fzdrTYu2eHLO#7BbFMi&Oq|%N*5o2Rn3L#j_YdnKv)ANZ_)NJToOrki z+w9fbEk8U=&ew~{qqm&G#5b%A$<8Zht~trc5MT=o3Ywl>>qvmRZfs&40nO2)F^BTEl`{YGqK*Vko63`MIy zKkn5FFB;v6o@N@r7rCnB0mTyC)9qDCxM!8tZa`74EkB;&HtACBUS4JL$N{U;tp=o9 z%$GM%^d448TWO*;%>(2U`}Px13c{%yQ%2Xh>t`K_XJF!q_2`sa(izF73JhJt!e_`j@!^;`0n3- zlkk1FF?Hv#20bnK`3UH6@#JVxGiF0*?d>Kcs+#3@ELn`W zv}R1HK5h7*#d8Y12Pg>6BwANHxGq;xh+W#=~ytGClPt ze@-0t{K&_>N~z|Ju3LKyUTTq=j+Lq`nLCU)mhVA014M!5+AJgD#~@f}`cTJXC57zO zE+0p-Ky8R%#A3@qeC|UfL5P0xdI^O!oE##*`A)mk?)QpKXEtKuZhl8!jAQ@X;O35r z?cmis9oH@v!4yZ2k#?Pu4Q`iR1tIQ}D)&aqdC01Vkid3;&upU5?$_W9gBGA2_Z;%P zj2@i3<~UO#F5$M2n_b0zOk(Pybs1efCppwhjJT1+Y4#9)s1Q!XpiyVOI_^HB)qXvP z2iVY>TS+?nz-5?33FnY25ko6P_e}5~U_gV@- z_J>R050!LfQ&nk_r*vh%mD+W?K=&aBF8Yr+*{a`MACHzjUZDaqMFKcRQ4{{4NJ83f zlVrZCk5G-ulsr(sSY>iR)fh>1k{+wyxB*F@aZ-GTo;vQ-liPpM@=prn;WG!S%KI)p z9mx!!r6Eg{e0-hX*Y%XY>zJ16rX*jxo9((Y6YAIT)7ujIY!m-D%K;qLXXW6CZ|_hI z^PSqA3X;nhUrgBB%l154avKRU=AXGv=XEDxeW|F2rFvJTB!XeZc8SfU;$VC>{E0=A z%?D$_#P(L$mGUW)AbA#{17uTYXKo$4Wrgm<BAj z49~vXqniRP>pCvBHYU8K8LVm&Fsx5&QlyhY?c(6GtiEOA=qp3I}}K-*QSxM}xwP0tilK z`WOwU{MYgB`j@2X9%yR;TegoPDN4NnhdkArMePl1|wV4<&TqBbX{2Xi$<$)_ELmem{Qaj|H%CJ!)n_~mp=13FR_)b znn(cm-Nla|M%KtA9*^Sps&~6|O7_)S=f>L1qj?Xdq z)M<636m_pe?``&!Ilkuu_=1ye)U4+!-w8E|=RoE3EHta$`s8?&EHNT~mWyw*1K2;r zMY$cM1TQJx%Ehg;vM4V3%bdN>N66MA*BC$&(CKajy?S@8Z_zFheVjk~UCeV`o|fIt z_H`s>M#C%5PD{Dqb^z27U4$zMtuoFT1LJf-%oHCXNwM<+dtJgbzXPBSNV`~DH(frM+(lb27O@SHuK#)P7UcIr0tPuy#2CAgN(Y6I>N6#}f|#PX1uhnt%qBd4;Kp*TpD^Th5_8IDk?W}hOE2CU z@e>sO2yR|cVT`!OMIZ~;VT#~FABv_3`-u+_AuEdRuW#S=Z@Q7JF8dj{YxwtA2w2b& z%ZpjVmBC`{8FJtB>x-krWD~3`z7%gpmv_rms2A=j7#bSR&!q%wzA)F>Zq(n(W09Og z1PkI4tn@gVbN&psfy_sh;%#7$rr0HY71-06B~Xwj>ne&vv&71mZ1>k4{>XX!s(Ax< z@B;4vjCaA>I{ z$H73PAMVW5`sZrtX~|7K0LRomB0GFJyf>e=F)T*&Z=#+41;BBDaas(*e)u>7cJQDs zrK-zvHOCoh6;_k$RTDg#7W(razN^p*sh(w_u_2+;mGn*!jXs1%VNi)K)uCUdtB+KN}a7&(lfYR;Ao-@qO5 zjI&$e*?=MXkwXT>>n=Vmf(l?JQvKe(XXoL0i>#cp8Yq5Co#ErpPn{W1y#lPZC5tU( zJR_BI)dTJ8`+kcht{Uj=ZcA@nggC7USH(Sh&jxpAE^zIru0nZrJQ3*E^ge;G>CAq@ zW*>F%@UmE+7=}%N2nRbjH z0cWcA?Sa~#L#b`a3?iS^NW?Y(ndGz_4QXiP%E@Z3fPT=;{_=^8r{qy3P4eKxSV$|MJH1T2!VMh{6*-7lGn`6BS3 zy^8)=g@cPbTd1A|TUXz;o6bg0Nw;lq4B5WitC&Cr3I+BE8fRw$J1bt|(tXPTS&BjW zB@~-Mi1piKK;bYuR~hpIwvB~5P#lgMXj#1Q&~`RU0pQK-ztpCx(5XJ9h4j((=ab3! zSdJIpV2ajD!=<}sC-6mS!7BDo<)uW_Dy@c5&d^_kAp)dU~7FQ$-0#Vni(nI$ph+{deB%iL5cgg`eI@YU-f7xhf`$BgHR>A7Kw{G1^6L9?oi}P(d z{&|89KuYGs6tsp0e(`H+YAQad6=ix!0o(A6phZ$*#8!d*PawgYZ-ix+Bs%NH-4D9K zSHt$|JfWbED%=>M^b>D;d0v@seXQb z796kNvvUZ$ao=I(*IrggbhFm5*k&6|2V$65;-EhL1nPT8WMn1HJ8+P zzO^N$c7am;$zX<43?zl%+sjPx-(Lln+yf39G6kkr|2f3JP1yep`irK*#z@}Y-n@?s zlrJMO*y#H1iid`U`&%jCe!1JR;-~4PT1|WmacIAU`pLyi1kRYdIZaqo!Vb4@q^eiP z$DMH@$q+GFlkfjMgwf-U*8c}!W=()Yfa-^Tv;XWb`WgV7Qx8u|?>QE5vDK&auG7B) zh$`>G!2ZdZi~T+zY^r*-JKNbp*TyWs0QTK1+yBgZFpB6u8}v+%s->d)TOfb`0f0AG zm-xg-!jNO%%{M*mqCJb4Fc_58-+Z+DM-u5b=M548AeL~}8)N3L0N9mA_IX6WC`Il; zhyAtU-hIDISSCCIpiuz4qowikJR-94=H_=3;?wEetf(zm7Pxi<+@5+tzy=UN-3Vc# zFMs_?&ciL>>IVd!h%-_pselK^K4szLnn_9g%WR&{eVBo@lu*pI<%#C%DW2#j27}-Q zI5brKzC|sj{W+TP-@0k9L4IP`xWor){BoUO>jROUz)$cZ$({L6s^<59r{mv@nbSj< zXxmvG%q^Y)rbUciSj3uj*fZb5y9xEAl`#}$k0pKD|~ zUeS2x2!DNVLJyqWrwfifm&_U@gL$a47>*^9vZ$iia)bj_`yX}aznKW&yQe3jjN=)r zxe{IV7UBIhj=-kPBr4VR%z_CAUl)C5d1w%QG-AlqLwvD*0CW*vy>WBLs;4;MH%-0J z2zKh61_@iLM3;#0PbvfFN4 zhZjPoF2oX?X<9{XXQv`zGw<7)MwesFe@i|(Dvk(t&w~dVoucUMl9xeZYUxn{{gHsD zv9uJaL)h>0Yoq0wHh@j?aw}7V+w!$CE$N}IJ_HmPV?TqqgzI0BK7q2vGXSGFUprUDL^N6Bn$7D;a~mSPKX3+rtMQK?b6LV}`K0-%nZQ*s zD`sQ#i|&A~!G**%&(Gp4+^8#j_PD2KMx|*X-R@FAM2*4@~W!-J4iX@t{{j zkrOs)OA0JIKCNBV#$ZIheJdCC*o(HmPJaREVm;0QEc71cM0umr>TN{e)}WvMd|y_d z+CL~Tg7$LjcmSB_o!#5xBEsaajn7xM8Ct82@mg2(m9W$pFfrT>r+1o4tJ5jtrh8fZtDjefA`I+jFAg zq4rO&{$SPLmaT#kW1^|iX(Q+Za7zqZ^i3f~Vh+ee&7Pahu9ESbr==XJNoi^Ns z{%TnfZoEndM4FLf+;`5MBse_1Xqg`g4#N8T8^8YmP|TB+d%Hp7Y&5L#x!r8E62K6Q zPWNep-*}v|N0{a(bn?2E(T`f=pQ{Gf|AbD)0ed7AyeZ8GJA7D@O1iHNoGFy5 z>PW3LvVls*{{pRSd`Sj$cOWGI2camB!Zu*w54}ofOzV_K7%?n){8;19wfx|z{vKj= zkpgfCmM;>0(3nT-W~w@YPS1ie$$+k_PGYFZ`A!dj*kvpIC*U3cx|A**wUv=Ue|V(s z^_mTZ2^S{qr+G8gq`N|FztT_wFX9c8+)oU)Ccf{9{eFk>qOS5~l$O77bAM-hq|{o$ zb!Wv~zRLR}qAICmR~&tf>tfI{Q|g-^gr!Ad#8j2M6rlmX%ldg@OwVpie_6CELKwop zTvNrBV3B)nOx!70Ki33oHDUZPH~%s*H=}FGJ)JzaRMgOWO`NDg3_8Z)^C&`xUBVw0 z7|$rE?pI5Gavz4p<#KVb_EIDpH1LG)r`bleh-xkBWKEqOg~vK7Mn9nO%it>kgO z8?7b3p7YMK?waY~<|T^SUr?Lm;Z)OIejL$2nA1Sn2aF?4h|db(-DD=9^^MX5GpA|` zM!+KI&Fn`YG|>4IwsEcyU_oqxy44lKAfa#j2W$ieRKaR$h;)!^x@%*rX+RqUd@}g& zCH{-FUw{_F?trav0$rl=s=&fE<>JaW_m`-mbTyCX;x4u`|Io%cHn4pAveSL)cO(H0 zSLue8UZe!A4%%q#57*BKW1^&7RlDZ6ZM@oP+AukdZ%P~I~n7>1%;jOC7HTJ1Cw;YD|Cuc^oB5`*{CoXybz zM`V4=4^?N6T(j0JEo4Wo(rKD%fl06G#|Ep`U4+Q``Re&f;<0_p528N>r$qOxYFol} z*5{;S@STBl5uMRVtanV*6B#WjJQ=U0(eejlCmtDTBV#L>%{pY4$k;E6gXOOStX;wG-V%T&Ty>jWza5?@`Crxx zz`W@(>^9ldCII+ZjE74iAHI1()iz^-xfFvkyQYf4ni*btrdt1uns#U&P#v`Iz+MfN zZ@v8DlfON15*%^jVhE@k3ZUJvn$5zpGs;~Ph9G<0V#X1{s&@^|*)K06UA5a$ z=2Ud;p_-^-*KwKI>7mO=h_Q0&RWdbkzK&)bj^^+OX6F~D&*TYw`?{>cSvCqt|3;*z zXY?$!S}fGiP2K)ltiGfeX$7RA1A+W7uoWt{8mnNxZ{SfTR!v3a&%~cszCCQ;r=m97 zVaX;zwi+0;0H3s58>iK)(bMaO;4Cj$An@_I&Ygr$pBT@4X%x^9t{21C30HHO4wXoq z3gTlX0e+btdEYX9<~*(crT$&jvPEp+AM^V|%Q_{=5ov$iT4oVmw}ppn(7=VLNc}v$ zTNE#((0KK@7`Vf1m|KuJDOv23Ss`#DtJw1d?1uRYIM$lJ3|ttgf1w$k>%UEzxk`(YBMm-R>1~BlZP_Jmoyx`P?+^ZXbprPsGHTBj_ugK5I1ct zx=)0p+nYxQ5?;X5t{6;8RlbpNi4*zBO;%sT+zx)9_76j|*BBnPqFK<&J?W^_MR6%B z+<3&}&PJnGt(xHrhWEaZ3s?O>m}b5O;?QwR@Hgk)MCz4!3|A%n;IeZY_p$u_6aIuZ zhwm(?*0lSb0hrbv^8QbZ-5-|V?-{+KP)y(p+UO6q2&&-JgG=Uh+izF#tVzSl)ZjuV zcCg!Gy32==?F!gH52%{V=kEvn-RfOa4(pF}ovjHZq0G_@z!7Z(+YkzR;cLfKoH%O{&p7k-5=M8>t zGF&M(TZaCP`CFU+X8pm(kLbusb)#-JIP?BHlOUN@ABZWPi_kqn_+h15I`)dA8?2+X z0&lpI9V{HQ%${Bdx3B`vEWlQqx9t-jE405NBeW{ADZqucejddm9ZFzCxk^2$n&Kv= zBSzDF)$LOtSK6Xg8CzgrKzd?D2b?f6bhd2)GNmZ6m140jDtBxeW6ms{DrHsrpG8nu zUj_MgJ4tEjEjzoFN00pTT@({G+6_qwWgm9PsF_;q>9V2dj;*J>p*m4-SF4_w1M##S zZlu){w02a262(vU;!jwPzA@-UuDG?5@ODF9R(&Tii*2WT}ETO?cD%+4X>sSU^h8GowkbO6@O^z63X_%q1C2C}1 ztV3C*u}uudU}m0Q?|VKy=bX3ad47N0pWlDqd->kiec#vhy|(XzNTv@X7dDdUE{D=L zSi)lmz3&ntIXaH}Dc6YFGjulMBjs6RnH}tDJ2CK;PfC{aQ-@oI z!YsuXpLL^~>`K4)D)#4rQR3oOZoS0NQlsJSwo=}@gH&`Jv+R+m8fk91MIK9l0zsKi zK6%1j5LH7nBrk0lp>hj4EH!HakxgRRKkIb_>Hou}n5$qtMPztRoc?-q-6C3s^%)S{ zd?hW3#Y{DX)YN6$wAue0J$Q=W0zCBv<;4I{j;c`fFVg|U6e`ss+GV0O#;I+40yU3K zx>H~!K>gCie+@9~Fn3}uflvCnT19WKMejQjAFOXJJY|Msj|6@kGCi?gZH}I6{HuKL z&w>5lg8A8Cc4&}3wZCzglFoku43CZDex{Wm1`w8OFXiHu%+e$MU&iU93Q+3jDb6Ay zWH`*1Ko)Bw{3xzD>FblGvf%*Ydl|#55km7`Q-IDhl06?EdmAwO!T=~U056fcfgZz! z;~rR<6gZ-Y??!8k(;{~CiTdIQJNZ@LiQ;TJI|c$_&0dAP!=UcInx?DKST7Rjgz;yV z966mkyvGNpqnHD*yfs8$B|NxsNZ$L-he!9D+y7&o17NDL&eR8d!e+)ENEs-BNH~HV zx9pSg84l}1uU)E4Zc@`^Dda<^9}Z?^g^Z~@LX zSR!r&up%aB`r||T_M_;1zawZzhvvHta^PhMjD}8@Or61*p43Ty1^-1_Y<)Pfck`7~ zRK)I!1l@v;kns~<2!xd1;0?USnqtG7o|h_(tv+>y_Kzi1#KTsK>HE!lJdkWxv9gu* z^&^IdfOkByh!NaPUGbryS1q*q{}0t){*=ld;4>Qo6$zpsM10)hjl)S=%VuC9KY8-< z$SpiLD!{zm%252vuV(g&m~nhi+kMJN!O5yy)a-@U+VHo`B*C{CS@B%atL`Avu=dG7 z+YPLZd|fB|0fKb|Od0o~ni8N63a4T1G4+j~=pP25r{A8nUp&gMb08Q+yPKe0p)C>I z#|&{pNphe$j517&s?KcF-r`F^p5%8kEwr(8_69CykU+RQ-!+_{7H+T_IzQd)W3fvE zaV#R-T0Zqu4KE2BmX#rUJijOX|7)mX+&!0eagFB!c|wS8r%pS9a&645i~6p!p$*#I zFH4gq@IUkNoDc21Rat6$v+4F)aHal*7Q9?}Hp3fVJMSfJ)*)%AaA>(wQwHdKSdS&I z?rWpjBpMPj89jHF$|NA|+V-YX37LY7WCk3zTfCW`7Nw#?8@DiLXi>W--W&<+G**oY z=ubSucW*pjb44R){qMeLq8 z*zMl}d(_xpa`n&8usUjd3U_A!cO~lI?V{hd(+h{QD5Vtl>TVG=Xs)|=C`L|o_&PDW z>Q%B^v)V|LVce&Pb(jMNFX0>3wby))6_&wcwZDi#6$85Rso(iW6j#XY;IFiLEBw?Z zoj(Q<64BRA&P6j@88I3xHoQ|sy-ck zMui(iiHUg!s}ffLEhg1zh9N^R^zLFkRk-o2GcB- zM=>=&d35(=*{J7oaUHBOF}QraA;gtRc^84${nE!)uwdj9kjgr14QU0P%U3J$!Icub zHv@Wp;r09$`}Z3G#?J3&<4nI8%=_k`-{T`lHKe3!8c>elR=smWxm~Xd)gC+Zn~HsP zn5Nevpxu9S<-`Nt4rY(wUd3RiR;&9}ymQ?DN^n_|9BGRV0Ap@mBgAau&Lou;HLQ>a zoOWMrZilh#+qOn0hZ6`K41#`Pl6X9NQMtLLn`<#dG%8VXx3`k|RecWYo@B2VQ|}WZ zJ=%z3a3{h1PN(fvT2)%~CnA&CWgg?)M(St0CJe-@!R62(> zf@p0Q&o$T03x&n5Kjt&T;(~S~WES@R#>xEGvEmP?VvXqKW}r@Hkts?CgSGe0iGgIsE>LP-PfQ~SXxrTdWZ&4dAi;D zGh5pRnH~!iyR->D!``G4kiV=>VYOLZlQWs`qSu3cLZVj7;2uAmLJbl<3gasP=jn&USPpSI7o8UYK(#k3sFN-qEbtMGfw0g ze~EnSuOS9^ZI8_0vWGH1eg*saX_Gg3M^xLaTP=i{Ej_rb7$>*1pa8CmfUJ$Qw@h71 z&bCo(&8COu-mH5Dus_OvSjd)IJe(JKZhKYxk?syk!fXb&xnYMTgfC_iueTtL4rmsI zc3S(8<|^z)nGzht#V7JQA3X`sx$(otgq${NH&+KRP#R~JNwu|zaXASMyc3fnJS)uJ zn`V3fNJbg#EhH!XCp~C|n09ETYvcbU@<6_H*cO+2S-1S&{b2owm+SUv2?>HO99no0 zJ{WhQk+Htg?B$xom8wzZgu=}RQ+da|g`GO*8xf?*CE7``b+U zdxIsbB!nIs4FwmX{1F&`)pM5x2fW?OQ*;I}5tYJDTyMvtquk>@;Hx*wCBF(K8`G>WiT|OE3RfXb5Ss&RafBt+c zV#2!QQSIplh$DMife7y0Dpu3d$|zxHp`4CJ37lFBrezX?OW3?jytR49RambDWImP} zi=GBgKWbLE;+NV!P)uJ2Al{oBg=2+&lMTvZ(L0O#c1__fE|RmCadT$VhR1=AH z9~lU{gH4DC6T)6pRkh|Vdk&;BD2Cn4nIW_A1}6-D5QVUm!g!Cks8ML3ov z)YD#*wci!xPqU_wgO8W=sO zc)Ug$$M&~+J*9l;)Of?8l)$6@lLC`@O4`;RNF#ST+i^&Uj9+V8MQ!Q`(tS zF|FYHu0C%eLvCuKt&=Tw5yjEmSB{PhScPEVKe5WSt)@DH38L`=BbXDHUb(LI6}1S( zC-v1VIQS$-^&_*(uk!Vui!@4*Zy1x&4F8n3xvi7)(BxiypNFWw_Xpjua5W{lQ1n52 z^H$N^*KI*M9qS2N$)Ezg+dPV* zl-o7Y#|+0D=v^D04vHOPc{6^ojwhj4<>}+=9IN*RNe*j;VjD&Gi*1rCa9AQUl;AW-d%1TZTPS?w>N@qbvXb literal 0 HcmV?d00001 diff --git a/api/core/model_runtime/docs/en_US/images/index/image-2.png b/api/core/model_runtime/docs/en_US/images/index/image-2.png new file mode 100644 index 0000000000000000000000000000000000000000..c70cd3da5eea19e6e3613126ba7b42ea33e699ee GIT binary patch literal 210087 zcmdqIg;$(S(msr9@F0N@B)Ej&0fGbx5S-xd?gWP+1cwkDg1fuBySr;}cNk!p;TzuF zecxx7=eK{r$2oK6K11Kt-PP4qS5;pfEH5jLfkuo50|SF0DIux|1A}-A1B37d1qu2K z-VhEI49p7wa}g1FNf8lpc?Vk)b1P#Q7>VFmRb;iV0|cp>iWCSaG(poaio#~F z*@PXRCE)yU4aoE7C&Uov-Vu~%6j6qdV)Gf8Dp<`EqB9IOs+BX^DhkbcZF}s3c2*xD zjyr;;qbZt5F!!jbqEVq1a1pfcVtW0N*u?X)*(Bv)Uy*;uv`Zh`B$SwEWsQVcY`;6b zIYGMecb-uwy&i)+T4=!O!;-;JA#Hs71S^QO0S~jT>=|5*4U?H%_gjfB(s2I!ua`6^ z=bk*HN*fG3vZ@>WXlZ^QFuR#yB4*q^n$y4#$5V_kZI}vUgy{sO^uR?D!SLc4v^6(- z!giZuFA0;Ob~|{f++*=3cuz@_O>dpEAy+uOLwk!Hj7HVaO#f)g1F75I9z-%w>&i`~ z#t4lbdo_D(A`9?)T_m#nRyI5&K#gPri_)i)%pfL_8eK~1KI`)X9A@#Wao>6McJz?% z^C%nG_BKk{A#!8b{&o}rj0h=A3|llp#aQQ_`k(BY2Tij$1ZDBjOU5D`<{2QbsAV44i%MDz_t`RQi{MM%eZ)@p4<~md=@#lv2~{@ z{F?SBfnrfz;r$$Sc$dLCrJAKJ1y&PBGDIb1F(m#pXfq(lj(A5uuiI|$6Th=dq2+gaxC|3cnq zl_zD`bg!qd>t4}&&2@tHwn-!|Yfw`Me@AZ7JE;T9MPystJ~usvA_gYpcUI`LSRs93i)v4)zKY~z;56H`Z~{sshpKSEg4^EO82OtcsWvg_b0bN4Apm&h=!eIO7B%X_yo9KL)GPXH3w>R8KkyDtyF zzD&_7t&zQDcjp7MhWqU{`xOkH9<0?LPp$9BXRxs3M1Hp~PZ~Q377NjBy7Tt(=Nrf>W+8y*FAz=8^V5D)sh7q2k?~xcL1#RS|1fK)C+xKs8 z%Od6I0t&Nu^I0vzFvZpBNWz5h#_i2$cyX9N?`DRM&zNyG;+2TYWfD#tn$h0F9e2Ux z2JGvb{C?Zp`9X~%F2q<*(F#c=Kv{1_jdT{_^isJP@32E)Lp0fk-Cn%@N7tEXa=2W# z={ov(&8>6u8{3Yu4XiVr2NIvp?cf{%h9h<_=w5s4<5lg*i(lD(PDnEh@N zxsrGx=13TmA}^|E3)cbB4f&4f4(*Qqj$JI;K8a_aX>JP(iyaFciwXevPMk~$n7!vn=BOuz3rir5CB7zB<{>c0+q#6b zn0Y+4s6`zVbMqFdw_MMT$g}r#FF~*Q=XgwozR!V0(IsC?tEBQ|@>Eqd=VZ7gT`FyI zP83#T`K8A+!i(aI>eOu0Rxw)SM1BUA2ciTr5h-$S4MqU(MXuqm{O^tLo36_az1j4L z5lFU(m^jVe*P8~BERysR%d%i{(ErVU-+pRWTAL;*VMuGlO8+MDRf2!Q zz%W-rO@bF!9jhD%3+tLSVC;uQM`eGxn*M~+(u4njAM*p3^NXsF`y>GX*6^le~)wOFlbKP7VA{_)N9rI?!*1o?a`SV z*Q80d-Bw&=xaWwMTWgXhANV)G96$o*1&=(qL1rLW@Jw(Bu$u6xmB;f7M=3|~KBGR9 z?Y=&zJgdq3$#|dC#W_WN&1|kN_B>d*Sn&)o z`)~~n4Hym6y2FB!2wb>QaRey<;tx%J?S4I|F1Xfs#=#?!SJ@SkoRX)}zESbiF2CLg zxZm%E(g)CI#`G#FS|}dOPPejl7^?i*Cq5LY{dN>)&$USt&L_BP=Xr9Qg|~<2ig%NT zGZ{7IobB~T?6T0!Yk&Ew@JdG@ur1_K7@J&;Vvpb+Egm(^X?Am-3&1XzHU*Hly1+s5JK-8YL3MA*6Z`7bZCO3sq+Y0N{ttnlk@1&npCn>Tx@en^v68& z((`3?M=~((y_9vq=Yr~L!_={6jLY-8%sXpzEy7~;vl3}}=Zxis=FEp7VMr%>+pk-v z>oRaC8$k{8jaoAR)jrsDAS?$WRmU==q0)RbaiK7DsQO+tro^?$p>yyO-v*yVML~5y z#i{g{)=GV^!$=gT?yN~ki_R;xhNi)Mi%&BkryowLv&h9|OM1;2o?UrinPJ@A{x*eW zwWX!=(#u-acLQbHlHw9^mYfBL`;|sqh z=K*vN@{ZARQypEhfpw&;e2Q*Lz`YH_X+y{GYcdvrRTp80%w40M0r3&M#M5MAuLBJ! zE2*XZpsC$azhoY#IIo*qo;64DQI9miZ$3^Le=_@e?p1)ZzHLp9fbx;75=hi;PVbOT zeA1$LEm_ls_x9FIdXZ_VeVKOi8>bfV=Z&x=QbRsoTQnPirdW@|nd8i*wA0~6sd`Ok zxl5fMuPo4v_dK|wX)5b-&o>>p1Z$IQ)veq6YQw%`lSSJxZ9G-!Bd3?fP^r9dh;=3#$-S5j9(SH-&qw0On5v9tndh5%6I2ERU_>6K@^dIi(GiUh-q z2!phorqD+R^hOhj@lZj8dqiaf2BBTTq+>$b`fl3Wk}zRztbCSO^5w|TgG;tZj3#e0<_A$OJmJOLxeyM1B)UA1Bd+` z=5HQrLWCSjh?#x;|L&jv`${{m(E9G@zZLMWHq*jzC=WGfnXE?t@6MoY;RJ#IwWo9f zXjf(agkcmS|F0oG^@x4jfc)R*6#)b1&D?DE;?@6S;{P-5@9ks!|4lpC`To$ZEdPiZ z_Wx_1f2&CZz{!8nG4^*MW~%S)y67y#A_4zx_T*=dSbr=3Un;erfJK04xK0qU{moMT z-w1uuJpP;V$mP*d$6+`m@kFNS#V#s|YK$>`wp?qc(YXF^l9d|WYo%r#Gv(Nu1uOKCC=0(k@xLR6F} zu`!Kg5em-i-SqhoTNtw^2)iOIiiP6!t4Q_W)y7J`y{{D(e-6FMIItN=@ zRFaaCk=4^m0}^u4!^7VsCrKI!)S?=28V)Tw{0)MhEtQWN%J5^E{Ayz4Te8tXvEgt; zB4ctG9|r1lV)L~?q|N8pa=w5HOM8`v$3{mZF-RN>i=^%gVByp;*mCr{q#5t$ zgn+Cr!R=)13~0j7j0!iYJztKy|91aVN5N37dF8HiLFLy&CG!;yF=B2B+|BHuPtbSr zap4Y{TWvOJ)SDaScbMKqBzWWZ7q?xuwBDa2(eIMRZIjORrqmo(8hhP*w%t28PKk_s zw{Kn$m6%95Fd!-lT%XQN$;_mrr=yE$et;Vo!Dx>C6R?)A55G`sFXHmsP7?v**F^OK z<(Y}f3&0)JJhpdR-5E;yl(@#6#~s(022@+?a5ozdiTh4McP1r{T79KF&B$;p*083S z#j$42AZH(EvD50qK~0q$C_mOGr=!|9T~f&bjTLSAprE(H3Y#4Q)u0%3R2eMD)` zmNMa9T`wRdG*yqc%5YTM8tzv+9lNS*ClG-cRm*D5ua)<#4VoLK!gd8zfT(*lx`wNZ z=_Ovm5yy#dY8Tae%dm8VMH}Azgoncpk`|hSnJ|qNsJdFn z87mj%#u)l)WI}@t!~=g(KEC~CAbtwlLU@%bK_#*-mtEAM_Lu$i&qr3=aYQH1oz{SD zg)y%e&}9~~rbr)S@^ap1mrgSdwrET*4KSG~|LA%`d!Ss{YQN8)4*ao-=Wa_CLUBer z7_0aC`Q{)(H8V>+pMpn!G#~c54(WO7YMKDA3>Z_IK$^YEylN9lO1jYqM*&DCSUh@nGOvWg$ff^`-N7TzQsvG z+=nc@fl@C|;_6IrFXvr{Q8 znXnxD{z2u}!ON$B9q}Gv8L9yDdn>7zOV#GlR`YuZ5Tsx>rrt<$1Y#4!6Vw**E1d#! zk@d2DMxn>&I&i@LeecnV^HQ|YDv8EADK93AM(imhP!Yg-1?*S-Q}oclK$XbA`_;vQ zf{&(n+l^F%AFfizJw!%U7&&=rrxTBoTmL4G(2@C&5h93|^Vbz6mR)^?fD}wh?nQP? zS)yj#4s*=DD%>Hdecs>7HZo6Km00AodaIpS6Mw?&!kC4$09{~hn)Is&3|H9XU*g6(X6QV1X{xBNMYy{*=$P^p(<>z=Srek&i>rBGl z2WsD>p&(M;8>QYf!J~F2jeS>|0-Fqq_RjxEFg{m-hPaU732h?v#&6s(oK2s9jQs;# z_qq$!{C~PA=vZ-JuVc!Fe>9RN*)Y=DEOp_6oJQ81K|2qR%8BW}B>QGtwryf zYYzbcq3#=I?^geeM=Q2UmXSE1 zJrVIUn3rCNQaq}uTbc{r0My5~hJZ(v*0Bihh7K2WS1TG?7oZw8;J;Mv4mTNf)b)SR z4i&k~uP?iaK0g7XKh!PPQW>yoVA(zfnOS_azyC?LmF6Ex&CMd`;NU<_%dP6*c-btM zDmgYZR@awd1YPWUW!WX0~zKrG>I$dcPpEOO_ z=b;x1^`$jK;kbHWC}=Ur{cLS{`$buc-6??f{_tj0oh8^?xV>m=sbFv3Q>Gl<2Mx#7 zbi(0MYfdzw0BOSd&5Wo16L84DKqR z>OD`pgv6PzdFmE)aP$wy6c;1yxacQXt<*Yo*Y1pJl844Kg+2}0EE;9oIE}7GpsZnGp^hm5=*gD7`LXf$*h9bvm&a^kORrw%U?U#=~Zl- z8C!h4N6*rYd$l%NEwbtuzG9iuUbYIn52hIgmHFdcD)+nWe7wI_LwvA*_txt09#v}R zJuJ%6Vnxd^$kHb}FE2?h(a0oQ%ETmqfAl>F9imR+1W7$IwaBF(pP-Www47 zlF3qg9OIaIe%@m{#LFrcpHX$F8z@i%(gq$-Hho*3I2nUn%iZlXsXMogo%81mn1mfc z_EcA+GBZPDxR#J^^`Ys3T3=L{O>hqeF{5P0TrbaWOfm*m9M_MibAFx8=-=$@Nu#RK~-&Ck&W(AsNBLLEvhf`rqv04b#XHvP?w zg>3qq;cm&Ppep`R%ALR0?!c+RX_NDV@ruAw2#zbHhgC-mqw8|MA34W(h=8jkJzdDA z)#-Qh#lhjt?To!`{|L7>ZGi&8&=oPms`|~Ex6b68t|t{nSUuqgM>Voa6H0zi*&mm0 zpR`T8r(CEkPHR8;4l|H=Dk_e{uQ^qAIWGmd?~7L{7tQt`a9Ud5m?Uw=jU;i0sVLQ& zFVR@Md1Z_j1i&9Gl|p?s0cZ#zhsEESF~G>$?7Yq|Kh+hTBXe2PpJa=qLDvE(+YmyU zBnaL;%sE$d-7LBNaJ>n@h>FI;=1a1wO~}qg+KQ4%;u;uKF0#6tj!`n7G-7XR>kU%5 zge2OwY%NvWP%qUisc>M{tUWk^_Zd}+D_KjG-&?L0CkQg`cpX2s0DW06#QE=nxVfxF z`@DBaMLzoIw%mC?oSQB ziS^+|sU$W^;Kqf^+`hj)gnHvK%nb8r8XOT!F~-d{0?1+h+|6#9X;P19mC?H$>e`ba1C~<>uG`fG%>#tkK+^yyRFt^aR989@D0ts#a?Mu|f7D{SE74<>*n5WMU0b46qX7VU zrB<6)<$y(tK7!^1jjrWjeY`==no@l!EACHV(fsKY3-TeFuT{}w(wg9-<(cQCF|Rlf6v@Tl6by*d7x=Y==f zqri_9uP9-aWZyhVakz>AJiI|);n}#H(D|E`;T1`S?5wOt_%4wqqb^I#uS z-F0RFXy?$-D6Iupxep4BDSJ%X9RA~@b17s!JL}&&Tykf*X*i?2hWD1 zki)G_Ioc*FTTmaFqxHv!F+nf6s}UH^>&y}c%erGOhrZ*wn8k-5h>xxttpv*@m!&yz zKvN6t8Wl|&{W|xP%Uza;$2gYw6lV+DmU-*X)ZVvuDwPF_MIg|GW5ZTu(Ob0BKVj0Y z9#PF#7e#*tZq(+K`z7S!rC1uuMO$Urt`5Cbg1^P5A49I^*4S*0(4*)me)SUTdne!B z%J>7o*sXnX)aO7>QB~zds9On<^wYgrL$f8S2)SX_E@zpK$1IV^1uW@BS(-kfyivKc2dzGOrH^M*lH%ejJZw7*E{SIQI!$Uc{u8@M3F zpXSojF2sTe+@$Z2x2(nn!*Wu zL*8^+DG5w` zry8bYql^iV`xs=QKz-vVAN3hd&v!&9b^+HKAl~z-wi%6$;NF`cE`I!-+2M zhxe^&_unir%@*VY5||FG-`KQ4xWup6ZC5>C-JOA1Syo9_f0d%9r%_AAyTLumk}i5I zcMkO()9tTMMb}cpQ3=~E&+6%Ynj?aDL zv=*Pb#w(EV2LI;GmaoCP-H+!nkqK?*r7XQ{+hO_E#0D%wjEYV=Bt0DM^fqkru09%! zEiyzyR?_HCKP+n$s;m`WS53v|M1X`uWxaWxxM1Rz8?^GUkjSldPtO^G^CikMo84FY z3`$dF(aiW=j%pz{hUR6I#()IUpLd>HID!5_phX=MliZ)~7TMp9oh774)_6RFXiu8Z z0$(qSAf>R&r@Glx%OvuHxZ{e!t`=?4!;LT=fwb}N&6Bi92OG^6zAU_B zez*HnjkhOx(QVlHxid7apjHm+MbQ&tKoWl`v*1aSsjKt`1hN8*Mjz#)-KxGi{z*$l zO40XvZ7$dHwKqm^rsbP_FTP8UrWvxgZ|axufZCa>Eby{?Vmfa?i8A;5dAb`6XCQy~ zIn(MAJRtZEQeC-t5r19R0ocI)&39@$A;*zqiU0t2cYUL@>^p{!PBX3n9@;vNz%B6@ z%pLKh_9p+;JVB}?y5t<6B5VV!v@flSL9g2Mkt!cT4N#Z;- zEvxs2)nb`A%I9uz*We4c^47Tp0KqpHa+Q71%;XUDL6Eqkw%K{6hI;^X>aAo^W3KB> zV9%;uQeEPfwVo}|w{$ty6K_ z^uUgM6W?^|CI80z;#E>Mg=xUDLrV(N1cG+>`mKNuHeadi?Tr>Yg~Nz)yg#k#P(`c% zg#9ii>-EfLJD8fDzCd(Ab_WBdZ6pO4k9s|&b9IdT==IWi#kW)&c!2<&<4d`(6JDLS z&J+~R2IjhlJhtFfuU(H+u!ibor^r?JB#p5}YWuA9lEkHnj4%>3WmK$6?fsH8n|mEM zSt(|=>=BmI<~mRr!CUd#C_66!b-cuoo?!^0X}wuavu+gqaD($@?N>D=qBz`--!g2pLT(;=N$|E3 z58=_zTi@RsH}#(kc!BCE73vst+Rl!mxx7KgOHk71Vmr;%~^PeK`K z1v*}I8-OCK&Er1Dyyn$whILTa*=-BZg%i%(Z7FoOVO*sL^|iAd_k!{64F!fLWc!=j zY1aVP@pl`Kp;ZrO%z;N}rpN{<0g6l-@Qh3Pazj0KTxLu#J2Y;7fY0H12FZoA$pT5;tjqnw zn+eYQ(qu%~U=UdxyU``$+{OjH7XVrTrnZ-b-WAqsCvw`pIbhY&SV$aZWW%GLC?1dB zrFo(F)fW%>&h2iZFo5ZR#2S~}Vw+)59=0tF=gz4@oA!in%&QyB@Iwb2jd^3! zXr5=Qd2?_~B&^t_tK5rVDVM6s;2oOTChVsXn$fqOFvtsN{#X*L@LTg{g%R91d@ z_lDQS<_s`l#~*Bv_sWONS4a$%iTluH7;+u{yRb~7iYwRQWk?t_owiD*Lwp{V;}MeQ zT2q_x?wpD?THQDPr!KT$&GdblvV~Vcer9%okWrh^kALU)+nhO8akrJ7-Q6MTc;iIF!L?>%KBzP{0&u# z3x=$;|J;J-NPe?RRdeG?}>)#dMn|B&L+ zTC`hBJoUKq&U9W9R_0Rfbv@hp>PZz)CzBGI&#E>(A+$SsT}f3uK`;MvF0Z-n$hqHl z6^kPx0RyQXCpH_AE?0%k*4;OAeQn#Mr;SqY$E2=~r5|@R6ed9g`aWSaGHV53Wfq@|7z`3-{IjGWrN~m-o&gg6?aFQLXsq28IIm{;pRl+Z}HW6f^K$LXY8hK)dcIJEiB<_-SLf zw#4y|D&@AOofp4dL8Hmzr@x#I{k&ER7smM;BTB~}+-Cl?lxcIDMbM84Ui!UuIttG0 zEMBaiTP-K^&;wLMGum3N)zY=$+fnM+M*haR;8{jZLS?Yc4scbtNJEkqPxb?39`~B> zq&fQGUQ-Wt_R#hV=W0p#dMz(4+2tf5Gj*p( zr6f*{)|(<)5Nj1{(WV;vQ)ydet)}hDEMul6>1&7E3TJ~14zP*Ir}t~um3zmL_dFk< z!-B=Xm^oZ^+X9IJy!mW0v}L~P>-Ko8@J?|}Q*Jk~DL1S*3^8xTvW^Xo)M29IY}Sl* zVe(4^U^Ymec%S`Y3rxp@MLudv+cT(Pumg$Nlq~1#i@ds?&T>|yXs&?dr7+d4axF^% zpe1eR$6_CN2C6g5?TIl#WnEdHq)gmBL)DP1iU%y_Df8Xs6dW=V4e5dbvp`Hi1RJ=z_tOd@=JHF{WNP*CULut z-OT|8=Pd-o;NR&IvvV~UzMkHMaodA5QNCN$%CzIbdq};HEaVM;LSN#8ag_&_)Y`7` zl0HGV7993@m`g(x{9nAYK8v~5wOfNo>!NE-ew1-6*(|LiXqj(wu+Y-h6;pEQ5;FB| zh{N~@0+`HEpC#Dxc_{g6-i}l3&GR6@xsJd0ZfqGX>6=pVDW!)wHVu|V1fRGfi?8Rv zkvZKsXt9fqxQ!orCb5`Mzs~?Ex)Y-x2{i#OJ2jo#tSNjJxDX8@bg2V34wi81vS`+r zlU5Y=4ycaZvzGltS5M|N_(O!JSv{|%EvHvbcvr3+qlyQ;SI>?DQ$I^0PrWo9k) zbw6Su*7$5VU|TgZI|=N@)8cHJ@%LK;7#1%Z(TF|#G18();ix~F3_T)2c~`==^f`=p zU_L%AnozvWKm@sfuua1OnlTm!bhuspK^_(zIlh1(lLem5y%!b@(l@DHqV`(&v+u>=3{bJi;cqH`^<35~0D2IaXGE5eLajeXYW0mv5`qEO0{YX4d_d z0-9$O9o+*3d#2mE(OF+sM556Y#sfIyysQ=d-N5zWvx`4{wR(PJrOB3z!Z=Zd3GF|; z2?iLd2zoG=kZ1jZW?SOy;k=8OHDlBP$BoyS*0xnnBum+)9T{Xb&~ri+An>p;FxmOS zzELwb%*!iN8nok~DZ6xr4uPz#u#sebr;adsCvo-#2JzWK!ua^nj9-?%0fs&@aABBFtHn5^SC_+h{=p5j*NKYb2c2F55dOR(`BG4Ld3CT-qK|?LX4>uDWJt zV>8m-q<>?Bh=^D@20-?p@Prl^nXEOCKWglbf}}77j)*kH_;1@1co$BJh-%Cx4AkoY zE%Zwoi5v?|%?tT`ZfxyXT(MFTg81$u&Mx8m!-DX;ZG z=U6HGG-TH8?2oAtV1DTXzBOaBPLuWG;7mML<$y>)`1?PN2Ixz8tFr2SVI-U}O*U<{ z#RjO+X>>rz2MDz!QS}Rizge3~6b>0%Dn>`C@Rp-RJA0rqbcpWffq`cV`fH~vx;pl{ zOLG-HT@GQtOV9UHR_sB{Y-d+f&E<3q>xaUX?*!UYuGX3MsSGL6`URW&2wq zpqulCG{a+1wE^s!u8x=o?Os#K&|1xW)*7jJ;pxS^a5{M#3<~OJsKe-Zhtl=Dl&pm8 zvnR04n&b8dJ5^hV zD5RPr`ygbwQN9R;^Wb^q)Pl$7-N@q5aIDW6Cq93y-58)Qy%Y@u@pm3tYW>BmlN-91 zo0BuZSk`KjG<|yljnxE`&AnpY4-i8Y5e_MWfiwOk1q zAS)nJM571PUCUsou<8mF_b|p2T4zbZp?*$_arz|$jB6TLLT{t1BlM=ij{FFE8yiYp z;?`@DX&4%BU8`=`06fv31ZLwIEmps}$P1<|%W&gpfmV}@;(zf;&6Rot_s(P0?TjQ2 zxb3EzM<;P&bTebIbg?2>koYt*9jJ2M0IZwheiFicfu8b$p7FoP<>gS5$8X5p04Lpk zydI=qmFERKPO0zd!#OQym7s+I5w}8lf`6E#d?SPcwEA-NRtv`-AifuTXkCxY2uGkGCx z>Tx>u^s}-hB4qzHA>I=0F+c1(<>k2wJt(ULOu3$Fe6kRSVw#rx+I5rwAD_C@jz|O! z*A5uaq-x<1ARgRvtV%!3a6i7kDDeW3M}FFC@Lyu=jRJS!!$KN>Ej-h5f#jV*}^^j#L5XB(}m;o zBDlhO)>%F+g0opHr8g6O1Wzcs(Jt3HNTUp{36T2$%o^6X4jdlSp(?tZm72)CbuXu( zrIwny`aD^xPqIH*%-;#%Z?NT&K@okMKE}rBRZAU`lau=e0PxvLkaKs}$b1oMPlXNp z&mk1u;Y@SD-UmKA>YS#8YqE}}A$y2mWPfCcz#FVHv6433_k;XK`}#z)6QV}ltCE&U zvko?R^?7=&oK+l39mrPb`Um+)2;pi~e##>!H_S!a;vQ{W@Od@xHA1741fKR!huO8e zN;_k6Q6ZkfuP>h@cS4o-`-y**3#H&xKzWXQG0V-KB)PN-y*(dreZL3nPW+R4#D)M# z%$8gGx8;ADiH`$r{};KaRBbk^5=nN)A@!tWp4t) z{>hhG=Fb2Ituk_2qdahWA?}g=yIDOQe#^pn-Jqh#kqr;Vt`Q)1YQ|KdhCGqA@)h9j zk0wojCwk=0?0xU#wVfYFec*{wJt`qp@#}Jz5IXfg!r-|XUg1}qB7)M1|6U`>2#Yn+ zk*y%4?mUGw>-Xq?;;LZ(A^y|9lhQ)-8gD_Cv&OWGXGC%sICfB`8%c0^d#M35Tg*uE zv3btLilwky)mNWvRW2OX)}X^`3B6T%6*moKkQK&%{|77Yy;mG{q4jKiMs`llld%+Q zw7gRY9H=i0{=N$D&awy9&F4%aIPXj?)ZXXmD0FjJ{<5TlvXKT<9WS%1;`U~%CErTQ%0`9pGz2{a11bvbJQ+?7|3TLaHj+&GNj>?U z2>qG8z`#}D-QK0az2XI=6ff0FI?uC-jq-6wIom`Nd%(x|JuMFg0>?iYg80C|8*8SN z+EUUnhOnkqhieUbHujeXEi2uwwt0DZ%J7_x`Q~u$$DB;z?k7d{|4XB^ynmPBjUtr3 zimL~xS=Diuef_^i(4Er$@QAkkX%ipR7uEiwN!)25!VuUWQ9FXiQtGrUmx=_PtOadE zOw3EiG;~#h!R&;F)gO&eIwU_3kM#s#P8$oQ^A^n&*9n)!S^kbS7r>z6jNP>|U$(H+ z`SdA6UD_oJ8bD-^@^=B5-77#xv31~--jfju;bRN+m%-(8{3VaUj%;U7vxzug+j4c& zu9b}Dua$~T(a8I9m`guvIs1X!;K|+#{n}|HD{b+Zb+qXm2?+a#+t7@lyG=*8G2Mo; zZyAWBF*vHRInpBFk)=x6<$7Uilr7bW^IiE!PL_tU5jl_G_$5AmaLBz|?^i)?ZaIgr z-BxyOdG1JqlP3EHX1KPu9Z#NJC>fRfsP7fs$X^_Bx2RCd-P#lFcw`X&$3?H-Phn_7 zDmeVq7cWZkM8-Dy5jLgEo|Q2- zh&_ILPft$|rFGS@v9V8$fVKEepj6{QEw_UwBHMce1Dxv`X?^|3yTDb?;nsJlH(5JpZ7_%cH{O1cMUu&X>?a z{J~$oi6}oWtP10e&tDdN-kbYym4kYhvrT=4%Ju<;iVj+dMQC~d0v!%p;v-z+I2zKk zkn&x~6Lk-+=84G13deM)#cDJC(l1bbnxBJYy{zbG;2oBWe5OG^T*Z%sa zka@+;;VU#*La0^zyJ&1xIM3Z*jIzWDg{i|A=sV zqWwlA#$V{^8b<`8$CJb_*y~t7aU~@FLpYSKs)P#L>Z^*XLR_->B>xt1=L#AHLZI?P zL#nc9p3A31xP7OCV=-MY_SF7zW_KAno_vH`UrGjIw&yrY`OEjyE#;8--a{g&R(PF1 z@sL8j#03H+n4-e`>-JI(;5(;Ghbz|igMf4FK&VMa1So_&a zLa$LpA)};VKKIv6x%4R04$w29ng=*GPiSuksAtGxGwiv@|M^z_+LN-{$^=by?V;`Q zO~qxP&41wpUG<4(&mkPnn>+IU=Yl}B=itp`dlLV(kOmz~0DrBVCLS@}G&V&`MS0#}y22np#|vf6}Gpqtp~$z(~J^OJF*)K5O2 zGu4M5-V$acK7ur7tKliM)X#PNWdPs(OUkC3kq!dSg{ai>6-kCQt*1_$kingo8E#Y*?(>`S09cgnFYR7H=d~Wg5^)}sX_u!L!Im6RS;Jya-L|-9NiU{}f@Ax>f)C0MQI&KrXneFxiKN3q>3%Ml~qfTfM zh1G9E168^Eas$Eukuk|ugM_h@fb_ogB5of@W;kuNXVkC0hg}P?B0{UR>AFo$&XBQz zl%&$FlDvGZ__GozJ1~A)>Rj33+V;|MZvU=zlS>!fn>Ny~dU`&1KhN0hFgsG4H!?65 zFQ>AjXbmPqiUQBdz03pH=hAt&1~M|FJi*A7)b&Ckvo3dyWhSNJX;qT;CCIfW_tXo{ zTE}LZB}b*<)~(%G$1FcR%mO<=bGb>WnOlcMY{XPUCNfb*p|kXp&Py*B^z%6O=c#2h ztQ2mV)z^VJ-7u0tk>6Lj{dX1=j0D`V4!m;BsZUq-Fb-O9>)SinN?Ui_juU*RtE=@W zZvJyyK7mnOI1(-^4u?pE(Pp{yq$u8yxPmB(aXJJGQXeIdlwkmAuHpSJ3q#ysz!>u`yniU5OI2}wnNBn}Y=p?RFwR+1b zCRqB-zNsg$XXUqj62Vow0~cuLqmy)pwcd-HD{IRh{)8|c?h=xD0$}cb?itNg!10QW zwSnHpQbWgV0$yJJS$B=9`Od+^555bz*-h_CJ>O5|&et9Z4QKqysc*KN)owHTI`%6^ z6p*gJ{E1cVoS9@O!yDdgRykSFN#}q>@~!1W#JHU7Ka9q2b*& zYO#w;;2jP3+^-6+KBa{+@y4qUY$C+m!Av(zhr7Lp%K`Wca6MPti#N7`>%8Z+J_6Qv8>3pB zjN^$MjO&R2;ku5k?$|bNdgbGAd&SdiGwT z$jC3Hu~>GI@af8q3wiWcPLjwVkfVGle%2D@N>UjN4LwoV)ed{4OL&hC``;I(YPYrZ zieq?e+w}nT9QW(^WPi-l3dkGAL%6pGgf9tVu&<*>NQB%i7x^g44S&JUeIPLCI3%JY zZ9ehwV16I0w~ASAdy23qBi9!#NKhN|yyg=VD5Rs)%3QiO6FqY`Y4)KJ128rR9w+d1 z%h-SYDZLd(qS34!?YaSTir$x=+Ru>Ih*xn(O-z%qGjWXrBSGK0Ly386?W5{y+YCgy zH4Z(HCZ*A_LZ*R;~!yn^i{%5zIMDxBK437xkK_h+Za+<*6DYe2+3*_C#Hb;hQdblPG-Bh*;ZkSsOnV zhkmskTh;tyyBRhKRxQ=$w~by8cpfqwzVV0Pv;o$?Q*whFSH|ya_caBc% zK&Dcs@8JjBz!=NfsQTR=E%_00fd}GX10LYC(a}M)vP< z90raR8wKa~ny(%VDO>CD?F7`@-~wKYaKH2sN}-CVw(!F5Zp@F)|4?u*+YYJ1z2*L_ zILyW47M=ZDuJTSi?-6InmmT3AD{|^CwC+Av1f)H=l{`c|a6_t?mAOSmd+z5&ThI4N zx9*%{uGMmi6d4hKx3;Dwm{R=>dNGbq<7sF72R1l=yN&<=l4|9|w!7FPKyDRsa~h%5 zr?yfZ_0gh5lx(6Eo*5LdIX!HcF5eqFBfWEryh~&xyj$`8%tPV<*g(@gqi>=`Wt44@ zx2f@IEdKt)HVkK24S2w}KMIX9@vgMlPv)(iJ~o!6KA=4O!qxnb?x8{M0o4!fp@aLj zhBhy{zT}YqVLqqIA9|4wiY?YkKkRDeAWVk!r^OZOfPHN4$@KjIZ4tV~w`unJB0M|U zNI?%hs#)!Rn*+?h>hd0iiqcO+uM6w-WV>Cg`O3WbXFu}KmTSTK=<-9)kU0bXKf=B{ zs;RDdSCA?qC?X2dR6waBO?sq=bQA%RCLq##CnNz85s{)ONJpyFNUtG)6p>y-P3Rp$ z4-k^vuTsd<-9ie+{UE6NDCQndlO4T?<6ep)zeWo|~nG_jGs0z9ttFuoC3q$sT`0 z*rNUXj^l(iqy2>C9Vv_&)=?LG`S0fpr1`uxEj6qD$(_6LDKd2&^4Z&{Ck!FWy53i zG2eLa$q`+CX@N!+!reM58BAON2o+0@a;F$RV_}DPvbvEhf`V1k3QIV}g>PYLF`Tf( zjSBBY>5zp+<$#X)4vR5yNEAIAF{obJ^2x9H?tLlxGL+Nd6obF_*nc=1J(=LzeKW+K zCl3qx0$bSP5Xn<|_c+=KA*(S*_(t_pF zcG9_5;hP(qoUVrr?`vw>z!uspLq*+7m!K};eudoruZSGeMRT&aQ7WIW{F?>#sc~av z%2p?0j2qC~zZkS=#vX(beu*@sHGk5)u0Wjk+=1Opym>K6DmKpo$`$%bH1dvq z?nDOA(xVS_0cZXAt_LOT`zFOPEo=lQg-!DGA`dm}%|0jKn#H1jeN?;tgZp`k_(6=XDWw2mIO zMT60CdOFagJ+#`+{C5^0xBK?sI%~FM`53bM=%-87czy-){(t3V{sV+QjR9HRmSw&! zH_WBSM*E!)W>|aiNL!9brbe{=9JqTU@*-9-0GzpZErvTuLR~aBd;75a{kG)|n6sRP z)S?^{uZobg^cd7sW+Ie({PA6?eEb_;OA%c>>xALYv-@mr zsX`}H<+|D4YeLrfU(K3`Rt|D+l`s%R%^n5ZljwekFNQZ4rzoAZxLN02I@{nOckpI# zzCR56kqL}DcZX?Z`oxo(iUPXp(^|fQ`3=QyvOm!>#ssdVNNz+;xR&2Qax70!Y$k40 zLu8YB|2FS{P60J>q{qp3`d|+xJvEfv+~;IVDE`x@NWMCI%h<;|23@4p*g)z0=yTJO z6>ol7m>>@SBr6Z|<0}KOs0#)1@G}ltgh#It?~gtTJ`X7Oxl!@hzHamlIZHJAFSsNK=b=$P1Jl`8|t6lu-9$g>p}HhkR-tkJ8r6gM9#+|gAO z8y|U$J>la3gATAG6)J2S+W6IAOce$Qf#OY}Azeq1VpCXmcD7ElgezLBxW?g!+*Y6U z4lG)Z2mY4!BwBGh+1VASz|POdM&XVL;3h>hv2wy?Zleq+`1~)R{(}(GZ;=sn%){zA zG*h|mFur30(l9`-nwzFY2ZT7(*O%=0ic9KPQ7d{1gBTpyjI&5ToYs}1S0$+xR)(L* zpigG+3alccTcd-{?Fi3Dxp3Bblrg#zcJ{Jn0?O%8sR1OlB%0g0bqhjZw`hXRz?&C# zNeLLzZDki^yX2fM>WzM+Eg~So0=&=T(HA_0C?2nLp4(W93cz8jfBHxk4ab@InZ)PJ zWF}WQRz>bR__i}o`HVRm^+#hiYhYQ-4F=VU`Ae2WL06I#6UJOS)Sq3`Uu$l*V0lw& zV=ZB+?bkaJxLGhLNyrgZ%nZqFQskZ$A^oz>eA%5S;j#3?r0b>W9tP#v%gQBYOSD6u zTV8{^;G*xJyf`|Fx9Gmr$ivMsHCw$UP2_Ld0dw*=ytI~E^?oPRe{q4+d909)`&*Dm!C{jl~? zv0@wa76Wl-?>jdH5M0hV0*3VX1us2od~U+M&Yl!03l7N1H4_~yoh?p&>z$K}igQJh zmS0MokYp8pZKSBmgt;C!X)RTUg?y2-kXiZ#|UzF(R?vBM{Yo-L>vLCla3KxPm=#>$?YF{hp z_jU(+%bdrBs-|q3uU3@eizl-~sMX=tQI}H+-E&adiXXM`ZU14d{wUeI|Lsr=u4eir z_S~YDNX7bb5aE|Ifq@Fg5E!&k^Rrh^grzo>zwb~vX5$)Wb=>GIKt^F}SNe;(`fzrf zhkQ-pCk7>q@VC9!aK~>h{XCg3Bl+9MTX-~N^15i`{SE8XKqPLK}?vtouTZez&Iuy1*an zmNmK_CQ&(}igsI09-TeU5T9ieCrE2r1@LAa3zqL2Z29_Nl%vO0BLEhQ3X?{jer}X- zgwQ{-etH<%A#?St1`FvPBr5<1jBYyBz6Ra@qdbnL0cKB8E|K{p6gKqXH0o0mto5}O zES-|M=j=Xtd-$hx-lQc<*VWIfK$Y>%lgFVD*B0Ky%Q#v z-C73)Lb~0|=U8RUl~IR-Ua%^8uwO*8xyLrw3V7DNQp*Om%zC*PJBFlE9zL8PIQL-D zGsG#)wl3Vg%=Klfsu?W9<;xGMuWZkZlqC%Ps5hdgwvE^oG`xk{o(V6p(i)H z=Yz}Q&7o6^8yJYi%C$b8QZJ7`-dkoJz+okxeAaf zzWMMw5+wyD4E-R?SP?m_V3OQsMOFQJ8C5DjKvSM&Us_t2fG~eBap`Z?j4m$=*iZZ( zOUADxYbjsvV!X22&ml56it9m>w)Dq-SZ2Po2^D^Qf)vTEy|FWqtl0IsD%kHt*d6E| zP_J{Y@(J_752FU>(RfZ;^$4%6DO{OTX=+)qnO*OO7_WLB%bTQ* zf*h3Nag(GUHsGL(i$;adZBzsHr;$5}ChO#F@rksTgy51srf#zfx=cBDmXRW*K+kgJ z2Q=H?tcWw@3pe%rdAng_B?@T# zc5_*D+%*eQSugVQU~l~|1KP1wr7BLXfbJXsQ2(aoOR>bk8)mKgTVD(l&9|aW-L{8) zV3VDhA8t>zY`XTBhSr3I&~1&koFhJ~aqxcBErLvZ-fTjc(?xFmPPSEAcr$_UrN6_y zC33iA;XVuTJF;tE5Q0bS9BX5`j*uU9fB#Zb8+*x)3XL~)#C=&A$`hO4=1YqO=+Uuk z#dTUe3T4Q8Ji6VSiG=rfZyMM%hThIl)m(?J8NS-S_-97SKWi{RRU9bRh_DEsQ|2Q6 zLGgl;tnXTpue`r)+%oNphV>duSXsLJan$4HddGy`6$r)i0;IQk2O@2dEidJj-r>D1 zN-u&l;3g6kyLnqy!rBH&4!^zN6?|$4t|{$ji9Z=OX0lhJ-fr0s=X>>TrooORvZ6@o z?wJouO_myy$oTVR2k&ofy#M)Y*$UAk!>a2IacoRrN}F<5ttw~+h##7B3~F=eE2Fkp zd8OKCWv_}(i0!6HIp18Gi||nVuh?h>##E=YJH3d*_g#kv;3U*Hfb_Xw!}7LXJJWUa zJiAZYcNAir9%XKBMgD+=4EwTq=RV$c2I{DV4}m0Ry+ z^WHNC;;li%_CRkh-Tnx}Il=>1BI`9fX;~q`b6QWQ5f(abqE{GLmVOo-vCJB{1~RXE zU!1bVOf54D31#UTZ-678x-H2wiJxqwCex@GTkpyT>){T9bsg9uq&ns;D#7*30yU)u=w% zcAWIo+i`$?ZmU9xXNpoBmoTuBoj(90e9|DtmeRFLmOQ`hu6_%5jTAK}>UP|L=*Gmi zeJHb;Yzf!7-E}gf*55&!a@VcWf49YqYb^5PKb^=QlgHIh!?ndX&TJ{S9|ii>zai!9 z;f_<)`n6>x#zEliRjC$z1v`v7t9nU7D>_OOAf11t*5SJQ0&2)cwqI}53*lQX_=RPt zo5QWbD<3fXeg?|wy$YmK=Asnb!S<`iC|_R4Z5FeDqI%IyO4mmWieB;S^@CAp_bm4P zC&-1$x|=rs_82HGM2?{K!hO`d{rXIHe8pg)vHrE~#;q7DM-&zM1C6@i9Eb&Pp#Rb2 z=ct|5F}0eUF#H3Q-$9!`boaA=*#qb4mE5oDZ!gzzZiMOFsNY^xxdz24n|~a{rwJ;% zB6QYYS1*~*`VS!Gx6cY}Ve9q5{_XF!L$n^H`}0aR?4;e@hDa}U4Q7F>AUr%gV~(1> z#dWKfj(UFVh3O1botN3DE%Uk?ajUK)VQve5sSm0BmIvg1b8j{nvBtmJ)B-S7vK5* zK+iA`FmvX$Xyv>+dD+m(m!FL8IN1_0q(>w%;U9UCA6W1~gU4^b+$QMpF>co4NH5=n z%fi3CDMq?~HTvd#Z6?IDN%#HS8Wu@v@F=?2F|0tEb@Xqi+2{wT9EwtFX9dg05W4qy ztwaG<*(kv_%{O7aq?bFj8vgny369{xO+k_b#hOXm5Gub}BIQ!b9Vs(dXo8ktX~|B- z434kC0rRvE!i>Qug!^)XkA)rDz7g+QuX*i)}-OYAraLL0K4f8Bd?_}PG~8&T{&@ii8Hon!o{FC z=Zc6qKWhW;qM1@SuLVyyAXA?>Hy5 zkA}^cJeE}FQ#f(oA}m~x&ZtTebhqcvvXJqMNmjH=?~>p+iZ}Dm{2*mc=ou;?zSk6C zbS5RPh5d4z!&JtSV|B0FVxC_03sAh~vvl?zAqnf`1Vt!_)ZaIJ1;K^x|3Jma=;s8^ zB?6X!l}tcuImI20SXfxm0r;EC^DP|UbzauZM=bIqRaeEedSzo3e#Fu~-yF@fh7~Av zTK}cEfRP1^geM;;{VnbH2CUUqK{eLAfcD+7Buc7z%i6+_e8YwRLhqas9Zz51YM0(5 zd94An05X3Z_M#=z&`A-uPIigI(~NSrfscca#lkPD{%$~Fp*ezgKA^~~3LSuO%xHt^ zYSIz-3Y{;{m;?O+AH3E?8!!#vu<|_WqRL4n4o?5H`T51)KW1>6hK6TrPCn6P0qFwg zv10uC9oE%0SMWO8I7)u;5$wqjgyro+I4RFfuz+S#=u(Oe%MC;H>SGwzOj+gU<_F1=U5uaOC)2=wY4gM6v9hIKUr^}PMEbGS z0J51JtA@5cj89S{dDHR3wi??sVVuIR#c$!&QF1UtS|0A;gQ;a`Daf9j6 z^9r>7?2ULXW}t@XK^~u6KDqZT6;DP+N}!E&_G8-*^0jm^4RomT4mqdjvm#M5Wk;d1 zKZ^=gXL1j$eHet9ESX9aEISTnN1AAI@$fL>O%Yv8O;oZ7x%-tktY$6rJVD@=LNjjRuV&+zva9jqCLh8r2NfE@n#;WlWrReBbl&iX%;Zk?ZJj)y~?owfi-)~U4hO3%o^7#!eNBm z?xd!V-0l*|n&sR`2mpjG?ns6d`=`4K*?muk^Xir^OhuF8;6JRr+HKRuEH z=Qyb$WpaQ2qk}>KslMsxP&d@nlW$o!u4%ESb6H~y<*f6nWCQ{Q4WrOiCR!j+kk9~; zwg*~$&Q3Jy-2ktl-nz-_+=V~R^4h%EM|&oPrcc$UGx7Wuzd@5u#1)}JqxyiB0<|i- zg?X&Q(#Q3R*eogt@v(0KS^ViaFHOh(yCD5Zcf z4>CEojnkflOBTPW9+D8JVLottGKm`0QCE+9u$25b-}a=aYHGtjO-(slblbas!Y*ff z@~3re+5pM4KcHk>=+$+bg!dR5gt{B%Y_wl@vQ_iSo20i41Jwttuhb}g_(~mX9_xOH zX^UhHHM+A~P0Hg7hkB7O)FqV7`!gw%imEXCtM>Pk!#Ha@+XyJt05yf3a=C++GD(YF zVYa{qBgyfofE~v}j*4$pW=rxRd%mz`cAW7DYrxV_>{%w7nGHcH^OGL0i!V+?{r|T` zi>sL8)7cWgmfIezvF(l%BAa*|f_!}lOwO9WG-=KDE8U!aws|%|oo!*vSE?<$=uxEU zQw=$PAz10fCRCA$QN}7YC)7~s5cT6OFfr%O<4W+GJ{cg2EHx#Eg+IJ`YdMJ`dkpu= z9fMOm8M@Lhan@9ev?ar8P5NSl*}5`2e6rgkuPn1UsZej1(7sC|s`0{z*;+}&lrboC zv)0_6g~xWyE5hra8%Y$vOpqV04Asv_wNwC6Q{mE1+=KE9*t>MwGT_A;(b9t~Z;wVI zp4FWM_4i-*OWL0%{H*U&C@ysIwwjLZea#(LjIjK^Y{K4uc(_*jcYAdZ=r)zC%^+tMW#@5oJv4vX!~X?#FjLKsbnVK(Vow z-{hKW5;V+z!HaxWlu7uunAlHE0p}%v=@iG(ol(1a(%fn8+nw#vn$7Ztud^A(((z4g zuY`#~X6A?A_9d{8x07}Ihf-Jn&)-$Z|C#(*qOFd}Z<&Vg#wRirl9N9$n&G!g<^`JO zp3#2ygAYKb6^pVt@E?U79?s8Zji{5amcYsD#h4epMW^fxpQn!WpR8WSWkwFB72i0z z86COP0%6|6B5l9d1Nmn_9ID$t!Kt@v(;YL$HhyApE!d=Dq*8M$fk!j0!^Q}0$1>(U z+t4Bt;?2LT~ZZ+@!%OTPqN!TsK+*b3RL7 zHm9gigbDXpX${Hh(;7KMJ-CY7xh=gFOZ7$0~s;`x(N3KXi2|(`HIHJehTrT;<``6GQp(yG3k2~fj(nA`aTkz=gGZ;8*L!n zuVarDB59l86S-~FN_#+-vU`{wcgd&S$1(M|f?Gvb@MW@5!+R2IfrU0DRAPT5+Gr9w z@q!ZihVEoz-1#O@(RgltH=;Q|Xx%BDRzz&Ql1)j)s~Zn)Qo81Wm*?i#q~>48{&qjqZ$d>E~`1dM7ymc-utSWoKZ z6z>!C39K!j%Gb`~)zD@G65t6hf$DmnTtK8SdthpU;AnHpC_n&{KYYixw>IBt8ym7l zHtCCV_6Q)wBA`3F-3Y&#FKF)}$HR4f2DJ|utbrBvKe9vL_5XAl4lNWO#P;{;ut_ay zxVV_7uZ`)1a?sOo8!SZNW5g4wj| zeLU7x*neGhxA5H0vpQ>Q{sCcCz{DtLRrw7Arv~Ry#LgNVY4X@aew*(Dr?dQ&r&j&Z zDv%wq9o-bC=w<^#K^gmR8F*PHMVgB^e6m|%vc%TVcO^)vu@gSIjOtw+cy@(Fj=|TA z(N}q!mE~j4+IMrpd@exs{w1?FeKs3M+y3if82c^lL<(Mijmz)3X9hb+yDbp;c4+S2 zm6W@qwHt=!&JeNm=#{mKZP(5{*M?ZJ-uuH~>zPnnj*mB^@3#mP$gUQ_UlaGGrd;@5 zE}jtiz6B`Jxc}AE(obV|nU)!sVw07t{_0`vQE>u)Uf!KnN7LHL+L|H*j+0e{qSNRl ztj6BP#8Vx~v}anakd*6s!K6y=BCEVhxO-Rx($GB6JzLx*Sr8E*A!?-=TaCMjo<@=u zJ1bB^r5*AC;-Be6Xx>!Oh%71gblmym*+Wp1@#%_TigLYCw{hje?uLkz`vB`TS)YpK zm!LGX_WY23(otFFX7AUM2OJznY2C=f*((DG6E1Xl!Herg3wvV3$_)qfLVQ&@N(cT; zC+dTroxqq!=W{8k1s##ugPqt&YUEo?zcH7!Zo`b+VWcwh;;|`B>Q<3sS9aZ8BKlHB zZVnujPTErSBxJsC&!{IJ@R!1yanef`E(K_xT?xbUu+7kTo?bt%w9&eFmMFfsN{P$8 zhfJtjE1kahYHa;i)>`FXcm{sjaM)-`j(yR4~upy$VpYtAW&Q z^5A{fSc|T%{3BuovEKr$70~i)5U2qR4R;|>^uU8KF| z#Jo(%X)ifAcCsywo;F#ksv#sZxm5 zR{q9^0RScDQJdlZS&7#XDI)( z`j&3a*mo5u4BBXZ>>n7?N2<@g^QfMH@s z&_PLYZ6L$KZ`wPMO_|}!cIdqDo7~t^p@)cvH@4g|$6nBM@?DMLHgRSSur4;%za6WY zH^(Y{@U>fO>U~(}(czqT@g3=rdi!aunw4A1w@e!98yoJV>nuapfGL|}j&MQPCHzI2 ztlRh5LXduJ-##d+^SLw|<9Elxf^McOH9+&hWM(MEiBgIVZ)8m;`_{me?V&w2?~ z?#KtP-ikYp-!BKr%W2pJV}!o5ZPs9sI{c%Kqa4g@KSQBL01p;7_40w=(Uu3ZJaMOU zKO^>1_L#nIQjB2s*M@}jxdIn_FU|h$($bhql7#k0ojNL(d#Pf4sNKS7@nT9k^8FKk zI_-Dqh=ZM66*OtL*A$LJHlxj8tol2W0`e zm?Ve)g>3%X?Ct;+u8VR9YZqN~fYAT1ssTS&x~2b+C#Sn2jnp?vDD%Nh{|#WK?WCs? z5;kMDzv_FVhH1azt!9Dt48&`)=y`U>pssUKU3TJow0_RZiF^E`vL?0S7)@V$q32l| z)Vq~IWgTgh82@?wO6 zZ6OWz+TT_CnRWZiX4_6(Md!WWah1DhaIl_QaaS<4Wd1&S{dhJcVBVxR(AUam%C#F9 z+x*Y=f&Wf&aOojwmlLA_BKpsr=^D;@mP^m4jo>HGD53H^e3R6=y7gBH!d=mo#wpFj znJ|gCo;pgzG{KqU5Id@v^4xr@el_ELFt5BIaYGkw`Y5P%<}90xbL zl50xdyhHlAH!SA@W2Km=NWEytnYPEK(VQ^f(pxz6-PX3RR{H1Lg1Gm-g18v(@_Y-u zp#CVR&t#f9_+;f4nGzk@CGxy-&@#Kew&*a``LDBR=dpZWl7Cs^FnT55%eQui{r(0z zZfvagx6I1MI$IS?T2}%GT_e%z#(tKgHGcN)3^@%~&_6%)$jB2DZMzexz&VMYX#>1^ zLu63;mNm_g-MoQ4yBCUPedcHJ&nL?xxsaaYvd{sa1jwMo*9UI;l!v<$Pd;7~TeYdV zcE+fan;IwO+tNaX^L{bmn_IPCjC$}INvws)aG(^}L%1e9=0BdbX*`2nRE@ay3|CmT zbgM|oRrmwV2VK1jHOt;F%-CN%soR>cA1Je>{CkzaqQCG=zs4g1eBjGu-b`lu85g$d zsOmtZXy{3{$-J&et)1ThnGYed>w_jrfB6haj%l5^m{j>$yh*aPSLAuJ%-sej;|NTI z0u(cif2Z!d8xcz#p;O^opVQQzftbe;p%${F_9T`u`}^zJLH&~EK1YFdTFAg-`w+S9 zc@2`E8iy!|7`Su^Z|+Kw>~FYprrtt*va4q4yae9y0^y5CZk}dmN&AIQX1*&z6&22T zc6)D?^6xo4lTxIHnEg7Kk@63Cp35`r#~Sg^P5AFU0YNFmA%4mssIh9JUXCN_ zN3^Ktm&l+?1@ikL@<;m!s!cJWm4GAL=em*;Q!pU?Vq~kSum4o%cm;LM$P1AyhuG*uNww#;x3!w#K6y|Di4a^(W)azR}T_YWamld6Ffqdn=vR zKE**&Jj~OmGmyKD7-lkj@o{P&`jGujVl?JU17w0oRz2OI=~~uARq-n5eGL6hUETzC z)S6nu+Ik4d{iB0-YN3x4=Ubq!6GmU4Q)6zl8cRDKwJY)!XtBwWMn#ZZQ}@rcf_xMCM}ES#Pk1@5S? zrC;K;Q*;?T&=swg+l*iVOYc_{qapR_em~4vrco=yMv;sN+H@((9))(zCb0moNcwcg zd`_c=+?CQ zjEQ%k4O?!q$VWx_+&o40^SbHAp3tNxW@aAW)=$1sI0#EKYj>38(eBtjSGY$8723=J zN#23nlYf}M;d7!Rq9~-3zu{%Z$rPk}Wg6uro0}rjVz@n|FF-g$`h17*5{@_?NAE{m zsd%@T%MkaXPi`~>7dF?Oo!vD_;Ibi?T0ixPNOJlgwlY`UuxJBO zGU7s~$$VLFeG=32?RtN7Xx9kl(A& zl1@(xlvSf%G2?~c%d>Ir69-}lLOgtnoRoM^?JVgz)gJs)Ht&Yc4?$5W8YrZH@(p57 z;u-Q_eAwg2#&OI{9MWeQ)_Y--LkzZhNBJdU|M%L*>-d&X|MxetnjpYyU;b9VKPstZ z=s3Y`sMRIbr+@9LMLMJfWnlmn8&9$bA%$C@HV@BOW~zCO1l>?6|s){wLlyzGaD5yJT0l!%26KBIBmvFhN`yJgLn zokA12*JOs}9;f32ceP`ePlOgtC4Rt&M?dF(7hGwf*v0SDwbxgDNfZIE_p6CVHz<75 z>u5cgkh1vIWYK}4OqM;x=z{8l+PZIdP-?7JhAj>Pr@J;|!pv%ClNyYzgLoP9dRK3F ziaI;IA8&FT)Shar%8NQegB?$rT&mFHY^De8z%tm%eI~0ERd2k{K`jmV&s?< zYk+Jn#8pY)Z9J@R8Jo41PMTj6V|t5U&ptN4QA>)bch;(UH{AZ}O=!Yo|dmp{Kg&2IqTWij2ZIfG1% zq1p-|bM*ZbpJ_;Ca zhM9-V!4c+sSR^)kqu(2cu6kNIo*$+K(9B((X>&&Hot)^nP($VNE$)XiXI0XuXbyf@ z{75%`)$6jiyW?8#?%d3FTeun<6>lfw_F~#SScUEBR_PD)%3ab#f#tthXAoBkNS)%) zOoJUIW|8Y+d^fG(VBL2jOV+*09Cs%)gKY1`mIZhw&2Iojn^f$NbFK0Ih{-gk{yF6! zAd(WjIZgF*_$YC_(rzcF30Cz;$am-D7E$J`e?Mu?`+CCY6V%m1ecw(#;@MHu;mz#^5Q)^BH<%{BT1fV z8kNiPJMT@wp3K+2hA5xw_0JZ(J|)wgVnEK2+i<-&d&pZ2U)a!1JQ4E4f-sb`b0NnUcoaZz`Q!aBf_>Q42u$yA{MyO{?{X2m=St*} zOSNSY{dY%@#ZuLHX_(zAJAKqzfYh_1oFU`d!d8wa`bX<70VpJi$eLF~-+6MR?SZq1 zJM6wTIqqn|a)N~kIQ>8u4MFhx1)o$S>d%mPlf2@aLM8%w>NYNEnwHNwjv`q^j>Gl3Zj z(HCZwb5S{bkn~(qha#tw?n-IrQk87Q=G4zp{3e31@bj&285xn57lzDf0unT-9zW^n z09PKYe|$8NlUTD}dPIzTy*Y(YqZZ%qO53zkjM=1O8_LPgw;Bo6cC+x@a=ZeS*)r;~ zSyDOe8GeJf9=jT z@U(NtJ`(?h{q4R;_+dGfr50Ujqj)c5vw_sanwNKB;skgVT7WK9HALJyCT=s>l-Oxh z?kr|6D?<9rp_X>VP4_HtyIxp|w@jL?#v1>Ah7`z2Gz9aP7qRzvL_WdMA@_ zS7optR?xwQPFOPNU^QmuJ=CW_kaTj2OZ3%AtL6I-K|gTF?#{G@>AG*(1vhLkmZS8e ziQSRGPqzw{xlrV`*RLcsEVm(Xl2YPb>2*q z$E$|mIi06QEK5wRL?))9+$Hox9_0NEhXG5~uCFvQks1bHeD(8ka_;x$S{sEU{u7$t zz7Pxu)!JTDQ&UT>MMY7ZVzym=c8+r!8!_i(r&>&TMacYB!FAC|P1V%3W7#KZ7jnR} z!M>C$@h`$HRxkeHeNrb^^UYn+MvU!iCUv}_?=7~#X;w?Roq(MS33ms_SgfT zkdJr|dCVngcs5X4>cLFuD)`|RF4>EAtDFKkm(K+sK9wJT8-DGe?T<~m*L}XSKy|I3 zjmE_FI=HN{SZga=Xe{9M7eS>n!Q`}GQ)i)A_xNZeC~Wa=2c;G#)Yt$eIjz1UWE^se zht|78o+n*XYPb=wC3`i8=5hh(JAa{WvON`4R%pb>@Od4; z*n8R6NcP2NJXz!vjRvs>R}saB7JGFojdKvVDPoiPi7RY2yU1hhi8k$4?tmpbdFZ9S zKx`2=)b;uIv!2OM$sqx_Pg3Uo{KLG3%BKk6yI!!C=elh(Qy@QRaPQ^wNByyRmnXPm z)X3OZWLWRn197Xr++~B)ViK+((8Y7f)vh8F{8HUn;S`5rkFD4!b_CN`{Ivox2Q#QU z|ALO+8%>Xdt-f5FW69DQjCEey4?m+#i0zIgqc|)xcm<}|;eJ8ndE2}gL5vvs*LxUF zSCU4?Dr@YGBtBn1G2`iVzesa#)EsY)BE)D9F`0JAog%Y2ZT?|6FF$@n>m2jfSzJofZc0+6PXVfVP^(Z zN0IG*@AtgKaXvFR%nY6vaPQ7Md5224Gvlu^kT~nb@UIViGILmv2~Ym>Sp2s`N%y1V zb$K!;r32n^gr8_>xFMfT*gof!Kh(cddI8Y8H_hC#jm185xg=s7d=MWtwEp3rRhszgEi9LdUhZs$ z?OS9fILJW%O#2eDf#_hl|uJ61{0IR|Ks^V#)W%*wYQc@8 z%H`SBhn0gc9rskz=_s|@A%p%%iC=Kn+owGy4VT^}a^ZK_*C*q4QrMh49hDILb2Ui> z$qYhqm_ngXQTvQ5-|SG^eNfotLYUeq2MKg0_&lXJNYOnu@eN`(*5F&no&yI=dw;NB zDF~VAu^t8zrqqHHeymh;ii$#0{!J0eDa7?)YsLiC!7Z(%{;zizqNw&OZk=1#9sNF@ zv)Ru6Rl~-39z0jfjGHk=+qjFAry zPYHq`%GNf@x1_=3C+aVreE@M&=4+t|G5!XjKQj=AcTQJZt|$8XLaaiLc21WD_R2|7 zT#Xp?8y=d=ri}-aQd5WC`6B;v9^Wl<)m2nT#AUebvp$-opxJ5<-ntT8%y>5ix4|Hg zG7y3kC_dAEn$Ns#7|bE>;5@i$h|Cm-j0)F&W({*w)-suH&bot_Rw3o5${kx}H zsa9S+8>(^d2gC(|wf9B-X5$>>CuYBUn=Woql%t@J>QC@8*{3N0;B1mW@KpIMlPZ3D z2Zy-qPQ%Kg?^`u41h8njZ77Vv?fBX`JBPsTC{9uKb(3`g{%)Z~$b~11YTDhbv(bqm zcIKl0B8irBdn;4=aUvExj;m}capC8eF?>ED=K@hh{57U*&S$$D`;A)MQNkQ>e zx)#SD$P#G>^F=BUbqL{`-e^I$w(O9SkJ4#IAw7^%+(0)?0r4_heGtH#i>*Qfz4h?C zJYh_%j_0d+a`m$<4ip2lPZMCO@(xM^aNA!x-szKt5jA4zq!o5@+3s3$pVNX4Q2GgF zenFi&5{e*r$w$Nq#se=AP%q0F|K6y?Ks;B!oa{lQaOm!C9$$?al6sh}5J84eUMvpM z?q}M0vOfBd%Qu5YYbtXNSG`c zae!!^v#-kEH?_AcrKtAOTX|@4O&8Vmu46azPLOyj!=1xH`x1^*PUac1i|l?zXOCIu zUG3~Xzis2i=ig!@Kl$QSA&q){&TchTL0(Dn7yRf*pLwGR9y$%>=NaKXB_qnkD3Fdg z0w*>3g7;UsBY7J!LqmeyLP8nr4w;hdKI;by2T^ALfc&^8cbOt!BynDrcV9>O{bzZ} zdU!wK-L;Y`at?O>+6a}9)6AyEbFP>u8F*3eHM$yx~yUzn(FI8q8kii_MX zdU)#5YXnwEcfMzak>=cxmnSyk(tx{9_nA8MkdW>B(?t)w6`89sZ8S6+03_)-CXT)k zrpf3;Qa#Bz3(3{K#Pu(ts3nJzW(MRbpGSdK&OJ>qaq{_LwyTP*RQ=VbU&|63cPe<+ zdqGBiLP@pcM$-bgj(eIJkC`T=`U!R|Ub9P0iGDZHe_aJSqEpchyu~`~L@gt6?Lz(7 zj-q=HW3Ubl<206nqwYowT`>O!g7X46&G;LC+h{i>zxSz@6sY`B912X|@OGNw~8H~O(DY8JPKJnM?rt6by#ueYS|A2TDuH$U`h)Kddg&To$4toc~ zRKAr4X!L-JE|@y@ls}?$YKP#}vpHYPFW(vHR4F_4EZmj>#H#A6Zl}j{P9@`_r=pBg zGag-jiT*w?nD9Rg>365#insKo@`n|+&}xZXca!!<74~c0u9&e4v0t3q#%_Xw(y(L{ z??GHnuYs|Re3yqsRQtFRPcufqcd8=5xm^~yNMA@1{BEU=^WK=h)9cjl(0_4fXUOv? zX|{3?EQQD^FMfAm2fe+1LOIj3%JtZ#`sG4v+iAKM3V2$J`;Jj>Ku?0-(cTEtB0=C- z1^L5RC1yX$5(0UM=l#DP5-<*hz}c~F1#UU7#^uDc<#++AuPA-E8+CIpG%%53&HaYeZ*v0lEPcn9Qb9H53(^2pS|{;M?Dq1P?H#Q^V{{`JNjuk~_;E zYU3aQTfQs9KH;Bg%P(HPewr@`{&TEtgIaiL=1oB0O~(L^q;G0Sy;_x^RJ>Vp&x5~z zg^WVxHX!E^xAnIGpl<#a^QKGlY?n;-WpgfLH^s4?=Bk*Rr-Cd-0l@H#g!s~UeGG~Z zvA}OR3PUyAX&5I(qSinH|EU^(H02WA-Cr(NEO9r0qY9@+@=oVc!-7}3XD(6h)bu|& zTx+~|+VV<)JxWVqF}WGTLdHI75Ypw67$2f%-$rikh5>)h$maR94Rt!LARhz(KAJ6- zCrQ@aF)EEVQ@(?@l1lC~kcS6g!9Ur-@_05jwq3WU`KLiLQ2H+U{4}3>)7Q}x1gDpaatrCTH(J770NB=6b!x!I+ zH$Hw}E=ag6ANqRf{%PuH47j=fQlDR8AaRF)a5ZH~YUH(7Qk-GwJ?AN`zhuQ5#_lSW8)hRM7fy#ZGGqZ;xlC5_^xg=`1MKK43}&DqEO z()fNp_wV<;|LM`ggLB^R*LcqBx}K%zva+NRnmPRBd1YoT$+|VQKVHKgw(>0EO!CP@ zQZ>o6t{II}AH{ud9bf)u5m-S~H;_bO3!$>|&OkZDqJf0ATZu-pd>Hq3#moMnSY3NL zGqaXhdg$!!L`~135{{(X!MYrqecWwEVyrp;Ii9x*gEof>VD?taL%V@^Q1`wQwoHVV z*&3Pab_cgV33zInzo^&^7UrZ-OZalN!#?@k)5+~iBNO{wqO9OM7rcvZdH+*EIiVRW z5keJz)MI~+nPa&)95`Gauzp0~P{q_c0SF8b-+k}eEFlVpT&Y^IdKgmq*&qy(iQ;i>f zKKu8;-B@LomS!R9FpdHp-rI#cG`^l9jeE><9Nu1c!R>V)fLp>Uf35qMqz?6W`rOZU zQxA=L{<(7L8LS+UhU-%^llxBiG6g=3rHN016aNNUu19EX0|$dQ*sX8K2_r-kFtm*x zPD!}6>+e^CfFW`BrK1EDf9opQ!R-XcT)M8kFu#^rNHjSAAn>GEiiao5N9d)U)aj@f zSLBaQBmP(Zdf)~W!JV_cky)p}AKpT_mk0(Aadd=gKq9+CpgAmG2Aps;UvBDt!{vVH ztHdy`kIc&-IryKQ7ngwg{n%e^CC|qM3^0lR(ARoFmQ#AbCVw|~*Q+?pTGOJM6pAZ} z)StdzbE#z4fAt+1+X0DZ{9`u53~1vd4x2_TTLbZk)QV}4$A1mV#hk#>8pnLz$y~|z z%&|>hxBXZ)CpvYSakx9-a!`w^L(p9y)$-3(`+JR@!66POyy!@KaeyPN^$}L0R#c%r zfY73^`QJi+@Ut%9)$}F5`LXO4<9VI{Wa|1MWMbyaBGVlY|6Rd>jO^h&5zrfRHS|Z*?8q_Yn}XGUY&-NMCvXQ-dD?@`m&MF% z{FamJYWYTTI1&A3clG`$N2QOmE~e5;YAW3P^n};h>S>g$%`C4aRqGldMR#|-C{3Hc zuP4APV7MJ?@o`={CQ(v=A`O>>4}Kln_x@6Ftd!Bu2T^Jg9@8yN{c{C(KVJak;M1+G zt-J6qlXAS*AFa)_!^#8ay0W~@0|El7y-xp$azQ*s`U={KzHS0>9+GK%0i@nD0WG(xSuF=>^Y@0o$z3C3bc}q^4PxO=jWAaMJEdedXoI= z<1GVE22eL+QN$6`{OZN3Zb7?>JA<7CmKyx3j|yvuBk)~BKL&ZmhER%)lP>Ga81^&I z#(O4TI8-XC5m*xtu-R|-L{+}wccmo#BTLB)L%SJ$x=(5D^LoNxG}rsn;E2R}p_B&i zA3TG4q_9;RkZ3_W!1B|4H2UJ+SKxbfyaya4E9skq5A=v&u-&T>6x@fjQO_>=tvkLc z`DW(j?QwSw5ES4$UV=d9DcfYa9Er0*{^Vzwo(HD97RX;~kIs^|S)++W{ z?KIWW(Fs?juVXf+(hhXk_+7kU>1w?E=&yl=`Z2XuBJm}vb5W-@6a1Y2-o)FHJCsOL zGbSF{-%ne)mOhbZM_dW((-q=4Jw1(4WGOmWF`E%S`Kf&5H7n<##^;i0CyH%bUrXgz ze7S?^=!kalQtD@X;go$NnfB;{ykUC)aad9^#qEOLJlCC7F!HNuE(f_N>kir9UC%l^ ze9b(2>V0Np5kzHZHI!Suue{5~B&@9Rm= zId5z5V^j2`!R5u%%02E|3w)@#_>u0~rsZ&vWKAEruZ|AS8~H5T;@WKkWWGM#8yf); zIi?0*Z*-nvF;$S4zozp}ygq3&-0?$RmTCFw?j|6lAgi)i7Nyo7nzJA|lNR7e<+b}| z*r08!Kk=C3Gpe_rL4#T3iM48Pk`K*R2(bmr>`#z)cwGbb%)O3@9pEx*Pa0YHX0;k1 zHhPSj_h}0qp)OX#t~XD&rAqNi>XkIhlWN!6R2n{V*YcmUezvtKiy(*BknV3H4&z&& zVl5HF1(H22w#hCMIjP9(4Q*M^Q$g8v&V7Y|O7cMXoUNVhw{kA+>VrM+2jQGpPLsqm0 z>KF~pw&4NNNS9E0KC^lDcirJtYm9Sb>c>cTO9q+G^5^$>vy-yA0iT9*UC-N=q|k~d ztheZ!uH`brAu)RD=4ps~j#aG-luX3x12WnY=)06<72W(cGz8?Yp1Uk|*)(6v>M!91 z_7(ULW|}x+mr}4h566?Q67<@+dRW)`ah<7~duN)I#CQ&L&{o^kh$*yDc8x;c=>pWI zN#Z%17Yl`7_jctO6^v#BT!`lueX=RW#R87de31y01@1D|kPGZ77%O;D5S%s!t|0Q| zPFfP%Y$kY8@ zS+PVBmCoAlMGG(@-P9FDcLgd#BW|(IWpjNg3WTh9%UC31{Itr5LB=%=`ho}GQF2;O zx3o{TU=T+P4aY7ei<>FJYKbhiUt%9}DqM7uxu*Wfy;B`bYyxZ6h~ADk1U*<-x^`bG zWk+9a&m1NZ4$H^yC zvJ`45-!Xh_@s%UWH=Vq(W;_(b>i05H&W+-It;e zY6ZZ$w7IzE_{j*2;=n;1_!QnB4f$)(J(;CDuM8+UxLIOWzp)dB-9JI$glbPUnTV)G z5Ar%%g3yU<`ixP`4~N{X@=WyK*npHB1hYe$jGb%)7+FHTE$wv5-jt$(g*CdG^*SZf zI+#pf?ZIX@ZJ^Qp`Dk*Pip$S5XTMbhy%fTz-#ZACit*8Be9INF8tRK6JXEHNTi!`^ z;iLYF5D#p-^|6mi`dMegj3cZVi>6Yw*Kim&7B zKO|YJb-*g84k^vOi@Cq4M48QeFfv1p$Zqef51nYvj8dPMtr`vC1WGlRdlNM7uyc!? zhKxMQ^jsP-mlq8zGz%n8WlZdwCC^tm#^m%2ALnpw_}h+s6BoqUO(k!d!WY@&ovbWF zN%a}c>scoQoTCxO3zX5a7Rf`TrDIW4DH#;WjNG(OV-*)AP?_(n@-XCOzg2(|`Asol zbtVm=G;v!E)O!Qj4{|Mv!q>+Fhg*IJH=vjaN>klgt}N#cJOi#@qGoAo42EA%9vg9f z3MKEFbV?tdK_V;!1O-*NvLEcQzoA}hPf2mgulym?o3qE$nMoQTBrvk!JcvAMZNA zJDs>!O0gC?y{0hpIorT^fKi`6W*cN8e7p=!G}-E)uskpycBERwS8H0w%Ua%JFaDHg z`T%d@lUmA~UE8T?VtPH@UpD|{IrG8LD??U#816|N9EC-Xz08)THK0XGOt}T*Z|qfB zPk#(wB%_AuQ~Daz3re#~M%L|&cd_7rHjT~Hi!P62%rS>w`8>QuO_ZS!k!9h*<3Pn#Pz zdOxQcy-4kl9pToEX9noRT$&>!x6sl;G8j-G@jo$h!%bL^R&n;8zJ%3G^0`yq8aDrh70a`QT_UgY`a#nGX+OEHCarLw^FX z2^VbbgGi;zo}=bQB@NZh!dso%qW$s;Rg1u|z23zuQPH~KN#DRm@jq6ywHb|I8}b=+ zDX)7S+3x6O(ZKTzpq7gxQy%`{_g+c$8JR0!klPU=7+0SdP52DAsQM62rn3dS*Xxlz z_O7mF4*vTIe0vtyvDH)gT=ic62^9iaLr(h2UlBcBe_Jgdg3k9@b49U{ODFUA&}qK8X7jzBf*IAjh6tj?Y8_dtdcWM)#l zUR>|rVt3EV#WUCk(`>tA@t>+A&D<6Xjvu~`onjYSJt#dg%K>l*R=&rxprquuatd?`Y)>JNJ26y>*Dv#DZn>u`PI4?bhU1?9>pa zOhjdN`H*MC!Uk%IGFZ(|x1fC0EpHa~8JRfLpG!&GY>t_1{L1~@Sj0b*YGKzUK1|1= z+4%;2g*nJ227!(ksZ!I#P*%HCM51N0<Fc0Wh zhT|iU3s(6)frXN2{MeE!PlXtwNLjhc4DPVC5O23VE3h;FMOBru617Ihda;I;&|Aq` zfA-n7-Vgg0>8?h6?cn8{fn(HnG40Xt#ab=zp?fL_LmiKB4WxbcM&it+ye=6`9sQn$ z`Wb?^Xgg_}cXQDt+55-OceCWGbB~sxDtT$RwnWK|Y(7fi*hPa)cOR;)GU1uwhk&E6 z6se)z_rh1#9EaK&b}PZ)dqH=7b0MH&^A+fC_S+X^f8vd zZUS{5QCYqv1v^G@yE@*;3uF=lIqV>45k(}n&2mKnOe_n;Zzk0z1omK>x=D~Sg`rB# zH`g)xRD9(5aSnbt+m`{NLXYlZXe?hDC_2?$Ct z4zgLw-7f2?4a{c2je9mW9>G^)I@FGMe5p{HyqNH72eZ8bT`9SDhx? zedw*yMFs@lz}5E1=JN%JmbQnms-M^3-FS>p82!-;XQe<7m9p+byKZ%;uHzb!k94o&w(ct9wW;Ng z!hdy(&S&4+$*tP4cqUB5?>(x% zaIp$c!S8pODQIz@Mr=`$!w)he6jwv3b00cXv$aJ7N?L=J-O{m`)5Gc?{}h*kX&m+i z{#*mqz}0qOM`+tYsrh0E&Ce6HFCDVS-FU>8NKYtHMaZG>kp4#(w|J@-YC^A%1n`kN z#~P^3U9I+Rn@HmD5j9hUEC!0oCXss*`F`%?R&&JA?DZGXp#NnWyVosm>U;dats|~A zD~GQhfR+h*kLwK2_qgL-eEJ3Al2_P9{NPpo%lAxS`0AyKt7QZ3YcHb;-8zpVOu1Cq z@$OP3Gh_!lGb<#|`Nf1|L*sjOIUNor(khgr@u0v4BsR( zM2Z#%j!pVNh8%s!Qw}pf^JLW)ml#oqVc974%{Kw)-sO9%QShbc;GGmIUg2Oz_Jm^G z8X}<5+HQm}GO1}Jbpt9He zZKbI<2IG7UHHfKLqe*aU%bWKqpIFnZW8)GyUjLc7gxM?OPvr_a#;UG}+BA2=37f|| zbOzNPg7em~FiQKx(ELG=K+)==ot7h8qsI^hPWd&*B`TKgIF}Q8-N$@$nr*h0l*>;{ zt#6tJ&ae|h1uYoKoX${lgN+(emz1K5)p?=4uCIw?P;u#}Ew8Y5?I=szDjU51%rbVB z{zWUCz9BG#^(ShJk{_&jeyLIv@6<7{jLAt09C2Ms##uNaudAsxLD7L)TIVw#NWH%) zF_n$8Dh2)|jKP0SgL{Aexk?W?x6(JBow7u^0W(anj_1%|Rr zPuOKGhWnl+i0yKP-sSeIy`tRnmyQx zn{nfrE6rCd`{(|B9dG26#a!%OWUtHtrhD}%`Q@~{E>LeM`2ftNp*nezw=;)Bp073z zqpxAEP8!|WX@!#U(b|gECCWpqRLPX?FBWiuP2K_~Co07)WUolPZ>OT#YU8mC6nSH# zMXsHslA|L!32Xd|M0|I!V}vY&u4S82XsFE|>J*_+Mp)M-q`Ic6J`H2n2hjfgmMBlP zPbUM`17v{nzg(8GbG$t~Jc3a94A=jXf z;T@9crDa&tdxJ~Gd(oZGMsJ(Mcq-uJL*iF@;n zj*K$oUQoa5VWQ30m$DsOpS4A0W|E*4B41;XLMz8qtQNkSCl)#1pec0VMMO_2a{ zGSoFLi;;&;l=sPZE_(O-_~80QNj?LnOVg@#1?(%Kq)$5uRX-}k(6!uc)f+6tZ6qPU zx`CAS!c`9;mRlURpR~e|6v6k?vuAmCGsm9V&M+(IB$oCO%mnRs3k2*IR0Y#>a zVRhx@hxuub?2VK-f5$2Bp-h6`Ai@4z-RvHCsVDfiqSpLQelvqL^iL>KDBe@u2w;>^ zn2{t`JB32!qgwgY&P?q}^ILii5yR)FC=n2Wi2|g9J1)vhw%B5<3gZN;u~{pj9z34t zI}w{>JsbT^-`sH^ekRSo17{H0jLAWW8Cm*d)@)Mb$xFxd6K*T(Or=%3#$8X7Nhlfa z&5pUMKj)-umg4mQN^PaYe$=u#5X_H5 z5tC9JV`}m}c5>vXJe%&(SAecu7IYmoI{S5x^T(Y6>DM z#!)2DD=XPKVCz-LBIpzqpxc~z((CCS{{6*I*jg0Q3nn=13lUAP#n0kFA||6O5gfO3 z@A-6dSh005T$gHKYU;%H&Qw}uX+f=4^y8y%&-S@5p@=5Ua_y>^y)V4$S@{~3@sALf zmMSJ9-N&CxviZDnDIdO5bajKbSPu3#tjrDC{DG}Qan4rPP+WJI$HOe8+y;;1`;qHn z!e%#nKK;bWeR%Nr$?J91pKhatUcLNd*m!$@k{`pvYIRGlYQ}rXp`~2a?$y{Y(l72+ zxs#11z7=*OCC*s9R=W;c^VDiM)L~RrM0HgmGck;jh(y=5*XA1MnlOTkkDqeG;JY=E zqa7IzF=ob)R-*p$$)^BjRx!MKPWRAcl*ZwD`lEdSr~-Wpf^)uWNfTnx7oZ5=-mRJu z>eawkdY?7&cuqi=bb}|DJNi=wtCwn}CK{sjE5ftWd_f+@o8Rm`rl&$UPT3u&eM}uq>LgSr60i3pad&#S?H%9NgNg$)Sc)?aQ;q$I@0FM7buEKReLXZqo zAl69X6N2rK2+BR8;+*_>E0?Ya<-((1sfzqqnr~7vAFj;-L8opICa@aRJQ$69Ne%?@ zGqQX4e!fwWCB_*u%pMmOvH%weQ*h zoW#pxk9z$hE9*qVseoAdiQ8EB^k_U0$(Va|sJ8F~YZC%=Vpr?uKQ3{wXwSplluOOL)Py z8c#~$d=4_;$k;0BiD;nwvA~-x)I=cMg67V~ukd9jbP6$=R;iMME)=+Eqg=b61$$E3 zX6`L!*3j8{sC!YB0;*jXg1cuF1&TraGgSET!P|h5SiLgF7S|J@ARACN^JZP_>YdXT zMeGH&9B+2Z*Wx{TgS8yY>EmHC!N0t zyBD4@jz*3K?sdy$24EAg+Gqp+g_qDs7H9521EZn_D-Qsc)=ckXm8qIN$`5f(Spdwl zKm!Fkx=i(P#A&~k@$cjA<96;3pzuauP&NbUfEY@d&nvRh9M=exXN}$*!m3i6UH~-r z0rlJ$)890fYSv!6X4KP3JGMaSwphRtI`}I8J%&bJ!?^tL`%^q>ydV~sZqC)=bjA`s zf$wJBG#5Z#0%YZgs5rCD5iCW%Zt3R}zO6fogPy$o?zCywSE=eC`gaG@aK4|`*|h!f zQs3rj(gVsyTj2f_^P@ZqG_svTJ#mu?t%XYPkjj zc5%cXFN_#|R{E5@)?_P>Y7gb!YFB4^J5vn8ldDl8!@REf6YLeQ2{D|mXlso19l?(p z*Vg^)!=esK-MNN?>iZfQxkcNQe=Nq#Sf)5ikK-L&UE>;%kA7Iz+{8cHGhc!6k!S*e zRIy9XXC4U`NGGnXb#`fcaL7lv)L>F@7^h)~CSxTzD)!i3wHrU!8H58LgoU&s8a7G2!q|6#QXl^j3x?Hz&(WN?fx1wAaL#l{7)rU$;OWPyZW^y0V(_Si%XGwF?Dh?&U4QSK%|&L5M;KS_YFb=;VeQ z-s48~^8Qdydl02=#_-4DoDIW7H{|*e=%r>w-tu#5pEFliyN#w8?4qsnXCxU5p1zay z@CVF4UtPJ>rb@go7xW1t(zz)SB)Fww)^4FjpQv#RsexY~=|SK+6QA>zy(xXl9aFhB z-5ju0VJWdx>+xlDrNOoiJGiu*A0gsIy*fw=GqVkhs>aJA;wwF7nmRR30`-MZ-KPn2BKUbyn8EjT;(TJh$ zw9ncUmkxg2ru7EXrI*HW;SjVcLDhG>6H!(jytb83K?zSlqdqez$HWuv+ii(1KG;dC zTAVa!f3>AX^Vy|-^KCPXju{?LJI*g}Z}kQLesC(!J{nk#HLpGMzrxz>I$y-0!K2Z# z!M$nJx6#?+N7xay3AyEuSUs2fubqGlxm~Qr_^&oeP8A?F7|*%|cSNaeCInm?4VSIu z83rShu&{D{S_mjf1M5;vObSKoi^>+vfs#V<>eWzhmE!Hr!FKz=U1HR4rJBjhJl!Ic zA@i(YkOi}3e@aw(rP-tReea7l4WiL%Y*Rpfy#awhFQ+nGEbAvy0)EI&?e*ZTW1S7f zkI$ef6hZQsruI*6L}5Ir4QH=rUDaPBO|&-(Vn)i69brTq4Oh8R{rz2gr{G>?b=+j_HXQqj?C^md=(er7cErI}O+>=nFlsD;>pCTwoS{Z6*g`=p zVq85a4jS>J(9cyL!G<`VNQ@n*cJJS+)&nBrw(y_$bGlE~sDiL!3D;yZ-5zonyU0m) z;lhIL!iKT6>%4u^alRDn#1N@k>i%Ac%~rr<1O)dQLBnfO)ymNQ2+-+@+hx0B(xy4C zX-)R}2JFskq|ZoT=xrmThHii;0MOko5F8i{9GvkFzn!j72$Q5d2?0>n+007NdwHqU zxm_ohsa`k-Q2j5AEh1qh#H7ewZlH}kh{1GUDj)VvCTCKlsBh{)mT%3zr?=k6r|({v z`|;6~-!ak&)8G#`?wVx0dOVZE6aWdzXWOu`Ua+l_xhp?gFlYmEwg)~X_AlPxEPX16 zNDV5J(}RmIb8S&x*jFBH%33AZN;VQE7MT?6h-1Tz67Legh#~i( zv)YZIYWe`bH6e{S6u_$E7U1TqJ;M!}K&KPX9Uu%8AyQt*Ld+0PP!7Ci`T8DaAb)Me z?F#>dRDZZ{3pHDY#_L0Vyo+;hPTU46#7Y3))wMm&@D1}3qRx3G5*R(v@g7t~J|tVI z`w_kee%q(r7s|Sm#%JcH`IG4$hD$~eM**W@e!Sc9!V)=8zsRlj#?k9Ce*MSO&2%&< z!Ayb{b|KY%ARUxoa2XB{x4VI|wWN%NnzgnvNp!x-Wc0;L?3<}Tqod`leCf@OGqyHE zf`s~|7i~`5@2_9umV0LI;^HEaT=b{a%2Bd=Pd-6%4Ije|o~km~+q@uvuMvAdO;i@? z;v_H#dxy-I?8`)NoY^Jmx6pQgU|0}Llhe~Mo=y%nQjl(aw$G_4h+%;OAnn@PBZMQA z4ZPGAxxsujOb}M;;1Z$&x{P;$4!`Aq#c=(UtHl}?=hIGbZ%1tv&3TqcK9YIC7Nroa z=DP4hag;V);xo8mhLvQMSLi&#UK+7$D)LjVONnHyI?8RaKZgQzaxCDv8kCLz_eIwD z7t>Kuvc0`2MEyw8P1=XO==Sb;<~7gD#KqE8bOW*sVm`#A@V%9FCQXL7Y~Tx;DSan{ zP3Qb#fSb9SdN)NrR0i(~ggOb+_$lu-`LA9QHa81rTQfi&H3IwIGXI1UG1iY;B;I4F6K7YG zK#L#g_^LY3UsQPlrLFE|flZVxd$eReI2czgZm#ml6!k+fy-y_8Q)6*FXCTJpWx%)f zqAuUovjamzNBu;UH9Viiv!i?pM9eu&ZM6||S@F(1rkEe!h_?y~3LuCxHcNESqAKMK zQBe#~hQeEBGVBZKFQo2^N=6nKq2q^0J3#DH#x;aLyc06)Y0v^iY0&5Nr+m;<^J*~) zXurvT#(ffvh~C=B8v!OjtNj7CSC!^7^e{{aL0e%^{9A?|O+>f@BR`YrIR6{;nr;uc zf4*q+^ao91TcC-B*?BemyxA#{!D8IP@GF0+3OB!OSVT(7#%H4f;|GD-#@0lP=G{== z&d^wg_Kj2@99D+Xf3+b2NjR(K%FbtZS8_6xRngM%*m9tOo+aXABea3VtWLAJjl zo>L{Qo5N$oH1XlyRWjsO_MM~*?RGdQ_DsW4>y;o8KAqlR){$AKo`he^k?E;hg5P0+ zTNXs~3b$8tDMdqznG+|D3miWmh`#1;Igco)LVh)rICSghO7u0_$V5uD1whrCBe?9D zm@kVt2Hxu#MJ0!9+tP&+j9x~)3)kH!A*FM3yF~q$>x+8sn2G~3>IvFrIa;?$J@9+S z+lwxvs)4=|OG{OF8S-R|{<(hnF-5GIM+BJl0H4AKeTk8jWbE``pY2>-{>nW#!a!4c z3a;z*5uHK5+Y+Ueff$;P;Q5!;9^RJjGs;nsKG`>obbAkfaX@}p_(LHXGn*WO5+FcvITn{#EJ ztp$b2jGcV=yc^;MQYhb;1vWkPfe(b;kq4t&&w5&N&Evy=5kjNyf;Pi1X!czuLMkkB z)^C|_b5w{cIG?wIJe|xdVit-ObSZgd?$ab5-`My00k3NgXS|HLBW7^c|HJKb7oYC= z>?HzH9QP65nYn4>c#wYf1;~@q@@JK3%V8s;O_VnwwK8a$d*_=A07Ox-M|4CT=bWMJ zgA4HKl-wczq!Y*B6IXkUa*F-!KxJlV_@k9Gz+F!hoO2 z&!d(t0w$BtZ{nNMaD(P)k5-M7Lu`K@(QY2W;V%A2W-_5B;U0JfnF)V0gfBaFChy=a z6TrQ?_?i0;9B%SP&;d22Z|_dAD4%lcU6Ig=>hGUwKB}d!EpD*HhfHcfNowgkLVEvT z)Bur4X7tv@Ns8*$@*H9^bINkMYT)rk1RiotuYbAjG5(zSdrg{UieIGv@>s)3pU!mj zw2`Qy=am-krd#XZoWB9z-59>(q~Ge7Ygrs`9zz!zPl|V13p@uMbo=yv>)>)#2iddlq3n~O_5d3J6-(|@p)AiZ55ZRb7Rb2K_42pg*Lj-Nzu1wXkgyXImZ zMJJt9BgsLmb^c^W&~zcz)t*qCaX_tOWK9|ASYqX>lyP^VnMuX~DrX|=)PcDAQo8ko zI{x|e(2~7Lnx}LJE@r0HRBY@+A=m7TU(g^>?g#8CYZ)jvqvccIelWICjQux%>^Q4@ zA?zac_?T=HUOV$2xHZoe@pLAVKkJx*aCtHQ%UI#d!>{@)G*i_M{f1qDu4My=eSBIX zP&|#f&_?UgawSr0ddmOBaNNN>2Y&r8ME%X#&fc5egpwT%lRST5CBUiw`+qkEK#^XD zO6jDW+jmBo=pxFnKQS;6kG`(=67vfj^Z)%5T(j%(+8w!rpk#B1S@8?398ek#IXnMX zIuw+Zvm+E8BCDQbp~`$D9*)xu&v7XE2i>?MxQMSq(owhf`nE*?4L*w8OK>@3r9jl- zUZ7@YafunEFdpj(=uw5y7nueV}9Yh_aEp$-UINz_h#G3KR7{Y z?dX1vFoi0Zg#kSun@ zh2w|*a}NOySdxhVGZ>Drp!tqxnSgth8N^t)A3dzfY@om_E(R8iMP|S5X#~@t2hRT} z+<;*mw{>u4&p#*ag>HIP(vkTWgrp7;7M?4zb?y1>OdM&&KBfBB79%%!D0lg0UsQteld-B4yu@~}kQ`Ndt_hQ^c9moI8rh;ni zo?V(6c5>RaYBbq1^KRdXK)_1Cx1YGRc3`VC;|a%sYe6?3?-p-;$aCcg&+|(yCatZ! zE~(jl35Nf^7STq*V0YJNEu^@(bi3{&jN0*I@3-kkjaQ(URr~gU5Gwz}&-<7U0)7|z zzKnvTNeipG3)^Eir`zuh-+uDXZth@Cbh_F>dNh-ozG^MR=2N*7$0+zrG`+~S;~c5hK4qlRol0b;~k*yBd>)4v4ksi z0am<2vOMR*{)L%N-^omYR|9Mu-+%7s2{v#Gzq=i@zxbVVR&+PWpF{xq_IIxJufqW@ z5BN@A@Vkj8K65eq9_j!X=s_T$VdwbaA9Vk}{vaJ(pXK=0tjSB)jO%n-Gu23_G+LIw zOo{NHWc>g6irw4WbEZv*HTVs4{Rs8l+|{e!ul@Ju_=j;}#opf|Yr)4jYQInSU*G-f zVSoPW->`lS+xtc!rcBHztM={lG-~CScfdx1&8H@2Af+o$ zQ4&G5nZo}b`Gk^qWqX@Lz#fGPv>fZuVjl1~y1`CmcK4JNvCXM2L#AwW$>J_(OMU5=u77(={{1VJ6PDFFn$Hoo~1w4tR_ z)0?ohCmr+M$k1csQI<=a^u51p_2fW@#S5LdPtiw0**ulo1O2NNNL}5RWE&sx{pS2m zUWRHsa1Y=$HXzR0WdtgA#b`jV&rNtvu!GAs<#tdh>>N+WjiRTdT_^c4-zWVJV0zM} ze;L5o6NS#?Q~dIUh#YpQ>(#zd&Px|8P)OS*G+`;~2%{g)WT!EzxR zTT7^bjZ>BNPwr>R*Lr2fk@QRE;>L484sr#u3+^KxmeQFw-_AS`4>WiPs4YRzVAhg& zRN(YvmA!r-{bhIiwe8RR4`bF*ePu(Of>2kdhmQ({SMMUgKG-{J2)UwbD%wN;5c+S3 zFbnV)qq2|8C=pZjbU0DORMu_q!H!_s?n%6yw#O>oPCVCd6;6LI9DIuh9K_v5&lTg) zLC+-y)2OZ{3_G)KyV)-kKn-9*iBW53TAJBp)Nqi~L3~Y7zbz-(3 z>B}zrprej#XmeDy1@MY%-lK;+_Wa3(p71oz*hL23nYYUh^xi)y`>kj4Fvo$M9i4gO zH8Le1x>%f{{o$wN3Z0I~);?YF8rtd8()*cb;r>e|$V5tiF7hp7CKusp3o32@xJ#39VI}hNidS7+ zbS(tqm#lOw1ofwWJ%PA*^isPJy`{Td&(uUihu+@%`gO)<4zZ4#;1CMOKG0xsKWXK$ zl#Ga!ND$h+(ta+!M7W?C-z&P;$g5`e<{y$OFx`PWzOWu|xy{b`aq-6t?ArM|zvH5g zQ|a4JKZm>GhRo#v_vbqY&()6!I<3zSl?IXVz0LJFW$F1Jajp~h;WO+8Yma8S<38jN zgsygCwxvdXJ!9p8XWWa`Ri-8;L#Kb88(fhra$v|ujr*8{&5^&}@Eo5TGIg=I zd+}LqVpN1#-kUB*u;#Kfa#*&dN&TFiH0N+S}IO z`v{nqqO~(VY2#IAN`AS4bQ%lCYCKM1>Lt$=u@33_bBDmY2eZnje-u}Lay>Zt=G!m> z>iYn{FHv%&vjM3qxZCf!bAD5zpo=W8S&0w)gr*nWy5GF58}j9}>3p4a&fW^af7ajr z3re1T@w`x6cD60qe){xWe`d9_yzT8TE-FQZoqKqc2&vr1ts_4Z9=E>x@L^)&%xU8) zvs#L6{yG7pb2-YS+}xx0V=iQ7xe}#U#XT>-(Y@`KDuJxDwSxtt0VPS*>(GQ#;A-O2wLqYL3TIDq5aCD?QUpg01ULTFhC zE~T9S#sRoy!Cz0Y>=qv#Hix{j7Q8NzVRphtP0tjHtCh2?Sq`;`xp4)qJnR0)U~qU| zPQMjO6-`QKgvr&^1$OA~Y@u{z;KR>W40aEl)>k9G_-Il(0Hf5fhU<=2*CT*lvnAZ4 zYRdvdS9vyDqW%!u#M;he7N_14f*?AWjbxt4sK0I+1sB!$l$#s3lI$XcK0il|i+koz zAv-_6bksYUm6LDK;Q=1Q;i_F;Q@{Qmn{&Wyt|v3>r5WroTnd9j9`VY-p!uXzvNsRKkd+D4Efy>WY*f&DS0NtrkiWV)QVbSO=>N1 z=ilu0JS|vz?5bA&wB4zGgEu*p2hWx@XaP7j|!aYJA?X}n)sKF7O*crJsjO?YIgv0{@ zNn(b%A9mXv9EF{8fwvFIKa=$f9S@CCkFa#c@UX@^DQ8#(Aqe}vzWjR4!NK)xZbV!^ zJ!~((H3Er+thGk)NGEycXh)f>$zPDA24+(9!%2(Aj+oykq*1 z(4OFLfM2h_aH}HA&-6RD1_@%k9@QIdy%}adHt_S+Zx4Gym}v`RxO4Z_Yej-fR~2O` z-vzqC05k~AI~&0M{mA%0&X&OH#S)iJ7V^gN5|3}{@WpYYzT4%UgczUU5t*)a#r9f~ zpPN?;jdBe8>gm%T2{!c*%D1{2a`-~V*jF|kukuO62ixS@oDZh4_Nn>yOK{!r%o#=J zS+VlIcf-cgUS4^+jIBcCg{ww;6(8`jG{+^w?&BN_8DkA;#cS36b{U`aakRDI211cb z`N$BO=|11@7mkKjtJ3Nx_D}xmry6q{-};n$S}!UF3NQs`FB4^GtY~`|HR49cG_$>< zh+NMBwIi>W`DlJ_&|mKQ9PgEVLs1%-$#SOg^3x+k8EkFYnHVcnbN$jS@+{+`ZnOr@ zW+X)G9Dc1};ohO7(VDm?;c-8^u_I-{^PkENJH=1UH$PsPxOJ!WCrj@ZaWdRna9D9r zjL~HyW!*N!>b<43R+Gy%y21A1ke7&j9cOnt!!NQ%odAl4w`LYPy;Ovc)QvAXU%0?XJ$}3aVS4nFbsHJ1CCeh#Y;4?cN6zAM- z#Y)8NZ;UM;q2G{)1$^lG^x>KW+rT*=#ux7iVf)m@A6Gqeih)9Kw-P>It;T%GPZ#p8 znM%aR%y;}!a(rP>N2`fU8gb$$^0#c90$Uork5)3n%urgu;*JZ z#~w=vM`xTrZ~QAT2B?M`&+7vO6JWwOAm%1%mn!_^phQeiNv-@Sa*iwH(ngl4Vs|&X z$}9aY2z2K>KvmLwL|R5IlCxG);I4PXoH$aZwTrJ9ty`^Ih>{5la5JXWFcuH(S>~R_ zV)e^2QDi~E?FdBXv1_ zwF(+d;?4A-W^jcw+Yyu{uW52TF1qk4`MpiEbWMr+Cf7F7KG=VkwO~oRH`Tv1*$<@R z1|XDI`QFgRF?pqqCY_GpxuhU|eMpo|NtwCl+pZGMS@vGMvh=5zIsx0iWyMseX`|e5 z+z7=P#%6uTX{_9&!tFVtBXYAAOE-)1Ko_aDzC5gcdp*l$9>4T#ud|B%4MhLF>Q4(r z3QqHc+Xz|ZQT7wr&Vd z^2H<3*Toz~93tm)a_M4vWe1zrqHBbdUhoaITYcR> z@U9rO2+;ciP}pJ5vV618UUH$`5>kc|hLMMK6Jzq7E9$TC3-Q$;SDI#4n`)kT8>zKr zKY*;m$MF|*w|IrzjM&$7K?|V40##)(6vLK;MF7LyRRDvP59pOA7AC$lfe{jHR$z>c zQQloh<-^5{Dux%?x;OqD5NH(I+pISTBN(lG>4p$$vX#0#p(Z6<$g)UZtKWvHF(-%| z)N}aaWkl<|o%g;C*JKsMXL98SY3}}9EK>nt&TlC|(1J!SQ)oE?gbaYJ(KZ)X>om_f z#DUocqdVl)7bO2{)ef82`f#up3&fNz+pLXWt8nm0wDDaH=*`Rf7*8Zl6xcl5)iIM% zL2bE4_EDRBmWdKcdfMTei#Tt5t8TmB(&b>b zIV6Itn|%2S`|xlbORwBXWWX4g(8obXK?U8mS~$TmblDXle1(|FUpzGU7-Q`a=C?j8 zL+M%_3A6A+YD)NQ6}}A}=#gtwjY#*U|5B(JV=a2V4F80QntbWXae!fc{wS#9Y=a`$ zp7nsrW*u#1j3MTY2@xv)!o7d+0q$!@B&R?h_JaP7_TYlw=F3`qRQ(EOVKQJSt{tZC z>NYG7+HIvHVvUC84q)EHPQnVKZ9U5R>)B)UQmH%nt23c;O>ykyo#eBb*kUCHJgKf? zL@Z(BJ67U6<2io<@Q($0#SN3cSYja#di6IC&E?qX{fSqdn(F%uH@kH+SHXL6~bj!jyGwb z=GIkhTT-!ybzR?-{_zx25oCPiS-AYVK~Kdw zYMxSt+N`;QV(_YUC6&59C$6G9VB)vo*gS}9o?PD2!=Go$-wUDtczNsw%ze@+kPgDG z{(>CT=VihWYsOA`RfhgWaX; zMRHVFUDVzrWCVD+trpVX+TU7F$i=Qi9#Av)R66hcd|QR)6{nV5vuPdY5q#>#e%iY9 zSDguZ@YJyTSk$L;HFYWbw{s(?1&ecydh-Z%#o@-!)6=*kGZ(U_jf*@h{TWk`b1+6m zLfPI~oDn72h&YgxFh#ixCsh~qo`KykC922d?tPpW3?~E^RN@uHqc-QUWfd9qDbf(! zC3(9oJZ2=Opf=G7rnpB`xzDA`^#e`^R!grnSNH3#TAlmYujo_Ewq|)YltUl*xm*`D zWFj(fj(dJEKckm*1(#e;un6#|`eiwkw|dZn?)H^y{>TrUv_ZmWCA2>!}_;t(~q5%BRV! z9vq&jLFA5&FHP%wOE1FH(PzGG?XyGdJNmQZV0`F}fvC>qt$r2mx)U$TBha~zM|J_K znC1(yuDlR2E%MIYr#3fth4dAMbFs4QyM5`V=FJ1IM&JPFyW_1y^k%C{eE-5Qp>&d> zmKmRF5tv-qB)3#sY_*~o@cCu&iPm6*PZG6#KR*^RRyGv&iAPviD|Mjk&G!X$yTSYb z&+$p)#A#g9&o`&^yd3Xzej^D&T$?`M$?w1Z$yC0mD9C7HVE`We`t|E9S^ueH$mVOi zebjx|CbPc{mOoJui$QGVVe`WycE`Yldq!fVkbG_?HPUs-2Yd2v7drkF0RNG z2`jQ`me4wt-}th>^d{vdTak2o`|9$&JiS*;%wjv3BRQV2=Z@7^jpL3Xcov0qEn|}} z1w7O{!V>s|#fZo4$_{4q#Y^vZFsDh1rIjB&p+P%0M%FwVa*!~B$wu;wC+wxW|Gdu4Yr_a>gSK?gW47piZ zOHMIKqF8m$qpkW(PsSJVpW;L0?gGyR2S5jm_TfBLvq^qb}m7HHi;W(SAdFK#zd^uP3m#s!NOT_)ez@JI8D z4X?g4UHlq?TV*@K#r%WiBf|7-{{PVR7f?}s-}^WY;~)b_Bd91TT?zt{gQQ4zx0JMi zG(&@gASoc-4T3bxkRpg6(lvy1cMb5r7%zQ){@>qPvs`z{%)R&Qv-h*>Ip+lH5}={u z6B$x#30Vyd`L|_QT?2XI$0E`>g1XA`d>=!6(5^nUzCNnExTkMVsuwEf4D11|n{Z z*bLW}HiweJZ@EY$3sCx4k&H&4bL#76^aV*zRrjZxVA0oT;0wk23C$uT zkz(u1nG5N)8VkLoTbIV^n}IU8N@Gl*Iw#B@d_!n)B7k_&z#u;#=>(U}np+3MncBCY zE^-@uQtc9_>V7NzN@@&IrPd$&N^M4azj9KGS1N9W(?Lf@!p^={eP$+;VmCTIqB(R! zs%pM3)2~s}9!h)c{KnNpgmRSOGW9&zl$6q^r+F^|g!5z7g{^rvIFMXe_Mxoq7}+Ho zb*^Jq3QHcq6plY1`ylZ;BE(|uxH`on2Yw;ZGoqAm=wj8jO^*?&r&D|Q5POZ+hC$e3 zME|kd>D`@AjN6NyUV58Dp|(@cvN2@TEiI)8z4t`wFgXi*a)s|rRQqh(fic|x`zR$Q zz*jfX5XxjRXT?4-V!MbqR8#DOoX0J-TAr)XJDo5`us{N6&>^S{*O3Y{IIST+^>m$i z2~tkDi@qIZm?MMxq?(fz^GZ&@$5RGTg<*L5f`>TKG|En$$)2d24BDqjys=TcHXDJq z;9~I6VdxYYMwg4 zU{=mmE^$O(q)!`MYEqpUXEcjf(7a^b)X?f%B30jGjVEuq-C{znHM8f_=$0NpeGU%4 zArQW58u|JgZi?<=mhp6B7!P5j3{}^+BHM40_h$f zJ<3Q3>BLQT$aa}hE$i{pc9|TEibDf?!X*?yIffB`*f6)LVi4Yu+=7q>_SEh#s z_KWrL4YEr#O~ex?Vii8BgVQo42y~m`FOkU~74@lK+}&>#PAGa5c2N?Dd__w=0UYHU zng)aNd+&xj!SmX9^m%Qhn>IRArYeRt@Au-4t*RtzF^rFRI|8Es+M8Do z^8j1FA6-$-1N}w|!a|L_J2#nhquiZrO%_w7xU1+k9r^mwpr;7$qlB_)#>(Rv9p_XL zuZQx1FLMsmO6K_5Cm`z@NF_rT(Ywbb*PW|y;Qn0xZM!!EZ6GZ88G0L;)1zkf^1IgkpQ_4aQ) z5R~!>ydD!n(HAkBl^U0P&$Od&u8C-yEiYU5#GWtg>@G^~huuJV_2YaViPKecD<`_m z%E!|=UMvml7K83~lZ|xe>uY+mzT?AP zHRsZdX)yHMvVFJBaaQnrD7L#x!&3JN6$zHe41Pz$8}74v=--jmyDZ|S(0YMYG6*z@ zC>M0)On?u%G}U>{-O9S;UCs}(aD3Bjf9g&eC~P#z0K>GY&yS;FZZ@6u;G{eL?AZ?_ z?f6ZHR%P9Tl(8ayi{X4P)XtmgllEZq$&NBt=SIQIWFJ zHrK@TRqv%PEDlSr`Pn=+C6EURjp%q2*2h4Tre89n zngOH7Fmk}h%WNt`dqHRC7@7qi7G2E7-QOZ()R6+ODARGsI6gBmwdmBno~^-?PpcEr zDATLq6gDzyei_IKbkvWU+XfUaQnK$Gqo+X$31?#g2l-;Gn7YES$^4Rfyx}sW+Nrv^ zOg{&tM&xBbbIn~v;$5g-C#ZIl@}kLps?%o9 z#xIQL6DRXb&9(PfbTTA4X2x0Q=vRtADdv+_aUJ9zVO{yKV~pC8wKUn(ouTD&I7yh` zxJm487Wf(fq%HPj6+u1>>AEdlORO-UKX@hgtB)@gwc+Vd=M33hz4&l5Pps%TbMM=% zyOoqMg`oEKi6lCK$Tnkaog&9tCq8}0mNrY$1sT%F9Ok(0#FH`eu=83OjS#)*?PM(a zwL{fVGV*Jzn~wR~1%ev__OaNDeKOod8qwbSrcX}ENT(0%u5Sw(2bOvDATT$oZNF*0 zskHr$S9e?SQFW9lciX;11yg#v*Yu+T^&G*92FfhON^QCdU!`~+n+dr*`Lq}-e*5EE zU4cHf$xh!UNiS;rtMNC=hzLFn#DXW@ghF*nRNXJD=ay_RzQY8aZ*p}cGAu3on`#>b zJ%$8e`#zv%?Ka2}Gi}jifE?5M5=QF1g{kgD(q7I!Sp0r#$Gy~5c88L1ms{^m80Kx* z+m={$tIu5;e9cX}zLs}0Gc!E}2+6Z_E-Lc|cDv!kuP=d3=@%&9SSyiIUqsDEv6#Oi z7os8m4Mw6dJ}9x)mNepA<2@*!>_$5mFPoa*_q=&_gRD7;hz<=GrDlik`wgKDr#S82 zrF!rTtxPGLhs{o(ja!+K_q!{P;SOu;4wRu!jgVStMB(x z=c+Dels2qE-u^zkw^^r9w?4|q%MUMFX?;z`4USCB_G~?c7mmgn@1EkcJQ3ur=;^YXgH`` z1Z+C!+gh4wFk-8vAsg()9q>$SMnl{lo5rg#yeo22oB7gL?eBo#LP?B( zlLT}UA`Y4+KGurvBiPD0wqYDJAFTNFJ$bH<=r|@inreHZfeHBHHPG#^xVYGFF#w3A z1ECNL8Pt58!Mu|D=9UU6#M zW9^MUm-F2tzdHbMq|&2NqEhL@FPOx5A3*O%set+LkrJ{zlTvlKVTIl=dRW%99J#)2lzZXq(RWIL=222)>3e6Xa zW=oWNiv~R4)sJmnMg75$mH5w%s1&ifkm{Ug0%4mP_f(+#*!7%xr4w1**!>&#vc!HH zemT|5NAouqx0&Zov)x*a{SKp(YS2C66t4sFv;q+vdwF3#5QaQJx2)|m;M5nmcIyvd z0Q2JsGTZe-vVO8M4nr2-B1Zo?Yuh)`FTECzm$=3?j2*1pkl&&V5JaaPL&TbE8NXkX_92{|jRkf;{Y>4@`}^-tL8vb~>N5<5<>VyP zwDd7FDfG`|O>^&NY2T)64%-QWcs$_IbNl+f>p@V}mx51uL2#?F|W7`Pe z8kOpE(GnOOgG92+68w=3LJ!RP@j~{Dz}`lbknotRi`zIK_vbW5X@RCa7BvM`*`gCu zpyPn}kD=%&v$e}32-gjCkTs%D9fd-v4Wp>btyu~G+>dGyvfO@MZ&192xy{j@hK5FP zgwOr@Z#BP%UMB3wnwn!KT^V5O7Xm0BSP;NsbQn&Bk0BMo*Ru4$;J1yt4Hm!i#w*W7 zunb0NKAkbB$)uSrdDIW@cmTVS{I&}SRlKtX-8nzj+SocG-MfYkqe1gUrfptdB>DhW z{;tf##vSp4dx{d))rx5S6(<5VAo|9KV?QvTQFAAFlJ^ejPmR^5p9wRM;7Bdd?+l#< zzQYO;nW;hZyAL_`^ZASk=NIoJ<3haq?MAOc7+cUHk-yus2-EjP&xG7?$$}{)C0as} zG2!2S&Q<_$GHd5{Wv5Ha>FqVyH(QHgZ-M6n)|YTO@ch+R3rtXpl+XEA7(>Z^o85Uk zZu5b$#D8~xB8NCE9>`PcCZXY+p~82{3{f|Dz8;h%n(zQW zI|L%8nm)`Ns+pT0dK~ZLnImlL8-F4tswg^8DltfKq_PR))AKN*i?s%h&L}A<C zG@wv=H34#XWbGaPfUGZ)<7T zydHJ|yn+Z7K$q~XN8Ok;e+|cgc`lxk(|*AOgHlAEr*oj~Vfv<1AxKek@T&4hZn&Y` zMuh)O4OKdWp_ozm+Y-B#p~sY=w{G5kA}!4n8~c8JsN&(RpHCJsBn$}+Cn0Ym^k!J! zs~p;Px}A*S+i|ZFK>7y^z}VkoDpSpngVsILq!x<^PJs5rAIkyqHYp#r2bHEcc^1@% zdD+<^3d}j#x!t^M-)F)`U3jn4`U42)gO9~nQ%^W|yT#TOu&#h2JxY0q*wtS|_yJQo zR!2jJQBlEIqRddnJjb{#l(txlKV#-!qd>7}XO}N70pxN3y8rte)1qHG;w?tTK42i& zeuN%#(!Af2cG2;jHuI;g{fdmX@1i$GE%Q|`|4A<2V zx)DX^W#UCJ1{~ww{lC-t&oF-t`33|b5s6On+}EcDi_xHTCH1&TNU-yl!~6BX|MM?I z1bE1`?OdboE9+@UT0hLW2is;NU;+I`m%o7g-!JV2&})|+1eBd1K#pyQhiDZTwMbmdb?%o z2HhY!x&kqQ&hmPo!L8k%qP$?S%fI8D(E^o^9kCkh)PX>Ol| z3U*?k-$P&VOS3s@J8lV0Sa7l+_!Fo6KSW&q1SkdKKIpoRZbzipd^8qHx!cV9*@X4~ z@zr0`0@1yi$D?VZ#TIc9VBa4rKl}^B`U*`z)W1~HrxH9vh0s#Y$Nc?@o!}W_#5#AOIU3wMxwFa%Tz;DQFKIi-W{P`^P~3O6uo-$`kq^wn*qzLFo9z1nOrm z{PpyN1ZnCq|G!lZVt+L})OF+2C!!)r8+&B&M@DUw_g!xG{HK-M4}7z)HLQH?4veib zC(lWR%2~d~nj2wyC*yic;P<_^xj=6V3Y>%{-HzwYC1d>at@E?Da#Gmf@kK_moP4sCop}`r87^;x_Sa75rVqG+g0U34#9n@ z)J`K~hLnh-1#U2*;pe~7nz`YtQ`wwx{K-Rf(&ako>Yus129jP&P(k%m#BMJs;JwJa z1?Z{4m2!^=_`3y?qE`f_c6uLwhh}_qE#`^E(uI5bvm3n5`|*QB>;{U@46GM z=7t)w`mlbuF`;hS^;?bPLA;%jImt1?0pzJOjZK<2e`aB68MCrvLeJ(2<*wx5B)!-txEN{(EBOlErYO%JQ^_c2b-hr^A{$$a!^G1Mi{Nsw=Zb zb=8o4qr*doT_<;^E-_Uqm~hd{I0w25X1Xb9W$8An>US31$qbQ3iQTfYve{t{32K}@ z8s0Ja;yanHS&TjRzoGwxn)_c;q_G^EUCa}fasHUPh!mAkr+6r%+8g8XgZ>2?Y`_C< z0n>hVcCQsB>W`+g85XzuX`k*_nrVNx1t;_2C}TA?r^P2 zu=RRP);g%jG!Pjbi9CCZtZ=)1lS6&bXs@daa;Lv0f#f)n5-M%Ozn{%lnX7c}{;_17 zzf!to_@p#ox~bl3NFkM9Y;D*s+R{td9L0a8kch1Aor*O$iudhSkgsT}qamOB2~VzX z5xsU$hMn%1*$^YQU6sRWf5!TqDUwt;6{W>5^{%TMwd$P9>-6!&0rT@o0 zkyC7fK8cObjNv#cWxCnf&OM9{hexrm_Udm&R|Oe9~!s9gT^t zjRsx|e3fnmo!0w1$!7z?;7IR@%`1wZw5Z|~{r z7b;1zvUwKAn$f}6I(#0-9G~DO5k~s0lsU2wq?pmTo!7_iFiGet?^!K!q>ta@w(EQ} z9A{=nzKz~?H}qy$PG&`JT|vd(VsBO^yWXWvt!xqOc=_RNb83UMzQom%pdl;X{yImp z=R#i1@xDJ3y5o0<6uG$@@z1Z-SS!tiM^_M2ZJr%p7`py+i@u2a1cW;8U&8rPOYxc_ zpWgjoMM`9+G;oMZGL;>t#X|&0^FkW%tFURZFU;)_Mlzw*JH|bqUW$bvGTNk8_Ba@N zZszR-sk^%#wQ=MtwxBq<7l#t^?Tuo(yyfny_xMU2T|Zdssy!tlCZ?5Gb$7PXiYwr6 zZr*Kaf%d(V)!QN$v1>krnay2%v>N$Xb2Bu@PH=gNoPwfMf&ZR{9jzu$dE8~vSGs6Z z@}Chnf)a9{a@^!`gtbk*vUr7ZkZIR1{|aESy)4xSYr0U|;h@c8Z3{KNlYV^&N?kPy z{eOlaGKyb8PSM)abboRQ^=Qq71%yM!_I+31^$;sG^bO6SkU_bK_vuW^2&0ZHk1_1{ z^uddEN2jI3d(5ouF!nGaTjyhYT^Nq%MvrETqaQe82auZ79)c=%pMrdX*2o6SIQq_}t7nY;ig5(D#M zSG2Y4^x}Eje6i@ErLo|U`C?;nz(DBJ7>9QnCMRC;LgSa~2axaF z(F0iquDmt*Bu@le9V|6ET=X>+`ZbgD;LYc#vm8nm*1T2!J%X<&)O&{S=5n~q+hUOB6+f+K{u;a~eEG%1Sj;e}R}Q;$Z+}~giLPT#k#*?x z=4seWuj9O2pm5WB7?AHVs}#W4dgqzgBL2So(51oU{{ZxCN%xfys$p!v=) zAcXfcO11OYoLaLj2R~>LzZj|h82{-lNAqb=I9~dY#(14onXn~kUy4A2)5EsJbyBIo zY0(+8Fn7F%t#CNDb-`CqEp}etdo`uJ(j(`!rzlBZ>`GW#Ugqw2vz`^xxE-=0x%m`z zO>9K)^m>-0paMr&ZwRn!|Lb1p_sz6%pBeYMmkZe=Qt)iAM=-wsO$XrnOK(-KX$pG*?_tRu0n{Sau`rBv*NjysT3uR1v!*E)Ux4y zgxl{B>D7Bm2;%Q4Cl1i9v%l>N;e)m8OX>g~^Er6~t|H_gRgJO)%!?}m5BCWo0^B;V zaX2QY;6519l*K}@%>r5v_TC+&%s!~wb%7Zi$?tzk+p~PPjSF=-#2@#BpQ5}9)jxiS z=Ls!F)&@$xp|DcTUa%RhildiLW&b|a@T!0Jo&@%fKB;-AS6}h?kkv@Te9E2DFr!$_#8j)a_{%~nT``lo+D#aA61vfSBKx_fBBwX{zS=aq);g^2zSL? z3&Zl?1d#VCjiQfQz7GB%ipXB6ti?Y&FSZw-Ku~Q@g=;B4^5O=^Xgv_CZWVm?3xcaXA(fTYZs7}k|{=ez=-}k3W#3I5tA0@%dTYWCXJVZove3CNA&N-{!^bn^tlU>gjUhz|}B4u?)qKdI4wU zu<5BzNqYA+cs!on6j|c4;A8$ZHLDN)-a1jPH<_Chb9b=ByZGQc5 z3WxP(2H~!1a+}J0bB~zlLg8&_>%>iTA1~C@wNw7=8UiQ&5!Sw3C`_!Sym?#tU~MRS zoV%!-%lL2uDm&q_8E5O^njL0bxu3jJxg3~B`FR}SJ$U+~pppc3kTe$V$~Awnxxpmc z6!=0t6#NhL`0!q(YxFkZ7@%rMNE=t&MsxjmVW=koyC(Nr_m-4WyO|{(6yE3QT zaC*s?VjsXY%hi2KCVdx8HjPGl-Fa56>Sxs}0|oaaBod*m&*C1DZhB0QNYu)N zp;(Juy-+hX)+)0GNiO9GJDYxCYuCgjzqq5jxn?_lqrC8WREr&Cf6Wp81JB`HVV z{3nI+QXir#y@WggK_Mdor@S4hWUXEW(YMVzNZ3_}{1#9U?1bZ2byej3rqDfiY<0AsF6=X7!+HudhO2Y@JyKJ%#7s) zbys4lmyL;Bq8dW@U95|oYk-L3)N2#TdA=K_fQ|mCxrnADBg7uM6>+bm@8zwfN(`!P zf%X7wz}r%b&(x1b3A--mSI-5XPMgyM`*_m~lEj^BOJZ+_{Yh=5uhBCFgt z=(P@iln1X&28Qh~HA(>GQr5j=ncF)-V+eGiob6K6Rhqp^9Z4LrGES>DkXMgcM0Uen znTm^E7kCsoWU@nQD?HYYbQ`_pw7)`+Wp%JQPA(&j7mLSsnmTzq7cmn|nXUjwsppOq zPu%b|z$bS23q0F|fVUI|MH)Rpg!??0P%ybX+lB1_Yd!p;M2x7vRMNQ<8xa}+sSsD# z(~pyBX6N!`?E0$Gl`O#G#mubC8+wcR-3vW&IktM^McIuM{GPF!lhT|$m9mrdap|TT zQ?Mv(k_WIh5f<6V@)`X)WsDx{V9V-MRT|lnq@jY5b3Mz$tv!pGs6307My?onl*e?_cV%LiYV3oile}Bz;>IO?n$zYP6uz5_3Q}_I-VS~JIhP}| zZjW>zGpDF`_i;;K>_X9SPX)!M-NmB%3++O6*c>l(;B6K|RaGde!M^U%gO%}$h;Gh? z%_&%jEwY(*Z*{2P2{7>7^@k4!{9rxRQ;%j^D0qOCIBO8Bzv4whkG#V5hgHkB!0#Ch zUm6tRjRj)B_kJjHML~dol|W~+#$vjCu}Z3$M}05Bu88QY9d@P{LkY{`XWaxGve4NL zdfO%ITYT=tZG|ZMmgalI#$ioCjlLDO=k%wj7?-LSXJb#tpK z$RD2%n1qoGFyufl2694^7aaUbP;OZ@4sHCZ6`r^GX>?D^6+B4SPD*^s;ODPX?+HJ( z`V_zWY;x(99{;r1&47OI)h~Rf9v8|?)pnzeDcEG}BGzy1NHizQv~xh*`8q0Jtv)Wt zAa&;vb{JK5-<4}C!#=zygzDD3e#a3Ug!|%K_YnDlr_1`d)PY3WKMgK4P%|yi7;ssC z$O*9kD|f8*v92ZS?put7_=xlqy6$4dsMVJQ9&C6~>NR+`6}~I^#>Vz#;-ifmP+j;~ z*wc(+6SKB9stFrG(fd~VvhjWb(u?V{ZO6F?ba@uaFJqOT`Xj$HC{M`ack$ufJ}e$) zYoICA%y=LFil>#`y%TntS~+CeG~3tc=kKOR6VweDc#FIb{-|2Z(|iq}$TR2hWap6V zY-kB#{%I7uEm&AwjX5Eq0|<4mIoNLae%S@vYu%b;mbJ=RLNu zWBSXEm-`pB(#qBbAFNB{eBtU5%#W;}LQ5!L^>7(~2_E22<0WP9)XOC5kTdVgf9A3| znUT_P)N{XKX?ExFBhuj3N<^P8FB;Xh49W37fGK<>F!SYy4v4Mf`PW{j*t7wAL;MV4IU)D&PY)A$LQ1(ZeDd?a7dovHkxKbG4fXV!MK4KL z`Ct=oggDpQ3%HbL%z2^KLwy;=uG3l+FLY$@k{En^i7{Tem0tB@sn^~+tcP=QqCQcJ zl=6#CmNuO+ufW~o8GRXF*xI4c%Xy!kF`>xuckDX3*^9ZsUZ<5|0>bIQnd^T86a?>8 zARQziiGj`q8sK0FL$wx;Jnykn4+ld6?xR;TqUMV1EY>~fT?EybHL?|i-IL|hgk@6d zkFw4&@sx_Hg_aW%kr3*}ROG`@xVv0(3{l^iV84H-*LE)b=H8v93+^<@=EKrz^z=J6 zM4O~ZRm)FJkawA;HW-a0;wBmjEAfDq7^_aLYvPa&QM_~1Ro~E8ktRxqAvSHOZmr_e z3@Y$($>>vyrPi0GGM3K!Z(oW&9sys(nCHuvyl_IbTb5v-9l+L_H~@&!4R0vUbdnt` z(`xz&3)q>cf+EZz1C8dpSTTj(;L0Gzy|9F-0#}(`M5jw9Fi%6@dwhDy* zBUoI4*7gm+8>G&L>OVoyf#_%DA*i{_eGc}IQlp*LM=TGXG8(k& zI_(X!)w{EarSs;On(i3+fBW`LTU)~Wc!%s}+z3AMo*0lWCpf&C*gir@>~FYmP$|=vUEddE{= z=oLcPmCR66Jsu-!??MtC9n(9x`=dAcn?ABy@yWc{Bo1&is9J;lw|(`YL`SkJ!8gkt zo|S(40VOZ9%%wDIQ61m&$yeC7@*k6n1opv`ax>LOfGU>QOiiZWjlT%LO@aQ2+4pGC z@2v>c^Q*2&-Dy>nsZXFkTEoqEkKOp?QiQ*MN%8^;iu;>7Wl67>*Au*344~hOOr~=% za1sV6VP6x{cuI8oD;02Q7lJSI47~G}^zUwPPNH9Ww65?_LTW@KZ%`PjrFm^Q&a53F zy_t}lnjhVTdV!OP9r?2F>`Y*7HJ>wod|~-A1F=$?GV!^t49Ix~gU{|hTm;+5_w)piN7)g+7HA& zh&0oHWi;>yc`8;`-B+!Sy|@?QeJ*6|n{@uQlJ24r>K)tcXSar8(5XIhsyB~hud+V5<1SS@%A}kf zb@!|Cq}^C16E2C>`fv`dgrwvm-Teet6{DT#oBLzOpOt@@Ye3UHg| z*bs2(@vID@X-TGQWQs~O9IMu>82#epe?>5$BG?8W8*1M?9@||)hUH6*ykhQI?=~-z zO?#-*Q>30lBY_1ylAH}G54ST>Os{#T5quiYrYBnTIe7EK1chC|$2G20>Y*a_>}PY~ zXgv>H^gFNbjUVU~=q!1#6;b^noBS%?%wPana=*8oq@EwkA>xHqXV<$e)7&^%E=*%ImVN zs_e@!KTwD!p&ardq$PhaU7V5T-^fP0;)Q?l2lpv62ZHI@kLU)D&?Ty){CVoKjbuk{KnhdS9CIOK+U zkr(omBrBz*>J@wLCork`rFy)4u#@X&S_}8UT%F9|uIRbjgDCZuY~BxWqRi&NQGm`% zr{zTkhw-kcDw1DgXg6xLe%t5!3Fzrn6&kyml1VaKPc`dENJFbvby&oILgVj90#O29 zzdT$S&gcio4nAz4T|>QBeb*?3stKqm|BlI1k}6@_A9>_kI;aBK0P2TXJsNX46RCoA zj!GG#82J{~MtMt}t?oUG5kG?2_Us^+rx-YkS$D@OCpNNF&V~mwHl|1@VRH5DMr2Nb z*o!rFE8Cs<<{O8fFt6)S-1%bb(i!Q|W-H*8QF5wQq-tM4>6I38RhRqosg(qZF;BFpD*jOz>b9K>=1yqEqkmxi|4;yYi-`b~SM#&L2iZ~) z#BW}^N#Q<;My&VWYyRI0kr1HsNlC+dr)S^pinwJ|pGN=YDu03@S^p(L>fZ04osJ(Y z_huU_1jw-L)soKt^A_cD!0>H&7}3$o2n}I{q7!6z$8Jv+eg6pwalkXYNSvsL z|CDwD7=eI@+I_yl`jRFpDT$fBy;Zmf^*^b}zjC>{o_Pl>#`qMjcnvHl50E#451KzI znV|j8`v8Xa-v@FBH2b8Ih!qS5z4O^`;aKD-doYJTBk;Q5-wFI5$@m;W0J@Dw@k1iW z=>2rouJTAM6J26|#f3!1clRqFpfPs9wsmOfU#s%z-HL{DS9kK*>coSo;h1IF5rr!BMmx>!uoXL(Oi#ziwx`+ z0W85XY{=475Fs*jNW0d4?$hY??$Mu@^+_W%q=p6to)4-*lW*r<`&qmu=mrm(ReJL0 zL=l{{F^NLX`$3{74InE_-{Pwv$2n*j9X$Eer=+_)Y)dlzuXw21A=%m4*bbDf-XT-3 zSjn!$28u?r6W^m{@e90wxPHUyN)D0s_o2L^17Px^tDvFpjdbj+K5nM{D_DdY2z93V zf#9eYR`&8E+29}7w?#Y?Vfl)Ij7i*Ea}OA$B-mGk!DpjI1n&P1rbsY@S$jcD64pxd z&z)qznu7|@M%EY+MwOzZRCccqE}mm8%>+gn@*QOg29haf#!r(lpSjqQLnukSa7kI$FB z8q)t~P#5w(Q@d1uboc=wqLS_1CjBr7SWyWiw|n8h=2edY{XaTkp2yQk|JTu>QRass zQ?|W+#Y_a&;N!8*K42RxKw0dA(D25bgNRmre1fr__IMF_A7gF035MU9`cFZ?IOwF% z_Kbi(W=;?eoB~fqVas2}acd-)8)yU6MxwDEPkka1@)gO&3V|y~7_ND_eRhIc{L{b4 z5T&~LiZJ74QHA`~b0JzFCfCux^`=j)Pk@z>1jl+Af%Q}~)F7kvR|cL+8u+;Wwtu;R zR2Be7r^Ov=VM|Z?E2}qrN_Eq&@b=3m^njz`12}sRmwZVe#{oW{Um7H46cj1s`aSvk z`+uw%6(CYrK4-2W^m90dVSr)lJdeT$>EEYfI|{_~e)XMJ_N3`CXVCwXB)F{sDrlv7 zaF^c~K27lj2aNK&0)|@LcjEcPGxyxEi75-V_JCttAvPcsS4%0`-~`)o0c2wfvzP*) zF~K2Su$TpCfAvzFnWAks=XqG^0X1h%R?7bA^ zn9Gh3RU_CY@x6gOo9oky;e+P|iNdL7YQV>2BP6pX%C@d28>|l>{>9N>H~?g05Y)6u z1dxm=0(gW!C=l$1c@y2TBv*ELkUL*Jq{GL@3&g}!pgdV)fRpo|s{Nk}_p7+^5fC?0 z5<>iknj^bRG!2Litzw^3Z?-XajJSa2e<{l-t^M(-rM8y?*+!H5`#GX{CW+d&Jh(B8e?auU$9+vZint{r(je#=miSnGd= zV3-C>)0w9tHb|cupz4Yu$yS?DNEQ=g0)lNi^pLaW?8u5=+VY-m`*W&R70X3_X5E+(m$$GxkP{0dlegd>e zHAfYsd&gS>n)T3uM(I!M!vJ}vm$GDNH**eFewk|5z;k?4_CPN(v6OT62N?x<@FqW2 zPv#wR!n=I#;dm7Mq7@Z^1=?}lv9waJ!nD6K*dd$cCiA$cnfXIAY5dpv)&c^=Uko%h z>3XM4AAMVX%kkJBI>QG8b%2eer;18MUunb|B$LOtuVm&I-g$3j2s?{tFSpBd~p&Sc{fYR zWIE#}%ZN5x)5{#}{VytF1xsFfCEa&T1v%IGv_DwqWWD&{*_Dp|=&<*#{19&JGo}J2 z5|@<^?>FOEc6>Q)z7T^2pUot@VX4Aj#vY5v?=f090W8-UB6L*P4*xxE1i|ar zUQoFZ)PPK6UeFW3=uAKb1*jxogRWmte9Vh3ygi)QUs?S5MH<`T(sJ#B-7&52IZtQMy4_3e2I$aNV%Od8YgG#(L~`c> zzQgjYGoE6DKXnctaI0HJh-C`!ds-^%c3nm!3}V%V%f=rLe1pJQe}A`O|!5)1#C1YlH_E*x97l_(B%zc z61rq{6Y1tbf=e-~Ad|qRla%zxRs`U2<$Vf*GwzXbXsuKs(%!$%Y`TeEw&2I1}FPbh)C2=5! zk&{ytuSfl0-A;&^Gt}7lgQbemTLJ2C&mGdZr0}rudFyKUU>AOmzZ?cQxv12uN;7L$ zg$zql&WdTrs6>Ahu3bNkNS<>PQ}3aO(M$GkPT|@a98^u>v#KuE&NoDNuZ2;znWgYo z4egS#Yt!TRI(N$T0jo``UR6k7TRA95=o;m$N=k|=AFe;-$;*EHslEN(J+v>um`Leo zncL;|-HBYm)_q6rBb!Z52dmtLdPV#TeQAAWaMXEg@NI*7r@&A$j%TJlv6=b#jL8QF z7=kVr!h*1i%)7~OmuK;erDcAPA;&k-X&o-Vl>+3%ReB>%Kyksolt?hu3??{x`*tVh zdgx~$&IJ3tMI*T|Y%J0wzA~x+67(h+3euH>CtHP5T6>Olp+GhKgwM4nG!0{-V#`Yu zda<8;SHtUb92Y~=K_|qxhR2z_I(~T7rKT~=JN2Q2T1FzYFPAWDI(h}V^o79K^M2Fe zsYjLD+N%;gsnNjRWE1_27ema+3ocD?=*E->g4J$jW>)+3qY}qAfIMVrsX?=Tx^-es zLDu=uQngg~b}~6eeky-uv%WNzRZck#YPBscQjq$o$zF8?{t5=RLtxeS{1^&S_G zq$)a$!|Sx;eiEwph*?HIaSNvpH}A={v!sPJ;>p7{2ZMpCEwj~1)A#tI-a99=gL1VR zqQI&R8Z?KWtCxU?$S@Lca2`&;+B1C`f?WAw=glo#j$Z6GA0H1a!?o@v!EfY0;R+~u zl*b8Unnu`JIRV;Q^FxF5;)TD+YgF~Ay@q#m`M}~e3yl&jyCFKtXJ@TFW#0e_b8)w7 z(8bJXL~_lyi7n6K&@ZKvm;To%asu~YS)-#jtS(55ZrXAzpVese<%b8J1)m=AFmv`y zRJy=Pj`)D4cEi)NritQs=edC#8Ev!KR(jtww^U)T_)+W>w!RImI;)73@aI8l&t87$ zomOcl$79xfIXt1;$aiSMeho`$<2Ole1L!r^v_dO7W?bBk+59Y?xae<^-j;YZb%VNFB zycrXs_If~_wfEgg(JM$v79XPnTD_gqo0^KeBprcW$aiN~Nj(B0tZQ7a?3 znrem~OXj%bkJt8-iN&Zh9m{nYecoT@22rVxBEIC@viT~0R>8G1WejV#i@{1-Mu6di>*18%{dI6Cxp<%WdCSETC7jTl z^Z6vLSDoSv-yn_NcT9If*TP2$NTY{N;ts1x4PHMBw`Akh4Uz zh^q`wm|Utu=ayn4ZI zByoPaoUqbcL0Q0>tWn+Qy;P#4zEJ15kXg`bekQcB8DpAU^X_b^=EY1#_}Ylw)64Dn zsY?6aHM{Zb4AZw$`yI76$b?uoE8%BS*AMbI=S0Tl8RWl(##vdmz+|3N+BKeMY;RGD zU87tzsZoHjdOEl7$-h1mJd!5vOniOzLa6U`CBdj1vCc}GkVon}y;)HV(}(ccjMl+2 zC@=Am+C|%3Z}d>6P#O-RG)uiyU%_VLowu3lJ>4fj9`7RLpv6;Qz!%!=E?lG1OT8%& zh!qFKn5ZZK`k;+W@06Db4Gq$Px|ltF0W<

Grwv=!9*TatX45e$TwCGb>WAQ! ztWhClx$CBr?5VFBLH?6oZwRd&5A$n31Cg9q<)reXZN4Il8ylD5kAG$}uK6BU=Q1e* z;^mI)1R`Xj!cvt=<3y>fRhXmpI;%3j3LLDO6?UilgxB%BdoZATFq08@m^`r2l}wL2 zV^|7bpH9IIJMg9J3ruJzPK^+OK)7-P#yDKP2YNP0KErAV_5I2G1$4|CsT(R@UB?XT z0!Hw)Iad`oUdB1sT)-Acdqt|$K!(P4go0?k&s|uG!%30;TkP-|_qNwJe#02TEI^IY zV9#0g;|9UCY&dHzo}s90b^d?$>%A&=XC5>gLGB*Dw-DxywOyg6rAH*BK(^7+r z9>{cC92`@n2Inxg&feP}Fu(cndZGD`^^qXt_3QC<<5~IrG~#f+N%$BVePcH)t+rW{ zPMoF6EqC*B*|Ki{oH=XR!egY37yIwc;ryy2a!r5T%;$oJw9 zG<)d{?9rE35AqC6Ll4)Tom1dV5Ju*#p57Y3kcPNHslt3asFJ76+K{D+Lg1Nj_9|a) z6&RG)c4{zzCy*2@wYJt;DdYs>rSnk|GPa3(pfDVXTtD1=5a%S6B8l&K_sLY+&E)YYY1ndHKs-;J)mN%vzkIa9g#N62d)Bg9mxZ0&5V zxV?()<`>9yoKl!qO=(Psp+%)v%gbXn+X79CEg~iwhj8$llrt~hOnj<__HL2FC-et% zih2JUrcsJ6o0a{*IgXz*;Y(7c7i&)`mLrw`8|+7IA0L~jZF`X&gs3?ZLM=j&W2%{5 zo-A`VohDTxqrYpk00^u>iee`z{xjq@bmqsr^^3=KMVUuPwzH(C|_@b zu+kZ0G(`#UUTV#7I@13D&nnI%6q#y%FS|21WxcU7aWEjJU1)l=vagVt=njaDpr<^c zUk;4l(do|+>rO+qDaog@ne0#Y30EhW`?j+C&UFjF%O= zOv1HKzy@irO_4%Y*$}&Sq^b04n#0F~1n(wqUx$%`;y9~s>eLEGjx*K8ta5vemR4IN8Z1eZB3;w5arROP0-!p%4-|7GH0zE z8<)Wv1>CS?Ie&ehhi%0iJ>@I9xNrV;>|W}M8VaLMFv38Y*81tW*<6P0$f@V0n+6~v zo5Rt0nxdzpZ!UFjyOqsoWDePPBPQ3`rSl{e$h=9Mm>AC$XfZHd(i3)bYham_P#Zfx z+5g~fzjlUQIA*T>^y$<0k=v#}5)DLE^NC>*)sT@pBU9l&`PLZ4K2s&h}y z>f43?GUaNw*tCzN>%I{U`S*kXe)KBuez3jC-fg}v)YwjqXKkR-c; zKe!b?`UEsJdTk+R0+<*QXv6Fgu=q5kFT3i9t6BaMU1Psb2ucDagEem* z7M+Zgi7@%FGca?c$ifZSJ|JQINEHfbOs(_8TtDrJ;kqA|NEn1!CFOwI^X&c7yr=1X z95zH>&)_G_r}>O6q&#L{)tqe+e4gf-X(|Vqeqh^0(b=9j&41nX4 z+0G&<b!ta&d;nSadpFW~N82D=3h8u_(oq*@I*x3V4}uMW1`f}2wK z*1e@!X+(DRZw&#KJVT#i4UlR@$QM~$7D!@`PxbmY{ska4e9^7o$}44Y8D#;oXf#6F z^OzY$RUH){)_;A_zuFc)Q|KKXC4h+Rq%sKk5DfwEmaZ|vj*)e(1aWsh;Q<`P=T~0|HQof>+%hLJgwe*R6YR#du#a;i7%$>Ci*|> za)U_km2^kF`fNg6YH~MNG^c!#E9)MT)cBW`pyr-DF`;EoJhnl~bTGxN70i?wKJi)z~?Tn*PiS4rV06K zRiQVqLA|PXnlR;s!kukL)NL5AG_L!63wGqvrG&)#VQ1wn#SB!aAGjQsJf!i^jXe9F zP3y=t(Hi@;AqQxi`Mm0eE#B!dDI?gNW>C4dAuXZyz2a9J77Mk#RJz$7|Ngk`4=!5y2bQo>4$QrWEK-OuzQ z9N*og9d3wmAXZ{^#{xyQ4#*%wsLA6kBl=rJ?1O!G-NX{pg}4 z_Hd3ykc@b$#dE8DKDfM>3H`DXXtgHF@hY~Ddtx+@C$2D> zdM&(to4UNDjcUrglejrf|MKZ_btR1X@JtxQD?6tvMU>R=x^`(tpr@T-*Q@grt#Ya~EP-)3{C> z^EM)EMj!W^F7tWx=1nR%#!m)xT6*Qw=9nFjR!m~N&J`*Ne-~OwTpR1=xMs;GM9z4! zup#2sY{MJs9Yx)Frz8E*@B*0fCdPZh-O6WEu0QRrMhMe;`ugQ8hla+2*I*+5K_##C z?SXGA89#)gpxqo#<-Qj-Skm7BAop{_&G+>Y!FD?XC?5}&N0Ko629`Z+} zJb$fa^~t-FXhHWt2mL>#uwUNoV47ODn`wB+9k>z@k2eJ9RazV2;!UY$vMN!Q0hR3y zRyBfgEn?%37!;o(EIw_pjziKjXOo~I^K8uk++mW2lyJADT|_dj^qVVtQAt<@E^E81 zQ0ijXT<=4iz9MNwm4Gw1&hjHV7F~LlU)|OkXCJZA(G<$BE7+y>luC~3hN;!nyZ6={CbRe0B%c6m_ZuK#_bmK7qd+WmIpE2`g= zPoXQ^=Bs`EBcr*UaMj*pQIpol?wK{u1~3%BUR6kyyJVvI(V~wMCi*RrzPGlM4$f+T zPpS|{1Ks8}HN)h`tONVEX-3=~M3mLZ%`FrpsFqIIIs&acdpJI<Sef~b z6}GJmNP$W29Ww_eCfJ3!<1$7-ADJ$m5T9Tk&y!u~sejUTsz~!`Wl|;E-2TaV$m%ICjU#$gUQbL7x;{O6{_2Nr?PvqrZkz6DnF0{f~{|>^RuKx>Z$x%?ka+2P|9-U`VT z>LAr|;GC1?>q!fP-eC`Y*WiW8)i82d=`3v8vLgJ6j={s3G8*e-<+$ap;L9>PWL5RmeY7Pbq^lbnShRWq^r09&;#2 zl5%vWey`$%xq;nV31}DoE5!L7=g6|j>h*lXVr+ntvsU#iX*lux zIk_mNt!Ct@v*CJ|PFRi_@ydIKmr$rHbsY58NZ-jGLvG0Y@8k9JjHN&4d{&rfRE3X) ztN&BM{AJlbT)I*=mEd&&dK3J>4u>f}YMGcQd+C^}U#C<_{>=brKl{_4`1>J&fY$+6 zTnm7TZ7159Eb(Bpvu7POMz`tw*gm5F3nKb;3-zhIk2#pB%tKXs|71+S?_bPfRg491 z%|R=l)^t6ptnXPf=r8ypTOWO^-Fa*^M>|gHyosf18v{7#@d-2u zptL*8vF&aazXAAMUiz1Rm*uz>uvD+frL#T1?&R;6be{e8d+LNkHroAu3=KVSxCiq1 zGWEE_mc`;PAg+3M?6^3Mi+M zcQos{j@twSNo@PjzANS?vTbRJiDw;bLyrT9&i{Pk_n#Lp0$0g?zO(AQv04aYb(IqP zB!otJXn$<$|Cb1R;Kbv1FK#KEJ`{W}*rPOhtY`|ZQF;9T`|H(&V*r}LvAp(2`)?oV z8>7L!a+7pbP4faWJhJ)ze=h0GPokcMuE#B5CcvNuVA(qyHyJdiYpnl0hW{>CbVwSS zh3&)T?@B6`1LrA`IK zd47R8!qL1eT!h?sW>!wl!lFmd&Ga@-AZHUwAau%7&uEgB9spesPZARVb@3&4rm3(_ zTv2orH%5wx8HpU1Kf;x0=meBgOB{)2rcYi0CbKBk)l%_oPgkVz5={jh6n|~_IvErA zeW>V!Vu(MLBtRf~8)d1(uaC9#V-U1)_`uV&3j=xT=HlQwi?Ysp$bIZ?hu6cVCx=#p z)sXqvp;R?w>EquS;J;`36tD%}zIb+=jc_h~)lWM&OH2FUO@=ayn_LC@l5RXhf$a>_=?z6^o6Y_GB4}18Y2SoPPIqzzZeSZ?0W%ulE&90qZ$V$yrwR_z9 zXthxVWI|-ZYduCV+~S%jCbwM zp3k!$AY7K0{;ScBMC}!A!w>z9n*6a*02~iMcgIif4G!49eqCl)Cr-Pwb${|ZwN2g| zv}hb*><7g8B~45Q1Fdh@)X6g}W=gNOra|MR#q{(~^#HD~bMAa6P>Zdp7BsTk!bH`1 zJaV>jrYRB)#*l+S(>IPGY10#U|7X&Ww3J2>9`P#Js`t~yag;E!Zgg`Wc}n#rqj9s^ zjrYH1^QGDgo;Lyl;)e~|WQ6avv9A|AmC26_puMJr%#l7^*jOlqb##rnXIdH9HGut_ z1jfemP4fA!ZPDv{u&BngId9SLom_hw7`-=?W@dyPuI{NwPaQU2Mh<-gR0$ZY!<(j6I$OR7wYK~ z?RACbI|gJDzH0gEm@Q#hXoC5iKcLq|Tas(W=~DXf(&;-uFA5;LYqqV$Ui0q(k=Pqe zz|0xTe0or!o?qNUQk8M{+R}u|CLO)qH3m%@0+p`&b|=YQ?;V`xk4>rN{A(W>;6A(V z-5Li&Y1R(WppkWEE&8h6|0p0n$_DS8-3KIF%Nmcl_pcN@<5D%`x!A)BL(;$Kxo7rP zxeYtw?}G#91I10GzJGU4GN#_uJ2vU1VONQHRcQvWJWLalUfYX)a(tNoEWgO^Ix*m`nf)K25Dqv8u9vJLzUb+QU2^yJXkgQ6IR5z_3;UBNR4U`sCQ^ zXcSHVb!aMTnx+Quk80N|A)Nh_bY_?NO3J)fs=!w01hE37yW4$vh?=CLy7+_PcDJ|4 z%Df<@+BO*74g!j{E1XAJX~kuP>5~#w%+((dy}r6TL%e(Zia;PH{z5KvYC6lhgPPlr z@P9Vt9@a~0FVoO_*$%#aDaEG!5*ZzEYVuC-fv#jZ6z2$4`loHeZw%bD(*j1!tZ#bB zxCiZg-B(-DDXrSt6F|D;HroHV(!9;Pky-aC#NQ+4T|=>XJTE(?Q4K7(LS!Z?8JLyZ zFGYOi#MaTL#ilp4TFG1oa86BYau{d*K$y{UyY6w%3=l9v`VizD_0gdBeNysEXbWQc zbsIu!-nNM_~dt+IiiU99n%XDMW?a(^q7+$^&R)CL$~hUaO!9xYxri?4Cdc9;-{n zUwO8CWx4KEN~L7EB=te|txqG&RTHkXAI^2~p^#N%1MD593$!X~Hr@$bAVyK_4~m$0 ztxt;c%e$6@0Xv|ZWY|a?WPeON{{tAvn*GkV|9+*H9)2+lGhaa|c`}&OO}Zi{sdp){ zoVCpPy7WDd*~bFvD2cBoxyD1ijj7RLAkqRrVOaW@s2M^zTk1l7#NzKVSZpB}WV@;K zd*_?U_BKta^&JU)+R_*BRuwtJle&*->LawT#7VR z)@7l5wFhRzyi$s1e)m)yv~r(u>VHYXK%kE+r4f{(A()BgT4KFhncKLQW;;-|DC{c5M9jDAnGS60zM@1BYRPRn-QjD-dqup zk0M#^UyOAj^A2c28W-TMEMWPYLHcE1@?~T zpqpT^rppHzKL%mXIB`x)y`;)Eqh5p~_~TA*Z!aUY%FAa)Wu$;MscOFSXcpCh_1V_^ zKM^?3#WB{d8R@Y!kO6o>KJElzFeY*tw%(-Z3Zgclb0H7D1UWnp(vYa(f>8SDKU)WIRF5v`#6=YJgH=6l8 zLq++!yMNx9pO#eTC*TBN6$Wdy%A(eQqN!=Us;3^PpJCB~{DK zTrsn9zlisNbQRRI%0%O@{-_=i&zQrJ$^~%Kw#sbU(>m-G2;=NT_Gi=doj$9n^`H=n zXYC-sUb_$lQW;SydOWax=J5dkaG{YAP#R_ev4+jW(CVz1)yD=C%Zu5%l=O`ZwAZ7m z9dJ^jJ$j@YyPPAoMBl`yV#u?^wGnN{s|{@9niu1|l#(+|p&N*G|51=%w0|6Mg8-G-ygHmL$7>I^D{} z;pFFUV&<)nU*}BstL;9Ijm&vMuW|dP;5zv=215Fi* z;G;^B<(V=Jxi~+c-Ob!Q!_qr)u}8ng>0ZwZrpeIgf#r-`aL#VNGunrhCaKFOOgDwQ zB6w#|9vTS0GpDYl3chFhWRMFMJ3qx9xnTK$W4C$>uEcx^`Tj!bPjt|hdw>gAO z=vIT*A;uSJ6W22VZ^_uV3|q;)^pV&8S2M%;q0D)hufECx(s(DR1ZLUFHgElDquIw= zvV`B^OyE_XUY)=0E*t(%RsY1FFZZwn1?}Ux_>7g|as2iANRu5T4xYIKnpylHLTy;rqDyGQ)3ca`fY4O&QbR^#lXza4oT9OuYzS4-+c~3a3A6Nh$-%2CHGdOZDXSVZe@tSSBy19z|{`nbm@$w;>dy4 z?FuT6+fvp9Iwx7P;+*x`jto=I3#55dp2S}g~xXXS+E^!hDE zYZ!qj=?fcCD2 zAH|JY1VWuL!-b0i($^jQ&YnP^9fwJgL+6MXn?)MRf+0n}|D(tE>*O(1q+L#+zC%Ex z&@jk$#^#m}_iygi@1y+-k9}hxEYZjVhWh>Hvqy^scVGC<+_VFaxx+nI&z)#I`B;Vi z(R#7a8}!-Bw*OHIoq+`$1nZBQ`1LWnWZ@)vNDNibZzPEh;~el(`EFg|Kfvxfb+(dm z$MO+*H*T6>G8V`+03W)EvQWL}G5Kf_reMn1={TOcksg2+<&O09Gn}BI_qVlmz45~y zDph?Y;cy{ZD9^sn3Vb^h1S9DgM5%gZ6}y?o6*z@0|3JMA@Ue8oPZ7!c$mLS^=W8~{ z{x=-H45WtVFFDSk^S=JD!>!GbywJN}|SON=*thPRbij zO5#+USM5JUy!>FSGj2q4YX=RKgMZrX%O?(x@uM+7(_3GbpH?K&V3G~T(jC%;RNKl|c zD)~!NuDVBhPZnsYLqGv`K;$v|ko1W6o>%sZ%oS-(i?mwjSWj$Z-5vx7*77@OK&3Uw z)^2Uvbk1teZN4zY|6{R$Fu+X$uw|3)`h|Yt9m2h=e1{+dRY5~r{%1U1(ykE=0FpUG zVfu1nzRY$xP0l-2nK@@J!rAP|69f!ly5H)?Rr4JG7c)@w*oC0^Z?}0HhS{HpF?aG! za+N4YpZ3)cE&>5dDKW7lSF@6?Cq)lB60)<;!Vfya zC{RUDbhir`Y$6}WQ^~L1h3(Th}Q^ggU5<-EL z>@w%+p~zXv%C2ib;UMP3ZLN`VPYxvXB#;5!0HZE{< z_7*_SN^JH49rV4LE`9p&>ew(yx2>+52&Eh#bnxRRDEMxjmuNl5H1}TRcJuu=7XCW5 zkJ32P(h&CyQ#?`YQAG+>at;IrrY(kIq6ext59CJWifE^!=JB1~4W(D`h z@nzzZASkK6IKfPu-|g{4h|$njtB2d!^^MI;zc}aNjD#|$i4LX6vMsb`S{y*XOyoUP zZZl2O$jEcI=;z%>DalFKE-5N3-HFqqji`q9%%D#1^uHqXucZX1Vx;@6Sa+!3qyFQs ze=d%iJ^-7!AI%W>wKY~*0WAHxCUU72{p{qhr$)$ZqWe#0yokZ)j3zS@E9~0jLY%*k z5eF?cV6|7RWRVD@eYBGCb2M>S!(RfZIoJ`L>I1?}g?HNRz%ij*YwZqV0ycH8`|3R< zk561LOiD_on1-`q=xde7hWgoeoJPK0Kv}#9{u-wm%dsn(Ej-}KBUQc9^(K5}t!d4Z zrg!1n*6ib0n&-00mp}b_arn>y6)(l*2(U@?S6gu8uGjka;`u_qi2@-0D z8MK3Ls=w~Ekqzi(by+BY20DGsxh&--o*PG2$cRhff`xB7rdx$gJMsSCi7w!ijUTXgQB znq3N7nZR&H4QW&mnk&88+wG?7Ac_Als~T;v>hCCpSt9M{VKLpd^3cWC=Uft>4ur05jDQPsFOn| zD+;UH)^6-)vCj7@Z0{25zwMsB2GOp$oE9sRWMY>HiCV-T=!(n9dKz&&Z}8f~L{Gig zM7`aoo{ZZH{QqFl7ps92se`lYjr|#c`9Dw_tR9_#Cwwa`MV7A=JWuhR&YWOm$PBn& zQdFnA$z(F{(38(}keLnvZkG}v(~`7&YHlRSNPJ&`&Q{P6w#?m!SuVAhD7ddv*pwl>nsJtrS|coc-i<( zY!lNwl)O4q9R`To1OKaZ{25uyyDy3Vj4W=6rRTW=^>kT!N1mAABCO*Dvd;HNI>=W% zf`6Eou64T~!d1oY;?8HVmlgu6yO`^$QHS{A>s#DjkRKhFap!e;)kl?@;RK*50{TSC zBCMHtw!XYrV-L`2W21Rq$!4|S71T6On&&#K{iCQ-QC5G#xc@?M^+neG#}Be} zSEsF%DQ_yrTg4{RHt+WD`{#+O)dMJ&_tF1}Mnuy&*OqdITw^W?L|!^yGklF6JLRP$ zSyjI81>+0KwtwnI@l%fpgp#EfuFGr%p__0G5D%HXm$CoTxaD0Aa3nHztIuPLkh?g| zu{!JAZH2kV_(Y|y%^D(q?Yxior3a%r8QETs%67n~1Lm<_Ujc2Gp<)GuAOC&(KP5Wh zFRY!*KhnIBy~8%nkD<+zhhBiNdf(BX7IUn^E1Hem+_63_5_y3=@Qs5BfUF3=2|45T z$pVVCnZBy@cYXli;78ghNp>%M&!&n#VFl|`tnNevwMCM9L7WR9-qiCB4i49>m%wgQ z5t(j*Dek?B8d5jB)qZSM;73J((KN*_6xQJXw9}|hJvv&eC{v^Z2pbfw(I9g$x!1QK z**+cg{gu|^s*Ru+siXLE(G>GX7^{F?F5m6;Wh^$rcj<`(+7-=;7_Mv$>pJ3qLo}Z@pJw*Qs$sPy8jX3P#er>c`i%pXqP- zVfrKUFaD-8{fZz-Z`AIbiN(rQ!B}V;>4t`D!{YY>{Lh)+Q4*bm?KK*7EK?ft<$&B`%%{u36?@$B!d#Ix@FtHK*3 z+G`619^jeA;X8w-*+Zw?a-FAZn-5>{*f`;a-y5|Uc%|o0uZ(vh|R%Cl_Z%CxMpxb8CT+-qLu zlP5ntcV0&WhWQCbUe?j2V;~V^6T1n}H~cKyPHEn~KkFkLy`M#12^>PC%)diUgMfDo zMRzoU?RhX1Ivpf`=O#qndD7mtUphw438mGhn0U5BkWzENrTY5wN zuN^C0fMFhbs6lN#uFGeEiCx+W@{liCIoaCtcet`#Q^t=#pg_P|;!k#k9Kx144BApN zsf(R)WZccDEpye~b?7yKakYL$>bM{96kjt$$G7CavIOdN zmjYF?t+-7|!elrsZ1Uq$X|jS-WZ^`$yIqABboqLr+1;p+!`~muM^@->Q1VFgGtmJApa?WWvT{d^FyuX#M~80*Y2+focl^it=#K!>FWg7`6rv>P4Ko zJ93!;jJD};@t%~n1_yRRfl&=W>gGOlMFEgB+n^&=umxmwZk!apcJmbq{|VsnVIXH5 zqCmSdo=W}8IY+$Pq>=htlofqg>eNEygzJY27qizxWF^~|UGYSdyHlr^t>y&K*Pnj1 z2wG*Iei7ug&EDdv4k=e(DUGNLtO`tsNxQ1R`;XPTcn~;c;f=XvVYp3$(s*1*l(K|q zsHLc0SJxT6;w=nha1mf&M)+xRo|HU7e`~C-I(t#NJ}os>4kK;`HA(w$G+uHcQpor5 z#}wbJR->TQOY(=1opOLX_~#eYglAd#qEU0U^P2FNNh$cWcD(HlTn+iJwCqE0NYMuf zl6rL_I*FMBPKjkplgfAi*1yziZEGF}wgvc>&b;%J(IZ-E^El9j*)k zdktUloxvR$$noU-DQ&X!L0NiEQMg=Du2UJc4DCYq?2KJUn1pB;SnQ|fGR&QsD$NaI z)J`GiJ_HZc4672teG@k3M*9-(;4JOC!t@bicT23oeE(E6X_z`i=TYG!|xYZE(}N>;fQsqn2* z7M+`yV84}AyM+t8MvqFkUKP|+-?<1-v9B_F?u|q$ERBU}3^HqX2>iAm`jjeg({R_A zpVBWnkbAUvVv`kD0S*0Vm5WBbd~h z%4a=@w(uifsm2%0x_#fJ6f*4lp;xOIIZ6w?r%SP$ncS07Sdx?E7VqT`!MxlRp418l zaq^UJseMn+M#e71@&)IU)fZH+LX8RIf=7#q=jF~fPuTtIX7>a)UP!+g$TbI#+fxI( z$#55x;61v6qK?F|P#V*P!#NOi2jA&H_xCI4ZcBx^?^Xja@~qor`*PK{GnIB}FWxC{ z*9%x7=}O+w?mfy6BiEttXflvsT{@lHcWJlSZPA8zOjXi+-MTWOt1{TK-+u*QQG3VZ zu>eQeW~0x%_5}{&iwU03)SOyTnU8+r?Nd$i zPG~N1MbKHKtDpm$beHE~C3%;DyLTe?v+c;qGozF6vf-<)n=Nn4O8YNYENU-m2eW(0 zTni8n*wLw(fAVkr{~mzof5*R~;z$-L9NV6xIu4@T2GFj_6Db@*8_Z#@x~dBXKKHo$ z94+SoT2cE+*2Y~~s(9GBjhU`$3jK4j&4=9qZ!Y(O;kf?iVO1%-j&7E6N%mya= z?XLHcgMWPI2R5LM-Ww1Wo!cGl6dKvJd6*mP?>MB&J*70z745)Zu96!XQiGEwPhRa> zsseQdOg*%lEqGLIhV$RhO1h=l3#gd)RL@MZTyM2WxMq{}!E{L5Za2++sr?$M`~@ky z{$;83CkZJPFh$V!O;&cc4Sc;N!jvUQ1*{WlS{wJD072~u&vO}q+(lThKi~^*`lo!s zv+l;=28qPwB~-CRAGlTg4FNN>I{m$7ziXM>&KJbO=Ly`~k@aD+0^y#xNl~)*I@)3C zTeYm~ya1@ZT%w5VKbk{& zw#I)eR(=A8+V22o_<^9cJg)#39Mwwo^C(*%EO@IMe?trBzakDP8(ESb5Ou9J!&NOG z*=2;KWt+$KxRJ5)W2E5aWLqOl?e?0|_3mB-GtjUFL4G?KgyG3EsjQF>#T8gq_9#sQ z0tGKLC$X%_esYCc^%^YHihNil0^1(Lz;2bnTEsP=Mr#?uCR$ z)i}AhAY~2I{JV$acvK6=j^~#NIj0eA;@a?2Pi}`KT9bj>?yxEP%wdOZfv$K;;ZTv~ zVE!gD-e)E?Mx+4>JND03{wXVHW61YEWz=h|Zh9=yvkpKs5jOJfAmg;JV#|Oea{zCd z{jR{Q{KB~^=wyH3LP>bj6X}`IqZHotx(M$^+x4 zCe@Q8@n=g100BO4T5m~x>NARv%wnR5sD|C5h0pXVfu;j#o5`2H?yGD*Iun$|>GWK+ znQaHTLoiR+K*AQIt)>>PY)PNa*7;OY`i8c=qDw~y{XguzbySqy_dYBk(y4@W2ny0l zr+^5OQX(ZFAT1pOGa%hvBA|eDcXy|PbaylK05iXX0+j@uJ9+?Kso*aU7%mZ+Ii|0OP&3{@P;X3&+}$M?2Q#jkgOVCS_$U6jj)o zz3ssMuk#LWoDbgiUXeG&wj2NX^W58tozXukb~OIqRO}oBVdcIV*yg9fc-iNu+aTExoNpPf=@Aw>yz&6&rvPsQo3n;v*G9#!&7N1MiCkb&9NQ zRgSXFgk4CizR!P;JSh%aZh<3nYwQ*p2;R)r{!Gv=dDlD%FcTP1IK3CuH#K&1tUxJu z_hAr<7Z9K@+S5ys^D>wzXSQzVMVY)SN&C-Ndj}oRrSWMREH0gzFeTV-)I|WL)!xs7 zp<`#4P<4t8&g~hvXLBU&#VMC4P;|3CwUpxld@-vi))dVzMD?A?b5WA;j2i)|Y?W43 zxJDwEo?Yq%bY9{OMJb@oLa+!1yIM^YF#^6l>MCoG=rU{%5P(?k{(R#-^AAuG8VYa+ z!v@%@gAPwqp}Folhu-eoji{}mRvQU#aiPsT=4RJgC+F7AjZ`Q&UKlbO_lEV+DA-nS zFoWNG@G^>ZT2s6v!|wUk7Lc1Q*8(LFrPWtDO>vbLZvGdGu#D}lvv~`0d&rw$)7ru1 zu;p19V9~{HXb{_R5l%qq0Lxh~`nIHKJyv?FACN5n!0mLhL(%tyjh17fVW0b2c}8JL zDY15uiXO)KRKM*|X7 zp|f;+xcbwN=;UHy1kHRh63nTCsz6D``JpaNucz;{Zq7{It$iqvx784ubEDhzAw^u` zsw-;<&-YoW?OtNBt5K=)e-wa6mu}uCG_p||W zo?8@*1#Re+(UU{EPB!pgH&JEnFd!ApDv3G)eH$72U8~Kq)Vf64z5N6EM`p8no+Z

`)W3@@#pAC_@dQ|x)(+$>?YRnT6W-0g|M?>kwf z3oZC}b}8B(?dBveWu2~kdk9qPQ!?C2KlyqNA?JrgcDvu$#$1qHb!(-Got~}11|jrd zMLG|}O!VN+i@I8T>e~X-sniYnF~(QAEtGC4kqw4=vY)XyCMN)YQC3CjSka_o-hKY6Vjm^>sX3DP=5k zi7@%(R1;(f9M9qvH_h6CeEAII$t94A&*Vo|1H70iyBo-M-L=fW?1AAI`UC6izMg~> z-k^yT#Pyo_!4wYGiHmD10LYwjJS`^W5X8*-?<+#8MMY-2|#O&z>jMe95SvF4+R+wfJ)@3d9StVH_e{a(VlMK zgrT?98$Cd=PZMHSPu=ON5g9U@^TXU*^er3Xzc%;Bz~+t(DWDat;w)=i0^lU)n&w0R zm{;6O96mrMBoW5z*|jcV?F-sWKh`mV6dJheSir6GjY=aIY8Gk}K3dTl=)=MkmimqF zxg}BYZ}kC^3B<#>DCpS6Nn5AtT+4Ou21N#&2KZX}7Var@63xx<{0gwzyoWxwCUEFf z^aABk(EyWewfetw%fOhsRZ7p%kY{y)pW3}@C*tN;8{Tb=cG8dfcpk{c7wgsbhb9#6 zt6jr30mXq?%@m|_%4}U9LL7X$$ys9*bOE2St&*IrvSYMwItkyNzSXD6XZ^UCC+#h& z&NeGt(qb9+^0x3FsC3vf8=T}c=j4iEZ*LuKln7a@v76A>=1xo7Hl`x_`!xAj^A;7F z=v8;OGfz$F9xwOSyBc`fGEoXMUW}-oE3Ixi(`UfWNQtp@`hq1%6}pOd1TX#0e_dTy+{(G=-R#OVaPolT$LpvT z*Sv?%3kE4Nn&S>q%SH0HdPv+nlSKEU+iG2#Lm!J?yIr}$52$}UcdbT(UL*`8n7U>D zEmQ&I|Lg#u!a*aUC^BJ4$HTt)>RtV1X63LF!jX{(I@$<1-=G)(@NQ)=@{Oa~)W{+} zSX|m-L!RSd6XX&(w_X_%Cu^~0{i_tG)y9zSV*`EK%~H04jDm9ib!bJ4J|*m$zXrOf zRe}8RO(W=rX>u+S@gF3|M~dsNm4I?WqWCHK`sl`w7Y%#Q&u2p(4}TF5AhYuv?9EWQ zZy;5RDz~_;u_aRmh%nbK#|8#RuD2psj{lcZ2CmSg0*yxVp*e};Ixz2?>2oWtk)u;; zK(OXK9^iz`m68X!{q#a!Cz$?olmXE75YlZ=Ps^DmR*l4w(27~ROs@MhxrV8lzpPm+ z!fAMA$+@p|a!|bET%X|U{M;{T4E>O%y<%KugMr{I0TT=M|L|#X6QoRXg6%Z#p+JE#5|L2I&n7}v+ zQk$fIW8MFIE-deWJC!N-pZ(;Y8T>O2%WZW5I$zTN`U_1w0qfsd0RK!A@F_sa1?PbK z|7A>RU`*2Jddz?B+5ehX2sywC{ixpaFR=eV_}L%7G2j5>#Q6+8`g@-L%;2AYd{qUU zaAm{z=>E^+C=dhVV1D5G*ING{a{-iFfUomi{PF)~oX@~G|G(M%_oDIrKbp-yU)28x zO-1pMr_}#2+R(WFkxx~8`4XAV_16vuL=y?-2IJZ08k+_is_~V7Z_U2396$uL4!#f* z{S^`3p>1A~afx&V^>zY>fx5?><=*ts-Pfhcg=;OuDygKyD5D^C<}z zdH*CBPo}DD=Un@ThT?z!z7N1;`jLKS$pMr#_Dja_1a38JzyGiXawplmVwCh zh1dBHHq_{%$DLQ|PmSwXvV2~R1b<-J`sy^IAdNPTp?RMqp5a??US4qM=g`n#ADn1> z_HUSI3RqIn48jUnkD8J2?fS*~na1X3@w%fa5NLdSVVwVTcz@-_z~B+kpA%>KvFo3` zNEv*8l6`BV_spS*!MtsJ^Us5Uj!dXZ{1f+eU{vnU4&;htY3NnFrb%nLOaWbBjnq?- zEy52?U9(#)hTdi?tI)JL`e)0$9=jonBj7w-3U0&rn5>?GE~1=mMyC&>tf^Bd;+>;= zM6`*EQ3UZ1Bugf54iCl*7b)SNmrPxuiyVF#)CW$IIZi)t%{DfEb*>+p$SEakT?`~} zocvxUN4g}bFj z4`_&qaETLpRWGHCcGg4IvOk2FlzUF_w|=FS#t(Ay&a>UCp&i>>0S^>{gZR%NNlc?z zJofFRoQE0Wpb=r#3hOz@^0#MpGu;~Ty?3l9kxawymYAI^yjh8Nn(+(oVu|s;V0Zzq zTtZH>YFP;+5KWu{fmTebT~RGI&@x?FyHq#V+G;Y-q`K_YAFaHk&;=9k+(l9krdlxn z@gacE@#iCW4KtqZ;ScHgARj3CfwDN3&r}NuvNHcj{N0gG@6(5{45KKd)ClcHcXv4< z8&dcQ=mJq(x{J6vE%$T@BBGx#akWXu{4B=G$?)Z#6?O!_bJy$ZI;p63kB0N@wP3iJ zVp243R-oc9rhpGgC%g2*yIQ&@3y$p{q#ISl%{ssQh!t_^I+6nNNZ+ICBps)vfJS~L z^I(w3%+-_~l^_6;F^-*n?=_n%Eo3nb6h&_ylAVM0_tM8$X>k;kDJf)|T=9h@V%UHE z!|;|j$==5p4f~WBW=wo=>^pN#d>j#N9}C@8*}Zq2MR0yO|4P*5 zx{5rk)#xG)C^mc@?LRqvbxN^5lv2sQvnQwVqjuZR(~Eesd-{8){0E@gtm1m>{ImNB z65nu7=%OP4PlbY?uIA#o5Uq)t)Mcu!6Ztj?&_6Qs#qhJp_K%A}NT-wdA)`7?L8e8C z7i;+L6{^_^j)FEwEzrQ)w}FI#K~+gF>YvrkqC2KrOg)iQB<-S-mSfo`vxCfO5^IZJ z*5j4YI0(Xh6^b8f4VP2~sND>WN4>ss{NbFO;j$+vV%tt`a(%emghIkg!H|(OJR?8? z6Uaz2xkvGOteEviW&8~B|8*}UL$q?xy7gmOiu71xx&N6 z8PUa()E<7y$gPWNfi?;1USCkJSCV!cuJBqMGQpxpkchj+90F>;N|c>)7h5Di0bx5n z4=|H|I;uK;r;X!H7IO-c>A)p21!zFej(8p5~IA-XW@vApz)FI>@*8RSu|msXL4H-3~yudv|< zQyD^CmWsShMP`LS4j|4oCY!vX;rl|iHU zs?yOj>oKW!&#WK2Z4NWurPX&?FLG~vn5I#e>kG6dIlf3N0sj~dz=K5 z_Ulw`3Mcs`Q2SD{s)#u6Y&d(bx5Izq`m)EeYq9HAo2R9JYU>(kYae}`J}>-3`ewF) zyM?T;)@QwU>EGZvmn2)K>2DcXyUBt= z@Jn-jP}onwql;cl5i-Fy>=Fgt_v+7#X7vJ52CQcvjxzh?O0h$NVNUDM~et-euD>uC#8 z#dvi{Wx44LeiZ$BTht9nN_Epge4kdY=~)pGYolG`fp%{CU(uldYad2W(!8+jOy#9Z`z>QiyVbM>w2?t z)1M)-dgvsR9xGaWd0Or~K+U)RBtf+OEg?r$WPW9>>ozUOWO=>+hQ&Hn7quUrx=3*WJj!KZ8tPl^~T#onRzJ#Za`YHURodv|${(;g#9=;;Id?}Do3 z4&JSVL{GiLiHh-Y>p%aB3&b_wqq=u7SPq`oSU@E#9?3K}JXd~R`15`L<4!#NNU9Mpov1tv!ib+F=F-OM8|zA%>;hIh%#J1QUe*%-<4YIO%Zy~Ung-7 zZUM;&Bu3E2eElh(*U{R@n?0%C_Si#OQCJ>jWo-dv-3${=pybh#JE&q^`&gGl=V_J+ zh5WuL;cPv?u-KmEAlSV|@-|q*mMR>Fxvwu94~J;)L+1B7x(Bs}W;-2MGc-=tvzl4t zLO->4>toy+Oi0un%$tbBGWrL;5qdR@aXGu>tjHvGI=28Y^J=5%-Wr{EbG4uJ)KD|56Z?sy{tcYw)nhXLwgD~4 z*BQO{QVd$>mIR8Ti*)%#4=%d8rg=MWLK{cW=mO-mGL7|}K3qsefqw4>Z1?_(vKh*o zwQChm8#hBC{ZwWO%TPcUKaoz_oix7mX4_X1Hev7_0woKtiD?7Uxv=c=SBO0p>d#Hp4s z;q{;m$sImYR%RR3@t$*Mx`=K-j_PJXbgw4C>cySFi@dtMNW`DrSECuQ)s|I>qhbXA zxIPZLBfC3lK9o2iX&=4XjoXg_a|_kP4odb~>5#h-_F7LSFLpy#>s+wgORkkC@^;dm z%vVNT1y2B8ktr;(B$#xFxw&BgM6aT89lN3XuwQ+mFfr-N1pCYF7%^QwVeo1;>18Lf z&#-kxXT$fwVakTYaz#Vwc<$(cLNz8mtU@86*oPqsq*#b#fUMLbMFfhx=j~hVr@Vhc zzbd|SUlO-HFwYgc8MNQ%O)!QmFQl;L-hWEGNfJnimsF(o9lypJ33Ih8ZkkjS?Y;Z1 ztr=NR(mlWT?!i|5yMz;8x)Cs7PdM>ySdaFZ5s^t|qp|8b7K2|Dg0^_mZ*rilkIt-9 zjBfgDAK^bga4^FIHkxLVvy~8giJsIYmeJ?lCVG9aUe=j&_r;IBtqwKvb>2$fP{@q| zs`S1IiN$!QLP}`QatuAK?-C6~+5v13Ov<7hz=*H7ftjdu7I{`HDzx8Kuku4{x)$YA z=@#_pDED>hWWN0O5_%#~)rOV_X&J7`+y(7)f}a&Bz03EqX)iaGblg!bAM0v4C&-jt zP?dDXrZTl+D^kW8CVWaRjE?#oup}%a%jR$z0_kv~fA1qzf>Wq&Q&^h*026r&3t%_; z!;Aq`#f5@jLBl9u`D%TfssU+~AVl3?oCKyIhtM0MkS(Vd9&i5&s3|cYAF+jR^)=px z(Z237cUGTuc~i=+mfM3FqF^nmDD{ei(bf;M;lhsUH!&wmttgzX8>#&|#mBYaPVy&4 zHQD2|tEbX~egl@D9;jpp7mEKfm;k#)(W81T^?r!0px)3O_iz=;LV8+KDqkVz>*w}& zIa!~5Qa+{NbYf-m64750&N3+nl-abcv|t94%~Br2)SdJoqv3ZST_?s)O=49&);rmC z93pZE{^Ml9;3l=2Mq;QE@x8D)lqPK7Y;HUn1*$L*0#d zOtM)@6oZNmJ#YT?<^J^LUX2jvQROZ(JPizu$ zEMrg_ovPg0F)kBf6bbo$pml##?557AP?CQtexaJQI#Ds&+UAzxqy4B#K={5=9>aWT zdG(%vBG5wRH)K3e_)D@xPMY3dzDN1i_kdfPME~$TPM9Z42|9H9zjP%o-2=(%S2hH% zC)_srdp&n}M2Yt>c-`wMo_@-`&+eXde+ZKMp`T`xcGtxQZ6aM*11s0^v{cW9l2((U z3Lo?FT-lqpoT|u-aW}Y%>5XHc$2LLc1m76 zyn*}dtD{Ghr$G+XJBH~hId;Z#oiLG21J?9~_rQj=3Z=xCU_dobr?CK0<=H&mD&4%h z8#NV-qPi1vyj@PF(kg{3#9FXru8NJx#|n>@GGX@762B6J$_L%7i!Ypp zZ`YE%mp5p&jmETQlgebt(AT%3+5Y-oL&UIKFX2L*d@+-+4v1$mUxtJ8$ZrDBp2_xk zcXK&Dka4cI0NtT1W~}^xL&O{a<4CrCvHjF5iDA0*U8Uhj;Y1OS*yySo#jlMytwoGr zM4d{}M`6#Wh*qOQol_UmU-ypg{c*;W`uvFam>Z4>Ds1BY(?056eEV?Bbh7WWcyg^r zZ}pva^~ic4j^1I9qrl&kcv|DJYFWIZSCBkqB28%%m&%gsTruC}j)X|K@;BIcj+{(? zoM_PaO+0ScDSVRFlO94zNHslWDYJ4AqVZv`ChdCuP#`1n*Stc>VebWLk#2M5y88FW zeCDr8O>s@fItIB65juQ^Q%2*DY8bdvt-6dbPwp1&cx=k6c?6=%oAM!;Sizf-rG6Q8^X=v z>*zmDQ_Rckkjgp7E<`A4Gphf4d8_- zL%qZ8K{Q0}=Kk?d)s+UGaKE1;MqlsdwV&1k4?0mijmd?>CeX7Q$zmZ7M^_vttwh`; z_k*lD1Yd6|iHy`u2`kaY4n1_Bb_sq>C=-5KXLo5&?`{l%bw+(TAYvzP)FIXL|IPd2 zSr=%{#s07qI05n*ugZH;FqMT&*v$@;Gny#5)BQFZ{L=lb+nu_=t#|K*VgfcS66kfE zK)ACFy*kvW$3JSsLD8ClW=!X2pBL`HFjkyfw2VDN$)Bu$EwN2CR}qeWk)XKo)$VnR z9%GtFmFvX;dJ`-{1Bd1n&o47-ll@jnj{=4ZWXK*NPqE&!W1N-gJZah+1jW;96~K>V z5y_tzigv|`v)*ppw|+bhK(hU>^gVQLuZg#C+W3(>Y;>_n@i1iTF40#|qyNqEc(FD0 zsI?%v~o2N$Alr)2oQ-f2ms} zE9k$s^Sh-Umn_>vRlinM67>?PMt1jA@LaaEa`RI}nEhs|dA@q}=W7g+`N2WJ)h(hv zHoTX+Q)Jxya4geO{0iFon(sr`+xl-^c!=-szU#kr&&s+KraIgxQL+B#_ebNUzmEG&>e_mb{E~ri!)XN*{iL ztw`WROMSwOUh72A$-0NsF&#BRu~u6o^}ad!FkneydUJF9*d@v@6L*+lSYE?8rc1x( z(Xc3)OTSare8qT?icF^Di9_40Q?qB>lssAXJ~_;eI{$ef0=wTuQmSA6)ldPeopR+B z`g3;(3}h9o^yTQw=q@St#R4c$HGYn6-Q0$sH^MvkS1=KsM(Xvk|KlK^UU0Ezv2&}h zsyI7F`aanPuS0$(ith0HU9kT6_-JH2J-Iih<#>(**KsT^!R5M-@d&uJg-FCqV;eTz z$+uCN;+eUYFWss2^#<|ghvVJxqS|1prRq6m6@yWDqvFZ>_c&aWlS3hjjodrD^um-9 z8OIUn*dBRqkQA{UWm=P7iS^1IqkdCWvBJWRSXQFyR2Ni9+U!88y=v^J1(?~5g0pCj zw?9@~PdlSU{UDxOr{}muemaBiN7!ro^QX3*yS+$^3@7awFCmAoKpA}BVW!R(spX}n z4Yve8%P70Tt#{wmab%}6yD(x zm20q;+dyXwl9$!V2gczn&0BrEQ;PZCirURBDW`Ia_K|ZSFF~@CYi?~0vwng613tR@A zb=06yl;;I@7`X0GkVj}IKMeqcT|ARQ38uqZdTKCib26 z?hD)Y7ExdQ-n^L6g3GhkT~KoCr~E3bDdjH$C8av$%dPU8c?dk>*$DBU{Hvm&HLkX) z*1q&h4|sxHcGeHvGpTP7QV67_#$v*|^Ukk2ZO9~o)aR$@?CFw@CQ@~Ryyxxn)?32JZAWPUpr?RZhif`K)pMc$n8%wj#L5-7lDJKIwb_$z$*q7vc7K} z5|G2`A9A+t*UVQXDKvTCUmU;ZW1y6|q%*xoNSDaSY&72Z z@>TE1loR=$YOd>Ww(-qFb~>bjQ}81AxwUWQQjL~rZukqrurO7>S+p!)nNCbJw>kb! zLs#HzTzY9)ylh&O_QgNXpc5ALX58GKBxqg4QM)eQYM zpOX$6q0trZ)S2#0c61)uiNag-qlFx36$-uCv#s!zw=XH=4`5!iHry)I#6^{S0E! z!6>!WX${^a6Hxtv#zxFSN$sVJo%=^R4Pnkthk$WCo}9q|OAfiDSVXa8_LLYl?keRN zuj0e}_ux7qIlxwVJ*U@5UxKp~qLz#^yk=iiHle8JQW@twmE|PxAPIFTX-pPCH!N6$ z+gdTPcKXbHF%`2x1ldwZF>B+u)shxIWp~Pc58(Swsm*d+HP;#?O!YJu2E1-0lkGMh zy_hoW>=|6Lb4y~(B&55@yl4|-G~Sm8ra-QM?A<-VGV#xjyBxw(K1!WTwlsrgI?MK% zj()+{emD${y!+W3>aTs!*y!osZI*%Dk-Y)2zQ`)r{vmckat|_lR_bz^tka7pVb>{o z41nq{=S&yGBSDg^&`8kJQtz(J4B-Rs{JFd|@&?8cYTv~LRAtoCgRWW_pZmd@spIo+ zs%wGm#RBz>Z%M8kuY<-)G*rFWl|C)62?U=SDZ$%LL4NZ}tUVHug+-VU7R(v~6k)a9 zueYj{-{^R+U*{vJ|H78KZn33R&&8%c;{7^HOqGH<7TxEepU@LG-O0BDeWHQZ2c8=F1G8zdOL z7|3TkT6U~=KN(@)Tdxt`1@rKJCdM%)qdcT@uz++ARG&-~m3>lpbjVfIN#cDnGxfQy zzQ*{uOetLC6>GxfXpKxz)IeBiQ&u6vJJW!^x?0Jx{y_B*V%5?Js#TiL7J_mS_G~-? z5q+jVi(gZxg|+`q;Y`2V!Ff|<8(4$(21s|{pZ7wOhS>Ta;e;q{5*lccb#DIIaCFHnGHt^cL zq-g1=aP%Ygl#A-tq+;}wZ4G-@;w5)LnC4`#XGtlcvRl+S!&BvuN7rzRPI4s<8_GZm zVpd9myMMh?WC0de#ZjV8J551f9K0w_bqDJx(}Ti@O^mz_IB<;-?d-A?_@j2eoo+T7 z%{a^!JLsKh7u%EM3C7G|Q83B$5;hifqY@>Ddeq{@Ar~5SZV_%sv+uiHzO+W?-eFi- zxMm})oTImX7SmVyARTd;{8KczlZcU?(NpP?wra)C`;4IL)1lf!%vNJR_+TXlJRhf` z(<*Z>8^}a=hDb*$aCQ|$4GhF{F>x%O!f%&gX){%n8hnzOf-#P2?yObdA7F zKnjL6Ln&>={UIEm6i-wkj>YBWdD6GZ9pX;(ll`_gEJSlJpX-lhTaoWH7{8d!Fw<_e zW@w0_d9H0vIxRTrBj#ZslmZ-# zPL~vE(^yJ25OArs(o85dS4u3mSNe3qk}TlGje`BWIeA}dH{$lIYGw}0#ag&mK>PTj z8NZ&YjOk8bGQx0b)rDs7FF!(f>qoY9PgHKN0YBo?TPz7ha3qy>?N)Q^?{mn?x>RHh ztdVEr*~H-rqpi#fU|(HMV@Xrc`=9!5mGZhH8qygiyiI;z; zZ=LR&ke`yp?GwmiGx#Gs9<0SWLOhXFg2ztPIdWO*=s0 zX__4@MAv6~zIzL)8zZY#dBjE&lc5i3KpPr zBS6TSDOoZj;GPzx1C5@N#kw|*#h#k4(Dk-dt8MVkGbhGp%aG2jCtDG2f#q*iuNlVp zc>dHt#eJRWY}FH@Wi?(*d8Wk{cOfFa$jAbt6Dx~S&9sEoH+Ol52SJwyOPkrU9U=448tptBhL*9uDnwU^R)eBKf^r=L zg2_Isr}b9Ri+Rlh2|B|ooc@XjikXw=*AyZzV*1SHsrk9u+fk{D!GWVg;N}7@a6@x!c%MW2Z4g=&Jg|XgsFB~Y z<88tzqYWQ_^jua zq7YjmS0y76;~pWN_ItDbcm*B@vS8zY*P4a;Ji=t#jraS6XM+Bq;xHgd09F_(ovJkX z_ImXZ8jH1z5VdJA2ul8##E6az&G4c)RG7%`o@>_i^X7(ET(8B51Xv=s+Q97N+0DZ^ zY^Dvf?4ewNksk z`e#AMCkb_AWDipz3%f;;K99wBvPOT5$}BMHZsQXrFB>!F=vQvqH@i(!Mkv5q;P{VwHJs_$@N#$S89*t7*#>CuH zh|7gzuVZrrJir2&F7fA9Z?K*kzioa~;7-z?*}>5M!e!XAi~+JNY+{056qXMk&SDLV zPC|!S;>S@W?@Q-BG<_vx3(L=BEySf5%8jdA74_>!6N4x*5zzhYmCqJ5U7mL@T`4wt zSxoqXGACoo*0t&CSIfs8T=5js$Sd#+`%VcLX@2wgB(E4SXf5|00OzTeBGs}*SSmWn-vQJOJ%k`W62OR6j7MGyPRm|25 zT)ntYho>(A0L=6ZPj+7ZM4!j!>UBjS^;%z4Bm1L-14%5_qq!uTrnBKmJG`oJr=Si( zS_!5Hg#nKaWGSP6;I*i=u2sPCy>*q#h#raznj8&Et8RTDq<<&U*c(y#g$_S3$Q{}7}l~>+0X4g^#>1p+|r@m-&CV-{@idgUju=Q zb4a=5pTLwFCW8b4Xp?vAZC3yafB`M&1HQ0pH%s!POLh!Bvs&gnn(M7mQ?NY_QKlAN z55evvdQ-kr{swJzrjT2L`a7+$Xt?;V+FWu%ZVu?~_bJmo8EmgQ9f}03IZs7D)Vc0>NuJn)Ojb@Z}5uE$X$YSbpfgtWh>LOAe=PA!4- z+rY%$c|NfSPCvw{aC*})aP8h1NG*uBd`_oC7g{vt(`MBn`<5%hj6 z(iq5EM ziM9xAUKN9-v2(SOT()o3O4f+Vo4)BwXX$K?pb=jrNlz-A$JuPg-gqubK`e{*S)hY6 zMgO7l!^t=O;*dBZ$&`HgPB+5uHbc7@N^OWHaGVpX|Jaj!R<3ZZ!Z#`aN`#K#ahE4i z$DQM>*R8^6i7pbx=BV&q&QF)Xw+>PB9T_QVH>E|qa&S6XY#@h4NVVpSdM%!tC+n(I zvs`JQTtek5`6DzfDg>j}82u+A@=7j%-%+(<2FfLdT!?c%JVsxDCGkwrM15$Yjt+-i zzlraQ-Tz{0cOl@^!ty51^Ai(9sf2c`t6s+V$lPu++jl6f^=@asWKC6BimrRk7mu@>=(wltJ}+DDYd_$h$)Jj|ha7w_fSkn~M%LN)>`&L^w}mb0 zcN!+(PiM3(Rf2-BS0#FU@7NW&3Cn23*H%vFOj=G8@%s-E{6H=1tIN%y7S zpVD1TyIS&t%^G}l%o_=DnqHMq@ks_$SQ6D1H;gkc7$1-hOqcAf(q1q|&#g-A^3Vwl z8=YaeKFWWfA*4A~bT6`@K1WNk0YQW7S>!M5&`lRHwK^!;qzbx@Uh9IcPf@>`H9`5*i_6BInBQNU0Xp1+U14! z@aZ^zXkmZvC}>gh6{VG)2sf?a70#u?n44!8E=rnVEQ{8kr0l?_;>dpS+nwq~z1V!A zwSSJ9fs?;)macl!N5)a+XnhJ}`}fJ)yyi>%tWStv_S!-YaJNigZZMf?>RPh__aKe) z4C~lYOSQYf`*0$fH-(;D*sKU**1=%-wnZ%A;XHHoywMWNojQtpb0l&VyNJ?)7ER2c=BO*E9o=t6j_>e;1)Qq z+xsiMu_8HGR0angXErh~S7B^A4ZCZbti)wh6R%mMrcHNn<72_6j%Y+11ygEA{(Qxt zhVJEt{;2yl;rjtq}@^OZGzvWcXLY`?(u5xP60?kh_ zAxG|KxQ~8)g)q7#tA7)|2R*|{Bf;pjJ1UF7q?u%Y!Fw2KHz-Q8Vt|~kk=eqK5@n5N z$X>;e>3DG_eMzVEvdzR(MZE8v-csxE9yh=vn>(c_BnsbnBaDS!4Rjp-wsp1}WRcuH z`hnV!m9i6xFNO!i?6&w07h6MmR<7c;uRb4`ynWX zu0ZDr6dRF@;6ZQokaTFVO6Y{#3tUC`V#AK6iL#9zm$3adm2#KYN3>4r$H>f;4$PQN z{ca?$tX$U7_@Y#cO}cBcuNJUz+`!&c8g=fW205aYP7lmhh|ANuubYWK{=3|pqdn*) z`maZAZfp%*(;0uQz|U;NS4M2@QG?94$Qkfh4kY;$HhC9sUBlSEWB^FyD!$yX48u$y z<_GK7tFVb36v#TxMdMFQe{rE!R)54i#4fLS8S$DuBtMCTcZai zxu2rZQ4g%E{XjNiMy7{b_PZDh9fLoyk1u%5N$)I7r9Kq~yMJbGjj*uAqHuQEj^Bbvaa~r3eYL+nIlx{|1GTfc~ zFR8(j{sXu9Pn&r}OCTGN>@1D5D5HUZ)GrFJBFZ$gGs5aT=rpqsKe z;?cGjMJ=r$hQ)iaHiFURYbNyAN99jCmhwP7jmd+f?BA$e2*il~#pRUx2skYwPSwbl zf36Flte+#c;{p0I5;W%#iT%qP020z|^JH!IvJ{JYwocB=?&rw|huUj~gH0zAa$Rmj zCbB&zP8@fCQ+aH8nrKyZYFu9>^;oOI+f`vC{^%9OZ3Z4%=p$eG%x0&o_!v?nX9(M; zSZ3rXqpDkNv^miDUIEbL`FEjvPvZ*kdWh4pVVHi=HXwGC(@B+>#QP8a#>4&?C4+_D zL#G&zvUWN5qfDPW@(Nm+vC%HX04Sc{*_=!~N4E7d7GNVy z^;JK^8ob%{mA;YVs`#dJ6oB2gD&3XL84(e@MPGTy>|WV~_X|^=JfP%u9T%*rN#-f* zz5zHbia&6FxCqhk@9=mGC|Oj?+vwk8E!_9Qrxv^^A?4a~(I#V9R|^znB9bKEc4_BW z>RRfx%8J854goZva|}q=}j1%xZM~V!Vg*D zGYBL`T!1vz^T9;B-tby9r4@?L}r{2l`oISQqrKEWUN zLrLqmxDnEcsI$ua*e??tir`08Sk}uSM^cAE*U8wk1<4NwQ|ClSS}b9+FU|yBJY`kR zM)a$`!-L@U$ zPi`SZmV`~`W--_yY5ePQ_7?FbN5+$= zho?1S=a(1~Ax46ap>Ivi6c;!p^QSL+*yA~VbXL8mNB@zQOt$LrZIipE0q%ZwO4?Zd zr9HV(+#__|YC_g={OPM#xEOxGY2DH;UxQ5x$2@FM?opj;Dz9EGGvb$#;{C(1?*ZvF zDVHnHEX>eEq-6JhvS!xh6GiqADHVyq_Dw$@@$^WfUni_Zgf{@4qg#4;kB$V;QBSKg zG5Q@2${qj}YB2y#%ffnPb;{~e-4;Ho#tAv%agsc1#i4`(pV3VqP<*hF<`rWaEq zuoqnOlqzldnpzOWI79vWhz55NX{Jc-AcpM3L_2@iD(bYn1)M zIS&kI0PBn9igPcyH10a>LcQnj3{6Tcbx4Uhyi*r;9`OAX<7h+xe43Q7H@aJ0_)$S& z<_6Cvee`XWBU9=>e>hdEG4HNb5%z!>b@BXddjx3R5?TNQPEIy@|My*9DYY`1=|>-n zWk%+@A16s$E!~^u>p<|SFI!R>OwoJH6Mfk8`k@w`iPz2{+nJs2nL@p<{m#VkhDD3z zDLc{WD&E=78g{iIX8vJdWXduA*A;TS4!q8+BUCvPZyyE%YmX;2FI9s5g3ysOQ?LhuFE+vpyGTNXQ4cv=v&xGX*KQf`JW`t8 z_|n)zh~&wQC%td+vsWYb`7yHKi{n#9RAXXUIG|d&mg9Ki9OT-=Edk@3b6ULftr9d# z&8V~kkqN{bzdqN)!QMW{XhdLCN`?S4)B{4bcysQUz6C~CNH$BWkdl8=XSVrTl)#g> zeAbY{4Y{3xYMH_N)BMubRv2u2J>US6@U;=}gGtFdp-O&XrL^p*N}B8V zGDW?yIensIgA_X%|AaIm;;K`ujF^bS3xj(C-Q0F!gF%NYZJFs}F$`Z;WTJ?IGla&> zH@4lmxBO*832USS7H6RMPX-x0Yy*FF(pRxJ(mMe%hJi$*;+CX+A%AX?zhgyO0{oBd z*%;J3&VL9{*B&8bOw`JwBjd%W65w__<^qkUjEFaHf6(Y0SV=ql%L_m_AS-aq%*Su- zVV^YlCo-v$4RzyojC?o0#%D>|5orLm;nC)5tJjTn-@Ld&lEKDcO0h794_`_wsd2^O zUYS8b(ZZ(@F6UMFhz}PxQZ><)IgRVqIO}KtMgT?%a`7k7D%}jaIsWOgR*t!95o)6~ zPsg$=20IavL9Kb5dp*H6ovAeT^C$k^z7~W?p0|68aB(6=k*AtPX%DrNDH^nnS4~AD zHH*93!mh;W$5Y^ZBaRfzR*hbEg`zwAz8GJ6lNJb#`904`ruif`wbL#qLEN2ni@45! z{pllFJvt=VA%XN2Yx$v2m23o&t2pU~?iRpnIp3-4lNOqBUo(0Z$yj7;4X+gX$f;vt zYW1L#;HKU!bCi$bdqZ`jwFqjwkX*&ILp)y3IWpe>k6OgsmD4O+HXM)LjEgk{^vJ18 zRNN(q<=h#4?8CL zZb#HWKgzSWd*2S9i028?=8@|#r&T6@L7eg|(1=;*-y4b0rl>UQwOEe67K~zDC%;r` z^#jBKqA_&@$@e&%Z)fUN;}!l0VSn$!03Rsv7a! zI!N&tJmGs-8_>tGpBPt3YrQ#=V)Ul_7lZV=aL)9wncuvrEcz!N#&uK5iX$q2_)S{d z7W(Fkpov0W*E+hl>%6akMydLbSVs>eaB_lTC7>{}@TCxMGNL2gtmB`c-rZ{;4E~mw z+jorAKnj5nSI3VM$A6=*?aPjD>|z4p;z7La!rX9?%n zkK(EBteakKF3y}RRBmJC~8)LpY&h|Ih zlqd!tX41K;sMaBI) z1Mdupf`o**2%_M#WXG?(EBW@h7oRg9TR3#@$WaEa=R;h?xRz-}?NSmRh{*hh6d5B! zqPy)QiugOxKa)V(S155i(-TMMBDIc>3KJeA4t?V=YIuT0$_W`m{XWW=Dii&XOlG0) z7MzVcT6vKAv2zWpZ*YUvME(u)nPtbCm6+=t`#UU1 zOZ+PBRtvntM)()UWen*^i>_@7l?5riZ9 zvz}IG2nR5CCgO*^c{s?KqV4aTT!56hGgU+%1QYqP-rV(VIer^=1piwu`@uX*6qjxG zs1uefIX2zgeG7b}T*CC0-napwmR_u{xT7er_|$1XEelk);d@c>uFQ>}9*q`@@+49a z5qRt7+H*HRx3az+K$b->x+5 zUe=VYxi69`qk)HnVr9*r%^Fg`_1xi6X)OFyd_Fe~;|hcwMd1Y+f?0Jx*Pna!ZJ-b;=#**oLe;ykVBbqyNZR}# zP2kEX5oK+05L?VUa>kRbl)at`3``L`4?k8N{m3kSO^&J_P69AFo3wN=t+;cLIL#R2 zdLE!LHVmf<)TJxSA2?&QyhNdsH!YalKstVM-QZsLTDQ#rvI&>qXMBvJtF6aAH%u2y zd=Fq@{YZ8^XQXh$M&->R8Zx4#gis~*Jw0ve8!7`Kl-L0I{>AQcy=2*~3*=kCKL$$v z$uBniJCUkE(wK(kI3YtUqPz=r@|#zPV(DN=%%X=mOdKc;?ojpBpppBY;};xbp)4yY z36EAQfvp|Eby4k0V@tSajx`z}VkWGE(((06IqD6UiuhW15xl>I>91_@*U44@=|`Py zh4Q}^Ll;fVjTb^hLox~-DBz_D`_+Lk2K?6pY~lGlFyz6n?HIJw;>^<;MJfCkDaj@g zgurUymS;AGLLp?`vhp3`Vm@tjdS-e$1tHg@{v)IL9R~rAkw|EP(xJae+`oq7gX%|# zTM;yxmut+y+JYG5vTvx7H(1Ed4DYt6x|ja_w%<>V(vhYm`*!>K-mv`h{C~@J|NSUm z@@vSaSVj@Q1quKD!+$)9o(7KlzyI&Q#`6D;-#_l)f6to#59dn~(-Z|XAo=f-(O-E6 zEo%HhBU?~u1ASDjNlJBJ*}!CR2<2v*8sVmO3iJ5Wq*8UlhGGqkfec8X?yYTQ#NW#l zJir3ZklV%o&o3Jyqhbm`=avYFuGAnB5o!U%P&1p|cTZ|ESuuZPK-w1<_c!kMCv+i* zF{aJv3bI*Z`i~sw8RnA!L|eiegmG#xCThl;zrRZYfNdp0#Q%V8&EK~_ty>`cX0vkm z_xt??K``_-MBJS04NFn7-o7B|h{8WM`H(FjRHvbOxu=985R9~WLxJ?Z;NMq@HbfqS z_GjRiu45TD;Q~ycL}!bc+8%Uw!xDu{gSm~<+nYw$B7sX1f|RX~%2@LIebLAOSB5*d zPg;SN9|O>^^M7$u{%dS7GLQ&&p7&Le05xwKHM+$m#eDJnbY>#|(8{eWku-P4NDB^= zR(j~DiHq1w+w3_M5ay{Lv1v3btB~Iu*C!aNI>eGEVXe~!VjjM~l z@p7LxmFt=eXnu@6@v*q}6QJr6sD^%4L<_zNevKnaI(o z0x$&50Wr|gY{;aNsQ<+mY;S9sk}QUqj$)lcTJf91(!Kd0b)Z8$b7I8ZVb5(T~Z9)4{WtVDj*fbNSaIJ%$C$4?yA|HIP6vAcJW& z@pFi+w*BUu^u(5DAp1_Wf#GKVHZt0)rOiDHfR#WWonybc41`78*#6F)_RCQ69qA&(_y_{=Hl?qM$G{LQ;W- z4rgSpG;1wK6Mfv#eGmntG=j6?KnII3uB+}~v!ph8S%6ww>aq1hiEa87J=$ySw}sh3 zg+whR>_$e%#2lRPrR903o7j!R=Q6qZMic&l$T)*M*Eg!a78%>wvJcOk10Zi78z{|r zG@Aq>EIy1@&e$$ACyU*l33LH%)3tJrE+| zzwyO=e_#0QbRG{!I*Vfe)Ch16UxdQ~&@&oS`{Lms>dOa%+P1W4Z?uz09Get!cd=j8 zZ8U8X#A8}2H{T1pFV5L;lKSqEr@WD~5a`=6h31*ySl;Cvv%X67(;kGTMJ3^*-`u?N1~pO}x}N!i&J8-59{DgM(u2JpNBLdYZZ?r%yU zRH;Nu_M@@2%%p#B)9rCgrl>yaaWL2({`rxlIL58!mGOk@m17P7O5FBhXp4fp{6H6J z%I}34Iui~7>*;|Go^zZ#6sG3TxKg7g$w$xjX?Lfbd9+8odLYGUp#t4{^Bl$6$cgVo z11)b*Y$-&oGbly848Z|^_{g~TO5ncH$#&4Mmai!?i*cJSxsg4m5!~_tftqdnxs~e7 zs&as$Q=`LjxzY1=z+$u};b!9FDgmn1SB5C$?rrY!D^={7^%*P5drjU)pR($Gf{rID z9!!rq&VyeO%$N9AeAquazily5+gNE`JO>nfHCqRPw+DeCd-G6`*IxKDO(M+5`~Z;m z!u0M7{aW*Uqf(%DZ4z)(3W8sM8);fSaA;*n7Q3EU%!7C2J3_NUL4>_KlZ9rld8C;< z&TZ08c^ZI5)2~W&RX&yn1bEHHhQW(GqY#$2o%x>!Z4e`V$@$K0lD=*u&rF$~$M4Oa z9sn9?9st_CeADWF9;8Q-NWpA71peYRY^Q2M@*CV+`z5Auy(2$*K!hBUXXnG4yvrPl zK5X!7jsN2(x(@ZzgF?W~D;j?IQ-J!q-Yoj2c8q3HFDLZb;VcwwP2>sCjqZ4YsKsiE zG`s(_FgXv-Ad0%ii)umS#awkF#Tdb7tEIL?4NrP4~w$3mk$FN~nVe!7Hhf4(+)zy&a5s?jtk z26j?gikP&Y5>kf2mV|z&gMJEnUsBLfNbN(jRo{vSnJ|b4-j$kmcFwSuw>$+b z?f9D#od~=NN7utn&ptA&wc=4Mn2-#;=b&j3H+FICT`9!PS;{uELe4r06NODq*ZOZe zyyv>Lx-oul!8Jgr7?yw&LJ6BW=|Ug))WB28i6!ULYMU3tTVf(B%kdj+Jzfqn^Wi&e;kNj=U$%a zv~+eY60X79Gm_iNDqEy>4L`S5dDP$J-0~@V7K(zcm1+MRd?>+{(UKaT{GA0CLEPQ$! z5D8U~zX3|wE7<1*t|2l2`BSl4%L zqe<3xFUscMwO3$<`yJe6JIq&UXna=QKCQ%Jh{Bm5lKzZ4N(htav5WJVEO937ub+PY z{%3*%H!Gm!(A=Cd(A+&5e8*bYb9N_s^>s*V=_D-!C6zq81qSy|8>?zLQ0e1rP0#Jp zGNXZBTL0`z0_4t%v4~EY7pYEo1Ayt*jzSWvrj76@Md;1fZIj(r_mqSnz*48t;jaCH zOBrLLj4I7kOeq}_XifD$JGRyhgE zKCtn>1NXnI>wcN!fX#Pkw-~zdWCEhE@vmRRt0V@E|1=r4dR1;?upoe=U92@`d|@qK z^%Qk1^*K(=7+Q?9`0dKmf+hBF$T%5R1e_&k2jkNyj*sHOseIO>!SS^~ky_{cD4F{7 zd+2zV*ud<`9=(1(y=e5cf}!=w<9J5M<1wiDg&PF5P!M{szq>gTQi>dd5bPP+Md2N! z3F2CkTacsLYjw&*#f&k-|F$Sq;iN8p{Bw98ptxT(9dD-GKeOL8qJ(&#=R8yPB6D4) z2JfX2T&1hS7xBce!iy^rXU%QJfMJ(d8ZB4k32?v!9p*6t7U4HgzFaZ;Od(}(GjkGa zygbUn?UD8>T)cKyfpfyMtsvLthe4D0AV$II_E%01K5s~XPag1@t@5$gQxv(wWdKR! zgJzkA(36J`*#PGk=tQpRv4Ck3E)j0UrufzCTjsBLI`XtWTycM-1GKmJAJK!ZeRP_C z05SsoA9)0~xlfUon4M$($`;l{k^jqD$LYu4taWA}IMqG8A8^9o|;z({3O0E%be+f0GDFtI*4qNp5PJ;Lrn?e4-`BHhL}j(d}O zO=~lnd1B|b^I&6?h3R!v1ih410d8rAi!jiX%Cougd+kr?nW}eInbBMh8*Q;(PArjo zi%K8Ds(TFiNBs+<22rEN;CL{&Q|6u&S_kdcJ65C4#h%%ykBH&DOJADNQH*m{bnP!% zIVAwOkmF-#Z`v6;#(MJYd4)ahc|&>!0S`-dHYP>%?V+CB7VXSZ3j)Lfw$nNZgm<75 z-Qx-`KL%4TU@bk}V5?Z(&)b33Q?qpwNHK47iLxz;F?+<$Vg}#CA+Ncd79U#G&{m`b8}KUMNRliAPQd@pS^5bXsN3DTL%_MNLth{sPNV%5i1tG?Z& zRFUcFgLOf0T9%MIsjzg7?d$}roc7jCSqg0)u2g}kVfRRsWLmj?%WEKokXKA;7AZ2z z@B-blNQ^;j-9w(AS!Iw@`EFxS(QYJ$IFFw{S^SEIFdRO@pn3S7_$lSBD!>SDDrk3Z z<2=`6QN3Q@3k4@8-AZqz!S6}k7LFHWDx`WjrUM$yd&Amv?;Jvw!=cCq$pxK4IoX}7 zU8{Qpw>C#VHUoVlb;}Lfxc%i)9!-?yqXmXvN}8`=gY)1Du|}nevO?g4W|E85Nk~k* zBlxBc34$k&z3F0NfSk2=QB7Gb1Frb*oE0B~?8TDz@_N=v z90axTRhWQy^VJbIH9}ywwNhh(EGg!yx_@rZ!fkWN4EFpnbR=`UFbwMDK+zdP`~5NY zfG@i)8_j%yM{g9Z?{W*(nHaN%@(%+M5MJ|IHQ@e8Nu-S|b-Kq{AP?VNPrvQA57yf~ zdsLTfSMvc8UGo_-eSZ3$SJiWAVO0n}L$Nrxg0ojO3*`7(Z2rS)nq`}&7^!;zi*DH= z*bQ>L0*ROWQins|+6Gx1C0XSbimbn=;B32{+CJ}_63;vY!Mz+|6LQ*uK09yKNQ0&= zm?_6D%Z>c};|sNmi>lk4jM6UdZmKF$9>_l`X}K;hx08_1H{o}hy3s8&F0SjJ#NW!x zcBlU{EOQb@_B~uam21b6cR)Stn;O2r+vx6n4_Erv$Vs5&Fq?LsNDM{wLm>OZLGqK* zGjbv_K^Ql;*cPiZGd$t`Ap(Ljpz!5>d#1*k(T0FNRljLxaul{d{85_Dg;g-7iKy(q zmn3S%GNQ(keGjECvL4nyxd>p|lkPx*z#X^-1Jua&iNU6Y4AZt@-W z)z6c{SQYBTXZ{T2WmoCBidq@B=ksTPWrw}mMYStiX=m%rDiMztp>eU2dXH&Valog= z7lr~$_7?qU_*CvuVprGWF@bn;(H{Vr>DKu|Lnlj>`u7@55z*~X8`N?i@1yxupSiCO z{E4G(hKjUHw-}4h8@b$73XLj{D~pinqPQ?C6eC{Q&3vqV^71;2x-;B(FheNUm^^eK z*NQE~rRUhna`ud|c{qInE=={ON^DIK`R=ZVypJ;@RSxASuKg*S(_?^tsq}ZiExHXA zaD~G|^=CMxfCY9>2v`MyqZN?IdZVAR8&%cQ4xwXL*7_|>_nc_`ieU6%W!KUX5-~&g zSd{niRz!Y8z4?|%Wif}ZwrGOwG2YiA8gv_~X@ym!nqQ1`L`yg!y%hw$kor@)A4;?$`)ARJ$?Fu?-{zW7vxt{o>zj08r%dp% zh1}s1r}w1!@LMmpEH*kuX>M}n#f{TFiU0;#^!nY?RqpM5U(M(A1{SgzeHNdLgG@H+?WW`9La7+*hMkfVvdz`t9JCtFOk>tF51f{NxwT2!G*mGy4vQh~GEQ-%*T%`>GM~KX zwNxuh@#P`Zjxc!f+ozsrWdx&i{&f~)6|!fvWN1xRLdBuD#B zENYovgnes5()f`05J(Iko-?>4{?JAaLOne?9d2HQFykrq{Knk)z!koK2r?3`scCwy z&!N5MR4`h#S-jC15e*C71}q3915~1`Vf@Ob9sJIhiF?CXCCn83>#;`EVq6Rzj zPCeq3h15Sur+0ws7$j&kR+kcjK|2x%_~4Q`i&Km88ngWcLtii;jr#lVjfu=ejX5`; z)9q2<1z_CvDlIExf6O+5!3?=eIdyB_f&c@H5iY>I;%&iw9Mb57LnPSEjKubj8-fDS zA{3B6LX=Ntfg3@m2!_p03oWouj5yXU?tDcMrDx(IlY9;YN!B{uvi=xd&<89=Z~&9% z0?wqV$j+4($GzE{+zd>7#-X_QOtFCH4@iph#Pivwh(t!3$f5Nq4{HtI_p`hh;-;Fb zc_zDk)=xC|VZN~5*xSAny~@RPh7DIanD&{nVU?38VyheQPBv=s)u!Hu$0k#f$5@29 z=2c~DP9ttv@Q)WG6(T?77b(1d?V0b1ID{oHmM5o=h0rNOM9rou0v(NLw(yieblIqT z?H$s`>abg3%reBq=;t7&eD;a&)2GM9>x;u0Eiai0aM{I855`2~ukD%JpSPSg9)w{+ z03*pQo}k~bXE!Gbfsx|c^x27>0__MvdypZ0uUPU{Z-ZPH)?fCyA!J0srg+FSSr>;3 zatna>S5Wi?*ZTKz@B|(&-Ab9jeq{}n*s1L=)Q7}nwdV#vkar;gt#qRxR+#G541v-< zKll@9O4V68gT#<}Cxf?dD^7P@AzF`h zUxH=nmvg#I+;Ht@n=3WrOMSsKQ%sUxn3%@jIP9K6LCdPa6)jg5{VYygr&8*os9~=B_0{C zx3aMc8%k1c^M<33}xhs-0Q) z@%)n|bo-XUXHp*=@o)Amo4r2hl0$J*=&%R;oZHySiO!q6WaIRR zK6pVM-q+D(JdtDE+}4pGK@Cyds&Vzv_nVrQYOgQnOhO7 za>D@pxvUwrx^90%9U!mPc`+5aznTdPbe8jhE)J+xxA=AzVt)+YXyYwt;q;`a8BZ4V zW;r~n8WjuF1@a5%Bx9TwWnlO+w@!^)H|p%!?v&fH$48U>mu7UCL)`yb>1j{~G`z3T z^{yNIjamOTvVUNK;qN2>O+)`S!!$kBLqOC{R(|6!9pUc{|KK9(rAJLv1XzP=GmKCv zhS?ldblhythTNcXcdKQAScTkwCSXq|SI}?7>n?!deC-V0Jz#gsZmjZt!p%!Z3Q!0- zp05;-^^mEqh?O_b=+!aN`|G1;mzF!&eHJs@_(~CZ3Q58YT^2(8*heFAaQ(YSX%t76V^9Ee9Fk-ZO*5jA(9}d28Fb;VEnTR29F9d4N1T&^$Z7F$uJ+R-g&o;lB&ob|&r3Z{_9P4Z82R_$}Q26^_P41mCXHy%wh%YAcOY znBktmLX3CCh6I{PubP-VGf7^*%Sfc0>w$u=ifD5Y|II}r8z<av(c8g{bSM z&YiahK>`bioEA1fbaE6HTs|*xU2kptA0=lb%&bZEYc^0;p`s z_t=>!n673vd32BxNWGUBqs&X{dUSHU#_xDeveeG>nzXtYg`+DMR2DOcY;ujW7p?HV zX`C2$o4kX1Y4xI8K$A%DR* z^0cM^l5p>NPyAmY>L2hyOpFAet7Dw;zW%=Yu&5q!;7lW)WXoyQ}&6nsP2s8|azBtZT6~0r4-i(U~CdLp}crxTA7RFdSuoCwrHWaPEB#9A#7I-z)Rs7?<3{MZf76)%>yc0tI4Nq)z) zMOHs3RQM6(@2J3kU8Ye?Z@;CKU$ol{%x3Q+DdUeCj&YJ3@Z;Bvi7oXKy?}2WW%&P%C3qHdC`k0UkrIpT2 zd-_52wY7ZqBp%tS3cs9Yo`3!Mk5vT20$35ED$zT?#gL)mSD};WGp*$R<$W(9!-(`E z2%Z(@w)#Vx>7XsRwOkxvcdgC0`~7{T!C(W{0ZFF5b$Qw}RN7-Xjk}ZM*AIshy#+Cw zXHkE0=rO}#;tAoHYYGu=OeV9#dm=6}{I}IABS!U%#LMhVJ#RMf65U}-e(LXKocAfLn7rDTusJ_@pIwv5 z&_J>lWn+@RZD>fC%>=-=Ya;+sGjd~0T0^zB@`+3ujIXa(&&W?&QJ5>N7u(nbdfXM> zahMy|X{nK6-fM5tGwU{*RG`*>=2NRQIK1@2V#hG5apmBvosJw~H-vP{?31^j zy%%)EWTMY$vDM%;=~MhISMOijdp84;Sr9Z!n!_2S@ackMSO1^t0mhXqU~0e}W~5Nc z!`n4C2WWakj@gAs)ud3&LfW$eqBeeP79V?keWmd!&F{s<#QPJspL>_IM#rnUeVfZq zn?PdcBa41K)(e&~Mj&NTWwl0mbD zsP{gq;)`cpc@#F-6}#Hp;(Ag`6GD^x^d{<>%v9=-Lcm$SYdGVi?ukjw_c^fk7yXvd zjgj}0+R5{c-*!ZWe6QbT=)gL=qVSkhcHROxEJ5APtON={2$Bx9QB$DivK=O0B?+f- zhQQBSaG`>_jjOoY$WW4=%AwSPTr_p~PJ`F_2!*>^v-&^gs`uirt9iiEej?|Mf`31o zRWk1Ag5kOt-wbKUpx%(fq=DJFg?LhOMQ(A7z=99md!4~@^ztN?vL2*PU(T(JKQqi; z<*fRKh-LI~`K9RjPo!@EZ4avgPnc1M{;knU)CI_O&>=w0IvCVYrBH0`dG*Rx=Rs8JaJQiuW`U;q~Gc#J= zbTeIL3j9p^>X~zFdz(_vT-_Ul0wIffa-cq=LZ|oGY05SY!BOII);~67T0K<%$#7Gg zt}x)`P-ixzqxR&FlA+(SJ=?-R@!(Fr-cUG})$Ak)Q%%w59%M{!G0v|=6-;Nszi z6$TQa>B#Hb5NWwhsH-OBeGZMsj)*C?J5>= z^Ii;Z?J$=0W+Kr>mPL0v%gU@U(`Lrnim%NGXeWEmcnlZ5Ia1R(PZ$+eWGA=2g?uW#}xkT zK*)r_gS?*N+FPt_6xw$tCo)wmMe|D&nig+JGCGr`!bp3|MW2(guZnY3ju0b25Dk+G~$+?_@ z0$iYHU6`m_Bysk;hA3e`&*Q_fsA+$9zH^ZsUto5WAV~VNm6BKjc*A#|6>7+*#?!on z3*L;Pxvd-q5du!xd2-XWIxVaV>X_g*8a`v#yMVecf%iW@9=x3o_SXcol7$W1vetuCzXIEbJU9GbX0f_k11A$J!?w)YZdL(V6hypSHO!CR? z(|NCNXvE{xS*Of7XUEzUPh4rT_AgIu2o;6>>@4Lj=;*!4`;>QCIviz+nr> zXLi~(BLtjIl*CmmA(pxp67o&CS9rI-3f2XFa2ojxJ?_o4>1!SXqi+`vfx1}Bw`*C@l zFUSJsvQ5RzV$B)vsOW|W^*UA*9K`wUCLT|r#vZJULaJ>z^-68W^YZGQzng}jE`d-E zmzJ$gmZIumBVNU*&7>SA?$bknQ|v$DX;NU3X$tOSBlI>xC3|z{nF#eUAr1j7$1tTd z$o;NP3t8>dfTkWB{T)hTnG!vRfbq8U&`vd!Q?dB7r`18I6&o^4b+;M+Lbn_t_Y*L* zn5@w`I$6a;WDa7Qnr|}LNDP>xm5O~HJNf42yM~nGcB{_4zJJr?@v2l5p<)6yz&Jmz z5c0IwNz`C^(6C#!Te;IWYdDz$Y4U}xXTk}13JuEXMkamlikGB~ralawQPe#U>$RKy z!kDjgl}!r=dgvTq99cBvGb_fLSw+5o*3Rp#R7xF87dx6O2k|6NrqgQ)L4bUAO0v=v zl7%C>ftzq7a)5e+%=>twuv>RPA(=((6JaL$H57j&!lJT>F|fJb&hxocfNBX>M>p<} zQcHDzgX!LF82YhRCOV1qVpPDHoaxm^PG+T%^=q~U?gIgZPvK4$o)acQM~3mvfy|^3P->H|$(Ln; zj;04s?!QT-O=29_VA_5l_b%KvKsO4_+43K{6#VZPsMnlx(a*zy;$wMY`w5g&t7>uD z8yUoh8BUKJow?Y0PW*h1Q4Al?lbeg&Y z7Cuyj{-#3JHt!%W`6f1xh-_XCRTizK6Mob!zv)XK0`~e(w8{zGvD_Bt^h$@%&b)Ot zzojI=*;Dwm18afZjIiXYLhY@o#>eM14|a?c81A)EgJTqwhk+GLf1ycbc)q^@_AK!u_}EBV|}VDrURpV_wwYkw9*C2@*^z$)_at6 zDa*F!jWexIM}zuj-CNli#%PMI=e|0uK)rdVJ~0(r<-hbf?&c?m;$L8qGSx0qpYzZj z_%QnI@cUi$IpTDS7IUqy`dr0qvG?he<)c>Ju(q2Y&(Lh=4N8x>r>BeOZK~Yi$SsXM zw#IhAAh557iRI{$@D9&&Nl$g`WlY=I~srt34?Yu}q`f$y6Q8^$Fv zy87FGuV}myLzJMpO+q~<9J6j^;vEj(%ReYQu?Sr`-z42RQHEp|%2={4&??H~jJp0k zsiDUQ;&g&7vlE>zG309tn$_OSPsZUlSOEPg7#(#l3qkDx-bzyUS!XDG0y-);ZYr~A zXV}0sKWieq819H}p^@vEpt<=*lf)*=b9vg;`@_g5k3w?nl-D35*QEpGTc^C4v6PSIGzECC@Zn8T1N`7hA;A36)o`Uc=r!eqy6_k?w=_ zZr4v@373&RQhlc|K;+fiW0bbi7BF{K(0$h`y;Xl5(&$w1!O}R;t@ZM(x6Kn#M)(Q7 z?G;WxXiikDS)l9`NAv?N+)N6C3GAh%f->9na2ByTOLyeLcqSg*J`xGW?JTWcrMhxI zSk8IGPC3p0oXJ(}bMt`_IVxf`_m)MpsU~@5=Y}it?j4z5na)r>A^dxM%%_~LpcF|( z%f9N`c@XjE!SIx->0H2aCAu5`%s`;LV40~TE!fn7xrd8xRZ)GpwF`RFE*OEe(aYTJ zaJ$TY2L9nA;bQo4?ogD~n6b^hT4JHkGLpRuA!e+CEgXg9ejtJN}0?od>6XH|#pj~(-TS_wJUN^?Iw~lut5PiA}q7);? zQ=YYy(W9LRe?_;3cIFoYeKEcetBvr2nBJa9b8c_Ff!vG82iy}&ItvQ ziVD6p`p3Kw{53DG(N^}OL{adqK}SBtDn4pjNa`)E zOtjG}AmnNlqo(UHd~f8dlrES!arH52a=+vOof|Zt78B10x~z?#5!as0r*yxt5VUa5 zB4x0ReY$KhVG)75Lrf}2GVRPvk5AH|1^M1G_wB8O|64_OW&@*>Ji{0Fl&nKX&iq~Tz+QM9EipGlF>@KJ->{mNg? zl@6P}fZ8&{^TvMaZkP7JqUyf!g7*^sfr^{m;ruBKmrC`4qwJfFWNkYs)Y{rsIjVno z0R%wBqTslmc+{YYA@#`BK&c;GtM|LW^i+63e@#LTw5sod@J5%DBhZywh|1`)UcBke=bU8zjwcV$2ke3ws-~P z8yC0MBX5n>vR1@~@^EW0iY-Su&3h}HtBWiulh7JKhfeJk4?e2mWNV9an_noSm}{dQ ztX_Z1pqU5npD-NK2%cu2`QZQ}oMOd2D9)GIW~SaG<_F$AV?9dKr(pj4^aLEN`=d9y zbO#^&dKz2+as&napOvoPbzpC)V3anNw=+?w#&pOuqu=Dq4`V^Zn9Yw+oZF~WglipF zU8(@5QF+*{*nk@yE&aK_{G(PiUg(n~oDcq}oZU9g>UHVy$ha{5ZmlTES2BXB$kIVS z5LhV!&f7^5p06Hi(N6J8-3nsO%~o6gxB;w1^UZDRjMo}9F>bzG7_$xlJR3S09P@GM zkC|NCYBg8T22~OV)v&w{44QC!2CtCiuTAq}FGjod9`ezxjkJ0tB3*_&r7Le*zb+pA z%#3D=UrL~X8L4-~^oDf8>MqcfC#U2A#kxZ^^l4?cBE|=N{WyVVI;u0GlZl429EYO@ z^_`D!*8!#~EUD1(G?(?BgfIt%~r1Z+WVVSvH0J=N5e7M;vsFlj0x1 zTVnPC4z#4B9yaef zzMO(H98TSS{PZ&%Muu(R(gVj>{i?hK^oPV%j4?+W8mn@I8W+c6zE0qGaWf`Hf_O&X zCKuf0tAP&|WOu*Hf40}@T4{=v8T0D8sZ$?q$tIjP1A5Af5&4O~P3-FsD&!@j2%%AS zF18$EzU1nZYU(Y`&ANVdf$A1E_=Bwg_CbL@3}Vb0P2Z$$z0BQ_A?R;$Y@?d#+H?g4pEevRGYEZloYHFGGk0 zTq&3La+k`J?e9~~?nPdZEEfdWq0O}bEr^2Jc#9^v2X3x!Phoi%M!pyQ+-KlH>JCf4 z*5-x@ySyP^?ewPgM*??RrD=1yy39}Yxw?!5rv9z93jbx(*mVhX25s>{-CN~eYA%Kn zCrh-W@)vt=A0GEW*QeqralJ1N6WbZcBGsQt^2MHXr7f2pJ|aV3YBM(CCSe;LNaz<|gk~+;li$#u&!YuRbj;6& z54DS74;bvDh_4M%tWFk09lL2mpMRl@4x!t$LL7c);B7XYj97fY>bzvBKr~yt@b2ER zpqHO1{-XsZib(W^7)*%SKAq-XL8Y1zHl86jf^XfezfrQG++lub z58Au1ReD^Tz04svkC;b%++!d>DXv{@0QkWbdj9x zYwM5`f!$3a`9eQBqXt6j59p|ZXh==7-&S+T6AumIPpk$=bUH7P>YS2wi3H0G-jXk2 zmIzCI;j~dIU|!4^ZuLcWluz3mq%MRl42(lJ%6v^=g_z9YJTF@gOp5L0)vni<2)z1A zk}9D}?J!Z^?T~@dx2y|yfz=#i*(l5bpc=)^*WZ$eDK@0Y2(CD(U8#D9CU3O1$h5H# zsqtc@CjgCFjdF)3}Zv5ICOg`wS*@r zwz_GT8MiRM$i28}*#P1Gg~D`lA}_arx?rNmBl0V8^!PvS5`uD_vPdcDFlgUKS56dC z5|z_`ZsrbGlYH{3N%A7pr`Uz{mRx2>Y0z(p3{*Bnv9ghs3=Hs9Mp0gk`;_hcmV%;BA(%*r4xtapGBnh<4`eQbR=p;|Ko)3YZ zWduZbRehnSmHzbx<9_~$^bPSIV{HLPBgzl*0Z;7Q88VrJfbx|PXYPS?j`F4U!`^8I z)j&4hTbv1!beYNbrBbYp<&SP2iBoqAmUmIlQL#N=-l^l-LbW(L^*V>s^(H$B8&$HN7Co4iis!C`g|G24rryXSmb(ShU&*S{?eqC z7EHlb?!+3B?@7PR5EI7&le%bL{^SZNwtP<7Z$m>3b#w4d&VZAt&|WR^mI^?eUHw$Y zPY&lRbsFc8SLp%aQ1%)6=2pm)2 zJ#d*t2;pu54Vz-ErWwWN4V508W;TV2pOs$$GhXdC0p{8ARytCl%#JB)q*D7U-jMt2 zi>cu7i?|Y2ndmA3Rz_ShPkNs(W|<8+zQ2@t!mB&ad|ubk42^K-L<+wlFA<=XGL!1} zC|#1ir8Q3$c1|muH}F?IIw2Af=P9uM8)DUEL{)PfJ^=78MFrghNJ;yJ^gA4^7m;`kk zX|09qKD(MXayhe!gklo*re&`P6?Yg^A89$_NYGfB(kIR=xI1Qc!hrwyrzz4gQIGTTyumAdr z=GXqJ*op^7Ansc2foAM-mNxNHKmPKr&$+y&gj~QB|aIZ)Qfo3x&X!x71#_jr%CKr&m6I zFH$4w3=VpB*W$a)EP~z2Z=lm$Tf{zoYOOJ>Lp@u57iLbix-oO)^GXC*qQ!@WxeK&H zOqMj#vwb|Pg?nA`i+mo9k-G0Lqg_*7QLqjOE~_hEw;+eLGc%GgQ)VRv2*=SRLj5_f zaVdf@6BJoJUv@iv#tiXqIId0f3E%b{)fT|_tDw6gwVgYCV2$?HnvDgkTc|Jk2q9BU z&i0i<296!kQGvr2wE8%3tpn;5+u(re<1HHE8*$?%P1Su(n~%?ooVj%3tAi&LNZp36 zoonBIeDN8FYBx!aVV~hnl#r;0*K%|U_cO0Wd=sy~`$wd~WbQV+@weaXofiOqXp&pN zyNvk1n_lr?TDa(Z&z@~-=}StQ{d&0JH5Vifs&C8@mCheTb# za%!==#qGxzfy$+*nG6>xO-75+T+?bcQNyv;$Nj2Vf-Gn=<_|ZwE7lq{!ZAN{+vL)2 zlvLeBg!{d#H5Lb^bKx(q0z1z#T?DGdbH#r)N-jtR@P^$i5Ktw6AYU(DWpaH;{h+=P zU2GjV@vc+a>--mAwfz>!TPBQ8$%dFS6(j2^v3sEe{OLGd&^=b_wvr=ho_F5bq>M4H zuh*l7a;SuTMYFR8xJ!`L95l=4p#pmtl?{?UVT0Jh#Zpb_1gv1O+&A^%Ig*5n6#VuI z%V})tKX17f&Igj4F0K{;TMA;GonKb_1P`COmzQqxjk3HLg$K)BUuS>>(w_#Wd0CNl zk0i4i4vtTNhotXMxpQl{KAYfdWGYu1za`bxQ`Y=4BQGE~Qs82`l$lZfV&{;qKuDpUDNT^ z7q(A-iXR-Etlkp9Hkurmkz+s(aZ3*U7$+xXOARTcFr6-UVT&kp*5!AsY~{kt6z5bD zH|17`kOFSTV&AXrms~#nTsl}S9tYKwT11|P*4Jv*#BOXK5__9hZ_%ZWO!(r-ADm!p zgyA3C)!lJ060O0@WIht{6u!D1R)@h9bLl~xI`xW1;(oQT{O`ey4o^BuUYA+x0?W*1 zT!awDyZ7QfL>7f%i=oOW+kRwBE=hKx@i;9o`Z(gVaxlq_90I zlLr&|lfnfAC14Bh1DEoyISbVK7rm7c^pHTJdkAnMv~sLSRm6$*H-7c$iN;~E0K{!1 zi7kWay>vhX4QDxF2%!LDmQ0X5#t;GHM6Buyy;g6=ZXXRdEm#kERjmm@NX>ij_1IJj zr@yg9fKE!l4O|W`oGZI_#9AP}7-JRmk~bUXp*q}c5(oa`mmN@X#LQhm9D|?css^g%l8O!bGhzudKcH zKXsmT6-H+PT35~U3#X`2ZA=%cQOOHIhzRmztZ+5)(Ru27t}(oedjtV^2CBZ4!5;VN zr~#iNk!xrc`J*)0vs%EZS2|b<3sH6X_C2%e`Q{cw@n!MKpT+b#*cL95<7+=-%O{c8 zQ9ZY;R-}&h@~J2guqt@jQIhot#aRUUx28cW zIqH5F;=KsVALz>vY?Np3f@U2$PHmYYgX6E~?k#59=pMRFh4GcbIEDT5M+iU&Pboc}5}6f-um>_hGDJx5agceN3!gCOLik`QB&F+Zh$YO;jT>AL>D| zGS87e9)fRKghU>-?meF}(}x^v1`BPrOj}vQBcCcv08qbz%AM;iI(lsxZZ>%r@0HW93d@5) zBA^!Tpqr~I(>^sKyFyeV`~J$q!4=5T)wy{zCEbmIrC%ZJqN7N?QGyxGm6-!s{F10V0tv?*?9~wt0-hJJf zMB`e{4j@9DPQuQtQM7ywwXz;%v2*pfE z&a$03T1pen$L2tiFRkf+W-YLMYOr5<=6i?hxg%(cVskk3avE){NK~8 zb@*l1;y$?u)^9sZs=aHG^kvMwI}_4m3klm9O?nH$987(Fi94XtXe<_f3IddoXoVd& zlV_Nu?`ky#zoHQ0W{|-LA60%lHCM2}HXxlixBXi{&chjiCHOmM6IUev#-(+mA+zHT zIGxWNAyU4N0BoNrOB((eK=O9@Hd=G=LPM;Z;E+z89I)k>p4jLZCbhN*5WT`eeXoj> zgxcRi8g$dbgyrp&K7($eMLR*k_PrZ)Q$^a&hjES)rxTH>-38jJ@ZW!hLGars1=9}S zW)k0&>$E8^YJ@7@Bxv)0L?FEZ}?s*vssIY^kphTiwzx#6lZrdnfs|dgH--`F9*G%rv@RfQ!=2nVWc8&`=&;&P(?na+GjpTgFO({nRg|l_1n~-wW-u?`HUXG+fczEiFaMza{7}8iK+6T6v$` zJ7@QkO$ir$dpE=4M;NdMs#~=TapcMEJsvQ1s?PSJrn8t~O{+6K-h!ly))J2eQsSf| z;g)vUkZ)xD6pl*4Ot-pKE&J)n1Y%H*dvr6}B*lQq(|U8sxc*21YWU7ut!*FFLQX5+DdNess2%)>Ex6(BVjbWn)D(pl+ zYhZ{b$#OXUOwQYd{MGVro8nfZ@2$g2@tOdbjjm>wePYO*uAJX);WhRU^W)hNNAdj` zw+;li0ki5Q=H*KlV}~qx5|%!9Hbxve+kVybyl_^7;&i-EM1)%pU2#H$5BaSc*>zj> z@!!-{<5zUXHn^`L*ERR=bx;(&Eqn|Z{@IQhMQ=}mjSkC)5}6ZjYY_1-0&z~ZhCG2O z`kz}KTHD_c@|W2UDuRB0M30J$VF}_{t2e9Yfpyn3xJKWl-aL#GW8BNEXmva>93wy> zxce^OCBk!2xaOMajiJ33N6C`oZbfA(71+~vTf=J+rWz#b5@_WR0>n%HdX*J)k)9Lo zuniNOGt&@6faH}-V|z_hp?CgK>>)a;wpl;++%W^38z&c; zc0EITKVsv|7K6d;XT%jjA7E`X!AJ0g`{Cj80_RrYAu*=5Y=Q=p((43vg|BUaCv51) z7Er-9`dh<}67MQVsFqkSRL7s=4V@BY(k3wf7-c0MiMTw(iFu>lW)or2uEzAHUiJNY z7VN98=TI_TnI3>-6ZqZz0#Zf=l%AsTD`6XlPTuh5yOTB`N%XLgo z-@6N?=?XRcLl0K`L~ZpM3315FFJ;*^FddV;+*j9F_ACrsRpH*UgG_rvB+B>eD;`=Z zNOz+HeXj|lgYq!G20D0=vf4Vx5UCZ!R8)o81 z<59G?F$}`nFmsYwUvc!y(WfF!79y%2inu2;n_KM(5}fGOmL!(AxcZuTsX_DL4Gn$! zXH*Gtdhzbd`A|Q7QsxI-CInv|PI~`J{yh6^_h&loLiOG2)l{~(>Z~c8CDJF_Stw6m ze9)GAg?Hp|J)FbY@WSq$)@q5qT&R~&k@QUrY>n~58d!Q|fH6hoyXFh>9=`3Nte)eC z-q7u#w+9y%4mO!*=#c9glu>J(@7>iqHq+*d)KtHYBfiA+ZDVT6cdoMODfY+SicQR2 z$*Xjy$4ir0PBe&CVNUFh=`qsOo#a7H)Qwk5#Y;O>QAEnmoK%==&uoP3|Ox$ito1ng%Y}1e8d}%t@sL+k0cJN_tnKM+AM=|m~A`ox#PMGNe za$lX*X$I9GPu%N%OVZ7eIQFDwz&HE`o{nnt}q{S5;s}*&F6k^G;v-kOTuR zS3c%md&Vd2j+dzyKRUj}@t&2iIXFPc8X>AnFD9EFb&^TF3qp15V_ZLhX`p+%i(9lF zd6W<8c)rzF zWjfu`JhmZ?5_@)HM1yf~BjR%-?+-EU_ZirL7{go3d;G5sS663tv|T})6-CT`pz}vJ zMn8Is-jup|Bad&uM-(Kc8Ro(Y$LGXi3oSwhf&`=Q53%WQ01&3QQ#oz^M1~as_*0`p z`5Hi74ma=-#0;o@bgM!$9h!f#S~LE{{whKeZ?pZwQ_b41K1+oeSMc`{MNTcUr8iv< z?%N>|yyEBn;);fKF#R*MfMJ!>i_1MzU0+tgCS6U?wBpF@VzLe;cayMS`K_aNh1Hv8ZB z=nxi^Y#26&$O`w_9}<}Bb&srO8&hMRT(*db%QPtNC>`WyxW$>hkXo+8zSm$))%0C+ zB{-l3NphX?Z_Aw^V<#+QOx>YYtfQlL>9ZR0Z@ym|$K4<{?7~2P9KW%;foWeFc(I8; zxbA4*O(H2EXY0@S)5QX0CFTrn%_-;x_Wsi*ruhIKQwYv4<;kd$sD|AFz%}uD8`wYp z!WM}?!BpB1_1!&2qq0UdgpZYAeatzP(b;nj*%fOoIL9oNr0r+L_6af-SQQjeHtt96 zKxGAsiva^zOu}@1yhbr)p9PjYg2aA%0D#vZg}-nl%E54~{*dp()))xdr`+9mv{fwj zW1#8D$dpiNc~`!n6&Z8-ea#LVzkrWci6-lb5rp8wND^-}Tr>!i(*v}0lBc`iBI(;* z-02{4q4I;ecV2Fg%45HU5Hi}9i|cz*b<0hH>PHz@KT&xi4AJgMW#16vR9yJQBuaS0 z)~cmj_AvVWrw^RwLz$B69Dab;4^{oZ>aGGgB5SeUw-8}4bM_{rSyGI^3_x1P8~UOj z87NEmEEhC=xbSH4>3h=TJwC9sx*~}oXRhjYw9E2p)JX_rPr0BlBxntYf0_Bu}>tOfUYYgWLp%RJ=Y#`r{?`GYJeC8?8 z*icosK68XRf_Y?y`>(|T*$H8%cTCSj0iczlrk8r0#o36{4I>D;t>cw9z*y05K!5d! zaC#9JZ>g&)Uil|^=8*?cTpuf12x(79rY}9-Er#I?Dxv9x86O5tMR$2x)vc>68Qet1 zXK((l(R?e{p~FWRHGK%lyX{$c>RvlJ&H1^B_Kby? z8dhmxR_8^R(enPbb&KWBOvt!zX1mCLxZy$AhpG->eyc%wuMk8`gbWlC%)|KNu`?^6 z@OUG1S&0X6H!FWSLuJ>*AOO@9Yxa1eJ6s{ zULxsqvfgL-QFJEkV!0Ne>`{^cTEb~(t%|#Bxp4YfVX2l=WWVr}>W!eSY&C?w=trLwvRyqHrI#S244=C|Gj3*Qn&rImz-J;k3~#kd3Y^3!aMIwbCW5zFP+ zZfUV};Co|(lerRFeBsi~E_-4AULb8L`B{?QHE0kYu&tmXi|9VAqd=qkV2pb>6Hgx+ z3|=dd$o#ccu6;6-Yd~=_v({Rts$Evme#&X1b+>!!$T$B}=W_cKPh%l6E1`r z*D__s&P%>tuKMv1yGYGe-Y43;qHsJjX}Vsf0qhuezbCL^=1`NST9nBT_xze8;JA1g zTgekxWv#)cm>Y!5viUl{yO*3o@J{P(vM{T7wht7FyXgjG_%_0icMrPSP9;%M68=gZbBPYl+~UawCgUq?Z3z6bw2d) z??KU_G9uyXdc+jf0C=WX7$!`V&RAA|Jl&QUK)6$<{hI}7^7~vX5iuU2f1j&KGLqB_ zy4pvh;hkDWK6i?5Xdjbsd_*NkwUwaBVayB^PHa0tJqC6o`~2B=5k(HeV6$TjWt2{? z*H${&`hw<*F-=}K8YOc|BfGc;+13ZuGtZB**zl=^HuZNZJG{sJ+H%-G5dH?*O6y4o zWAvrH{Ipjx%C!nhOY(losoP3P;g5?VELO?$T}d0I!+#@s8CSsRN;`!Xn)_u?nN@{i z#&u3*j-e10)eo5M>#yvIBkWW9M}7VM=`&SxQDU2%u8`pOH*ZxiM7+Vu!y3h9#?WD>C3ni8d?Am>mN^A;+N5#t7uQq7pzf$`^4UI#T&b=$0zJlf924vmioN_h&?!4 z_MAKQ4H5uV-g9d42|~D`{PwFnea3b~z~I6-(=!63-DS)ksB3`gyeI2$hgJ*lSCYWV zavh30BZ^Xdj=RIuqF#(Pvv$>A5*c5cr>K~mwLefV(|lV~lC*W#px%9G#b&JDw>(v= z(4h#-=&pQKvtKGX?G2y0ISsr-jsE+#I&dMUp&)Lt(xb6q5s{Gf-vxa@W zW2eOwf6DOud3r!?5)crF-|Ovh*rS})^c0F3u#!Pz+HnvUiOfrG9E*Lx(R;7D_=1$t zE_B9zLh*3Am{}HUXAUtN42nrV%U&#cYneYqYBbWA6E56)InaSkfY+a1q9%Of9#Hiu zqh5X|YUihYe8XF2IOTd3Xx_hJ8NJ!4fpbx5ZhQ&NBZh_Q$!y67%zGmU6(d9Y5sP-H zaYLbMjsilGp_Dabf}6gzNW^b5Ik-{$B+h7iC9zN0IllhjS3#gP@bc~-Lkyy*w&eqzGxS~;op{`9ex!DFV%OxDAl>MG;XeO zz#!7Dd;iTxcyQp`-h5Tk#hau4prI_l(r9FFr33eB;R|0mL)n)Zch6<%ofW5@L%x+T zOw=cTrS?Dc7GzQSqkpT9bcgSW{q#8gX1jUu4M&>dXOp2;*k)XNTYEAQiz@bJOHUFE zB3^J52TL~czi!G#ZM?g1aP-b_zOA^8zFs>%(<-7ZdRq{m)dCs*xp2vG2HWhz8Vcvg z5TIfH@D>`j))n#q4(8NpjWE<$XEpzngWRo+#{4uMl~mbs+~6rN?|E}nzC zPq+fKfdmyBQcMxphrge}jko|veo>9aa_3@TbIo9aZ>uVb5C##l{3_Z~)yN)x%DzKK z862cgaWt0MW*pgkdvU`~*XD5$b@$*r(S8pffqA@mLbS)DdOC6Lrdpg$vuWAQX4!v? z=A`Jk06AQpL`0ji{o_DeH&9!r#oe=^`(VFQ)NBZIV0moE+4KWDx8S{YQSyqmQ4K&N zk(U*FwlC3+2e!qrnv66%= zPHL0Df|5!>aJ<5Pjnz4pjcG3>uS>c&i_q1F+5Bud6!AN^Q&83Mg15vTOQ(#?(tn{DT}SRb;noQjG2T;229qAwngpXi*;Ui zj{#A*=QAam6T&&Q2t_=qTh%WwN|`u56Q*y<4*bgCwa)lhR@?v z3r(7!i4ujJ&_i{rGMB+Gg2DG=K^GKH%! z?>bzPavJW3R3T^aI%H>7`^})-79iZKO=pC{PiLk(ZoS;j*w42HoPY2OW2o=~NL(J1 z%@?<3!^-Ftu2U>8>J!@N^?FR1oaA+jlZPyZy2*LW=dE>EJMccGG|UTygs8Ztd)>(kF2*K=|F|=0}Q)`-J;9)Fjpyf6;sbL9!Bv z(2tk1c1PvqhAk$cdqXJ<(dHvxhODD0gi|%Z`AE6&G{+>@IVe8$J_jmnXlhqrTJoks z{o7~htksRUx!1IA$LTzv40|5q!KVvtdjYkG+V3)N{I%Z?r;lwcmK$^wL3l0XS(-FZ zI8~3AZ=;>|P7PMBuRdSS(pipWlV;A@A84=-)fD?JDkHW+G>f!TQ97t7NB({ffr~)q zc*dkO!kZjVm&`|(qa^5(`{B7`)Zs>24?z$1@Gx8a5rO*6KC<}4Y?{iEz@{!=IYHU~ zcq>QDcb0J3xrBfVt~1L{ZgIC=6;u^RB~E9k_(C;>zNzz@D6RC!&i zGs-4!>uQkuK_|KA*UV+O)XC%Y#sK5k@nCfPoyt4+lPwn2>5i0NN|{LTljUm99%B{& zIAI*t0SVh{_v|6`hgSZ})a-YbSs!ItH)xoJUN{~xr$XR&pEAl?hO5m|na*qpJx^Q9DlzZR=f-ozpfVTHK3!?ye2+4e z`}fNI4^lfV2I?>3Sa!LW9?YPn@ZTC9HV&7yG0x#32CXhSSi}`S!llyFQAgQyn{0W}pOXyalQ@b-4sgQE@C-rrW@Z_tj=!G{a%~%eL!&u*`Qv0{IK~9AkBMF(}dk>|(uTjB8=FJS#tTBe7 zLJD8rJ-gyrJrM?LvvKJgOe*Xkhr|h&7<`Lj0(n0<#59wQ+TI^}VLjiF3Ypw36aKz5 z_h2O}@%ld>F-C%?D40`T3I&9cM7IxI=?ljMSjMObQ(u<_l<79L#&*S`Hm8*ECF2)s z5QZ!UTJk5J(QljpMZnbo7p7XJM0UIY-@t!z=R87DTeZp-o?0l8rd1&_=-*nxEwY+b zN?vVJ7ILYUA6Huqv94HxDs!$#B3%mMr%5Y(EbTR4o1odiRR9_)B!r z5pj~`BgNDirH^Crwe4-q#l$sxF4Y>0jWEySv*+RA`{EzERSz9t4f^%}v}8Xf1hu=| z7Pkv+WyNK^y0(@dJxsaRcjZoTX?Rr^bukz8T~lxRX}f=j)UlkYA;NY`T(4|5;4Y7$ zKflcjMInii5aUD~&LbJ#aa(JKe`~)ExF5+pf^`v#`drDk-xPNoFP#z1IbBuySr^%p zoZwDhY(2lrOX%5XK^BFDm`X8hUQ@JqL$aSKmS^g?EPYoJodF z=yZ}dynCP;^thabD-CrV7Ec0k^y_UiYW`S22xb`!k{vPMD{q^%pu4pX^lAF8*5&i7 z{UkhyfOjkmR0KK>ZzJX_cIN#zxDj~ayOEG{(uEJDbYGOQ;8TlT-7<`i{!;GkJG)v4 z*S+!0OGNJ?ke<{&P?trcna}n`*GSAYD;QcdefCu9zH>@AE`tp5nfH4`H(VRN8Z?YK zU8E!1cc9E}bWL44zEUUWPtNNU@VfP0W=DO0$|0TUbnfnJH3?+$4a>ygjI39QTkr1K zqH_Sw!CKGSc09L)|x|+E5|!o+{cHv_jmNs%YhRaBZ-Q*(&30t z1{GuIzZGBpEUSO(L>v=VqWaxChMUU;w_jp5Yz;B_ue1ekk7tZn%mEzQTZ$5@Tm!#5 z*Srz4j2ES{fkp{{L|uo=U3G=hEAa3y;-f`(s$qt3u*Oh!weU3B`$DUYd)-XA7%0#8 zD20NU1GEycc8c>F+7P@_?F5ux#X7#fYYZ8UY*FqMegY9##>--kq=Ky zzkja)RH^U)CXm^UUV}rlv$8AVwasQBbPnC_LB|1!8tKaFw{e2FyQ--YCbdRm!&{3h zT7{hFpV$rBU$s-*M!O>G@L0qft_Uk4=)R$Dq(wa*W(kJ+eqYin)}X^E<8>Tnr`omL z2jy6Rmpi^2<&^tLaT;_u_B`Qj)_sOj;|=N=iTJ0v&p)CUa7WjbIwjXR)5YUNJV0tB z_ia*;0FMk|OU<>_;#$8clD1N!6zZFX9;+2PcZEy#{myG0 zV)poWG7m%!wt8bZ03%^%0!WEg8LgeTJ@e!{o^pWndrSU^T8bmZ(*N8P&0;-%(&^$# zNAhA-o7?Q>zG zlPQfRr<>PLMSTy2KS%_wy^i@W&%J-}p9$3joeU(c-km)-nB0q}ygmE9d{i&i@Kkk& zy55_!g{ao_yRxg=Xqn9)-9c`DwI9_lA0|DwUz!KqVX(h3@idmt6p;^ID|*wGSZdhv zM%Vu$l-5VppV(=@{+oX`ZFE|BDqsLQ1O)1RZJ_(2PVubqn?mVm;~Xv=&u%ELem`WD zLY_Au)%DK^_AhAVzhl}B-B&a}4@E>lYqeg3)mNgg^sz!BMRfak&p7Kj{Wj(mZ`^9YrUPn6 z{xxye%UiDg5QWvnevSF-^Jd@PeMjnRvqCL7P5=cYB|JwFo87@s+Y#A{VD|JSl#i+0ug6E2SLMiOW z`t`Xj{x?*TVgePeh_l;^?dhG!x9<-){F0w*OtWT40mb$kfp_6M0@(1~$0khsQ zI5}}F3jn7pC^2LL4z?*oCTDb7S!(3gEA3H?rAuH&;6<|v1-qA8e|ORpe;i9K zoD10+h}%+&%fd78Dlu=HbcNN~@WX0r{DuUr53AG?K-IY9n*Z+MIJYKkcZT;6Gx zuhJ@|Rm^ZtjgDH5d{Mw|!q|FHaV)t8dzw@@n%Z1ftkA^6RN8n9&i_H+@Tp8!q~|eg z1qcyF4O`karSb{AO1@Tpm@$-#a}E|5v+R0Rvro$OKW{dRoHi6`8f-nRE4?@ob}!?> zlEvPcLlIw1>tXeOz#KnY3YCNFR=)e4u>l(Wqf)&2R@XwYa*xR%!nZ0UnWFBJeOC}o zx^He+B5)HxoMT20iXRjGk97A(QT-oZ=>f?Dns&r1K<9?#I-`E(9DG8X$fUwfQ)kWo zU6EvxUU;)Aj*Ra$b_qrakI%F=jGVhR4p5tsUagqvHZkptu?X?5p_y$H@)a z9V4TepnFs;s(N~y>$WD}isRs|J^WaKNvQy*c^N*H`m1FPqFdSFdQ|UnpBI|q&=|Jo=rYPdaA!mWI zTGxPJ$yB6=x6guMhaAvMd|&#BuY5{7jbGlg@a1K(UEjq3D-zcISjZo4CjXr{|EPut zRxxF~u5Kb@zdHldF5+yi7Z|G@?Ec(_G4{9-mt)bCzKl_Z3*R-RP^zq@I| z5I+!z;7?WWtx_9WFExHjBn-={+>$&VY?XRpgE6n)-)W>Ur zQ}jsJseFY4f^5vQOiZ-LEN<*J)ktS6SuJ98LC4p#Umc9;im~2aF)y`2@vnB|A_-r1 zw15dr8`Wwm8V>n4v9HBLFni$6d$4g-a|HjGA_tb?r8AqKW3J{;zQW+ z*@+DptETz*6?0FFTRy#?BOT@3Ui>7S$gY5L{Gnx~Ay`DWrT18ODZNcjv6e@p)b~%> zczN19LIMBt;@-HDJj$oIqb!4Ui^V7}_|}9by4v9#p@$oJ;ybG#!`b#{tr|_Sy~jeG zmUB<$YQyfqr=m1e8w)j-tQQZ;jY5yX7wT!yi!Y9x!&jmUug_W?Y)lhS>2Z*iWR=rt z+Z(Y$H)BylS`>*DgkL<_P;+8fBm% zs~4V}|L1@Af4|>fM4-P;?vL~OzvuVApUeOM=Ow`1qnUga`@bLckQag98pemil)TLm zpOTz*HdH*po1T@g9|OJL&}6$I*1!#+&tKxbd;m(1pTCj0OrnhWFnz(BCgy0DUKcs1NSn z$^FlZPPD*W@4~Rg#)JVCYD8H3q9WrjEr5S?6@edBk+>n!LhNxY6CYn^W{_}Wium1YKR0Zq+_e`g zHPlLf(D#>udf=Q)0QZd%kw^A_B{y1ILJ6srPVHHNqVJlA>OmZ~V@0lFOG#pSZCe}o znHdW1q9MdF_WonpoNevod2;F8o~7GkDWYWG&=3>f5L9Uw8i&IxEs%VrLe-rTPT`0C z69i;*wG~I;6X9y#J!(O}`+N~!u=1}$B?iyS?PMUZdHGzc=7_yDJ5EK!X2ny?IRN{n zXH0Tn_QaPetZ@Sp1D1#P1Jka*%Xnb1#t673B<@eH;Z2hr|8-9*f`P-Clg!%t+i&fU zgTbgl1+NJloY!6dTzQ8LMR&+1gf+P;M#J7CO?UmAoo-orxVwDXZ|vKN`g>leRbeNRPktT$5r)u?)bi|j&m|f2Hp5A9uudN?U(m!*Vreu>89~o z4%7Yy9=Td;Tu$}>)huGL0i5;l6qc@|f81I?;E)KlT%`42(sf#KYpV`F4kq{#U^}vO zKxz@S z(94lH2K0`48%%7L=Z3ZMuH-3z%RJGh*z>2gPV?bYwtlMAz##QM6DN? z`Ue_hXzKmJgaStf|5eX}CahzCS$Qpfc2mUoI#~_iN-!C;_87r`J zJs$Ma_>r*#TJe=&3jQA)Me2-cU#@lco9TM|k{@1M4=*{$)!cpM+MUCk73+mf zGCb6Qeph^)0)RJ^iXdVWCh5Bz+VokAH1U~&)jGav>B46YHvWSFhQFFE%xq}Rq0x_@ z(*jVTv{@SWU)LjAA7WaS7qMDt0O#x2PPkJv8Cj?XumG?`8wlha!J&N2f~76l~j^egVxfG ze9+*?YgZqw243gMV7lDTFi!6y=I>}KmRhU zzl+P5-$rL-o&>Ep|C|s7R;qMTCk|l1cB4*my<+&q7nK)krMJUF3rK+#PY;zyreph~ zZ+qwm_9SI?sZjBmC+3A+TVmKUCfbwob_hR-uVPu8MD&pkSJ02ZFFvrdgIG zwuX0|g!dRCs(Fs0yB z_Nq`SvU;X;LZLABG~RtR@;R{NN^1=rKFpRsjAE?m2k5fT<=wDbe82&C6ugcX%4`ai z2k4wG+d3afS>S_{W;_7Sj_Tn6WiUs!(a@%>xp7A}pjOM{ozyY6Gcnmi71G8%J1o0T zH8or7C`u(w13h|q|4sTPy`D-KlAia%Mt$7Q2L|}RDM81K4%(G8v}s!a8z8Cn{bfy6 z`+TL0MLQlN#@jwrB<3f1ve6{(ofziNLjRh9eMJ{v`ycZUi~yw8D-*DJb2(lh2yiFz z^|;{ISu#qa+dt}Uv8H6WXb}=)JVdz|bQ42wJx;KiPHBv5$8c0Hkz_-6fQvLa%ZIz! zU8%XBttJ6mS7U z{j|$wPDn{F;5#Do{_F95cPojqE6n$NTYq~Q{J*7Sk8c>}pRyK>Oja>&&>e0WXF~@a zavoP=!$5~}9TU~{Ikm~e(UwCKfo)}Pf5Xk^a{bg06=hx#^#(21GbXlbpV=&52D04z z3gvatzx?$*>Y$zAdq&-zGO!u6rCK_P9xqKlPe+x&ATpB=T7)jqd6802NTXj*?hy82 z9UW-{>&gRG0JB+NZg*WRsOhVQ1?-l7&3%aF&}$BfxoOW~*L(4HjTF@gV^RLg;aL2# zD-i3O$xBGuX_mE85wOBi13}fA-~nOV?foh*rU%16cug!ivAQeSpx)Tv^o`l-XTHR_X{h;pyu*v6u0tJ)Knnfh z8SCO6Gv^sQtS-H97e3KY}y>nFkGr&$XYAg%^s;_<5XU zfJFMCG)c(mTo@cmUl^;_%}8b6epjReHM`gAIFk$}AJn+=B^JT2_hi_%0R&Z(tNLV^ zig+U5(aQsYtE7Va4@?ZcfDVZH@p#YniWANLmayBcZoRE`UfkJ&_tza5d}~mW5KsbL zVR`?P>`v&I8!)eqJGxlRte8oH0oYtUDYxI4l*hy*gCipY6&yyZ`oC%&4R(fB-wjN; z^RLz(h$Z0=r0*#sM(JKWVrKX~Qv#y>T)4wR^OWK$R+k?DW^49KAiSut9NFA^ypG!w z0vIgy!D`XpA$NJfj}`DX;<-MY-}gBmEe{F?RPy#WvyJXj_5UW+sPcyc3`5ernoVs^6@G5*Tz+4fB9kJx^BcZd%#au4NG#v1r7R2;H z4Y+`SZ7GS{@@M?>PDj9v9p<;(e8M&+9l*c4#)@>MU^4RT<5B9F-BLpT;R8f1N z1Vm{_6F4@Tt+U5{H-rUhVD^C6=2== z+4O5YM(+0TarR*vr%_!qEqOI}#56#s;A_EK3c=hOVXxDr<06^gGus0nFVrBZhBA)! z(=Xt^OtH2(Gb^c&;-VPK4o;5YC9n`AX!oAroBqw_U2hudReumF-YsI ztX3cBD>j)YTl0D=lcm$TlK9y7_)Je^iLtNgl9Wkl7pMLkWi;4)*wJ*r&9?T$WYeq$ zi_-!p$p~13y-|pk8+=vAYJS=(dq~yDTr2^;~>f=*eCeqT6hN&Yx}l`pyP6FbHc9>pmw7jxhKk zVORrItS_`vc=vWus-{s$(21~-7h@!4wSrScuRbYf0A8-eL0cR9dHbU~a|kaE}T$3rVDb#+3aPF;*+eT?y+F`V3of!4d>i|GIu zm&)Qo@as8%XxF6CLX!(zb5|eMF!H8iE(5{G-c>;^cZZ5Hxaew?%UE;&L`_O)Gptkb zS+8%wL+diuum9`qikGk=k}%aLiX8}bU7Nawqw`f9d^h!9$t1NEUB z3MI_sF-_?FXa+&rBVGj43H)Ymn3R$e*fwEPpLO+81lIq#pE|@Q0 zv5_9jdFWS%?yLEp%jx^CQ6tXbuPl}Y+6|9bmY1E?#wq6af-koRQ*~HuUy*N$+%#81 zAvD_KfTb{FP{>&SuSG(P9O1o=PXEK&Gw%R}ADvIemH>3aufuP#cM~h#V5RR^77v7R zpnu+Yi64ZR#vH$*5yK`t4A*8<~-Oi8ZU(&4^6*V-xd-G+L$ zUV3%5EA=CUslWTLOH0HC3**SBUOT4I$-oz#L)pYP+{d0TC$?0qIg{kdPWu1WA#Sk`n1|7(lu~N*V>Jp`~kRq>=8S zyPFvpzQd!>`@a6ZwdRju&El+c&b{xwuWRpX-#gu6a2}o3y{TE0I<>*ItTmw-3LH{W zjRqXQ5YMLD`rx$c)#BASaJGWbcdc0W@uo%}6117B{c*J~45jN&#U>e-BVd`RHF82jY*27*Ybnqk|?Ih3f5BlT%-K=&|cGtmO7?a4nnM zLMaDI!!zLxiC!;VS z?h&GK_*(${f<=_`2Ye(W7VV6Vo=Q_l6>c+kuYRin371IMnJT(A<5}x*BHiZ`q_J)) zySw^2<=ecWbk8|`=fCB^An?Qn3kL`REX~piMTs%PpVn)X_|O^)!W4{oO{_1BEh~K z2;}6^av;4WRB{o5{glaoC@hmW<{J?(zFI*S**Kig98t4Mtb?5wLythJJ|McHK8=xS4KhNyk-A>@RbY`sOQi3TA;>0q^?T5sdjnjWB=fsIwRLjh3*v*w*SKC1#8Sui&u6Z zzdm4_e#p{+>pdG?*43XD3J5-oo9@K58EB~zCZ9Eu;^z7;T z$Fx&I+HDUb&gdlKu2L+TAXsP4$&0! zhS{^z`zWbyuTw~8u74E=9|u6HT)Sm@dh3NLbw} zdOKSZ+g%#E4(@_4Ma|407`(ndq!Z+e2moCg^x`|wZG#az`SY&FRsr!t#2q%*k@fqz zXGGM8wn`(DLz(lQ2rsA8%C9zHea!%6&(X5nnvt5VBPb*^^llmwQJc$xT}cqHB(m&T zSy`h+wsxt=!NM|9Y{BW~Q#$Ay1VKNe_?|#=gFGA?uM>;3St$3oJo}T^H`KW_w*l$o?ka7YuXQf_#q&3|sddO4A&2Dfi!t?- z?f7GXo3g>CJV3V)T<=oSeJDR&JXvd+{;kvNxn4cQHzj@7WnO0Y>4fW?Qy}Cp!?QTu z3u(e^Yg7p;_Vx&brn`Tew#bllBJ~zMwLD!`0ztew6cp9ezPrPdNXN{M+{zYrc#unH z;8MQFhmIWz)$>Il{j3{abzWZHkPYwfclIHuTOxK)har$O^B7&qJFZu4Rt9E80`Iv~L4Q8k zQTxnyD4g`Fc?Qpkwn87=?Y7lL?LE#?Lwls})~~>g(eQ1teL2z=(h}ovXfh#K_Sns$ zuqqUS0yc=BdK)+Y zRO<~EE9@M(H{k(+o>NtNysVAneE(*#NytWdX1}f^y|g(yG)?mm(Pe(=v0tNS5a8(% z$i@$oN^=>{-zq%ryXz6W(G+cs2n>eaG|^je+qXM`51O<`**zX-?kWjihPSd%9aJ(~m9x_(vA*tH?a$45Jddek<_*jr;m2 z;>&0b=bOQvJp#sR*HiA=%cvjIe}o|<_OGy7FiA{n_gh_#A{BZ)Q1lKa%u65v zbv$0{*Ut>nX9PXOk~T%r{=Yb=0?SPX8o)vo|q=L&XG6QOWx6WaWMD$BMN1p|9%O7khH_bkN2eA=V=#D@50is2vevX8@NDwm*m8CHRW;dw272^wo) z8E`bC*sL0{dAlFW!6`dY+;4?go^^XE^D){(-0elDe$LI>7Ek#)-Q1SLQC(DSBkN1q zOk?Z^)uEc)i6P8#f7cv`R51AT&85OH!1Z!4Ds67PlQ}9>G;E!-=cJ`@kIfdp6G&~j zRZ>QUdmY;}^$;|p74jXm56)iY{u)AREFwSozC4QnZ5H*CV!z5*{=(Ae4NA#);K0RB zwR2zL?<2m%&^8B&KjZyfZ$!a~qw$T0vu2fkZpuQ3OTq5_Rp5_i0b1^AKO4UID=^ad zex^m?e?bq&-!46t*56%B&TUilq5Mz(6+j*w?*jjr_9-pxKLaPl-CP^cWA&0y3$?Mt;&B1n)^A)}Q2oVPGA)2Pez3|Cj={*=LfqZ-<{MG?+d21jKlfB(C_+5VZMv+Wu1?TsHcD7j zT_Fh`+iEwE>h^AQ;K>TvyqitxGY$^b?INC8z+0Exoa^#>oIO7Mbul>BNl{4I-^X>@ zTd`SCTrAo(TuainBF)!=oWY45t%-{Gf~7&}5jFd?IxyuaX4APar=UfCQT425!`8u8=LXk4da`y6;5pxBdAn^sO$8Ncc=mF z>cew`9X-{oDwuQFtEP;P=H_oSbR1HW59ZFNbyQToj#HJ@=}UF352TcIcro%V$+ ziF-nfH$~p!gf=rhEj<+~eYCaB!z?BmuXa9?GdNVTnn20AIuQ&!zkCLls?xz*HMN2k zzXUm!mma=^&FhYRJRB@ni1v28bir;XwA^OJ?{h)140~QoJEBGxZ%M}-(yTw7HI9-pwsp}Yg%-%>`I_e4(M^CKy^+|y5!&JlJF-KE3 z!Qy(ivGi++Dz%F*CmRoiWy{_#N=$evDDacYU=Xv*es%suhpEEP7JhT`I`MQa`{}!@ z>t8bhj(tBkEE0vPh8DzM7dVZ-lN7o5sd{vIA(v88w;~^a^Ej}T*+-VDX0hBLk+M3c z*~Y-NtEMJu3~mvo*`p$Dai?SEjC{ot(DT^>gaWIiN@S0f~4sR`s>5swm(`wKHetYkgB&*Z83EHSby1 zYUk_alHnj_>o)w@+yd2snx9NRsLshMf4tK_LkSKl=#g_8LeejV0G8!h7`bStpSzh4Z`sTZ5K*VW!KBasdfF+Soc;eH1DiJ6w{9$XkVWD z*0cJP1hM7vX>uI%%%3%uOJ|7b%yt!Adj`10b@kfd(s@^n-5?)5`p>Clzb)g5C~&q; z*nYBI`>nrJKEVLP-IBEpG90o1WSaa zW+l{DYFkYrRQffXQ@jdp(zR#w)EpT~aAZ)r^0SeY2y3(@q;w+ekzK?OWL=w@r?eOQ zHQod0d0T=$U*fmqIRy2W5Tcs)Lw@nV$@_^cA+#Og-xmMz=xwaD5Iy>$KFpwud%EW_oi8MQOgMa8wYNKkC-?x zI#xeAF~E-wAb-BYv_n5k@yDS0u%9?a=<1f>FcebGFb=BID3Hs1!-wnnwhjQ)QDzPh zPJjql7n!N|bnj2wC}Eb(dOv#~qKSxUdDY4xWV=9gxZ|vpbrV(+9E#XDIrJ8p`PRh6 z6%A>Qy2js8#niScQ^^MDWG&y=Tepwvpiw+Eiq)ft1i@Y0ow$ZwhtVjm>2z!d^xLJC z9u*V}&v6HQe?;%Q#IR6O(?$0F6#Pz1>teQxWv0io<}8Vh7!|v!!D%^moNxDk0~+wX zLk93|hq^qGUv5ph?t!o0*_>oiZRfKXhgP{IEGqDunX_3dp~}t9ZRsNr)lF^P<{1`U zRW^otHMPfzO{Ktmk#^-f@3h6bNx!H_yN)u{eFNxB*Z1X^z`#4Fxg+dDEK|Qa@4hJX z?D~*6E~4Df=<42fI$`eY2SZT9HH=q9nK%l)fphrXyHe!k`9nZgFC5kx*Gzbt9@i|3 zXKa@MCAe=dRt>mK#?;i6GrW+M)1D)zV{+*!G_VZtzMja;n2eBTtC6Y!hR%fbT;Dv} z7x_y?J|aECLsaw4meXDWE`6=MAk7CKGfCHHBhgeB9jOU6N~>rCCJ)A8GIm14XI#r?mIzqzAHh z7knZk$)~Px_(r_)K7X%-ArqsMW%zA&bhRRTz~^hv`dIj0l!>(=IAuT^nA&;mOZM zC`_rSqE$J7nXHsl+*5Mds%ly)V%96D-M|aE%D+|YL9LLgsRiC3a9Y_vj|#rX%5a~r z?H2c2Pq+KF1p|53VeUmE_sc0N@*=O))X8%qMYgJgs|0542D8Wz)bEd0!yPy#tgI_q zrs4UdQ1sH~5l$YHRyrz9rkbuE>F4L$VEUWRk|w0$xeZ=MwIn_=g5?tw{e1SOCizJ$ z$6@LC!Psa^`Hi$E9UVC0&<5VAz293y^<`zoJ7ewy7UVi}q3yV>i_Zj3C_C@6!54#m zuz#Y*pHV@wsM3Xfu~|twdxOl3*~Dyk?9Gh=xgh&12mzav3NQfO{p6$Y+aEIXu!S#A z!nC>&pDpvkG%Q*}>4xaUXd3m~?I;q6^iO*4lNH7iBo;;MlU;1ebOLYIBCL0Vmg$P|orNXYYd%f0CTgGz_8d|$9uJ2{M zdTjG@FK&{}|ENO2xl8!C{9zD)pR(^DdDjB&{eZTD=y{aS|s3Wg>wN7sj8=` zZIw%W*q=@*&xH;9W9T2LscGcvy38^edhc~iysa8Mt0rd32ZFwOn*A{0fb^Iix5`0F z(DEEZn+W$;+g7fi8fX3M!5PKbu9cn^I`(S5?7gqi!<=WuHJeokRr~nX?9T5Ruy06s zqLKYPHCl=eVo8)q=l+Y2aogyzs)_nukmV)rPuNV0%4{Y9#533<0IL6u(0vy+PO2E5D0FMB-PYa@jLu5oB# z2*m4U6?tXzZbnj6V;OY)M*PvsdROv=v73ghvsou0-3Kf~i|kv|DSP8)o@dHHR(4%Z z3(=0seIA=D_^|hvNUSyM>h6TLO~JU1m&s{y@RX)>l#gt-zExRSh4;2q8@x)zm9_iW(@?QEyd`ygnr zkI%aA3ag<`D89z0%IiYFy_rLf;6E!Fn%sPgr3^ZXsTZL0bv0A!I=Pk&i-O=P zJ;tk_%SQB7Iz+Y_s47IQTOXDtyiEpg5ITb(FNFnrBOMyeZ@ZQ45kXKvfkdR>x4ui$ z>!-QaYkf(%?6pTZsXfyY1QN8I^uotikh>E-_yhvl_LK|io$%!InfmL!pgD}br_7X3 z=}IynZ(&Vg`AVV)Zgb15QLx8;4v6@p^)?&73cd*6%SNW9dOSM~Uix;nd;X>vg?GD% zX|9buv2CALP8xo&!4PKr)StdPf_1?PI&0_Y*|R#>G}V9MOYNkg<@Tku>E78U)(+IC zJ52_d>PaUcwV}ZGj~`MDh#Eyu zXZD{OD+Ti!+b;!b?~Um$?B6-tDV?C5-#?qN{^~sArfoWJc#Jo?X{x(5uxWAFYl8{4 zY~QV?5ky7>JwY(sXR_qPIl71)+^ot$eyQ)}oN$vf&a8oTKU_*VT(Wt`9*b6`Ab)Q~ zb$)#L-@MnKEtzQxD7QIgE!tvkQx0OX1kIhPb{QjH9UP5#1j9?0(4Q~!E-`9I#8Z0N z>MPDHkbTD`d1)@Qg2V0dhd4IqJ(9xyA4Ak;H2Wjsp#hI*Lf475j63X`3~s!fH3%|k z--B+3x6N7XMn)b&I%(aRF=2@>)FZpR4w%Qxa4Fi&HHa{!w{*ERn%;wVTiyyo3bE1{neGRt+xFoyftvxbqgJJ>HKRt zf*h*n(FH}Y&r&QC97a*YFdiw1H5WysJ%F#x$Uak8=o+C;)N}ixofH+3xER<&dTu95 zFlROCaK4*XV`KH@K{GM-Bq_?`3=RwewnzzisC`GKVmk~`YM^w#X0AN;scOfd_qz-D zc1R4fT`8BVr^7)--Q{F1-!d~OJtORWlq`Dj!r=O|Qr5pP$OBR!1a&wy3cS1BQXBRd zRGyCE?mZ5spNQK5nNm7A;KKszGC@+&{&Q@dZx^dY zEL4*~$pf9&Jc{xpr~HdfuoCCRsi4sP35Hz~_wJGVh)|YCmh53DiHrv_Hs~077i>Y) z=V9znl@k9wdTXP|V?J;ipI4yeeH{Zx%`(YNje^1zd!0dn)5s?Y=5+fe77Ka}9ml|k zs!h8o6D-q;iCiTPQ$GRWOm9$P6=5pN06uMhqL-wkU+c>vO*~P3(JGKz1!l zZ6w!TL02f^$q%TonGaYkr@1!FZ)V6Z98w%FQ4P1t2nT66dB(t|${pcWB48ti3^P*5(i-3fC8>&l~DG+nA7`&~re zYFk%*)h~R%bRaHjHGK~gbUrycdZ)(O_X$kr5ffz40nb+9c!6GNGk)tPw*OT;^MT|- zMJzR>xZ#s?hF>@7MRE`}UNwiy9OBD`K?%XcGZ|_1o8{!0?a$8>EZt;RwT&!TzK&C) z&koLd-7H(#A8T70i7AhZYKt+rV&cPgwVQD zQNJ)cHg4eD^lK!nn)%>Rtx~OD{ryl!eR{o~4rVig^sw5z(2V+@aD9uL4sIiMgGDRh zKOQ2Hc?hbn&ug62r#5odZW{8+GC%~oV`{cFv#6r`a)(5xL6byhwnPk+pO{5#Qri$U6IX|$1do!|868W(8LoFD3 zJEsr)KUvY|ThmUL8d4)KcWWoPmdM9O9aT?H@m{2<8X5T&6-14g45KJvXnG@K`kWOn zy!iNdbI!(}C^<|L+{|iwTm?rsQ-iO}D);krm$hP^r$0LEGT^y}l3XK9>ekCvk3y7AgZKj+8U6+5|6W@q&Qi$b&q(NI@cX6ud~c6epa z6zEo{VH+%d!Yr^>+*5KSY5nrz-;(J^9N+myAtcT~LU^nb;FJY}#K6!4A=B`k_J1_4;GLSqW;xy9 z9;I+M@m&#n)0DAq^!{S?Xf@J!?a#)-e%`;?z4=^riE4P{7bUuy@@84H#Rl_aR6_5L zdZac#SCf)sNiaD(c<6@g5zQJD^<>T%WOP(=>0B@gHP2n~Clt@1O+! zJIv|s;=7D{S%QN1fCFGGn|t8X)9>sIM*5Ry36>Ngs7T{7E^Q&)5$Sp*JH49N^wHfVxy21H;(mn=Z^h^@7*0qbe5m?mZ>Ecm=aR9{PS?>eMvddn|rli|OuXH1&ON zI*M+(=H1SWNRYg)*Ib(UQ5-c%XMK0z_A|?iWTi)Rt1fe9DX9gl91wp89L-uh1H#tY zQ{24?zemXr$tlRw7Q(r0zmDkm^wM3t$zWmsT!Ad+>GtArAP}8 zjytov;Yv3|VHKu#;ie$z7GQmX7Da>{IadzUOvFNFsId}+^PYS-?8Fs+Rl+_*2%Q*7+d+!ttWy!ceUG)wtr86_= znobF6rrY8br_B*MUT5B%d#ZEldgxEfTY1Y>?C~_7Kb%rqw*B8E`gAhTx>Ci|bM_mO zyuoRJXH+{7ZC~FheKW4lw!MC7SJ-fl39bn#E7nb*5>{qgFA}RdE#M7nlhV4*d^?viF|KSUdg=a?B1MO@+dO!(z-G{> zp;CK7~a4ol_@GNA`>9t7}&BJx+`ryD~#H>^^_xGnB0KC%80~Tjn7M~^U1`~x z#Vz?2EAiz~E3_J#W?xv-$-Ue%!gS_v(XKk!d#73w(AUp$k`n7{=zHgurzjX17O^3B zBfR!Vl4yWqtGo{~zmMuGUP$^^Lo=|L?ku|J^P(CyK|^mO3D>c@1M6f_S4+uZ*6As0 z8nEI$8Hy(vRXw15{OSL)wBI;C2w0b;TJ^_IZ)52L5#Pvn9e8#`)`pg^-Ze%=cyK`B z>ijq;HyblnjkU$x9~VBGr|4T1=)un^+&hj)CT4RKZ+cZi*kvuLcfDqvw-v^AT~gHM z2)xd3`itF<4QfhTYUMBXJe;P+s-LVxd>;EH;&Gy^1UldUx#dw9jvTQITCEs%Z6}8m zXz1a4T}igj8cpoRaah9Fui94frT?qD>^k$4g>g&nga2mfMaP zO6vGjoehW$JHxfyKch@!6@7x5Q+cI=&%T=%Ekev-N8@P^nSgZ2n;%MOIs|FVtykps#bNjx z%m`-rV0p2k@?lP1&j__-N~*IVfvchGQHAQclHR4+J+}orr^)@ZUz`2yA`{*)Uo&ON!E8cPuvJPhIQQ zyScvN7F5a<7H1&Mtf0te;F|50yI0RuMEk9~eCoijQZNpBIsZ@lrE)6J{YJX(w48Ul z;WD4#q}uk;YIP`hyxDI$k8$53`_V&&ESb=^sp*#!Q4bZfX~Vifv|9z^RFm6%C4_jD z)L&9n%aJdghIAqIFo1{Zvo0sP(6z(IIHv$=C(0Un9zlY8c`$Q+#wzL?O%08Gbwb3M z=aTLB*oO=EO?vy8pJDGq3WeoK+hWny)7(Cjv$3QU;gSmPqUfI;mR?tBZ;YJ?Oz5Dg zD=*95vBbtjC8hK7Y%{*<;-( zvP0L@z4Qb?^>y<$L`r>=A`d)0-aRI!bC$p}8CtVAN}4wOfp;Q@B)$RCePZc(El}wL zDexQUmnR1qPhM?vkP0EKteQ6A?lvfK84MZH4NIur`qNuaN$R;_#SLAAC7$XqU``?A zrU>L64$vECaV%lh&$}6R)372|3Q^LNOMBZt>FA+(3|pIi-t0N0-eWU{i2+TZ?wBcp z?Qj0dqhf&Xk6Ue~N1-08y~qc8FmoXRH6$<6%94Jn8num83WcORV!12X;Q8jL@R4=O z3q2;PCDDE^1KXpJAjoSe)-E}NLS@j+lEqk=43bafZNPzRK=9Nq*%h95S@&jKJ_xeU zq@a%bkacv4zcc!w=tJm%Eh%~#7)PtSBnll4Y4Sp37mEh^JjS22oCs|e5Jj#y#L#JW zdyN`k1=Dz>>6zUpya3XCswtKg#nak+(;cxywHWeLn$;{%qQ?5q1P!Hm+mYl`E0wZH z#M@3|Nw>bD@XD^J)p`7cZ06VYN!W)o2XjYa3!=LmGAzQnzGL%znf+F)-2@k-H#4<~ z+CPLuxz^nHQSXR*=k8Y9y9K{DdI#*WYeIs(rW>wZQ90&!m*bM2Nz0p&ruwc5Ve4l< zjFV>84)-xW2{Dl`v1i$Ww@ZX{Y~**9YLe=vylM8HD$y_eFp@qAqN2!B*LK3ZOg)?M z<~guATXaLkNBO=# zhFpl4lbSV}=2z|{<`g`-w`u{RX}D&Ehht5AMNvnjgr8{J*|I4pQc_{7dLs3t!)A_~ zW$#xH(D6a}gM<_n*3Ya3?ePd2I2yM;=sq?)VkAw?q2ZI5ZMrz3pik*H>_8GiVft6@ z92wJBj}Mc;-SPY^7u(f~E5uQ@#Av41*FEs^CenIVCGw-vLPB=aVD za|0p_Vl}J&=Vq1*;oI2kuUTnh?{m`*r_;{dsnv`CDeH}6L80$p)T$n#2zZwykJ7-T z2ZS;<-!CpQ$;AwP^$Iv(mCq=WT$s~y>-~Hqf8J%au33tGSHa>oq-o;Ox#GZ2?0*e_ z#Wpj1@QvndyMtVO_B|ow5~h;uc0pCJm@=q4{nDMvC1*>N+ct{KtS#Z= z%s*8U-}C@U(5RWWs2ljksY-HWBLu|`{If6_+_VS9itd3xd4wdIw{FV*9! zY4Uw_Ij7L9M%A}DO0>nL{XlK3j;>j9;k3me5`r z(87=9ddSi7m4MV9)$8h#q22Q!0j8)@nU^wHo_kGVP0c7M%viEg;%^1Z^$Sy z&}igR6*Idz++_JMUBu(()-gTFY+!SL`2OB|=YGcpjd> zpg}JOB=$D1$14;05g$AmB-b=Z`{v0jnj06aNTM4AOK*EyPYj)`gz(wJji=VBH0-9@ zA?PmbO&*WpI{d|+S~E=d(QRQLXG?j;JG~*B77RkcaM#a{s)vO{6&(0`6_)crfL@M% zeqCNq5w&p}yRqlXiQxqf?1F0f(v<}tf?~+rp8to0F%UPqG*2_%L}qD?$VA^tuiAVO?^a_hEV>h62@f1nXvg(Q_>V;Ealy+ock|LEUWNl zJ*;%rVVzw@&3<3gO=Ojzff{be6}Y#ty>)}4x;j%c{Y%=PkfHrp?;|Gt#5>Iv zCAztwm$x5`^itXT5i#Pe?qt(gKn}g(i&rX%<#@|06mj?Xd&F5&{jjuBa0(aE&qM%L zdxn$=<2a%Kd~+9f>Mp#BGzHLDgdDv3qh#B}@DvvKjU;uI6Rqt`i>;CPDh-K)lP*@WI}19rMA0aU%~-<$m3Gi|1D zF7OeE!jN{yXzyh9FPJO`#Sl`VFY8lvt4)vjF!*DVJF;>0x9F zeEsVaEk_d+f7M%*2cLs;s3CTi&5`K*kU-hZ6pS3_w*U5qH_CdQqes4-tC~|g%d`kaybbh7BSZhne#u4oe8gQ{?Z>@0c~SBqR|T#HI%HA1>Arc}76r=7!Nm1X zns2Q>nLG^t{TAgCnj743-zq3L*&^3O3t)cQ1@%7?;(Nh-B5yC7zpIUHSwKQ*dtaw< z6P~|*NU4;HQHKE-f$kFb5yM~fE7J#C?8#4Y(=PLtoScs>p`oD-irH_}PnJ9WeQLmk zRR6jh`H+_^Dk5DV6YX*d)q$KbS?@4$NcQ3LI9W9T`ML+xsPF}jCW}tsxP$-6h>9gL zDnCW@*7cH16al3_`twT#qi-EtP>oip>Q+n`$Bv4`d@^S-$n%$m#rG4vvX<6Ya|=?E z*%|?wKlb%R3JupndLhw+8CX2A&08Tz-9)!OWpBS_QnCf#Xq13~0GM5D`jVE*@V z%Y`-r0yRK#7ErU!1v$6FA8V0jSncxLX~3^p=egzGw+c&4FE;^!zqaxdRh=E zO0A)C(YSnTx`V+q+{15g)JR&^e1miabQ_PgiX`pZHePIPhfilaLE;OC->KyvP#>Fn1|0o@$Z2%> zuup$Um1@~PPvihB=-n{D7q-?>J^out>qv;E(oK_Gp8G{N&IghX;8RFOZoje<-Ny&! zr7Y+Cnw2SayC*%e-V*1IyZq#QnRfVNdNDg=^wad2TT_;#BHeiR zRe-|QL0B%tZ!jgNDRtC7l5~zod)NJF(7ZLSj}Li1M%TZ4g!Mtj9kdEhnnoUnpIy&- zUrRO$ZjWni!0@ay4X6SW)c`(>;yk_{ck7qp( zSkZDqAQo9Ve)%WJ|JOSW?P1UeWOui&mUEX?2Ve1+jLfCmvr1)L5u7Y~d;5(QP_kSs z=<5$R^_w`08Si+L=Be6zdAbz!X5zXQI_;?NBeA6mHk$7@c*D3)_*kf*TF`Rh#0#?! zW&VpZ?eWgxV&XakcRxMev}mVeP}rEOazb6Cob;}}_eHHZrCYb->0H#yxD2k`Ri&74 zgnNv&>Ebtefn#GLgPw{zr`JjrA=3tdZHWGGJ(a|X*@8q?#iRv$+c&(*0Ow)8i zM6&e@H$m(zR`79EQ{{I`?<5m;_)Yf+hE2EWQi3W160vgJa;xr}#5hPqId4k(u~r~A z6_rz0C((Y*G?kIlQ@h&TFQU}rwsyYG@|`dl*a(6tz(yEoLfgssJNyLrFvC4m9L{{j zsavbDleP-(5(@N~U@oIo+QM1Ka$S_B!mI1%_!ld`HBk1|XOP*A=V;seYiGD1ORDDh zmUic|bmmy;_N!{$!L1R*w{pVK&4ap)Rm_iTT1_O0Gm0@ zn$arw$yVE!rY`H13^0RA%{>8b09s$+#Er4vO}MwBpU4Z?*mMYr^*qqQ(VgXot*NFQ z+)|8PVf4eyD~>}|;YdI0>aKL13iz$s!jv=40czjB;go@(W&}rlGWU>^E&%Kzy~GEa z3qA{Pp<2;nJvSOnsD2qTgVD3KbTmHXqv40$6#;MB6*K}&{CBD}9L3+`V!Vs$xvffW zPjYx^Njpr`bFY`d#VzS#4qxh;o5MPyj%q-AOXm5tt6;u7QAxJD?CLwSwiOYyy-fk{ zUWp9M0a9Ux$t*>N?4mRoq9=@K9Yu5r!|_YCf64yE2nm}9ZP^AqjQFZt70bsUIavg( z6|*pZplPIYXg9H&Be^HECzE4SnNAHPSy%_l2fORDpK_+RVDf< z2S7h{;l5jv@dE9Fs>P<7(Phf5^|$-JBpj-Kjg6L&zA8dx?HgVF5T8D(H>r}Ays+MY`HX1pKS48FVO;|Z*i2$XuxPg#N3z; z4tnn_u%*Dd?5=+D9M5C`-75y}Q}a*8rr4Kt#q1>Al9==xufJ!4pPbP1f2c_C7ia!5 zgz_74h>_C1x8aEy>E;Z)L+U^{>i7oi8fJG+K2pw3H7ah#Hja3yE5BpU@w{VY?#jmg zQ;5yTyS4iKShHMU!epieFvb@K%qbeYjrILF3oObfrb1K*7YT`!^R7F8&}U{1w(UwMs}O><%Ki~8Iq?PVZ?<8|x=kHY7#&HPb^y_Mw8vl&dXp-@#@ zg2}kDVDBWnp}mBZLpE>SH(4rRKTI%rjnO}o{h}`Y+nVOHGV?-tZH8D8q|@7dE4w?E zei(wc6mwv8a9);vqf&7iQFITXoy=81Yqs2(7s$MBGuB{Y!BN+ zlv|xfvy0SV4&QfkXp~-`@kM4eaAFZmM{pWmA6jhEJu&6=TyMe}^D9pCpb8h@{FY2f z{tY{)W`HZuazCUa4{j>iDSUg`4FR+e?`(%X(Jbe70w$flFjtkijMyngdUBa6O~a;h z1uxx4)#B%f(Mbi$G~c}Y29cF3o;%HBRs$mXVNlf-Pj%K;Xu6+)M__l;a(mMupGi^; zGZfwZL?r}{T(0{P?>`${kje27>o>?!cEvbRcrI9op(=q&$d{do__ZHjdfMM!FDtFS zT^|pRzG1AXQ1q{&!BzKXD1)th*+=v3u~F!!%K9wY+5Ox7E5E`xLiVO=(moI5dht5Z zcDV6{lhwK9XXmHwt844SU>+{pJ*o|O2nqJ*kg1T)6*#3qrP8kSehTF;9t4eVu{f~l zCTb#u!`P*oXS3^va7f1T7qk_xSOT!oTY9cLJDJqPy`Xu;Pi|f0?e@RK=$qqRq}on$ zeD29ZyxlUin9yvQgIh~)ChbiVo(UexVK>uQ?k$?ruZ_|c30>5Mo%-2NFL~hw2Vd7U z6w}%0(OM|INCv;*HoG0t3#=tKTDSKY{kSGw!o30h=zTwV<~8Q*2-*4mn>R42@LsBs z)Te$m`U5L1ab!ACNm^;fEjCsdJoy>VHyE^;@e<}fi=p1KHXqUe$!AtTlLGEGJSCT? zKQSUO@THoXn&$HI$KSjDRVEM4!n$*;)`twQM|JPt#Ejl6QpG^JrQ)KAY!)M)+k9P! zS`s7~4a5LN!CFHSevr`EJxk*n?FKC9jfjZWfm z^pqbA*}AWZaLW6yhnU_w*_D7u6}sgC#}T}di%pl{tgujBg=3${3u>LI1o0`N=|%un zQtT}4CMvDu!~dhT*8j{SKv!XCXh>PWZ~4`~Hyn&D_L{TRHSY_Bwkwe!Ws_NdYeLlr z%cih|le^I7FOZ-3WmIYh#WSZJo--=SbnJ#wa2d45y9AxxYv6t7xCri5Oz zvb>q^zQ|FIUN6Kkn<(fX0Vy&XAYIZhr#=t#Nq*&osoOQalN5ivYhyJF>otL)BD5Q5 z?sSPqixC12MjwMTM3R*fjo{pkemi_SqC`aVDa}3{tNwC0%j@z^8<&JcQbCE0a(w*B z$OxH6h}ZAn?y#Oal-i6Gp8HOpqkS~uREVCq0ao!ENiT*HAcev+b}v92 z1(qwtV*9nMOyz1{LwVsp43oxJ7-zvkQ64`k#Zl06gf571i|6%A^}XJR86`<K-88WtFSOh=MwSP=jh}Knm0#c2 z8i|BFeKgA5SXWo9e5BbbXV9V;DZU`nY-0_Yd~qMv33TI%aRcF{TjwLAuOh9o3RT~* zi`@Rs^E97V7$3}_1*H5U7fUJw7V(i`FaFT8Xfc#ziCKjx_8cC0_TjSZ#4{;`SV&ap zg`Oz+=OomM0KHQN-w;9H4(~S{PZFL4*1QNY4$#+FrT8anBKDM!(}nan-FCk0o(KOh zqhG@-Tsv*yh)_cZCmrDpq4fK>SeQGs+(Z^Tqx4|yWqv+1u)~OnU9R4JXTA2@+{R(Y z@=*Ol)KCPai^$>kmb+&gypH`wodgOgB`O41(pFkeWr;2WNU!- z&y9W_uZo%?!H(GnBRLozG^g7V235>U2s)4)CZ5?Z*mngCEyAz*Z%hQQ_qssT=z{l91S8eR8k%MG3&g7YW-3fq4KYTyo(#l_e>`e2 zVC5U*=R{ENjfY0NDaP+lj+V|iSy(BivsYauh6!V{znOV}&74!HMVP7y*+!-x@Qpj| zxvf<5JzNUF??KGt2^u~x5}_Dp_8quw+jF@5;X*Wo@$6Z<4eP!Z=9yXZ9dNn1{!0?q z^%$e1)ez!{Orm`28= z5b%-wQ{evp>*+fDq5l84bhfgxv(CsUBO)Q2$SNFV9348l{A_n8oXm?P` zv(74!(HTd|CY$_D_%Yd zkeLM{3-W@viJN-6F{G%3_GZv<~_(zO2Vjp&q~lUQlzB5E@Xrf~o(rI}i--$#N)73BPjsUe??^G^Ij>^qAs36$p;ME>RP``VCTbrlYd&R3 zD56Gl+!LCYEX6-mP{a$?`fYB`dX&iagOwP0f&{lx$q>FjKSzU;hrBe+IE_ouzCJVl zvcPZ-AH(cLbdm{UQ(27zV~KX?Xs+nGTfYyB77MyKZa~KM6_0VwWtNlQ(sqZx&Z1Q} zBYj4{JsR05?KI@u2OJV+<LYalh8usUOKv`tBcFSc- z1}};4y&OP3QpGw%vQJD#)RnZLjNc#<S3>IpwzM3e+#T6Wk@?vY` zHD2XM#pYc|+Mx}$qIGORWO3D(GdsMimqyNNqA#J?*2_*?Dd(F3gGUEfh^+<#Imrom z$;^T3;+^UC=iMOmE^YHas{&O)HTYLgpYkmS#Jch-(!u=te|jJh-Rfvlo#wU#F*~TF z-Z7A^mu}=Jp~+g|dlM^*ALGBOV$G1MbM9S6l#|^QB953)RXrk({n8%Mcv}C|)aNr| zwtLh1EG#+-SyQtE^n6D$Y37xxO$iVU)vjt~32x*zOo~o8EYoVMfeRXBR}p>+!J_8m z4?L&`3ZY$JR#m*zITc&=?{Y$VCSEwPaIf!m$2wJJJji{|wliMk5r8qd*zk4@J^Hu3 zQUlAnB#TRMcK&5aD$4x~JWdp>6|ASxrjm`XKbb|?+X7+v)a;vg$r0PL?5dzOD|-TL zqVywe>YwLo{>+)5M|X&xQ0u(oP&jQ(@7Xv4lTtvI!jKvLPEAqUrirVcV~a9Rce!Sr zbOE+kGrvE_^bs`@p~~AELfWFUE%IibU+f);qfQm@%7)*^MXn98Ah7J_$B=eSOuPv9VnBc7f z8NPR%pAaJXL>Xk^BFs<*^6CJ4O&cY_30^n`e?UF?Q0`|g6Bs^&@S+u+_?#~t# zC{|n|EhtMMI*x|@_A)laZv?Jp?r?r_4y(;9+>HBck4U!S_zk%)VCRvsp3XmdMWzG98?b{fo(;$z zT@Wd^d8Cs(Oc6gA;7_+eksJApk9FUNW=zXOd86Mgq({8`B*p4|jeKt|0iplvyR8x9 z=2hFUM}~NXfX>ZOk#*K>(xP0HoR~{6)o^zgoEziPfW_J(Own-aHEynHl~(xwN*zPy z~v@z%k#gufi{l?CY?VE-iGbx23P5lO)Lb7Av4 zullx_E zFZq6-1FV{Bfz~_E0mFjjY#zMii(^ZoS`OrTU)cJd%%x?;E(}%^G7{IP6HsH*@p9OijhCQe! zC1Df?O_oEdH2i6(&IyMv8=8OWI7S@c!0Qh`&zC zcOTbWN-$9ae&v{UCA>DXc2mDloW*XKCs^~woViOlg{bS{HP#6#x6}4(xN#u4wbXuC z$jJs;(45pgXWAQmIN$PBc1vxXK?lGcs*+kQul?iT;0`5=kjXS4Z#j8~>e&a918+w? z8KW`Rr2*^2G|U-VV-q?w)JvQ#R65Fl4c}I^=W$(d*wZE#gxFXlE4`@HGQ2-ED%EK1 zMfpNY%G>1-TISF&z4^r5bruf)GeeUFv|Ik)K&_SJl}uNpCms~VhHq$?y9M26LodAC zJyEc?>X~o^oZ?fL!1k8m)Oh!Pd2IEvRn{8DoShCBvb7vNqfnv>1y2q)V;zz@$#0tW z`QagMm31Ck3XVbLy56%m!{JNY6T)^#z;#lRs83DfUl5Ojd`E-)B`kyBC!<&k;EWp4guGzh zy>{_R`(qt9rn8AHnLIR;e(#a`3o;&^|Ex$_Wo1pH)NkEYN`LyD-q%Aw zQ3o7Bbjv}%_ORW*9K;vXPo&0`vA@8_UtImV(}j()Hu8#EZJ)Q9l01MVhXwKEI)e~G zEWIz9op&IsAsT#izuhXjIz!oD%vN#+!pSVEeT+711siE3gFcYYCQ^zslfPw&W4lKQL9~q zMs;{P0LBf&G6f|aKSk+U4fh!ozJdO6{qEmO;ZAXS$I;PhgKq_=0L;mYa_w=gyvEMH@#WqDb*DsonO^?>0i{=D5>e#Sp%(fhD}rt7GHb8Z$eo{{Nei&$FcjC+1S_juXMtJvEwdr`0p<2Cc{%OHmQ1x-*QCi z<~x7VP!!QuVl`ZOUN)gqjTV|uD-f8uai3?!ym6!X|A0AvPvV&~+vJ0mWLkfvZ zD%kz$?~*s%lq;#bfqMLV#cA?ht<>m;wzVL|gTok^wlBVEB#PX3OtbDx9-T9Cw|R(kjxh3;jP}`u&ms literal 0 HcmV?d00001 diff --git a/api/core/model_runtime/docs/en_US/images/index/image-3.png b/api/core/model_runtime/docs/en_US/images/index/image-3.png new file mode 100644 index 0000000000000000000000000000000000000000..bf0b9a7f47fddfcb7969e4ea05e9d07b800fd8d9 GIT binary patch literal 44778 zcmeFYcQl;e*Eg(-h!(*&I!Qei%#_Dy^WS8h~B~& zL@?SIj5=d@$M^T#&wJlbvex_0`{y%j&CIo~bDeYc*?XV8_xbF7z17#%pu5a|nS_Ld zPD@kGfP{o%hlGUm^(89eofge`PZE+VKu1+oeJxd0UVRT&J4a_*5)#d~DJIlU4f|Mg z%nkS`F7fi-+82nNATdyJAT4C;QP(7Uef0%z>Ce%Jl+&WDHTmWIQCy6G7xs^xf3ndE z^|w8(5q33DnhshCK;c*C&(Azp743(z&8bLmG&yRC(O|N8LDA%IZ>a8Tloa0A(j&dj z8$$1vH~gDT^QX9Y0?BOm3FdH{YX6PTOJ2&PiVRESB#Tc1 z-!?Fke8_5kZ*)7s=4VLkHGxaJfwDtJOG2``CQAyJb6-E8{~|&XKM9q06d++w;~T!a zWUoROV-=bGl`P>O5;^78E<}^mB2}ExPEW;pH+18o!k=<1G4O|e;(3vr zaf|km5w1Y}l#IUO`bgMM({9?Rke`>97~Nfr3ZwLf8Q-|^g$q+;-xXRGP^w_q4ScTb z2Jg2okW|!Vp{*Q__xLGv#;Nx)QaH}rIyLFL@p{+LRgWjQ_v-G(D)+w_MT$hLX+36i zQra?#XXeS@yKb8G^)35OLjTh8iMa}=;I#-Up@Z5FR~$97`n&E{KYP~s^V?k}8SOVm zEOxhE-`Nan+<``&{<5eUP!R{XD``BvC+G1@l>GH2vMjOAcCw_WQVVmRUE-fanp3cxop+-i3iA;w# zOoVGW%$XOw>9`hZ(x1cFO>PiR^Sb--*DLQdjrSveiF02IQ{=n)>Q#?>Xp|M-?N_JM z%EnMHitADH{9hwEcPN(R8A&~UHR_0K%l{)nBOm^5jb{{Ev-`ZWwWbe$uAtZ*Z2vR< zt&w*} zWg~vVoq@-B%o`D<;Dj?95jUsO^;qhqPJJ$+-~KIpZeNhkIHk84ZnUeW7WA82v8?LN z!ZLdpd{tL8+>7&)`I1SZ(8D*Rwxs5$pDZ_?)dKr1#;G?4uWy~m0#Okgw2QJ!_Il31 zCLTa^_Lm_WjT)-%FR?Z;PlbJqVyR80ZN6@>N3xcD7o^^ZJY#ZQ9O@S~U9{MoL!UmJ zF}a;}7p6h3+xareC^`Mf_m`i=Ud8vZV6(%;Ewd~wNaJDF$<@^|wNV(V57W5pOe{4? zXD=R)XMmwT#gW%cGXu{eIz5~29X;QSyPbrREL`CsnxT;B z#rrMv+uf7nlvW{DJM@9-jIUphFjBn;dCzYgNxt-|hEeh6KX28R;;-GL>r?v@c_lsk zhjM>{?RT@6%m!gw59#>?6R?^94_au5*Ny4~z1!g*3*<`0!7=n2rnfm_l$b}{9R%c< zMAT6qqDLkjq}rG(HS|8Pjcz&!;>ggSNs7bQpWD5^@$J)tr;@2rw$BWlsf@#)Jez#V zHAR8JKB-{d>;W#RWra$(YqY%jyrY&Cr}xEvk#@K7*ayz-+Ecy6uw!+~5vqQV{{bVM z#^RHMioo-$`N{gqKeZodKA_7j68!|=Pbk%B;Pd}vFaG8|C`;Qy(?VkoB(Ll&2;@IZ z(7Z{WNZyxr4YZVSS*xf}sxZ6occE~h=oob!`%Lnd3O!#*;@9P?9+Xh(lYdSwpWHc- zc$oC!;cl^-Aw}|?&xoP%FDbb#wXe*WWZqXD%c8K z3+8V$1(#1AmS!6UjvNdmE2U&J3q+~=$ z0wnPb|Crrq_|Ul5zzRl>$W_eQ@yUlbsJ_)abNeUo8|OFHZ;t9| z^pAVhBg&I14Xf%Om4Hf2Ow6Z2GFrZME=Ajqk-7>x!)9^iY30pNU32H@I`mZEMbt!G zin#lafy{D${1Hy|fPDWA&KB2xP`wGe|BRi2WBH%EQV#c8>?1j5Ili&$-nVe|o2{`g z9n#OuKh3vQJ)-M?yk?)__<9{AHS0V9p0Ia*!3n-AZe`yrj)m>wkEEUqUhCcXd(9)1Pxsrz2@ z^!cdssFNdDw%;SewsHj`9K|vqGyqPSG%Y`xlAh|D!d7rp}P?BoG%`FLvwDm`Xb=(e1i5EnQY1ntZW1Q6Z_I|g$Q|cMQCBsj)WRHG_V?BT#HGcVZRmTi+St4hw#oex6Pd~CE1kon$QP_}+Wxxx z^;a6-s}N?}w}V>yg|%8zT9~A;#I#$!vD`p^+*mczjAPf? zH-SxFTQTm^zXjp|it}!P+s6gWYs`Mkhb2s7iQ_(nLBAeiKSG1n=k`DDTLB}wqRv$q zd7tvFvEnYL(d614+ua@TOI^=wm9Ga3D3vvH0Bar6BT8=d^gdsFqY|Yb3ezkIBf>QgYa`xouOO^-G4R z(dlRrOP6&Z6O@FSSs_9%e%@uAg*sI+?0Q zd_l-H_d;CDb}Pzeo;TRIil7RstH@DFea81To<= zj_MDmxoYI}z2n_$D!9vX4HgjblOr?l$9@LiKGpY3(i`vjtQ*nHB@Qrv8Xc`I*-Y4Y z#vSm80q1>HJU*aauJ&mRGG}12*n>9A9yvcktVfQchF)jMdZz{*9?LFxY77PBDtd)_ z=l}Z9`xR$=G!@p>ejZ#iSWtPEh${LvV3n3RtI@>Mz64uYp3Ez^uX3-pgmZg$5Y(4q zGP!I3a;}$MfbA&(o0I4dh+NE%wnr`IK6+TIuR#U)N!U+9ZTolub}cNAx{~2H&ph-C zY=6nU=eL-pXYNRjk-Sup8TQA@Ou}~j_P}tjJjS07y4zZ^uUDy9ZZ+E}zhlyEUD;6( z&~Bw=eb8lf{B3ctsoBLf6aNCY-8B$!bP_Z}9marAtOMo{l26yyQCilT*2b_#{OrES z5(0S!KMSVO>o+{T{H!Mb`gIccJQaxuB?;AB?&IFuN3hGP$pOZcWal)(N0FDYBzg2_ zUA>3hU77SGht8o0u~I!ATBR#=ArxXEZEf^Tp&c17tnbM1&rv(;oMz;t(VxZKIm%L8 zCSb3Be}e2DM!`t)=L1zm6Ystt9w?k`pK95GKqNxM>q{gQr0gV=#4A$b7f8zSzpvFv zZn1#1p?4eg5-LmPhtqPf+aSk^k2<>FbL}Jzs`CC;rj6 zX_^vGGOV{QzNA_PH;HE)B@!()Wg|b*EhM$ywTB$tyJ^=hJ(T-kMEk1ac@G(XR1`n! z-H;F0lt1R>y-=k980+gRe~IGyI#28K)k^WPa+EfPLM734Ue>POxAl?euWgUkQxxTgj7D5>F+&Xkdpan1$sV^ z`eV?Go|hJP0U;!0mw5lrAEm3bjNOlxGC!B`{hv)2k1LVHE&QF9#HS%7R#z7|b?N@v zPUd$?|Hot(BgTIrqu5D2UUdWiNer=_vHQ30Ul^lA5_*%BHze(gOoXB8-`Yb+xVy;z zI^xY*5>oihPT4H>zbv1GNX#+U-$bFr`C&HpRH|NljJKk;~O$p`E-U6qt2VqUoFgV0h}_ss2n4K-+4W3REdh(H?QkX%)AEJ1A9KF1sM%evjBo~VhRBasnZLA z%_E(=7W#>s(I}{!^Pk(CcZNdfC7rx4v(OU{!=79-fSAE-RuZ%0laal*NyYtWUl6^S z@tjVFzqjpBiK~{2OI6^0AIb+S9nsJvMev%@GbfLonaPZjpFmmvNYLPc&7NhA>^hWL zcjd9>PjyEpr;1vN%FG3pSR(6t(-WlY2L|?VQS#!({C`@GI|BOoZ9bko=)3ref z_~exX7U+u#wLgu7fldfXPHR59!}v3kmTQi!RC^Zp^p4DHU)YXpt5RF?TIy{&r_u4a zHwK=qOnc}ySKWW9Z{5CC-F24Ajbf-7uO7;Hj-MF#9yy2}={#|1*}g2>Jk6VaA1%I) zz;im*ZVoA<*yONN2+V*g!*ES>wG5s$t)IvufXvj5KUq|ElQA>$u`JstyS*#?G})KM zI*5>MA%#;1V63ZX+k7@{$b$<+J2Yc~8<$*ji$UxG<8t|pSRbd_Xv?P0^oW35m-NH0 zs4-E16fJmlBh_^PMbqmYFVXIkpk18pCMiYgG*K#=9jo~5igRsw!yIJ#MZM_ertl%f z%%8kr)MN_1%+#HU@3r_wD{>46okM=OgnQ0AZ)D>kZz$`EgL)vhC$hrM`-+ zJ2Nvg2GuTHtt;V9h1$j}qi1WU>mMI~w#60?j&)NUd07@a9)1kdOZLvdjr_g^e8Xn# zudoqxY4IWGYxtiILP>{GNR6c0F-iTt&;AOl%gi3j{a^EUd~kEy_CHrYIIcU|%BexxO)KfFF`ogtyf`xk$gp8q95XZM?}Q0WxlSZb7?5y z;4c(@ylfOnk`ucm=b+>oz>w>=0mWDyu1~5SBJK5DgFSU)v!{yd={P2iQX?{-wt~IC zJX=T=X?^p`cKe^R?rihoTo>rjl;=)$v>k3k_>+wLVBZYCF9^QCZC&kj3Yyl&1hdVk zZgjKz587u1*JVwV6-YY5Z6r;#Of9U91iz{Og`~^>l62mwfStSRdN`9VMoGCd7u8#$ zC)5OMZ(bniw`C(1&{y|c&w~R70M&~2X|}GjHwClDKI>Xbzlh{q*mVFX48QY79DVxb zgKI4*)utV8@00<39&RxCNQ%*?!)N^UJq?>VF=-B$iUEgVf$WYD~B)iIgyuW`*)jVr}2vf?x!QmSRqJ=k1>{gVD9n z%PidiE0fKak<-Cp>fd`YU3X(RSR{FZRvJGcSZ2lhDP6m;`ACB>2T?zhxvD4S#wpwO zxdPX(ryG>`^xUp}h?ZHHSj&{%l!ETRwO1>#95aJ`4&P4_dUBa8+$B4Bd&DhZ^b8w* zWlI>-DP!)^33(NG{(0DAAvppaM{$>|956%qVWx$Db4 z@^N%v+E8!6Ky+rQc}bHF99x_>Xk2Z|A;}}>>E8@g4s9RTmDN(e!gIU58O+q+SZRhx zSH7=XJK!~2^zNOPY?V!b?BRQ#y^T%G_ZAM#WCrk`x z9yMo8-e5c)W3mg#@icxMA>5!#{@S(kpwp-_rQyXL6TWZ1=uq}pRer}g05s7L>G)Z4D=wop|u z#Ogw5_2437Rq?_*jqM~m5J%eD`{?r%lkHAas0yCb{g}_LJ38Y{YBON(p6O0+v-F~g zLy>FfREB5JZz)M~J2h>y8+!Py=;W(GK?y@)D}S*DA89XW87%&mO*M`?Lf<*qn51bd zVP_l{D7x^g47xB2<4QBGhTs^-HrzIzXU100d*gO`de2Kx=4*x@*})rH2CW{pGcE1~ zoKvfdc{UqA=Cx(+s{5yP72fv0_JJ}Fa|CrHIoMa;P!$MaSiLDqrn6P|Yxrcry# z%zqwQN+ZcqbQlE{Q^YJZa*+5CpTn!DKwPS%l+>>#*SeBy7-kS*{;lOMz)22w6nuT?h5l9jO~hr{j?Jj}AUT`<9u@w1nq1 zdu+a*dgauLsLGH24XmE#$uL;QsZ_^6|I{pAQwoDlsvMOC05qd=rfH!wV1~rXfexN+ z)^vy~7e>d7E+SP81t`#+uA7G+f@Z$npOhi(8_5Vc6KF`-%aDQPXSwaoyfL;O!E?4A ze;`fF$uV}wHrMaF;`~@l=b8B9Y1B}d-k`&i~f ztbzz=mC%mZ-ljCUMkkh8WBeE`h`7E;D2bQ_=s{&lNoa@W9mIaU?@Y=xw0t4xH`F|H z=WCi(S{Z|aD0WpWK0t8psDiht<1rS7q!;E~mYaY~f z)YI!>chu?AgfK1XM_`7U$bk*u8I_!S1^v>lQ(qu$pxoW<8@d|V9e&GFdx-PZSuFhd zR*J_jN>qArN9GC~sQSY@$RW*%WT!SlpijB-2vNLAgOr`0X3OBg{%s%8sJY0bgH701 z4je)-Rc{&mdH{4kk>bHwaA~LT%9Y~Knu1t`hm!YFe>fIh2l+0SihI0B+qT--o)FD0 zw9fTJJ8oWTjniq@pgP3ZqP!)|ATI!i;fM}&(h2T?%wKr^;gV6LK*Q*q4qC(Ius64F z8PAjGyKw$}+heVVdNvr{FtafPkV;Ck2MT2C0j8(EjCa4M`Xb6cr+0`H^t*|eo3k)& z4i1u>iJH>)bR>k+)%g#y%a||QPgdL=<;_y8H0w|Y9`7)Z7IzH~2F`S8SGBL+zb={A zZ1mBn&%6pB$04CT%Qd5pqho)bNvqf58?_c^HQtIGju*|26Pu84uBaLYxC{IlA%s1$DzQOW#9n2UTO+`eKS(9UbDGJJ-dw`$uK^Ib1tX(kb-lxa8-m zxadu4OzrjI{v9ThfX(0H?gMnwn${8|Z;JJRgs};B_yVmw4!h}ZW`7;J%R5u~RrKuZ zn_{VBnUx-m3+KRUHobmq^KxyGb~!$bzGjnX6W{1YQvOAjlvnd;DqClc=4A$H=N&8C zP&pX_%}DuT{T*GHScZ(NP@JN?sU1+k@^H2wh45sQJwRsOBW)RlLY#nRGPmqMoQ<-( zS54)umJJ4r@yHqhGYJ`UFbw1?6~-%fwh`Q4ly)VxwJVU%#6tRqRf*{I{wxRPdxn$~ zxy0&%nssiPIP>4Ax>o8YwIlhxI84TYJpjGZm@wZk3Ya#-*v0zih_@^;f%Ht&BTEe|DafW{h2N^Ykh-g_EEUs1M&rLxR z>Oll>syP?rb-DgBFz*2o7AiVZSiSr=nurvW!aV}954`_2@B|wX!onTB7yirXl<{^i z5sT{ozl$Y?m;YaqeVP2*Gj5;^g+leNv~XS{N~v^>EQJW$tWj7f{ttXcUx+B^@XFFW zu3fkR!K#*-*vgx~D+e=ubT7`D@qY-3*FRKy_02~WV_Djctq>K`hoRcR38bBCWiwYV zW_UmrN@AJ!&C-px+91E1y9XYH4d_Yv@7RaZI)#u0$uLg~mKwdcw^xw|GgW4+_G;OB zL-h1UUZJlREtMBVMFzdQ%xZM9FD*aWUtq1``R_VLjOt2BtlV})iD$)zZ!Rx-`BXx(I-ivH5;aY1~jCK-Eix!J?QU{FFaotl|+*-bn;QDCyp zqNXabu*i|II?q-9+^g`nfPZUndQxVl6KuaVAuZP)Y$u4j`<;O}xuyuI2yR@cXa=R^ z@87l-hjG^z6-`B{wzhz>ZjxO}mIyWS1Ad8yf1=p2*cl1}#{jaeY9#gM2?Ex@lk{hn z@Zi1&qC~bXQb?Bi9_R3L(U-n_^WZ?9tM0-Ez>pFd zYr{o`*GiS`U1`gGpIui57K+O%6wz%q8CwhdARnhJftU^BME1!a(YDG1qUXgr|6)AX zZsc`bt3bqyMVl-SV50Yy5v|a@UPVeVhqeHT@rJ|G*p$MV!#<`j<6^$jkbb5k{`BV= zUbgSq`6#xcT4yTMiM(lgtaNU?@B%FFDsIW@Feau;be1|9vwreCO}kX}03xsTwL@RR zfXRDx+KuMg)H^C+zO%lg=d7#!%V`g%GSAQ!W59Gdu0+0e8PN6yZaeVg!r=6mcURNy zEjNla{RC^pYH-Ql32R&EHtQ(VROR-L!0Q^O~C` z+kU_Ye$$&RI*wh|{vVKi7s^&TLWY5*36C%Q*u5ed>D#w%R}+mM5Qr>+F*^UCr@-t# zto5+-kVfuxzrXMY1{TjeW6S6uoli%3ttKmkY>k=YP^wYqAKaltCTvqAuAJzzy4m8r zjkBsANcxpvq0rX79mI*%VEe`sp}ta=6q@<3>)dWniAs}qo0PDmHt(hNiyFG9Mh`B7+z~OHu0J0tI=$G}YZU#s6Go)z zx8U+*+Qrs1AYnoXgR6IQtb$`Lm3MSA2gepHw_xcH_%GDKfK(*9QHMAaSft+a1T8j> zl&(FU2%K}MpG9%eTv!<1RtgK^QW{fd-fGM9T)hPXUMyyvqJQH=zpXxI(CQ9bw;^jK zxSHIU;+5aY-gYzPQ@Ll9LGvtr_b$v=a$`tppP6)fp>0#wmFfEmR(E>gjNq|RJMWz< zSBWHqD0$F05GQC5FtHO{x?*WXoFpH_ebLZIc`?bJL=2Mq4zB5gYd|QwABW6JUr86M zMRBb3qkjPfO;J)9v3pZ0GBkrVbWfh!zC5%iOQiL1nNV2=O!{XTlVP@=?_~z-Bzz8ZQ8=su6pwmIQw?#F z5g8HL7$RLXdEYE4b*5=+-YgdJrp;oZ^u>D*Px792FT?hNM0!un|4^h&dZ8p!%l+Ga zIi&+(C{_6sS!v=@YtUqGgMrTY>#hhU1EG+TYq;Q&M(Kv=E09+Q_@!5Abp-RmxfK@# z(%$ug@oy??M^mjW4g6mQF z4w3%51=NTuaGHS(>{PyK6Go`|JZHVQeZlR2e@Yq2h?`)%{aduvn`7h(Ka*14t4sRD zCh;7Dw*SM;tA!?CUR-~aGxd}AU)B_+D&_d)F#e&`E}ZYT;*o@6n?MNrof#0x%CeiR|FzCR2O8l+V8U z-dCYsmpjNt^}W}gjL-%cct@GTZP>q`XVc+mU9`(7ox;Td1l^b zky`z9WktWk|4G^NvjvmE{zRb^iHA<}j$@0+;z8Luae(<$I3az`{{wWnY6kWwmpv;0LO2_ zG_(oo?E@px)o#&O_n@fP_~kF6$d$Us(=(^R$Gi5j4QLt<^=4q5$K=NAEb?@s`97s+ zF#ej0W%KW8C*m*tRREc!RS-2U5H3^XZlvcqJspxrt!cBJdnL zeAUz*_YF5U;V++$ovCvhRdnx9|I8nv1gDgC{*el$Wt~~AF*SYivcxXNwl|#4Ii}qK zJ2eY!i`y>m9Vu?~*HMtB2_4Lo8R$9}Atic={&b=edH8&OV0ocLJ;Za_5=B+xvie;8 z?4do|=G}%kc?BQS$uH$5&9QR)fmD&^lSYLPpig@u zq6P2XDQ)_&DNW?9n^yyG5FyH>3xK-UM^3dtouBBQ`_n?EDxN+g+i7iu8hEy|*FC}t zG}&1?R;xIztc&kIB@*ZZ2QS%XuC{YXJHNvT+HbNk7IicwZu<@|Q`=UP$2KVTal_kA z`)lW0S5X>k*P8zcY2wmt>iy@zeqEl(!AwnWu*Uab>c+IpUcF0Zy;sOC-K3L(=WlGb&ts0ku?E{0V8NlOgS8%Tl{zpfv~r8fzP_5gg_*uRX+bGH{)nJ147 zC&_sySCx3DqU<{Fju$KPh2f48vwgs+Bc72D?W#L=V}0M>pjZ5oDY{8B-IaU(@HB*Q zl#rGKGm^{kD^}>}H+Nc3IcQBJp4HUEYStERS?yG56@-ssp8iPnuXq#vaH_&mFe8HM z;6*mQ)3g@rIo5h20zFz;J|`P&F@bR140oEUkh1<>Q4^u%wNq1w>hzct!dM)9Z89iI zjFav<&V@k+w^jzor?-ZlPYqNfR^+OY;ug{`F5O@SZ~~b$UxXS3104pkeKfcQs2k_4 z_qkq_tb&q`b-!RH_M@RkS(m51^-eujzZ2_Kgf7JICg|OL8M48SDTF#$)pGuPakfCa zav-x1voVpG5&YhQT^6h}L2kzB>1BSA2=*}vavv?W9=%yf_?4);+5GhDtsuev``JAa zk1~$164tflK1f?Z$2o^yx6H3EXUwn$L2CG+rHn7EzWOG@JI&{VDaa!Qeh1IWgHpxb z+_^zf%rwg4x_IET5$PYqJ-J8Y5V34F@0c|GlnOJKvLSw+m1~0?f%gYKi#C|L>pA_| zew~8e$+M>UH!-Wg>xXi@7pk;(xe$j{K6eQ^IA}%>bNT9}qH=RY@#uP!y?Mr=D;Jj5 zfM}-0JH=-!8YK$`Qx7F2tA`@&wW7b(wx%ou9~BU@)8DS<4&ix%`Eksk*cNNkKoXuwIdO^1&_`ziTL=ZIi8%d67pdmhS8FX_bC7*(1bXWS^6H3ukBq zcIgCVVb{%{%Z#S^q?!Dri+Z~D3YwqUva&fCnM|}Fnl7Z3piT8m zfrB$2mzNurai+n)S$GodACj0+`s(tTs+R4+9L0`Z7d{a!+=yn1_y#1u2(QFhG>cGS zQw{CliLc5)+#9?#_AHf{V|mdPGWBfUy;!QRTx=+Qih7%^K}6!;nCKzx7l92!FK!kv z`dzb(z^T8~u+z|jsK5yrrp+F?S8{@7^Q+I_gtPE5KSb9Dh564{?g$M%o zQt#M%dXYz4Ku_F=$u4Bn4_pTWunCmpU7%5nG<=Eun4OWrQ#o`+n7lw^ z^Lwq?&^O`rTB*2{@~Ozx8HWfe@!^l7v2?c2&gaDyMXBFqD-9rhSO|PXH?v1uIEGcm z+n_yz<7(Cls@eHIi?;SpeHy3Z_J2)c_4GFDMXd6WSV$#IR#I3#Wse%Dh|^~624IfC z9&<>=U zy?@XNQ1zt&Ec`kif9TY*zH2A%Vf_>0l`s8yGJ^~XP0)Y_ZDup95ED_C%sRg1I3VR3e+Pde8iyUpVXmg? zG)d5H8OZF*b|%g8WP!=nKE#8GL&9(x1{WrWxX9} z1{rvg(CQtZn@4x!$J*KX`(S)Ss&^y!Ns(#2z*zScs%wz3vz2{lEfLIiZ0WduFdg{Z zK&qdl!U`jgrTnRQerCT{A=bV%C%+MdJgGEopPI5TZS_czbejEg)&ev@O8){*SDD+h zIk8<0u68?8l9|AsBE0fPzsM&^9GPplEWGDB4k-0+ z1T(>?z1zE12b+&dTi2Uw9h1H>HPiWK-&M_n(;5|`R0q*bNdVV&E8yvxAztP*fG6;# znYzXXxsmz%ALjv~cTL-V068);>-ti}e{Bb}?12W}^T2Rl6 zTKm}%uZfY#1EQJG<)_N!32fKe`@pGFr4{DY3RM!BDS+;X{In^#GGkN1?lH$_dViyH zvd+zMxjPkLJ7|an)`9}T-dEU15L7!>?+-F6S-U*N!=406p#U84aqK2S__Wz=d+wmV zbk+fv#|hkHZ$J4SP2C6a)vX(6v-Zk=*S<>+J(-kwSaL(*6#b3wd@aoVXJk3NNM^w; zhSOt&bRG^{O6bjLs;@L_s>#OSroVccSsz-M5sQ5ENidbo3u&;lyD3#AK22hUE&w7j z#4AfVm61jUB5F3RswVGhP6F4M~bIUtKh3t9B@zk zM#&+q%+&HI+|&)?c`iMBz82Fw$I_(tla%1!SNcdy=>9prSSO8`apv%X_@Cy7iN;5V z?_=#*e9Xw*rZes5_BC>^m{p5;*l1XbB8TKe99m&^`PT;v51#kVKKvYhm?kX7v0Zt% z5C&U4zui0^94C+aGHz9};qp1Ns>^Ih@yJtKlb4SjVNxlQjo?TiJ4ir{xLiWj%_M{^i(Z}#)4d2VCi zGKf^bY=d#v+>Rh&&+!z^Y`dWvhi5}Likw$>E+y8iG6BEiFeYOQ*ZC&Yy-AEr2 zI^PkQ1%JJ77K)wvvc_G(ma+%Sx~l_(ftwO9yB`Eeb9!&xIb`(b=8BW;PxhYfW3vuM z(>p013>*;)G=#r)NKm3*+U;{aj63?tJpA*?>=(Z&&$x#NMcQk%?=?t5C$2@p@Lwmq zXak;j21i%X6m4^g`k@p{sSdp|9SpVhVh`WDwEqV9#H$0gpV?z__iD^JcRg_h5?gW* zmw^4p8ftT*87#Pv^Hl_o^O>OcWqABraBl|nDzWOPPR!>l;bnpBYD0wm_pLZ!eddPd z(}Y*PqnyH42+18O2`kq7pB9lvx+&2heZzD8{}}6vg?NZw-M1+-`?n8l($WS1Oi^n}w_qTl$oaf;)Tva0_ zE}n<7gv%Z8*TdNZH|DV_5-%{bD){wRU!%`c#)r3Ar2R@Zr5^Z>-ZJ^gCX&qnYCD^) zZDwtgM!GB&mW|a<^X|so<8JGk(&oJHY?Qy4?NTB&d{AT*7M5VpdF=sS&^zEIuErZ~@HVl5_v^lyeG$W(@oV)68E9Rfb}*=L>|u?8UU*oil|4~QmJR80@SlBgx< zIP?oA$_;s%P$ME5*C1Ta>|QA1NigA>*Y+BPU9ki4f zF9b1Br!w21&@;vBi)TG5>j4nP`YADzmyY*;8`8|PQ1|*%2)$QwJE*?wvs>Faw{R9K z8d5T7r|LBS2Ec|?Z;mOJg)~4T21;6NXOr!e&aoiKTk|b)kkhBhxDg7+vg;>1eI?d4 z5nBt*_Fr5v(F2ClsAcB_cbPdC87@2U-NF*K(>8F*-Y>R;v|w_L?^3JqS?*EwCJBPe zwY1f<-L!YaI=JQ;DHE$DkRW(zd>G|7xPA(?dvEugYnI0`li{QsBPIo+wlI4JHg-s| z0%hxMt+8T>=$6*qcu0l;_#^vswJpD_wt;~at~mLomG^R&fAJNws&@euKwae^H7A7Q zh-Q*>`4|rfb@;$g7r6ewkfXdfaBn89X(~;0`ha~Iu(YG<-K0M@`#fW4Yxe6@{D@dS z@MQmy`ER|<@r~hUbcvvJ#{I0Zv_?;Vy8TyXf;ny#cbwuRK=y;Y4_7%G)w8zC#S&(- zq#Npe=v*9$fI2s_Eh-jXWsBR2Q>+)W%MB9mt)$08dyz-b9@R?sX*_(}!wB&ig8h9o zpDSWmtYAm(`rKEvN2M~T1k@o4UksGp2?-kfzRh9#&2V^tc~v8cisRsv&}h?esI@># z3*L8xeA5-=ywTn~2kIo8X_$XDu(rSuYg8ib?+s8}J##Dk)JsroH*pb(?4J^oPlE*i7zEyT%UDe}smA_>D*}_~P-32ekYm ziUEfY22x$+^2USlRO%gs?TFkhHcgj+O0DVHtS;9bd*TCR!Ow$T{^#P=x4=$yraTMmafG@iT0;o zb@k*LB#rJleIGwD27Vm+DpE7b>KVV7({3HSv~C=T`2cEzNRq-Gu(8q$Bk27&Ke|Q7 zF?Q#)n18Oez0HO#m7(A(_PNlHGL{)MmHE5G5|o)|`O<6|t~k_giD$ptlad&1)7`qOQ&IthwwKcZ8IF?~SB!i-ZtIo00 z+qtC5=bYw9H^Hzy>vjNorr)Jn?Ah7k)LOZE{~~G7mK;HYJk>tiJ*olY_mQo8&)6mV z_@H?KWWq$*xH<8C;nXNB6df!7V~Udo*<6U7eeJg@!GP!_%ys(5LW`~GD=AB~tP2UG z@~}abgI0WOcgG0iXu;K0wi%y3w)u*W$!U5W6YB)F%eEcq(hUSABNx)nI zP|M_gQBVVb@}hXbBmT}N!-Iw}%VOJSg(NvHc|i_%YVWbp=>rP}r|G_J#my9%D7wY6 zfywFlm(&L%*|8ER3GY3wL#RE&Cq?3UC`6A~GRQu!ydFXEvUP(`qWF9^$DLXnrhT$Nzv-hH|SK6-P+A|%j7%d`D zXO;(X{&b(Xf+w^NIqDA4^vz-#XAqT=qnQ8(#7&dxqLQ+i$E;f9o+&{=;JariYmDUS z;N^Wpx6BXPVxP`(Eo*unsQidv?v&6kbJ@5U8j3_SrA#%j_DkMRD=T}kQt^ijCKht& zWab-GJE9!o4f?I~z`}ty%4~+5pkg^f4)i#U3%?#QY9Ub!Jwm+26%uMr!S9DFR!C_k=v z)xz!mSd)GB=tHlXXg9}4E_qb3RhdBVS8L6&qK@?yzX7=rdqM#v_z8?|F~|nY?_dxB zX)9Xal4z>*DNwj{eu1OAU;Dn}9l)3|7+NfC=`gmZm+1wTEfG`q-khRAeu<6WU}c@8 z&XErC7AtcJI4UDtZY^v)w(wZ#7;*yDXo~6MT_NE`h@f+V-q>f!-Ld_F>ht-vR(pf?S)mu)v8}C=6drUqFR}H@+siM_YhNSHtI&)s&w2W=CbChr zsr<2O?7eR;-B+8tf^pQ9V|L#sKU&=i@;{0X8f;bD+rtb?5|6>nA*oE;w6PzCjHu(C z0$%bmmLS#xUT?}NQzt}99*0qV+IKxaC!9slP*HX2UrN)3ZcCKM;<8&?vCMion9XatOSd3rXyi7`R7ir z9AWBmxQgOCNiNNUN{>8&IFZS8Bz~4RpNh=&Ucn4Y8WqEq4Vfio_O#t+^M;thd8wMe z_$c230aUl0ewufrVqgTsshCZ#COHmMwAk#yhd_0m5ZKQzbZFniaTw{&8=31HUBge z72;(UeSlJY)?zl@22@aJv$rfqeuAZc=p8Y8G5gxYsojx|HpqR#_B?&G!*Yz>!L{os zG&AN{7M0oUsVCzGa`kSr9k4vO;XKy8m}AE%aDM3W8PgQ}ya0PvT<+3=cXrHxLY$)Z zUu|LZjGZD{_J8Y9N^v8xeU-F?IapWFC#s$O*?Hv zm#O7iZMp_;3WWjsM1QVh)DCyF;F;+lA_Ft4^(c}lZ}dHq%gbwOY<{u%x*Q2O$50Y$ z?q+Ll7|TSCm^1@YQz5fgOz>N$r_t3bxv+C)1!`aCX3NL9?P&H^&RDyBUz6X6mHfu4 z15>q&Bu&)SNYN;x38tqg8p)45mL!-#v3d|mb$GM&V`gHYp@r2|n{^sSt^mj7?YiS1 zp$Fn(!+qyv?Pu*8GhoD;W>!<1s!~r2r;kXdogpbc{|TeqILW8s=MQE%4@Mo6bh^|n zA&Zy`2;Zqb=L)WPp{oOCW_?X^L34dTcEKMMEXM`YDIDC|BbJn@TXYez`}wwSe6R** zX%MZ&*jcs(I7mqu^odx>d;ngsbbXcau;ZV00^93wD5 zkCdER9byy+*M0h)d_Y(POcud^gpGxLPX)&U$FpMSU9!BgJn5G@_;iMwj_)f;)hk3h^hiQF=apD`!6VIF@ zl_*&vhpi7;uGK%O*Ez8bRyOR@ydrn|t9!&t7Qod?Zs82lyOKPEfXD0a#`T)GbQWY* zo&(hI62pUarcQ%3PSy01K_cs|4)rrwe}*Q)_R+S06i8 z2sQcA(7mt0Ru-P_|ASzY)4QZjsItwlBfe7M9C2l9(~dR>Jm@GgP;SR^48~fgv1TxR z&Xj^I5%1&3VP`2Kpury7fzFzO8khS8Z1}E;T~{`c+h|wN-C{!H_UT~EXD)oexXiQ1 z>kWpwR|%tby|Y^I*n6iNo2~)vYnK|%3C(KU@2{vy~ZZY{zaBKOPl* zbmB&=V$ux$f7pA^aJKvQ54f}_sye8mN{3ODwpMGa(i%Y#YJ{p0 z8^S%j?p_I&>$A*}W@l2Q0PLU^2*p)#AwerEf#)+82EA?0s9R6qU5OXxT#Qh4lxEw@ zqqG-r!v;$_<(1Wxf+9hp$X+w?i4@8e!%92ndY(!~x<^2v#E^m@0jscA zsg!rS^?UVO?WVqi?U)8lBVtENPIsQkW}aKHS6q_mG;YNhU&BbO^u1=XsJ>9ivqWT> z?a?2HuYYy@w(>YA5y|qgXQr5Pr-4%hQVK1l>)YAu$J&-gZ_(~0uxkrtmBW3;*Pb$y zPbmSJJK2X`on&(*M-m|Y!0nnFSVGC(!$2=_k8&;u`<%*YNNvw4nv|)6O#|KT^lK65 z9#sE5lR%ibTg1Y%dRx|r4!{e6Yn+d}sKqrpXLj3dV2ZkBv^A3Y#2Fr4AV~Mj4V_~F zjuLoy!3)8FNhy83SkYe)DSA8>$DR;gHkc<-Go9_ueHD32kPnwvGFo?bgnjnmrOOw) zmbtZV8U;ePR6TA@ghn4-+8Ekfuq)49NT&Lm3jkvvyxNYiAxPNE-QWUJJ8iK;FeBgg zK88;DS^1GsnEH*=Xb{?RWhDy<;6#S2;t5X60FW$p^ib?)q$A!u-~;?V;*#n|VQ|x^ zV?i9M&vKB(lFWouZx$XOx^xXmFCqHrZC|FL zzSD!~zEt!DB`a;3PLTUMno)(WQvZxbrLTsJ>%m0VNxU^-xpEU{Hq z4(xE%JT7D6*Eh8EJdoI1<7>X4e|eb#q8f?-yx%KD_iC4)vVL2?a0}SmCv6y$scql- zGFGx1h|HUp6n@tysy^`8!_#BLmUKT*DUbQ%hMmc!0ps;W-QnIhY)6fy(Hj9&*6r!t z)l}cMQ6NfRgjK^_O@xb9GmAC7RTP^xcb^#FQ*}xUIGmq8IxKzZBXP7~?}TZsAKY{+ zsM>zHjo_~5itLGiHh-E*JhOzX% zxAM)uU z3)$?Na|NsDrl;hIo$2mP*Q^{`kEZUYOFF0P5>Ul?-9M-kyHXpUw9tY55XST~%TMYZ=VeG+0P3gul+jWm;*hO1Pc9_?7{()CDm ziY*Zl(OPd!hK(R>_Y&z)_E%%AY6Bt=GaEM`p14I}jU1W7bU9VzV5;D3iV`BIqAnr7 zf7%9cq^Oo0$$H24O`GPsBShs!QpPt^LY`jse}5i`$;v+#n;8QT6ltnx+}GVp6$h~k zRgiY+N$s6=Y=O-^!0q_lOh~^4fpR9EF#BPA|THm)Y6GzM}4}YVNhAAQh>0X^1#H zF1J&ryOAP&-`mVaH!WC~mSfn_CBEk5Twp;S_PJ1ZA{Sd-syF&t8~YW5#S+^2xzN@~ zIq#`6#E`I>wqdWK6Y-!(g&Baw1Inlpbi1TOZvHGaNO#L5V6QPph=yU=qw{C!C5vKeC?8h~V^ zrll{?!Vh0^5e(E&)CVsfJ{0nrU(N1b{gPj{DPMjpMxxJ@9z-+WCmXjwJD)Jw+ydaK zOvSKF;H{$hO>9?3X`Gk*5k7)CeP&}8@GhSoOFE5n9xhG$5j(`~Au=AQ z)8ww0B-0_a6N!-#c7xeP2u*dXP$HN*BN`Kfq~jhMNj@1~j_4EaPwGtX_zKwk%#|P9 zPiSg#^d{0P;M4g8*_UEolb|T;BE|vbHPWS$HH6sPsm7|-XH)`hVCr77OR-cmY?`C; z`y!h(wi;go3j2-hj>9Tn`#e66L;Cx;C(9KB5MIqMkDt{lBhD`PxwD1KU?^8gxVD~P zHI-UDN!R_@i3z41$9vj!9pe^#kp({ZOTDj3h8l_=5>ItiO}r~Her=KD+aip0SDx-L z9y3c3;a#dZ@R0cS66LV)l5#H5cm4{IZugoIcTcqHin4EIV&5PovBdnNjMdl;VYeA$ z_Juij;-RP@$X>yFz}V*iZ?{+M9^ai=-<6JUlVx1W0B^w(+^sS9yzee_Y&ngfLb|1+ z3|BxMDtik~B9}HOrh_cl%MNzu#u!=Zuswv~!?NsEQ=!CCA}7HTzw`z0)kQk^2b-5H zZg4#+xYjLTxSu3!)pILzCb_xC)FS@fW08U)HLEMe{OK7P551N3)h@){kC^$ciC>;g z3xD5lE_FjUqC#of>j2W3;@nfqC%``KO~zQ#-6o7HOsQJmpM24bkFMr``?7jl^~5#f z^%p0M7eUswiZQBwOFBfew|%KAS=OrOi+rY4kmJ{mlSv^EFcty_z4W@9 zMx>f`_^jnJ0RFT*Zds{rDKzJtR6Kh9v+ZjBVY0cR$F|J*(gNNCy;6^`eN*U%163estfrjjf~D`1l)IOK5qn5ee|+kCzP(=QejK8eWYDHk{UE*j4Q$G-jL+ zU*IUHxhtQ4$y+`J=Zbr~k$UC8hDl_6b0aSMaIVw#U|}D7xF#-C?63nQPdRT4tnUG; zbwu{H7buC;Fk}lGoC*N+Y22pZjc3u>$#$6YVD{hya?z)`N7oIaO(nM2H)ZPspYLYq zQKNe*LdzZ<8R)jmZ@4?Q6zgmVyN?y&ZWnlGjoz2qAHc2w1QJvUsPHD|t})tl-%+B? zFwd|9x+nZk>=NB76CEBF>^yeYEZL-u+ZE30t22=axptR#g+E?i2uDOmU!%qNG2<3dCXJLu)%NLzD9}|Imz8kL^he>j zPbG(OkB{69CD)+t4Z;f8_1Yw96sGB50bk^;i#X<2v958L+HL@6=B7 z@Ipt7?`ZTu9<_ef=+9ZgsCx90Cw4=35Xb21znm1PzsxNtz4vOp!l7K3PZM$V zOcn9U?botVw3bhiR+DTXU*n?zzo8_e({m~Yt*-%#qK&pCXnr_ z48b~QDo@SU9E_ISd+1!EC-yW*6k%J{xOrmv(yb)Nvj7oHi7!`NiHrf6^vbHWm8hy#k%2MI*6Z=_k*_1Hs5oGxF;e3`sYs z$G$d~c0OMAFvpA1+oT-V`I6g#zR%r84kVx^pE|*-e5|AE&;mL)Uw1(OEQV=-emGm1 z3)d%OX>-DelxC+;5=7yG@;_6HQKCn^2hMeGp3f57A;*)lZ#n=uqP9maUm=|J=!ut4 zj9L62R42!n9qXuA@p3rlMM@X@N_!oEXza&-fziS{oBI1DDk(?A*%LX1Mn-swrExr& zh?-APbDK()m-N6S4cxADDOLQe=+6Z4R~3GgBTxiaaI0>9*>Q5qr|s+?`FI-vjH+_A zcU50ZD&jfBge`-9}dHASes_gaJ0lf?bT~FbHfCLE@9cO2ez!icx+pPqMP!F<1 z`V5pQ2^zxzrugKqO9^xRADQdufqeL!P?it81R8sThpSBZyu8R#p@@lkn2BF zn;|}1uP^kZewS`As@f`=?UfQK8yoH>p8PO&PN>+Ulz*`O)Z6n`$H6DSjjGZ z5!KtwI#tB20`wekKr?ju89uEG`j&9WJ#OWs%xEY}Wr?fsSUXknxZa*BE-954hHY7*JDmE0;?l{y-8$`j(u;Lv{v{_g_$jf2a4K;mHQls-xZ%N|Ex%0XylRFVqYG z;8fP8=;uQH=TCpVOHK!jW@X-{&0qfX*T?G`0J@gEu;q_bbpBBQSpQf-l=^kSQL#n< zB?xp=lX|<=(swXHo4cjupS>{Sh-@IMs#`Gb<)wIxc44K-Y^{Uy|Af!JIRk(`D9m8{ z_uXU4E{n?~S;(J71%ZD)?__xqE(X6UdENt7v8)`P)03|>ku)(ll2ZW=wll`xD7 zDkf3H@h1A!zTJ$&^QM@2=COabGDu=tN-v0U%+amD`SFb2WF^R)rhIrnHdy0V^cdrOHFrQtamTSB)3v^TU0H>!XPmIlQ?e))KhT*_)ko$Y;q6z z;T`?fB>;mc8;OY_WJtSxV8ddugzhTIL^>0oI@9VmhKp@S(|grskjcp#jmK z&3F!l#}0dD86Fd?L)N2}qs;T|mZvLmY08Pj&Mr(vjw6J|i)eE4%WI-Ud4ied*3E$)9TDt2G-dKD+o zOhS=9o_pE1IaI&yl`jd_ zSH#qLN3Yd9uB5H6NTM?$$WyhV2ye@X=m3x`p2;^WDkgG9o2S2Bq5zJ$ENQ5s*37_A z2Z5nFHaobldhw{eA(1r64xW79OH8rJ7mdvI6yj}ldSx}Ti!>bjsxEV!tg2~40Y|79 zY_-@$+!M~&5dH~f{H`X8XotoxM#c8(o1>2x%HNJn(B_Qd}$rjTrrD~U?+i9 z0YHFGT?|HF$2vaY(--2E$K5|jz@qnR$$5Zqlo%RVj4980YSWJTmbQ)1nJYO=&$7N5 z?tE`2@{*MyDtQ|JCH}lQ zFi1S0clu)h@%7ijdz3$q^4$MBCXr^0am7+ASC$~b_Ep39gpg_yvjy9`o4vd)8os2qEY`ajSjyj*ZdOd{2vq=VcPdmmm3*oK&l7&6ogHOmy)!jE zefCRenotO71E(@}#tGewE~PAs?J|pTlYSo~xkI|C#wh^I=b6LB33N#mbLq8}O@>jg zbFR*?d%M}TsZO9I=H2M%Vui8~(~q5)LM8X@z=NOljA@Qr)k3<{u~(RjJh2R67L~L= zZjgW)9&>x#bH#Kz z=lV#icFuR0(g_#HB_`5MpiaKKg%%C%hQnB@T5~kN`V?x#_5-jN{*4w^;QNy&f1+Zy zxTeV6XlCgrD9GJ1i3UOq$&@QEoKW8aGIrM5n)Qk-6`rokh14vBTNU|+3J)S30YsL8 z5>#8Lbem?v%TDO8U6Fd`*0NENoqP?y(L2Ff!SB)bGt({XM+~QQQbIb-i!I=A7>Ly} zj*3~-Zw5MUb*zJeFl@mvXouz64XMurW%h;a^t?A|m+Rnuykn1TDHZ}2TO*ueeHpg5 zwcAI=S2CE_=c5uUs0mE>q_F^r{@{D5Y>^E6xz~s_GD+`&dAfdCBaT;DB5Nh0M)_6y zW@hmRDlrc{qM1nBrcTGd%+n^SoyvX$XQLAc^M4voJIqvYp+J82nBNx&eUc)vsMd>MVL+K~~=_~YzjKzAvcB?`d##&*txof{(EB^gXS z1sx&X#*XWrna_`@3mV&3?)AvftVbO0{|K}s@CxPT&bhN4ry36=%qnG;7cqqp^~oJear8D6BzInHK=0vvn>py=W5(BuP?^Grs<+LKvf z%CY*k{+Y%8XB$g{4fOh|Xsam``j-QTnPkf{AJz0No<#>Rb{(_pdre&wVnJ?HKSzq@`@3E{gN`}*8j3>5e%AS#Q27pef9?Z2*CYj*s z-BbZEC-fzR3$T829}jR{(xPb18MINT@-!rE!GEhQK$?Hh$+0=x6{=O|5sC#G>Ej8r zl{gs7z+VVd8?-EkMpm;U8M=<+jKI^ZB)DN`yEiUq|DUK!4h}{xzpO1B#p-ZLU!9N& z;YGi*jy0}=#jP?SpZh0Z^SvU6hLHO47yj&0!Pd&T3neW!E3fiTEGtj8kzmoJz3}rT za-N(K0Qiat-F;;dH{h8kvby*&te+ylH3)&hfwoCIOqN60KVZ6qB2w5@^Oh{T!h4YNG<9<^DbW|CvTv4iatgXDX*Z zKKFVIFl*_RY##x9+^<~%`b7c)21tGQBmXc*Ho&pW*-IFE{`*jYZk`0liLogRGj?(nLw5eIyHub zpw)Ep3^@(Teo$K5B(!#P#Q7c>Tvdq^5d29d?urLU3uurkn-l&TkeH+d0WH(j&Ly&BI5&)%>)^tsxeq{VNrLyM6K+Fj zkoNs4p^&Y2gIw>d8iT`L^xV!GNUBxk$P>0|D_?!lMhzkC)b5Qec3(;z@V>xr$S-8< zr67QQk+;9ozn@Ha2=T5i5z1sbT_J3FPv}qg-8+gqR|}L6L}GY@YMTJ|^oD|4Dq=G~ z9vIFT-*)WQt^g7X4i&zt_2=s9P6JGtmW1*M>78M_(do!O%?xEm^kNp0ta))Us{hcF z5C(e=z{7@2UBwSC&bis_J?r4^rerzMi+jzIXYkn}r43Q5akxzs+;P~nC<5@?)%f|o z?3)1p^|y9wWHUf1+^L3=n$WOk{mXgq%VqfH!lzC$$%kq6$lT`PMRT(Q08ijCSTxnW z`#F5%MM9?;oRXfElY|b%?snV&KoDKJVJp)c7Z>`FThl6cQZO)dBEETP?cn29MpwAR z`{!HthiXirLmP`y0|xmRU%~1LaeL|R2i27>RmVR1A^=Wl&@LT+xegHJ!snRLX6uWz zV!i{|7-A;%m^;ItsP4W3xwJwLSS%4TzH(3U9<_=@&IGtuPQGE}3*2NEpCz9INC(k2 z+plt`hyW9XQ>9A`fU3ut`5tVb*6;fICHT1Xoyyii$3B1l<F#Dbgc5NkF}Q2Mmr0={=)PBBvGG+X_%9vRthN_;+8Qeu#zq3Q4H%|I&_r>Z%&#g%KBHtp^(KQHrIiXJ?GXE<{TCDVvp*4f5D8o=~@Ky7zxJ z3H%IlDji&<&Fo850b>30T-SUKHz%`FMUDY3$MVsO*CRklKPVWkCBP{e4?wo|7d_1eKv>WjhXN@WJz8M6feJM0 znSZaauP^lduxMqCes9h@QB#sr?3ey&q?`Z<$JMQ4Fm6!u6xKZ;+UMCSz$~h9htERU z)wK+{<3b{(jLG9oJ?PrOg42Rqr$LAef9j<HU480;}g)=YM-Z8yH7+Wg?K z%)so~NJy=J?0e+S*%2f4>plQf>7KKsnFOfkTL!w~hR|g;J2YLg?=qdxNmp*dhQQt3 z*f!a-bW>xE5%eqyy>{AdGM@2=W72LjUz3s22a76mK1RvD5|D9dtd-y5VRpoTE&fiz zp!evtFTJYRxr;m)7OM$aq;#{Fx|!R z4Zq!tpD}N_i2{}AL6ZqZXDI@WUM7ZY&TNe@s#kw?QjInRiSr^U&p_t82+kE7WuC)% z!*`c>?Zt<`u-a1wggm_5&Qm(>_2l=O9E@7Y_klb>hakD*hod9jb#vg}9d?=c7FCbs zYS~98Z=`SCe?)fwp&^5b3z$K{aHDo%KJ~EQg|X}FQA0p}D2*@Q=SR!bSER07zkVg= z<+{+iF@}7`rC+-D57!M)K5>(FXn{b?ZT5Q4PyE+<~0W$#e!TX+stRX%((!D09vQ$+hRv%0%(gli`D$fw)s;ovkyhaLt zw7l%QXd)2~fx%AKv?6)W`5t0@mGo1XE_K#C-kvFtZ}mnQRMQJCn>Dv+nVIl;s=-D# z^pI+U=IwB%Jl0M@$Km>?6!w2O1s7?)@YZ2JxQr*yb{QYEzb(RxaKB|oI4rPx{J4$H zw3!}w$(TaFENH;I`+i#|#=?8gaQZ?Ht_@{iUm^X;E7!KUp9sP`G!}Q-;mhy^Ii0zc z$GNRp)Z0Tsd@bY~!$c+b7n#mk1c2)@3xV9iwJ5f$Ch__5GSDH|;wf^;FbEUdlGF^F(^mgs*C`gaik*$Lqs#nT_C3h0T;rp%!3t}XUa?or4W3sR z(>ZE4QvBhU-J(hJh3f{XML2!G?XA&68Dm;5eGOC>KwPd%O5caGtzDR4nHS;z&s|~iP{R1N3>KP5V z5%W=M6l9YkP~5c{=2`p0eDt6d+o7n->7|1ju0I~F&bzw#c2J3XZToxmsHlSI7c@RN z5L%!fKV4>E4+3i5oDVOcN9vb4Y$-0!w0SmL4om;3Dw1rrEI)ZwwWrJ-ncbGar8hF3 z&9-uWx=)@wy}Z4;g`rC-$71^C*=fL`3DySBj|&^mZOt4V@nH+)(KKu?hBn!zWe$n` zQ=0%`jYGqAMx1PnLu+h-bBrrZ?}&$|PQ^hg?cqjq#^tW5t`3y2$n>@g0`q%L`k>Q%(JQyJAO!3`051wbhgMZLI}rjv>|1iSpTT{LE>>5d87Akjho+ zB+qdX@~U)Y{{N;_`?)s$(&J~Sg@*?l`0E-+ljhX?)P#1Ef%Azg=UEB5(VCg?rhp1!r!72-1= zd$fOeWMFsDVdP`nzQqT&?)jX8SRDWME={o)3KcVgxttj*wZeaw z+^(b7A|!q`98aTP5x(&i9NkpGTgD>twqj{oVz{jKh>(`45Nb}nm$A1}AvnT%NvJZR zX>-q)K2C>f#JhjyE$^j9Ld~*QW1FW%%DWkZK6McO)t>PPca>hlUM@FePC6Ir*wj|R zHPj+_?sNmyug3^*s!~w^gRwRV6a$yi6*Af;h`ZcsY0BRtMt8s=Xc%{OHQY}Q#Awn| zS+bi@={jeVkC5#0buO@$F|=1piw(x1!_g-sNZ02jmfco!6Vj_eAFP1og zVeR;>w&+?BykNyZiCyoVC#4QuQDfd_)%RQ%%Mp?Ho3-Nfg5Im|AGW+~%Rn#6%mppy(X2^zVP}R2kC7+?LA^3cbN5NHyj}RN;PX%LK7{BctQx}`g zg&jd~drm`0F?gWphc{)qMhH!T0mLnzJge^V$Ie=5h3l!H;#}VC+&0&g=3wi=qKe0` zet`rD_(-)!W_3x(A+Z*}8(QfqI@uJiibm>$7fw!PXR{L4nVm3Ss7>(}K7a7uI$n6C z;1@SXIeKJ2%7zJc!Z>Jjnv$M+OA^DEhz1l>nVXu=}m=2T$>T=a|RR*%g`73~F|bPMty zPvP4kXbiVq&4^(8wz$;@+qecXXRzoq#I^bUzcEZ3CZ-%S>D&d&iz9CRoCBo@hR}2F?WP zE>Vx7e0}GwtEO{SGt+;(ub9Lpaa5^jOdkKuH2DJN*d$AAGT56cU>B=(-&Ys-K%{T5 zSgykCZ-FJ96F&ast<5&5rhaZxnab+DJbsZU?EZ&Tt>i$MzN&cL>|` z+#2tBz7O1L??1n#|7{`tWIs|oV;`HFmRZXkrHlrD%2^5@vM)W)drE!R(0RV6et-3~ zF+9F_I)2Y*<$ST%6jZk`OSv*GaW5^GMZ$6aLASTyj|K{-ky{wTMW?|s!BHZ;Kzf_& zp6bv|Z{G~}?XKo?qUcRHb$;b-lPm21eWL%q=g*yZ#x!Q7DuPMP0hnvN_q00pi89`e z3<>;2rlk-I;Z~@qZO^xqX{6X8l)KWcet%qD9nsA0d3SbgXKu*uwwNIdEE;le!FIkk zH_mZXv8)nfY>mwf0|ioU2rW&DT?7Wr_d2Utq z?zr!mx6PWEPi*sP+P44X|Ncb<@xSBG;+pD*wfxoU{8uQUm*HD3JnO68AC4_B<9;Vj zuDG54)6ma+Y9K%12L9hw8^=~p{jo`nAc~VKIV=COh+^F&^Dpj)zdk0@(4H9skN?lI zlzz=Zzu49P8X@Baj=J22|6L_A!s74c{Kq)`c}Zt9=zkY&yxj4N(etm5U?84UK=Z%L zU`i(bvPA#-m;*$0j2QoSRZ2I>KUcHVHUR`0@Bgmzx#C9okAeN&`u_(D*pt^Vvssvk zo5MD}fB$|UNh|Y5&kZM>Kqvj5N`{W@)AoFN+&IfEBiQj{W`TRuNBU^gfUPDam@fNS zf@VUI3FPtP*Ee_t@=tGtI{l{V0l7feG&G7cZ}5B*a~JIibqaOygH1etN4I=(X=0zM zuJpI$I@w+(9-dEDP{s?|2j`{KEiSemQwPJE)ft$pyZMcE6{%4S7stci-1|? z2;2Xn)P&02^^N-oaMb;}{UbS&Z?6GAcQ*aOnfGm#zHnaFP~ZKbTij%aaNO?!xqH0H z*_X6GFrtiN3Tak)Y(=M?06Cu|822-q>>qpG@hlJB>YIrd6-(l#=f|uX-jrt)7kWfh zkVww|xw&73`Q5o}Cgo@a*yqk#@wTD@5yJhyR~U0VqtSOh?iupQ2{#1=PDeMp3dP?Z zCCi~h?%3i5Xl2Ewtnk!{i6>H3hWx(|`wSJCne=y`7u8coLsphO=i7Qwhv(I6*>>HruRo+;#Q(O$ zz~k=m%KF8niN&`5%nOH2zwdmH<_>Mk$9rP3wX%w(pDMN7C_bu_ z12l)Oye%cv_`9^`#jPY|lu}0CP%$j2tK#9%;qj2RLeu87(2m{We6y{@h_^Oak%ZRG z1aw&bksg6A5))ln}zuXK}jZmqCS5PcZ2H{w%&C0J|)Q|^3we;i~G^>_BE5E_d|EP5C-pzj4EV?I{e$E1ivYi@Xm+}~{ zx@Q}Xj!2Lw6AGkScWv#M>JLxwY|C>G9loL%anVvAWs|t>V1ObB6GtEueURJG&sJSk zR^5c|n>`0^hcT#OuC$P$WhnR%?Hn;vix!@f@|oM16Od#ZSiXZTbqaAd24^q4dGn?% zv;YKaz9jJ?(h8UUI4DBKY3VA{duTIR+?Ftt%iNQ3)8{z3j-4_B`t6d{=!n$0cBjq# zt+vOkHS>cBl2)pj@`a=fsef!MdB3gOGA9ry8~Pldc8it5A)T?8|U<4?!uGt_5p zHM!f`6bbMpb%h}~y%^;<1gJ0Caic2QKT#&gg^K>NDAj2;(U8+@g-P_>`w-;*(*$Od zMq`l!{EaFt>)tyXQm(8k06FnLro{gaaLgih9zQ9zGhGQ|@_rkmQXi+&jfk)jlP%2* zV>OqY_z1f|1%MO#ZVr38RDZl@ky7!Zt@Gu4`dcc#p{M%;Cgx2aVVHJ^Cq~I2M#Oen z!hE9q?BG&juj}$n2%~cXp+75xwyz|uLB+*?9R!9Bmh5iKe>f|@mo=;niRxt$$+JJ+ z&yWp=)H=cDU}82JGIzOAllKla%S)Yi=ErzyJWB#{ER5W)J;@i7t}`fi3(Ebn&Sy|& zX|71eIwk|Z%38K#8yLhqvTf8wWqUF6m$y@@L>DGX`!JZji_H;+p?m7m>A?Ss|W+g^(^HB-eYvm?Km!L zn{CvGuxi@Dp$dGgk=n|=;2gn6#fGLZwi2)+wC2sDYhyw3)JW^0$_!*MpLa>mfpaJu zQ5`g%(7*=fsb1}~VOr#ot$^-7N`roSBfJyFeEVD5M|7vnXyH&e684SX$2X>8*i|?0 zZRU60Fs)BIS!%Y}46VJf&-3JLb~OeLZhA#qb4{&UAAU5K5b@5@Y-!HJ72Px?Ww!d) zhkoS6-8cyn?MpiS0b0>e5p_`HVNpSMHUP~_X^ z;IzIPRrZxSc@AL?RF8eIGwAz8CH|2Q=x3ZG20;b@tN}CsYB$2CY$%xi)4U*sE>kL| z&=j^4+2UKOa^!+_iV&JX?&R%luUN;=42fNta7>n%A6pP^fIX;+>2*2PMFSF(U$AOyIO<=b7gOWP+s{+mh~7@mKh96PxRJ$N<_ODV$88I6S?#x zj74N6gxUzqsH3fIb9#`P(M!TkKEKzxZ!aR#oye7`7u~c(Y`SStnFcMQ4eN@_VH~il zWn{Ma9Lm;Opp+)W7z5{-Y~0B3U;p-KDtQU%C0=50Dn7_Q%7XJ`wPQhicE@%*^?Opd zx!63Ki5m<&fmn#|DtCw(e4gz&9q7b{4kEZ3UIPWCmEChp<(XYWE@4~rg*+j<1~vo9 ziJsHKTQD|%EuJ~k^8=x6+XQ)bGtOO zv|6rK>n4#$Q1HS>kje!v9ve+JH^+W8(Yv2MQd4Mqm$e1Ii?%^&%g8-QL_SypD;^F< zLqf=J1ic739jC)|;|GhF$E(q=)j7EtTs|6GTHvH?S{tA$EkQ)v<~|wtVc4aOtZ_V}b7dbN>@U(SzvCLe3jUCV~^ z;*`;IT;Wp}=%$2qQp3Jbr5`(Xr9tU|oHqSRpFIzfrpM8%{DM^GQJxb$m7C`lK`eAp zNy%HhffZftF^6TRi6cKmAPpcd1uu8JWZt~`w`e1hw#QU^C@(PkapmI~SnR54C)#@-Nn-3a!n z(>qNDPBwA9fp+9Hm{=CO`Ez&8b`5pG({Z`b<-xej?T9HOMsuDL4P9k058eD|?y%_G zg-B}E#A|SNYkapGit-SR>}f-emziKi(*%lGh&R3H zy*Lx{0gtdy40PQ4V9!z)V~GO0fH!-6eUjc(L8pfuC_aqpZZ+%bv)gy!p1)EJK2&d`cKQ5vh{pD)9d9 ziyt4X@HrI{_7pxjby`@Lo@}jF8(fu?$;XnF3f@aBGOnNE#k~`0!x#+Lmj*GG{3UOM zDX!O4*X|$JjhT_L*vlDvXqv%k%~GHWVvIa*!!+(3t@?6HH#yz9#~XxFC+S4eagcK( zy^>qq+WjH}f?oD5%3C_7d3Xm*`O}YbwIj9AJ}u@4i2ER!*wNMV+yVI=;$jqrm7{Q; ziHf)^uC~;iNy~=d=OC~h^hL!E;3-E~R-sl-u-6o+789foh8OG%YE{{de7rq0>TxOF z)=qPrOq`bl5%h4vz!B2QGbpuMfH0kBlyC=2@s;Mu-o{+~x(~l`X`c;Hm-VsSP^Vm6 z*1o>&Of#%U(Q}KmZ;SO_i@q_7>#=nm z3h!Rsb3O&8n3veyTw1C`K+W;dWuaCw63l@O1(sSIxt=(jYj*Y2E_GKg{t%?5nCp6U|Kd8h(;A#s0WX$77wGQ2xLTmhK`(o2@8_CmzfT@qCLyu`ZAZ}5KC z$DLW8r%roIPwi42pRbqlI-vg0Tac~^p|Xk3j#X$5z>3^7-Y^L=Pw&0!o1=zxbGa70 z>4d-ENS4@;_qdU&TuUP+xu8ljMsGQB>Ub5XY+e)pdIOs+Kg7YK>_RseyWqh? zJ6t+mRX$Zus^<7~3H(XSZoM;|(PMmB>l@1cD1I;+h!)$X%}b5y4=6oaH=MJ#PNplx zmag!Z#&o%e;bFhVG?RBvo*7GM_RPpHW{CzNeGV!WCxgMgnfbvq!P+;O*g`Q$#g&vx zIEaUowBPDNLjeoA7546wVE-Omgopw4hr>+svstsurNGIl`5$Q3_FtauI_UNWC^K^K z!d5SvX0$p7W{Q@WA%*NUX2|a(xUOX9NRaTTc zfNAo?gH5CD;t2P>b(LC_hBAunn;stts_{)1jPNnO_=4j%`=?# zF$T+ILTT&{4o*=8&kHQ_jW(E8m6n`;woNVUtQuzKCJOR5EXrK&-j2@c)c%}*n=(zP z9;!(Lp1$!J(<4v}G2KYZOxA%o_pa|hY%a&1D#`hxqT7|_EuvLuBg|zYDs;)3{TBf& zy9D_860$t=G#=nJy;(&{N<3P95APat8{U@uV5U84(Aplmz*o7hXX?DMIMNF~Ip>s9 z))#D3Q>oJXfTdM61`MUSwXO#O)G@+vKxcuL*<4gkQV)0qPMu9fq$k9#AuEq|1wy12 zMLyvin-*b^yT#Lubj-QEsZ=^{aru~!L^}uD`TTgOp#+6U+Tvn4yy^k|sz@1}-m{xc zW>7px3I6E$)Vxb(HUXEngr<3^NT+hBno+dXY*d`@JS={sR{+=ubCt3o+(#OnObJz2 ztJD=n_}X-$L+O|A5R9K_U_bG}jKU#$#leHNi#~9ryNJv)`YEoNIlVUFUqyd$1XNrD1nA|Zj0SYP z(*;U7b-<1M)`!7phngeRc){Y5lu0v~HbQ}EM6Fk8eqOIjr`9Op{bc=(OY<&PR*a7` zCxdg#8lulz%0P;Jdh!#5`>N;rx$Qp82FQhxWp?HlAqs-EdZR0sb9%G2hJws#s_wXm zA;0Z>QtljOrBIn?SwJjv5o8v?rsU8IS}ycPCg^q7xq*@P9({=_r4=K_z#)0lC95Fb z%idaXKkAQrpqFxL_VVwGh1E0J&XG$>SuBX!P&Wl)YiLTem9{qXV5;SspDiLnRf(A` zxr6&K-jWS`&Oy;^E-yc93Ab<|x8dec=H923d&Amf?xG;BoSvTNm&U(KOC9A@=2@1S zZ(mx6RhuoeuUsu|Yd6oE8n3+>^ufbuVX(G2H~&)n=W1yid=0kE;B!w&(|x;g2=yfJiCND z>bV(O6lwFM_?`)o~BY@d?c3f-Gd^R`*%YRoz}s@-nlZN}&VN}@@u=ttXT7Of|aYy5m!=W6)Zt^nrU-q(b{Quy2Zy;d|1!o_(Fyg|Y< zIAOlIk2|3d&W{K^=OZ#PSHtacxPsWQ@VkJfSS=XF8DiP%0%r6a&-^TWDH58HFMnl> z&%WK+B$D?EnC{}~;`S46<7v!JJ>j|2$De?nPgp2piYM8jO;6dZw?@3{)mn8E=Pt1D z=Z&!*fVQnW%l+ajc^=aHZhw2MTYnKE(7$q>QNQ2~Dxr*~q-4TA!c!n^PeF!HyEU`r zuAX0(MO#E#*5E6!NiI|aEC{XOJ#>~4SjIQvV#C?9m(VTn%>vT`mGMP!&t>YBs2jiJ zVUlWP8FAuaLD$At)KcE@+BW#Vs;pB}?siejO3smDlaGeHz1S@?rzcW^cgQsuvO`BC zMr10TWw?~%vNVuC{K15#HJ0CiV15Bo@;rleH1Fz6id}eDyei{o%V5YTd}v-QNQ5Rz zSnhsYJ0U=cbzjs2{CTbsv$D3bCyF$njM4Ym$;1pn7*mcGjTdk;mOnZV%@qgTys(51 zzWr(}le%b*+;#b9H-1R6cHU6MK0oVXyVc{M6*Dm}!n)|bd24w{&qN*frp^KT{L+X) z$ET(YTGYyq^W#%2RBAC!Ik(>G&_L0jH8cst2q^jxMTsCHLg*#HB%zZ? zLWc#Vw+JaTK}smHN;QISa9>;SPuSlwpUj=P_slKle9!lOuVOi(_eN_E(PK&E{yWKU zh4j9{&xo?6Xn}BKWSm>q?kbu5jq~RbFy&*iWSthlC*67G1z8)1g&I$5fV2aSt05mUU#Z*@ zs^@v7OZ;rFH^)(Vi#3fIewQ5u9KSw!rzCKMskF>k$$eVNiieb+*qZ!+tn17GDUl2R zz`}$YV{uwqswP_Y;+4Dszj4dXC&pc^l{V})1TpH<=(~f~WP%W!dL85+$``WiqNe-BWshP&p>4l>O{qPZBA^^+g}503kRd7G9$mk>iWQ>3*2 zNb!oS$uf~e*aoaJtY5+}SNn9dot|<5(ZIUnG96*tWW(YCOxG23NO0Lhq(dn1dY~lc z#Q)~WQiFK_;1VT)%NLenAu(XELSvsZAmtf|+yfe56uB%z5j1)}_G&`n zS)o8z&vKvcbH6||tXbkFNOUZ$T;LNtdWpLTm>l-EJEbQx6x*kYcOU*^3SpMR>!`MB z&(+D%Tbf&PyX9!_EM}LtL=TJ$_>S13ac{V5f;)8E3F>Iv=*Ho7hK#f;^&S0eB!aD+ zVbTZ9&VA5RxZ7Gh4yurS*bYUZL^9;RcGN?G{qskz)r`lbG!xqbE2Xq9nIT^TJLhfP zR(%&vbB^9gU+zc_kxEh?G9lK`I#`3Cs@aoQz1>DK|*t5m3&IN?YZs9L^N5@4bS+Vl5P)srA3{ zm0{k8Ner3cSIjSo95`bIEF~~ByIUx<=JxR0T^E|;j6GN3spNMVnn$ju- z5L8wgnMAt|vLvtgzFj3N$v*$3PJCbWMhb_v0+?XAO;umvmpT*F!Q6;OFwDjR zm%9&mUH7E+W(5Rk=_$Jc!CEiPJX1X^tTnRp(j-PMIwj2%#3^ZiVS^DR(X7 zaH`L`>bG;pfR6m*c*7=F#mTz#)9i|qoL`uQAG0Nz)PP&tj zSJ6p0HfR0{$#2w~G~>eWBvzE0ZmB1sT!@Y^QTnz*E)aR`4hS>0E5~1RftCPR_!vM! z@GEXuKtVNWpfbOGEhhY9fa-Ohky2~IH4&9wZM~(>h^4E=3<8m~CxBM%-4mwW zO`=ICw13yN;gp_jGOd`)ay@OP%Vv4^KRU7tA`aRy7FwX zklR0%b8Cl_8cN{Qj{du*jq30N8FT0Q?*;2%=P^e`uboWY`5ZM498Z?y&1M86__b|(oW)H)@GQ7snjgyfzT_Fo@NF0Ce8 z9l(sfJ=V=Dy3mBGbbS}$a8-AEr`~ag3~)0>+C4Zt||I6G?%z0y?s}MW|d@w8uX-kc;IM^L~%7RJ}4t zK<>5<9wv`g#)k7{n(aHuv@r=>{19phcwjA2S zukR_SF_?N4iq;EX5nXrFuuy37`WzwXsW7LMgN z238@(E=XC(5ojEvZSdd}AWxX}PdscdI;9B3S`60mHJUsu+G!$Pq>al%Wz4u&1bt-5 z)+3MRMqOq+qv3p=9=)a#+5^W4h2Mn{fQL^BkJM%y8^TCF3b`>s0tyW z{xEZGa*gr=buX`Zo&(}=CwZiqQd^wR_cC4I^m*?y%s`u)#0B%DGtSs06`A7->}2On zPDj+TUK6MC{cbE8l5t$1C10qg&ux+dKab-(9d@V^@3CBq_?4hjUWwu?(l{Irp4lW3I$aEt<44TVu5)&0o7;BZl zZw*%KZoKUtKJ%<}_=(E3IAnb8elw)qFoCHR40q@$JPXwt$)&+?B`0W)&K|C^J#Q~! zW1f7Hb+I%y7$gpBfk3FRq+!hjL&zw80ZOHr7CVFmcek1!E+|#uyg&&CZ(ly1615{vcL2!Wzw@DpS~UNSfkIg=q}VuIVTK$g6D(D258c8{wCT) zVnbc`iklHDjKVGi4nMzU(eEKMv|?I6v1IeM9U`r<)tG{cm{BD~6&10PD|nnth2aq; zpZsSoIzbmVOq@GMU#&xeuoxwdT zXoIQa&<3D6pYQk}_NsAf&)qZldu?68#$;Al&BF4ogIy>YyKLF(F$L;wL4@W&Oz?8> zS!h?2I_UAcT<6D@lcq?sp?uDID;Jiply7it^<`Bq)vaag@eC-ncdcTU!e?&sZvwSJ zS2&a-l4GyN8?a(nrhU`5WBdU->GJiS|zWi|yNMH{? zZ_t{cz2M+?Y{Ojq<1#DWLUK&46(mgWyKS%)iTP&vSEu5~O=a?eaV-j>TkVpQ-lnJC zhn3s#4t>OKwrwR}6ZYjwecJCMu$d$h@`@lHFA=^Y&SMi@Ep!>8_y!KLu0-)5Q@bR9 z%m+WTS1M!V^SocurdkT>Y<1RY!hGm)6d6)VBHI$91l7zfPWDD@GPCxfOA-Sj4S8|S zk=>M!Ew(M64b78IIpb0kk7KibV(p!X;LxsG5@v6-X!DkKy%!4RqVeFKobS=rn~lCD z6sDNEOZkKPnC-l1{i$l9N5EV}gy}`gV+MBf!$W^ja`}_7SXoo^;Q*B+r-FU|1ZZ298HWFoNwiZ_x!jl+%H@Glrb~r__MMA z9vn?*x>KUs#yWy!^+_OS))7tr3Hc732)Obf!q7i^VEl&y%Z5e@DuWBum3}W((VvXc zRN6XpM%8KfA$qm3&h?=EEtT@#}#l4P}u TKZ9(JEn7^^o1e=wba?PDUY&?8 literal 0 HcmV?d00001 diff --git a/api/core/model_runtime/docs/en_US/images/index/image.png b/api/core/model_runtime/docs/en_US/images/index/image.png new file mode 100644 index 0000000000000000000000000000000000000000..eb63d107e1c385495f3ecf7a751582099873c0c5 GIT binary patch literal 267979 zcmd43WmsHW5;lr!un;_W2p)pFd$8c{?(WhMG!Wd~Ex5Y|g1dWgcc+ot#mKH@qz(oK9149xQ6Osc1gHr(mgC2x~0lh<*_}v!_ z>=loxpx|e5K|!L=cGgCw7KUJ8VqasGVO8Y&Fw@lKh@s$!h{!J92Tg*>37CNAV0DOy zLHNDaBPy636NaAofLWedL=udT%Bg28YcY$3NY&q{Qci6x$2a4->AnrzT7G)8-{LhM zPF05iyMs>?iV87zE^V1)OK{P*yL0kqLhsYb*CSznBOKg^&J`!xc{r33k2G0qH6}&YR5}!^&y_h62L|5iHdm)64}4DajP!ymi$d1 z4Ao^t)+RoFG98`P*`G9Mte^a^F^tIkDEED9PF#cUS2fB91n48HbkDRfu0JmsF8q2wLL`aL`@N_9NHa?u=+(|ZyhJ9 zGXcsl&Bd05j!PsFr4>pD0vQfMlE{}9f5X>;$ZX$q$49w84YeIoE&|@KUAFtxH!-D) zpjCQ;H@>{YufKfhu=NSnBBuCqKf|x&>HvikyiC#^^!5a5jSCgrZnZ{|{u38AEj(Ai zw;jSU*YZ={)`s#v;1f4*yO;57#8(CSN}nLfJ8yFO1Z<&KqyeahHiC;N*Bzv*^k6Jl zDgN$AgX>MatMHkrClNNOqvG|<{0bxwlmoIE=zD<$W|5Ejv^M4iyJ4_vt)KC!*4^rf zZMv3p-m)B_kTwa%W%sLVVQk6F1Cl#XoCP;UZSyi>h-1Kl4@XFu511EW6*p=gD|q@f zCSdo6a1L(SdA0-h5Pz|+8OvDk)Dm)rq;?JKi7r%aJT0}zFjZZ!y2h)GoJHTXgIC87`GJT-t`GeuJa zJOZE{Q&z@W5qu1jHFKAmd=3lN+5-dvQ2^VW;gIDs=sb^st@gDvH@lx9e=QQXN~+~- z*xdSnt-N-)$-x0b*8#U!_0aH!Jpl(N!uInui2##pcS-QX)Yz8$2rmg)#ZO5AE7AU` z9Mb|Uxx=X(@cuy@EWhI-_q_NfVIEpM$8WDgNkKfh5_7(MR<5~U!>wJ+tU zyzj8T3av%Fd56>|)D`q9F<_9tKhm&Q%>Z4__dpn#m^|`ajMEO$)!TuzEK-Id;CnVl z0lir$vZyKrUML^>sIAF+4m4Vk?W~Z|DHG;K^b%2-EUdA86Y@KV!yjOI0lT_J38X(e z*;JU~f(>=#EMSxZ6m_Ol@TZ}U&lQW&_d9sjgi?GMZAI(9{5TOx376?I{)Kp2bK}%Z zVBJx+hH|2HkLM%uk%SGk8D68)MBu&d>&%$X{Ij3f#MqG1b3b%)l0+7WRuj8*8q@nH zNTqzz5YrG{l;YyIAm<^uiWGYX83oxF|3+#p5p6M&W2Z;nO@Jd8io5nC=+bnD_l;pRhI=6+_GfFLU32t}3i5 zEGl%IoS1ByWHmE0+cMKKt1u&}Q8g=?x++MOch9(z)l^^7O3mg`(#&fW^Gv>jzf|9| zHepI)ruB~dM4dyMH8mCZ%%7$r@yN10B7sHGCGw?J68Td3%F609Qmo?6l~%b&vP;t3k|S#2 zMe#*-D%R=CNG&pg-vY}6;R0!}%qAb} zjf3#!@qXe;GiX@5%$FM%41Z3~yw5ZgyhduV^uwLQ>&B5{p0}7Zn>4o2duv8RuchzN zU*97%@~-{ef`NY@eYBN)Kw)}euY>=t?ZmdE=KJKt0gWLG-K!*=ME}ITL6*dtL{FAF zdKo4<`V~vBkuPQ)mA&OEx?>h&=B8%s{dP%)C7YJi!59Nn17@*Pszuk+EYp3{=f!x% z>2o^^HkQGbW|l|u;}#E=v~U z;C?g*d<|4Y*|$+}HsHi$pXT7P8MhC1pt3cyabsV!bJ;t|{V{_4v~_Tlvftc&I}CS- zwo3X*gelbJQ17_YxzM3u>ytf_ZR+m)&Rzddhty8(@?jq556>7N*R|1g89^F>TsTL# zm5js)(+JfFavU;sKuv_>aL6x=9p;JspczqXlw*No&t^!HFx&xKWPHMqFgL z$B?IMYqAIDLxPv77v2NM!_d9!BnyLFz%g&#=#UyRXkN`*O-| z3Oc{aXOqvWpViBqZ1hg%e*uZNh!crxh+8s`GUWwj{Lnf{I^zNyZ9*G7tdQzh7a2X= zTqOM5ggmgff6K zE9R$yoSEF-^kgf2hrUwSF77^0t;0d6EzA1*a8BN38;_%#Z1f#;7xb%qwDG73ryS2! z;q&jVp1X?|-!HUy0^5S01W<`oh<7mW5aQv}jc$x+23+EHlN-3II0tx(NQy+e&`K~` z`jH2gBOZ9jIgHh%0akWLM_hJVYRkkGwQQn`QPxqs>}{HBM3k`((ids^Y04{kOi5R@w{rIj3LaDdvvZ~@ z)2e*#gyFPSQy2zPPmy!UkVJc9ry+#Y8c!BSr{RMYi4=*r*yhw|u6)zd(?wN#!iU#) z5|)J`h1J#iX(P=@=cl(>x0Z++SjDO*C6b?=G8Y?~v+f539y<}+!fqTd%N{})Fl%V9 zRGPiuZNIwoh2}m=)X`0-DK#IAoyiXDD}PjuDRF7C>+Cc-S_kRoa3`&gl%~-L1DVw%0k4{=v>11tQW=oXZvWG ziH;xAfpz%woN}%T*E?(allu1ImxOdY%gzFJS=$C%eWF9?NyjO;o_lH%77`1)K@;1< zektsZah_K<>?`)7!|v(44nB^Vt64qWcS_gOzHLoUUgbmCC67_txjzTA;*;k^YYCgy z0Gk_A8AZmWwq=^l1dc5aB5R?^`1+h2)(BQSO|kC#Q-@g#>BobO67}j%GUr;|p4q@D z!0bas(?s_9j&BBR3CcR*vTGOMV$HT=oles}eKbvhi`i4{d~kCv@+jhHV5EoZ*p1lr zv?2dOri8ahYrd81M7dqNq$S(ENlRS&vQ6ve=dYpKIxFjBpx)h4+kpG^t>+x9FUkUM zB~RKz%>C}}wz#&Kwi2KQIDbLAwy^Zr{OAQQ(=UIIpi`cSg9FyQ3$z%gOGXB}vhZ1;E07^XqW4MpeO*oa)~(~_j8 zP032K2x}pEpOhAl{1{4kox;2EaE{ZfxYRzf4FJzv_7MCKMWY8wC@c(B#Eqn+z^Fj4 z;lQB4alxQLufRbE4>;a`UW}&(D7mpyT;zupNAfm6S{();BCZI?ZLn>$(|2zak+OVU|@V;;zIliF5m}CusUzMTW(HQ zb|;f*ub17sVSMRfCOE+*Q8EMGyoPWkKK74@2MFL6`ki<~2_R;|z@?%#b|irP{Q0wG z4Py3}*5m-d%ROo8Xjj@f>L`bEmvh+E+1SX~n3p=MX=!B$%IhI{Wf~qH914{W?0-KH zediOIe2ws;v7pPn1&5%y5q`;qL4)<+P<7BCgY;fD6K^m?FW8sO!v`AzN~H@d3;TcH z-?Ipa+P|#4Y$P9102nwXy*Gu(3q%JR0$2Xa`DG(fgNXUOgD+DF$LUpz?D2M3cnQCw8CACy19PV5Z^0f+6=y!iI*KWkZyC}U*Y z+{`SevQqBqaDk4~`6x!S$+<2gH>^b>5Wc*Z8D;wJdY>E`T9^d}^SA>O0*>hU z`v3T#M1%?_W@MCi-aj%Duhrtl0gmN46Z7U>GKJwpP98L8Nb2&67%EfDKFJ&7$*ll)KV!4ZvuLoqP= zGPs=R^o`I(IPT9ldou<xFf3ueGaL<{9G>c`XKo-3@a zKG7G0QjX2xAZ9)s?)d_(pz;y*K|BCaA`T7?-Vx?{Cd=%Cg@8i>C1hoFfe5w^FY=lu zazu=b6h8uyEGwlOvGOdug`K2pU*WP5EqU(pzDNfG8TIulpBkMG%eRxFz%d==$RB{0 zhvhpd*e^L9Up52q^ho09h<`fyE*FA^sa)V`b93|4N~t8wO}i{Qay)Jnk9IH zfjO@j_X0bBBZ?4#=U@DO(W+Pk z-$(4251LIZRzomqOdrHUQk>w+N85uz7o(cacE7In&W?W3oKRIW)W!Y45KF%EW18Zc zjOEGu2_b$zvBP@v#)VZ?^nb4qJ}@69^wBx$BuSQDlF&RP(<(HYHTcrciw6hA??Wkt zUUY>?%--SQurgnkQIgjAL$x9)2%0sd($8OCF=8RU3}lQe6V0X=KbgKZh`ay`*@9Be zIWB|Ye$o8FsZmXiCJ9YZ4I(GOLgJv5Cv(A*9{hE_zQC@)AmCq%ox1qsK&w>0t0&># zVDf#@S-TVwo&1>Mh~;8E`$Fk2Q8`G?Vi4l+>Sp=YC)^m&8dO|lA( z38mb1ub2VxlEntcL>BRd9}^y_oTxslut)~yMJIL9M(BY~LIj;ea-#(S#RxG+EJC;llIAszhHN|u#ZYU)!3*LFXb6vw&`)$Rgf8@I zs&4pKM+v7NQ6TAE&o{S>526Uxjop zf`!b+Uj!lwNk?U2Y)oXmg+LKSK=_4~NXhtIc%j)@rd~yer`i<3VISe81u4C!5MMaA1Zn6z4?Jm{wE$(N> z7_1H`>Xz@2@C+YeLsAacItAEQZ?YFFYqbhX3nOJ%09ypLfG3`)L>ypA?-f4hW<2Yo ztTgAQ3%kdb!wOH)per52b{SyYY7 z0HCnq;WLZXU6(5o4#VdUTu;SfiqTQ5+e5CEdQ-S_^{41_)wk#khcaz(Sy`PqZ$m`> zVZrmgyOKxBj;R)lDyMEY!PF(P_`Q}K%zS2+T_m(z52bp3RY!-1hEN z7+!cJYH`1IsLL*CzBt>jbb9EqJ1=qWZ@K0r;eFb5WN;Y;%fI{1yuWxp@$CdG1cO=` z*e}C-+2seVGC4inpMKmd)89Xk+S<$t@OkUGA-TJdUssiSK#0d;6@{hc(ak;BbW9ha zAI{UcWCZ<3TlPQT}a zv**b2L#5+B*3nhzlKbN2_wQ8NiI{INF~#`#Th)!1?n2Lc)F0QZSe-X}c<T0=O%XQh1$c9Gu!9&&bteB;bo;TSr>;(1Myhq*>#00*`jqSQxt!WFd5wtmT!Qg zYmdyvJqh5Lmh^7I#_*-1u&Gtv8;+pVS9` z0LuVCHSMtmTh-EotK-eIOlt!WjK`1&d**iX==Mm z#Fs_@cpjSR=;lR2ak@K`9hf357i!(Bk~GpCyL9dR60d}!!@O;KG_P4z+X@9z-*4Te z-+wQp3hi&T?e4Za``W|OPY(HKRS6+fi$@6Ky?tloEle^-QZuFch5EuL1ew(9(7A1s zN9^lHh4YG+>s=xi35M`svR%6n4zHDD7lZYdZ#WmRqb`b1^Ja-EN&@@wg5B`#^hNAn z2sR0GXRIYTacj*cf803TGIo<1FAuTeR;^$newImhQDt|kD6nd-B&W-$&aLnmEzM*8 ziS={?J&N#Ex##wW#5Nt-5a3FRXU{i+v9=_?t8d?* za$j?AXbkc$dNf(gmiOJ(+62H$!pSVuS;zN0-deagO&)fB>8yvKc|5(DkX;ZH0tb2W z&pAjS+)PxGe`(-mnv@^>2&+9MmJ6AUJ$Q+V3T1 zCJnrW8S-IwX*w|OLS*IT*MSc%_kG5~fu>JoMQxOciTv6)c)wk1sxn{x)95ZY6?M^D zzw2$lj$W}!jR=7!%^Mo$qq%&bh~V8Jp;c_1)JMO(vV>ipVfqQL*^8ekb(`T&&Vf~h zJ~(yD9@(^?Za1q&fcYHkhcW9zE0b!4rO3BC@CsAvbJ2wb1?5G4*ggR>_<})&EM-+VnLp90pngkzBvlh~%r)|t>9)a327d%#5$_@D7QNCZCIs`O;y{O~QrU_pAhQ|hWm3IW z)aGbfbtUDrnFFKx&NboNdr)9Qno*rk_q?MVUOaDmT%tXq+)p{fH8^WAFSTIkU5hZ4 z&&?oOS&~uKa@Q}lZfd}5Yg~21;daZZTHs+M>myUv*kn2FBE3#<#G_XrH%{dudPf&2|PnsI+3ZVn@1D5 z5?Icf%FWqi{^3fiMG1O%5^<#wPtA4hhf_K5>jtx<9cQaqWM|RoPt!vv+U(bzqdR!3 zw@$#Ufi_F!)J=VqC0Z?BZQX~Txmf9Rx2~>{!LQS>;AY!89`i<+j>($k!v{Tq-0Yh< z`VX;#ajkW5gf)9P=wWdthb(5^ul|(n`NW`?`@BAdQFZa9mOC;SZx)n=q?tFq2<}C} z7+;*-easAe`wnowG_slBc9VLrjE~!|*=DIRUK8SRScKJ(U$8QGnK$sEp&fqsWQ$g- zS+e$Uc6G>X`Ys?enM@}jVMd{`ZLsdRKh}J)vT>e6doHPF{~a&yg}3Qk|5Pgj%WtY? zUP-3`-ogE{rY^-9+^ZOClKS!M&Ddj+eYnJ6b91cge-gW?%;%s6ALvJb;}%O zg_fo!Ek~+&`f^K-^Kn|qUNnqYOl&Nxnf~l3zci2oA}$Z`090f*o7N58qrEdJv{s9i zm*E(mgcma{d^&D9GI`2lHS_xiVWN&kQd`*gYyL#2ghMGEWs#%L-KW3=i8RXpt*2Q; z7@meM?PYRZTk~q**$QuShpS6ed^~jzzIzU>Q=EFOA3@6OV5#nkQ0;Ar>ay+igRFDg zBY9h}+Lb*c7yIvYh{zR!oqevAk_6;eTG1~uL??H--+v=qf9tXHi@&mqYZP!>!%lnj zEzAgwKqu1&Nrq=)$xt%Nho?C8!`pQmVu>s&4>0o?J|n7Sjh>^WPbAX|754EOkC)~oajuNaH=AEV z+E#~x>mL`}v@k*8>&fNh!`KS@>pyv7J~M<_9R!dR;#qi`XCl8)XH6lku&ORd4b4VQ z5{0}={oDl>Vi2YknR70dY;*}eS7vn+d4Dzoh#`F1ac3|%;nQAOadmEODgnp7Bl!&! z9FW0N1J;tYo4VNNf`dO9$3VDm!Aj!NDQi4pG*N~>oU1ywkd`#x-Kl4hatTYU|NIuE z{j>&HT5GUw+eaiii8IX=CZvrHBz(wMe+nRc3dd83h>j*zRYjU8s*g=l85^Ve(viWr zDdOttkFJr~)XWReodiR~C*JFh*Rt1AyC&zoYn4o85LXueGo%&5Co98tV zSZj`A>e8%M2Ze z9F{?O1WQeQ7rrR3FcQ+2Wzmq6IOY;tJ27D(aGYvX2S#`nsOlE4X5L>@Qc=ZJyzlMI ztJ!CrRW#nEo%n{&>2%xQlZf`NvML!%Aroom(&aF;LHeWzp!p!hUg_l)JC-d{f4F~v ze>Ywz&^aM`DPJ&q-=Pa0$C6I=C$j;AScPCo=!8+8H57alaHc|AQqG62FkH2)>2Z>; zKo-aE&MJhgL(KwZO^6Go-B$S5s~#8LQK-u8j2~`OShxG6>h8O@T{=%ifwhMsBWc_Q z3t{-fTOf&&P!HDWH91h))rNDs#l2-&*D=-%2TY{jTh6m&(ef@;cGIf0+Qr_lpBgV7 zs!;p;L!X6Ge+W|ywfpVz*`S_bxi+n*4Pe01x!>{0GnCWqOG&v!H&p|$HSoY_aohzadb_UAE4IuLBy!%}S>bVwmv@)@=#Igtb#f2=hjsTNAi{dz zsA$3x(Aop#-b@y+<*ADkW92c(>h@TA3ZrwRoFcEZ;F(?KZ@5`{>A9lT7o+ zY-C)l9Y%8Y7edpd{7GrPnvFl1v_P@1WusKQC8iD<^=yXJNMk7SRq9itmP6Qu*lvMp zV4ri_InrE(eg6IGR$FgW?ttf1tJ|*sb2g!!wv@nrjym=WHu)e3y^-5>4J!GNW+ep} z_8fADmY(*vI#O6|avFNFd7f>+`#C8y>kA47X>){uPh96ZLk7(tqaZT9ltVke%iTNw&f01Z@PT}NH5-9AqTlNjw zKp}>$Cp-ZF zq3_?nZ!A{HIM-UYk-l-s#KZNj(3>*5?WkEXWx`LCPbB=~sy8JcRfzCUll%^&-1S9z zj{>s-d3n%o1$+5al6Y+JN?+l)K3jR`t8!8xL4}-;>qmJdwXGtjES!cliTt!*1LszQ z30@IzKT@{BYH~L^o*(sdJj;d;7jDqiTx(M6@$fi2DWe%9wEx`Bo@}~kOj?$~A;z!g z>9)AcVX{}Z^-u-x8Nw|BMoNYU!mMO*udas|e*W@-Ono%hieJ-X6rS2#USrXuD=Rp` zr{m;Qe=mu>JupG3Zjr>KTdWxiZPZwHk}&_9eLT+8niylFuzaDq z_}9ggr$^CRi@7~X6gheIeWi4+idFE4uy?lCsjN+g6-qoGQT)BG23DG;jvCHoY*BQN zj=B(ji?9gF>2R6!$>m^I#B#^?yzBEB%E2X`gHc~H#wu59oD1tG^-Gj1`#~m*c>um*M>0bd?)+dc9NYHA zalCdsp4S;~rSk}n^%RTdOe9IO_*2@|(bSK`JHdRjv5nOC27^11S@KS%OrB+KVTGWYuhSsKGp7)V2e>9* zF(Crf-#X2;dIm~jv;Hg@_q%Ajh+sQ5BDAm+^8a zIhe0musr%oo*s{a`N6L4@RPW>^6ePPlLUxOSmJ8mr6jbTH3-rB5cmpbV{+GMp5ns? zv&zTEak<;G?iGv5u)348GVS%HZ}pF(4*)cLe_y}nSwvv_z!G>2BITUa7G$G)=!v`= zEfl;UZiOM4z1Z8g?i)7@6|E`XbZW4b-!Lt*X!yXZQ7xq9mYZpb^sW9N>bym9O{?u8 z%$+nX#4^J?{)5w$Dgh`W4sso!;&8Y`&^#6pxBpQ3URW3_91<^9LO$>CBVed@R&`kW z*Ar)vypv1xBmR}NY81{ZmZZln&M*F9uXQInCi40jM_P6)_L%TDH!-5jzle$igdSBG z4Y_H1cXtXSd15oT=nS^aT)8yfAMab47!=4ynPeDVTahh)A9lMwBC}f`0ldR~cYbDW zzi5{#Efg;GM=ngv+Q)fCo8q|jxDBh+*x>wtiDcI^u5iJfwVgzD!q;JIjQZJqySJas zm)p{hUIj}Mln75IH|F>N2kqM1gNZbv_e)K6=d?AA@>IX=j^ zI+^?^RbdIlbgdUj^0n?G1Y^o1;~V1Zn;^dx58!&No>;t`@lJ+B&J&DhJ zeg`la5~yJltJYD0)bM4DsZ?Dg`a}J5`avxL-hFelrHZ%|mCP0J zD6l)0-IT%zWU8CjWkik}A7^k@R|X}qWhTEeECKQq6}SRXy|OJ;Rz6-#>>vuiQ}xpR+Rt)fS`Fs{F73o5{#PW`y$5 zPHV5jsJn(fA!&|G+lWpO$-3Dz^H{dv+xB=H9u+xuUm%$pl~d>tfs-l_4YS;GYGe`< zx+)lW13_aXng0InlR|&*Nw`a7kW*Dv6&~|>Uw(RntLNQ?!_4{lnGhsK9F6+XhN(@T zFjm@nvH55a$jW3%si>&P$jQ~0J@)aP0h)5n_W*Mt2rL&?0coJWl- zGg`%=*4nNO2F6BxnOhtdi`DE(j{Sc`IdY$glx+Nw@}4DRdMx{k0qMux}8t(vCkFNe&hkAiLLHomsKtzecNgw);Jv?4#Lvx zO_b|D!<=T;HCU?G_|FR(8^F&+Kk8U{E%Y zlOM}UB!yaja0feysl78fD^%dDaNC)0#gBk}|KqV$G3lR*FgcNY#lH~Tss6(5xkGiA za^^VfPDC|#JI7L-P%1Z^?Mh;NZL#WdGq#1+# zSnHp8nxXR3q2xV5yo0fll^t)&`7hf zm{HrWc!@rqYyA)e7GOy6tGc=lnx-;s&~vrx4KF%xtcpI(DNLXZT)6PI61P%P40Ei% zX}F=@7PiP|PsDwtmp?c5JQvz%tW<_?T2A&ZaxNTuZx$=3T777cvk5Huv=VrVu2cW}p6 z8;z?q96hzjxo@8B=|ipr>ND_Uy|Ow zhn1Wb6lJ~c&U;ZDk|vNHf3FZ+M9CyJtq(0b9%EzpVinLA)@z5(i~CBGg`7%ds>d)Y z22hL!*~XKBPtB~_1RgiG&T1&mq9gsdRc)-7ds82UgSJrcID{MwP?6q`RohLOk++jj zQ%jWIO)IBz0oO0t?GL|KjrhS2B@=WlCI05j{?76HNK@)SzPs2!No1L8&g*h0cRXlq ztahb$wpvGUOXo#3;R+U}l0lEE1`ENQE7ByF+@H3L(ES0^WbCzeD%u2&nRL6J{vhH7 zSbvIMThqG*`i8TwSk=b%gb{$4>Ppws56quQ=;&CN_oTTwD@CfDvKW4YLJbIx%Fe|6 z^hJXuQlsGrG}8MC*&Sr{I*$)@ZCix&9SftX#%X4cyltn|q(1^#?jI^Hffq+)V|nSI z8QN`!X#6BxVDsLs3ceju)_GTH7Tl97xN z+o4XqDQm);3%J-u%2H*?Nh&4mM~wb=QUtvxChr4c^$vFDyb&lcX?u+EJcrlZ=h&V-Ovt-dR7jDo776oZnHf0KZ#qt zo3EF(P4ar`Kfxdeap%Us5DFpk!SU9^JerPv$pgiNUa2-VU%(MS?PM~h{ zGR-Cm=bT~SsC1~l~TQLX0fcp>@x zRKb8Ns2)bm`CPzrB4Igd9O2DUB$@xs*Z*xT)MpZ{{!Tr~fUcI9hGW2^99;!%kulj#O^- z2;({#)t6r0W8NPu-`-!e)4c7tE{kjRbO)X#RoTu7WAzq{g|*1(dmg*a2%HMB2fT5&JJ}9vIQ}vZ|`FUDKwx ze=sQ!%S{16F90pM;X5}5AmvfahZkSTr|iKUb2Z=Afk7@xe_ z&8t%1tE|lWn+Y&H^Aoup>QiZPN}ZgXW-`W7wIcW&`4k+sut$ zv8h{EO}st=S}`9EWQTJfTJ=;hoR|CN0@_+n6&&{zl$8gbvDDdVfYcF>=HMGOti+72 zym5?pO_RrflGjK~W7XZ?kR4~ZQF;|u?!GrIB-?K{>^Cwv4T`NdX_f+4v~SMTD^Ii2 z@2S?TWVo;D?rt|pFo6t#uL#~5r#MB?(M|u%;D=11!oBBs@4jOG3xI{*|3a~NleSmxfcr!BV5;dUB3`Y>J4`7*lqT5jJRXY@Y^j0Q?E;{ zZ*ETPYJaBL1ueh^MqNtSCVDdjDCt_U9SG+iG= zwsD^noO2v*tZxPnw5`RIXf|mPxLv^4Y1u1RFKr$1dd&W+e~L4WXtxA3DD!AEcg@rg zKApI?@}YsM4kFUhwHoJ#WqsFx8~M7M&6%xUA6V%)!F|yPx6PWmMFvgwb|{9UrK*wm znKV2Q2zu?=o$Qbl^>n+66cwMme;}u%b#(NdpK_Ng_^Ug~pKDz?fqV}{7#w;I5Jr0@ zb@s!Pnaok?5Jo14Y+>l4u^ZRS?4kt)1uz zxd84bvXs>3E~t9@3!9)qG^%H+)s_#h`&FTw1wjlGEB;5*FAm3A(Z93PdMy$381=bz%=LnH~EdP`DxBY2g@xzn!Hm`$If30G7gDOHQYo;e&EPl(?eO`|pfTKsiR?ATQEesq^4BD)e1k@7PcMvNuo3(=Kbyt8q33 zwC{*^a@NC7xedn6Gwpj86wq0|{4~vfODB(AGn-ZY>q)b!&+% z@f*YnA*8t6kDYafl!=8OEp5a^heu}{4aG#2YU&YWDUT4GwH#|v<|jNIw=GD9R)88Y z61`3oZu{d$9Ip?b^CE{%`XeXx;vM>RDNxVbMtIv(!~JF$Fr_WgINdert?fCw?7 z=Wz<-Ca4%6|B8|OCjfB0z%ckNLpSO8=XJzJ=8$MlR>z8jH=zEh>PwX@K|CAhoL%pWV@D)zTANgugIb{j$bvl5D zm)2g_$o`c{3jI&@M zMhxYsc3SFTaOl-pQ6Q4eWdoRiEyE&W4fxsWrx3NVIh#+I0k00RxMh-mej}jd7z&gM z!K91*ivF{vmrv7iBi|(3*RLhBsS!-7<3sH+ z@kFhNISJ)0B4mOf5H!Okol5v_xH_t-$iH_8ZQx5tDIUu^^IQzjug+0ZyZ^CqG=>K zq*`>t|I}v55$+9gC~2kmX(at`h%>1oA!+1HlOkBKJF>oNBkxc2!PZmI^v3@{NnVCh zILHEZwg#xZ+!MF9+7sdJO#q_9Z8Y5j%tczX9bugW>Kks)XRT&_GduC z=LlNxk3-UMj9zXi{P#zH3;3QP4N)qz!_7C$ymzoFJ2wQs)AK(fo&S756vt=QgZ+-Z zJqtc%(#bSy{msjIFhiim*_4QM!7*vmP;xi3t*dv`D$xI@il2p3UBNdy`BpbsmQ$X* z5M3dn94w^3SOWY-=MqJ1fp)Lt+(PF7U!n)12oIl0gz_NgR}mXcHc)@|m;WVS6MsHU z*bs0rGfWn(eCQB`1o{zkLM!U3Z$1KNYATVO{?kO&3Dudc33-=B zDuS>gTGUge{mqE-A9l(32qETJ74yG_q)Fl{1|Uc6jgd>GfD8ZYjz^~aBivj&+bn*V9*ZT26wItUp4e6Ll*@+`FaB8a?`E`oi~^L1 z0(;ROyB}0tNVDHCg>=szh;fu;vzbZaXL0{rNqd$XHJ`VC?0>q7DQR`|i6P#{C}yh) z4t&n=$1_;vhFr?z@#-wZm@g!LGyPL(HlPJq`J06MlRxl-uO4|rtedLypX@J)R)>^C zQi@YcW$+gqp6;rU0n=NOEiWWR)qmU-qCn&*QvQn{wk`5s!b25r|EV}@sDjU!Cafa; zHkpO*2eK_%&j*EnXwvoBLKJ%I-*`Ke(?tLeU%n$oL}{zI1N9<gbxv7zWqEF zO2j#V2*AYX0S{M(zUbdy{eJ`=K6m(8<1(N3twdWbOE}9TD$ZQV^Q_G!dA0iZ`Gua8 z``!>&aqhUOYEck07w&i+$sS+8sT7vaGdbNJR^N>Mp;fh1pOnspWxR6_iPm3a>HZncMuu!V$ zwlDe}uO#ircp-EO_LyGfKMwA{m6#z2H_00^%@<+qZ+0)1Md5>ec>xrBJ z8&UNwOIB=)g9FQO$G?pi98uoO;f&6NL$rgV6c|#_swSH5gfd7M)5$G8nBn3k0Cxbr zgZ9TCri?Mh3lsWG9+AbaJwf?|ll&mNl_PD>ddzcmC$`b}*Tqf9eNMK2HR8L)ElcBn zTGl-bd}Dp$v&CMfL@-TUWIeFpMV;*Egwi5~s2OuoyY$#&J9KaxcrAYZW1+^;!fS^O zHwf*H$mraX`=PzyVbYEkk*@CY?>GL8%0dfV|72LCW80y@ zkSb%Bk7UzcW4PQqMil9vSf84e_1fQOUSi0_*UTt%@7Q%XDn!$Xl-9ez`1evZ$>ZDk zA+aOx=XpbjfLrinFgA~Z-3ahJ@o^d=$WC0csBAnevs|!Z&EL=T+(XGuKRVc*0W?-w z209CgmW-lNG0m-;X*QK2?ZS#ljm^Wwp8dz)|M&I+y@$#A}{Jt2&lBpmnPg z7uwq+V#zdePK6fPfHDjgR55ChcN*Wb8&mz4V8x8V@lbyuA=~h}RG)PT9NzG<)l`Z* zJ6o#r^11Q`HS7=FSQuT~@_sW-c8j^YT7q;_DvteNUPF4vi;`5j@CcnH`D4vAO!4oH zpqbCb#@}9BYGi&f{5;N;I7Qt$4QRFO=7uudbhtE9y5R1_p(!R$nO2)LpUwa^ukH;N za@7j)mhfM)!n@g|{DbpNi2h8Vm;SsB;eVXUZ zM}qyiCv+S;jgqi#F10kWv|?dSn0mI;B>(0Aa4x_9`oDd7SM!pbyPE z+vo(aWN8j97i6BVwq&9@F-Ta5hi?Mxn#L(Y%?p91#-h7L^p`tvU;rmq$#~3kod6$5 zh4~5U?5<)aG<%(T0U7?@a=tXs?)CMAmI$Y!sJ+@Witccc7dbh0r-Q<|#kk{8cnpZj z*W9zJC2TBhnqwh7lp*-rAc?Qo?jAmrrk?|}z`@xU_n~`n(VMe$lCM&7AmTkgwHGV z52e`>Y-5XCuP%OkB8jf^d}xo^xvJqlmmL&R=vr|Gt`4zp-%Zp!*v(ITxqDMIvU+*w z^#@!?)abp9sd4e?lI(Q3@*As8R-yN|gw#w32ucdjHf{v1H;uo0CqUp2U#1-f!vxh2 z>7AHEyLT-XoC)xK&<|9zkD^&ohk(z%G<&|RYU6-~utQ}NP9csW&`%Bg^;-Z;hk+ti zv$d8D&V`0|2r?{m&4(v$e^4V1G!u3w42KDr@Qr8nsAt_)pHRKjOfyVDC|?o+s!n7c zR+!aq1Gs?(_4gi*=pg>vzx}rp`hVZ^fkH5NS5BMs$Cily?`a%4V5BMzy#xK?&uBYU zCd1s;htZMm#~TQvlq%_%@8+OiK(U~n_9AXdOQMTacM>nBP-bYt5+kA({LZsP>6mL6 z#rQqlfqh-4kBag&JWNhrwH%eQWeoSoV07czC(7IYiYVr+HUk4=@u90FxmuIlm|;~S zpH42yh8u;xgQX?*mHycM9EW9(!Uw%3gK%sTXmWg7pXRsu#Ro$TauqIxPz=6NjJhcs z(!4@l#pYC=d*X%5^~2Q@>UPN78ee>2IBy++*PrGlP$Uk+M#zBeI8p4Ye0>yqB#w%T za==};LXQW^$l!-4&%4dn2V*PMShppU`RQKo+wP{y47Z0<$7{@$6RpBLw{Y?$bY(N` z-yVmD&sayWTepzWJpjojS~4znCq%u@7w}T4tXps6ns3(H41WZKj@L+h9ppFz5#@_j zsg=oAjjD09>hBE=H$jXwb|N?M6%MOcihQln5bo`Q=k)-nDLd!qd6TT19LbC}AoXH@ zMw#c)#rOWQ%sF|p2gFj6N24UrC~Md@Y;0|T8f3W0N4TGw4_4~5K^2J)qHlTR!?=&I zt?QO9uUFFPta0{HFli#VCWdtx^9Bs#wYToC>vcK=T#p*C45q#-T0P_`UEBk&kI8v~ zb24Nm7LEm-mE(5CPgqY6*0(#^xp_-!>=&cE48TFg&6exsc-9pUd7h_xM1A#lf%r}Z zDY_{kBy6mX8vXx`o&J-xiRHAXaRXDg8e`v(SI#9^dLF~_+V7>fomDWHZ#mq~43rJ@ z^$?Od&95yS-6R*Qw~ZCrgL;~K$+vj6+8!SQs+Q`kt#6l|Ti#?CC; z?Xh0>fcj)Vkf+-y*gKmmKHzg+vjMS^w_y(xFItVaKRFJ!XtuQB@!(5|+63JzWRRxd zZ2wNvM*E;HP#k|O+Jbh|Cdy^7)IfZygmsAh(&zh*bOK88V3u4l*;$NlF!}qSx54Eu zm2hBa-E)9#z}g zA+KTN%oW6?5n2%ZunwqGZVRch79<>`uO{^2sAB3Fdvuvcylh;}w_Mm&^;>6}WO^t# zme`@WkQA`yN#&T2h-Q+}&a@w+QP;bcN8HciN5lPCVSG$5*J>GPN`JB9dl+AWXV%&( zxlnBVW1&{Q$6h0!GmUV1_DV3MI(2W4irhsXSZnz{U%IlgQYu2M0v&F3ZfWpA#r%s< zza(#+M5yYM#Q^49gL%9Ig3YgEY}BL!Zof0XE-YlDoguE;BGFX>e1u!RUM{oXl*ZyUj*s=d=8|^0#|n{t3u^ z^J2KIP1Om6Su3wwY}fesc$=dY{CERurNOx1@A|799tCgt>6)+AZ$4)}wRgDsqH)E8 zvoN|J8yicj)qL(|i%`D959V4f>&U$|6f1PscYQM#OV7??w^DamG26M!7Ct_mhgN^# zOnxA|_JKW%t#wK9@y(7!;Uf$<`i>_&$Fra7Epe8m?CwykCAP}=vSrf3ZH~UaA6`uT zg>^12Ju#nhNCX!7{6mw(&4jdkt8M046koGzG^Vs})_VBsTs|j0^&~cRmtn@CcD2c( zHgH%7pX@KD3NnM|na^4ElNiGrW>8Bo1{_U|dFp9uB0a3eTt;~7hD?SjPihMVP_y-% z7g+D643gTW5_dEg66$#hnKJM(3y!j1;h$~F(MvzH)vaWdWZN>Bn>J^onrh7oSqcjo z+}@c-n(xi)Z6CEhC!f2tcfS4>ICy)3KL}`(FrTc+#O(qbEkf<5Myy+8$+FlTz&tJ1 zdDz{%K`&^PT@kvQW0idKX7~$ss(vOWq)@pq5DwVrj(r;^mn(YXP+nAI1tPZd3t4J* zOl?z2j^^|b>aVI*>6Q=V&kMYopPy%R_{zQo%o(NcbA%Xcy}kifl(b5V1GV|8 z3)^HSqp~nOo&`+l-0Fu%FNB1rimU%o`oBJTJD^ltE$>%-;yRu9XxS8a?=(ryXgmgH z)I1jkavj-M*eJi^T~F8>$j8W_21q&E5Lq~0w!Ay9rYCwAoU--Rk>v_=kx|vtb)|+j z@_;AdAdsz`sWh<6xsOW$M8mnvvJnP6^(htMQ8o|{#*x=Y5xY~coS%XG%_%v2+`aWf zeqECqvkejPB6nDMX2ECI__V!3Z&Wgc6egegk`vmW+Et)QZDQF~ScXBO@93@R=yg zcG=l&JQ0N@wjUNPD75xwulo9-YQLHH^_LyxZkkQ-?0YmVm22TAvmOutIiSrS7HTA_ z)AHOWU6y${#hCc8hJ?0`>JD%Qiw#n(6BnjsgLy_}?-qx?U;>qHOyDSJhAN#|adySF!1F%B;e8b&MWz9@#da8@VccZ=F_)m7 zpG+!=z;zfESA>kpP5aT?@Z2A#er0*V-*6z*``L^BI7)>yR@WqPYgdm}m&>TvdVS)- z$#l6zMJmtXa8?uG@p+|u3J}l??Lq{6pG(fhlv$`2J!E0_=oTox+X+G(OgGTciC*8@ z;J640g)taP8lM1)%lcotV{_^k=^Gew7(BcHJoY+nS+hq{z zx0{-Nj4{`Ul!a4|3m*i&Q`{pd6b9?5Qq0pzOaq4>K9TE7H~1kDWM+5F`p00lN3=3P z1vBm4W#YiINN!5bQ~$$UZ}dhEjuDPIx!iR4QI4h4JM0fiv4bDmA7gtrT=46yExK&? z?TwG_{ARenk&F8t5FuXo&|AW}A5@|RB0M4t-mm!EDsbz@UOd(|+_7s7eAw=)@kOw4 z7#xIu5W}v7|CKI^ScE}AU2Zkg;>q^4{dKxHP?bUUX|%4cMt&gCGKxp0m9}hawU9hd))A&xT)ho!cgY}spEtS z^i*V&-0f{DU{nS}3~fVmp!B_29ugX^z%qn@kG#@|< zDu28FOupGYbRCCTfstG~Hf*hIg#V?ztApf)*pcM$;X#b;{s-vwb&fmXTe#H+0!ybZ(y}#NjxJw zJr(L(s&~vxM(^I2QzlL(V<762%8ZUbz^@b#vl~Z{M(b5voQb6MCtDSI_R2K^X!hCn{6ctleL*%!T7Baa};<6QNgeAvsi_DF!#O&?kO%$HV4^>JtNIV$5 z(Nb8zNMFmdd-SN!>>)eOUD($$yw!YD<2W6rri-K7#QRQLze3&YaCCB! z-Hd5+d9Dzd_-qJhg*bQlIWnO?HSFJc(?xt@{}&{ojl0>$iN?NdTC)Qe9Ubql zPM;Cf7Nv$B>W`q7at|i*84Q>7mA;g=)29#V<*6+$XT{WSwSIRP?MG6i@-dRN&tWqhx|ayvco zCj0#e-xhqt=Ix*zS5hXuvqG=GVHHCsB<5Qy5qF@#JG)Mh)(^`jq7Vt{!u!QS1|GLI z-liF>!#dp3WIqK>LaD?gX?1}lW7=xSn04Aimj{F|IOUESassxq!iu_43KBg@yS1|+ z!z8qeu{x35S*v|K9sMi>c9OEwgbCMacxGEytZ3mP)Wvbt@qw}Xhqu@Z814qC<8&<& zDPsC(YiD{;mB)&~%U2)>n3(D+HET}nLlH5MkTUByreheib7r4@?0+`b!To7T#* zg;$T2g2!C;i+*nfG7hbds$uD5X3gGkJi@c(2mT)q1F3Jb{GaQ<;WAQ~Jd~G|yaoKL zEou3!FGt`twIot4YBr7pQZrP@d0v5|r=NG(F8t+*fdXej7OF5i%k5x-fywJ_X)Jq9 ztTeUzu|;YNF?zvVOFSL7k#JO)C3PJ815A?k3raeZt@^b(wm)pk!NA))II3l=na;C+ zKyGCa;?tNV>-zr`ix(&Ljw$Q0*_UJbekcz#9+Tp8 z3Mp^f1Tw0!1iT)JuxM2Niu?dhio@Ec*7bP+hh0UVLw95m~kU7{qucdz~Q7TC;wZ^ z{Fi&J5b4V{v(Z5CUr|B+0T>A8hR*xZ2o~k{a_?4VIgRXKilHG@AV8a+Lfdu~h9j2b zTRmdq%GJrIvH6gSOY20W>Fn)7(w-Gch_3vQFI4c{E3A^6o!R%infNqNr3#lOz$c za1Gbs0=^8nBy!cyFT}GtAaqG6Am20v|1qM@p97KB>4Gx->YzrFzZ??UTlmRaBQCx0 zm*d479ajI6BB6_zOk#NXLB94-0Y$rVu*RU^Idl>d70el_<33`4Ca)ZbY}@B=dHpM7 z@;ALt{a!kxcuI4RW@b`cQ#pAUi-#enG-EYkkwL*vKD_9Zn!np{(PA(h+4}M{iJwSJ z=6@JBExvJU7Pk+FXhxMZY|k*X3;kgj7m7UDb_H+!xrM;~!ZK;;1o#s~3M=KY}4AzLM$HVeGCs7s*F0 z%j);>ik_81c~xql8|RqlbXvi@hQx^92RRVMWB3gXWlN;IL=4`cful<>MogFfft+{{ zI0^$&)KG?o$T=D@WGF0&N*q0lszC94<9jNl(wL?y#$sxECSq+jI(rtDnDP9N5KRuR ziB%DcaeQO{(BD$ri{G#yLQRfx1MOW*W*;*lG=aGi3MWs{f*FiPsNum|1BjHT(2)cl z%wJJef_O;ACR7Q-M_&2dTxGW=5h^})qUk+0adVU6se=`20j6D%O)n4R{dwwm8Q-?w zCpqxHK!!>gx2vvZ9EL)lmN#ufR8omr@gAYOkP#VX*Gpf>c#wc&y`+MM;!* zsi^WCjx>wsD+j}rl?{lusLyy=Q&mkzu9F1Su%}Y5ouA0sl@pkXjPV;jky9*A~v3s%ne;? zA9-?>NoJd?>)v{a7`2|JvHdc6am5Kfhv^fxx`<~u|4MEBOD6wg3+t_j2sAy5;XErZ zZJBUB*3!@*zS)G;!IXL*JvOND{{5CzPY$uN@xb6PdM!C&6dVA!G1-!-|J2NXT18_! zkTzA-!**!kDEde#O@h8)Hnu}keVcFel9Q7m-q#$n3Hwr&eEgVSRWgu_ESx%S9cI4{ z<3p_$G+<#34Ri^>K#BivV*l5x`|}eJ>EC|#m<3%>#Gs&}sPv8t9NeMCf3v<9FB+n% znvs+|&_sY}J}OZKAR2j3!tDPd+RICZz_G+oU>bK3vEQ>MD9v%J9sn9RlBI<&aTB7j54 zOArvpa|&twpCkP>+<(L={$m#Z!I)pC#~}j(m{0d2{a=zL{`bOvMnV4XyMK`?ULM2z zC95$cPmKyRn(U%|d_es`iBU@ncmK#p!qAXZ{}+N~28n&<-z6eM34&^CYvqAf7(j}m zypj^x-Bsd1+rypyY?+32{9*liMkZsj9~))14}n&2ufxDLPMRg65a{ znUDkVlLZY%&Zn9L0z?V^XVlMM;%v7hof5|&I{#J}yUDXU(TPRFfEtp3{s4IFVY6r6 z=qKZHKVuwOzOabe6sXkuYsnm?quMZE%lAetP@78*)SdH&R1 zRe!QqI@lyh>6gHXD}ZMR=kNIEF8(A&H-LsF3fQVk)*mi?`!6@wFPS2N-o{83jl?2t zbvchh`=~Vd!~HJCjwIzN=YlhdpJ1%>ZO~-D*DvAmwnN>XEzgunWer~qv1@TU|6OBQ z0$NDPCi_^SLVO~?RUk9?W6}f=sjN?sQ02yEC+ql;V4>^Swm*}>ydhxW@OjlmlY^Z9 z(AA@+wPWtJ-cVdC@#2pT93+5daVBpdpp!R`7P_PX*oxS4KRZhi_(SwY;F@i^;Uo$2 zLFLBJl~NQ*JBXy|6kC=`i;GPZYfDSjcXxNO zHpr=ROG?y$CP!*ZX?3~uPc$M*Yu{g=0M$=N#y#u5-LrV{W5b*uOp1|7r_2S2t;;?= zrwN)tLcB?%QsrSd?49@NInrE{ed_)9ehmGI>v6_|4hf)Xz?R*QZ+#i@K--izK<(YS zkL}x7VR(^hy)8QD^Wd%g0SINF0j|&p*WDDl&NR>|E!B1m75w*8 zgbspW-f^fvpBSJ8S#pefuEKAH(oV@6=pf8NtV*L%Ra1>@wfsb)J~+zzP9VY7WP&3R zER;m6Ri4SEdn?oO%@;Yq$8-2xTG;fs6#-o$9a>4{eB5TdNv|&~YB|p z?G3vm*(9lHQ3jZS(`kApCdE5i9(iRf$%WwfndT2&J9k@JA7T+OXk#f~_ywS)-i^4a zwm3P*-Ys_VGYFNntO+(hZALhwI*z}6#%#=arXln8cMaaM^q{xJ(c{tLt6VO2BkXn4 z;DKowgYnjW9xGAfLvP>+%w>P#=2LMa(1Rl*Di?=mPc~Uwv@j^{G6jih#rkrjc5oxisghQYOpUd{R5eYe4-%>;kqw|_dGP6!t zx%e}hx4b$D_XpmdoHv%ku37TbG9C)5b1b1U+vhw|Fv;(iEkz(VkW!1U;^;96QMg&(j_D_ z74)kEe({}o&sHi74Xc7MoY#slFt9*F9;Vsy6%x)ijQIUvU|R0Y&`pFONd+X|kSWFcv}ef3vR-^eNMn>sC|4m~blkTq zz$<3>91M>$HtRg^PDJ~W>LcKtg%H96;qq&*2O;6E(y&AR-u{a0NRlJo@P&ls8ilqk zcv1oga35--EW0Qgz$kJfWZ^6}iaeBylvi6|SQAR5>ye%bpZSv3cq9oZ!5FhimvR5V zKwQ-+j4GfIj)^oVIN9D%DNQNgw;l2nRif}I-$!t+$x_l;Kpj)i(#EGOa8jOz(a9+* zlRsv*o`Hqs>@FNj&_f`bLP>15UJ((rQVyR*$NwQSw>CNCbj8_w^KxL>6`yoi)t;$)#@r}#1^lTvgy zJTxCrvnmirw7_CN}2 z3B5+SO6>?ild)_&>pKdqnN`Z%722gDH;yefT4$h;Z&p4EQ4S8DHHwIsIPW#b{TDv$ zy0KgM_r*rDsYQn2B>~R!u5xNlJJ*(r77Gg7(VykH${ZD%6B0tA5kLg1-ywW_eTxAE z;BAAE!iRXG=y!)CnYS8?*+gUY76ltyy*{X;L;wszk`5dRjf4R)Tj2f%63p8lVP#42 z+m23uNIMJls&xNnv54{D*us3bmzWck?$n%4*F^;xJi*<5193J&FR8n}2>{_qS>(6Ah%oNmn`p!ZIMpHRImc7~hEQ^+RrSvP zJnZ>W|Q;CHx({AOM7%+N%-bWoCUky2LE;U85V^Zc66& z*&ga+=c*yFVNn-qEew^6mVGu1ozswrS`@>a*EOZ*FE#Frmu`nsxDT%`J$UjfhZ@wbrdWDzM>A1WmYmAH@SOkglb-twi$GD%cTx6sf zz|j@5jy0G%s&^$&SE{m3KvG)w{u6D#J0v;5J$e%n@n!lVGYLE{`&^+Io~Gnj8HA{O zy_e3Sh2c;!lzc7xwkKpG&)EhlN{!_s=H}NX13lkDD0Be`NJz|q4;rZJ8>#P#k>B@6 zVW?8c``vE~BfP}_Kvu?}h>UTcU97RUX*asMNWo3?RaUXU&g3%--4ikV()>uuR+{SS zX-!bG1Yn>)9lVlRlJQQTho(ot;FBAE3Kt&%n6*FtSK${4Ue zW(M12G1C^s&Nj4(OMY&S@v0ZE-OT_?C{fVnFqYhW5~JZD^bC?7Ajesv@Ov=vCI%(% zn*)SkCg&v4wma{hztJSe3gs0P#BptWI6FdLsNM21%wUt3XE7lGh<3~Lx($6NB^3r` zhPrbX`U^vT$5`OpqPXjChVxYhI}5dD$nZ{89ibpVKvtmm)xTLfPF=d-ES}UY;hUlo zdfx@W#RJ9i5ZHm=ms(!b|GkM;syhiG1cw}$Wd8y-jLK%qA zgtw7--bfYYB$~67D-D+zeqR=s4(#ZnVnOxsx>s6YIjc}{G$@H3wwR` z^YagsQ!-3YF()4#&>a0 z31UI@7TH6WE>fFzYBl}x6Y_lb_})!rs=An>A_}t} zihqMyIZ%GC+Ip!al2yfCFdR!rEx>QxH8NKA3FDc@5 zAWG(G)pU?>$U2LRg{4~*o0w?S&iP}p{(9n9T60CT-LA0`bSd0Ko*3nU#j_> zs;ppb>_9(hUL^z}syIQ>I5z7{vJMQ4)VcbhkjtRp&(H`%!K$OFsO)KJIGoQ(EL1o* za}Mrv5;F}(U2Ua!Qc6Mbm#PjLh8GDuEXwUbN#gtTDbj>v%B zkhdS|USc8uk zoF4|(Y)3B3A63kW65J^BK|udoJ{#&J2pFw5vJ@a0d<{1zf)L{^$ux9OXoSM&r8XZ# zir@0{gT+(g<}3Hm06J~j9atE)HPK{%QheukJ@67r2<{5y2*4bAf?=i2S4>6Z*uuJCG&xqk8c5F`1rs zkMzNN4i-|2_pNF^3g+Wbf1btv0(=}&#BJkeQ!oC)Umno0X0{ll^S?-I8FELpm@;$@ z&rhnonxXi$ztP2Pzj%5mfIDuTl$~t&Z~ngFFG=~2zurUyUPe(OV3gHF$7o?-DShI}$deSR+86^S^`Er;zkiq21<01u zwO~PqdAjV7e3h=p$-|@Vjyw?WUlxRx0I%i@@TR@wyjibK;U~2K7Ksor2y|ZiW(L(K znvd~7hqpxE%RE{F;S7H%^M5}b_$5oMLt_-QG#$B&6a6>kGXPCG;V%b%-PaDfVGlN} z`F1>=gz?`%^hm*MBc0-%4(?WMNO6Gv9vA--Ppi0p`QxUNXPCjK|Z zBFkn#QW+`VIVkg$jXXQ}^;F51Pix6YYp>+LQ1Of8|NNUqg1E2G#r!wg7ee!^9x`(> zL;Q0#|A(GTlBm>F@?kClf({1Hi(10TPFTnDZ>v#7PsTzyME4y^yU@v9$h}g2+Jk+k zN+J^kL4{QS*20q*@k5iYzSQyKx~0gzuKe&nxBTaaV`|WS>DW7O;KXCM3U^wXJ?prbVz?!Y)|es+mDw z(Sh?X>`f!ix{{KmAB|qtFhBaDQUnWZA8nx5x3*v_MesJZsxx6#BuK%$MI>6!RM`HE zCjUTJFChpO<7jF>bVNrEzhO%%A?V<&0>}tut^iL-Kg#MpAc@A<4 z*oJomw*`KxNaferMr?7MmmadDdVV)78Dc^Rc^@g}x73e@Ay_KxOax!Z&|}BMKm#y+ z5&|SVNN|FW#2e=G3`8NG*3++D=EiT8|49e*%S`_LLyuiS_cX8hAG9aD(V8b_yme&J zo@Rp?;Qlq-_efGF`p&q!emo!0&r`1tY0D-)cvIQ>|QLIo2TfEFTv%JdewOQ8VvSSa>} zsOn-c!gj$BBdR~`#7M!>qbj3-_ZfM0`J^kTcDVX$wJ}#v{C-eR()9{lnOm-5Vnz>E zun!=E)CWu$opg=$O(^QyRh>&@?Q*o8c|_&8LD&mmxc6ROQSsGdJ}YH!1SF9Y|AFP39K1ivl-CyY+?mu^hEZrn=2J~CFvvdW^^GmAWhmO8 z(CF}$uE*B@*^T2V%=pJ=H$N~4EMA*j25H1&F_5kh z$BF`7=RHj z4ETk!uyF)gZD@cb4{-u5r^<>g_Ys6dO`-${Z~yScG*Vq30-7LraSC8>LLmpiE(O)j z*7XkxfBSU^nxDjQq+82|prXQteBSkqQbs%s#gh#I?n$aGq?Sy(j7{s@P_-rG8-Dl@ zPc;Mi=aoxVyd%TN?~>cB24X6{F5y;~_h}~P-tb@QdgFgw7#vL870}k%urN3qt7LX1BM*pqF@0dm^h#$l@g21(%oa=y9g;e#IPH~QV@T5elXIS2)04`#x*Mn-=5v6bUeiRT|_Pooe3``2NIr24vZG z<%Kf?NXt*64h=sU{CgFvusj>jFpAWMetP7x1b-2Yw#Snt1zV4T_nyW}fc1%~MNkcU zt;kwYin7)}&-=yAfBf|j2O9-@I4;nY0RF)Wy&71^igcq62}MKmwTe3^m?rAiy}pz;GT7hxPW~n#(cz#T-{?h`IMjPeAgg z!z~6Z1N?T~{PWY^Se;~b9pSwO4%#q}IKVQ5E|I@H9*p2_9v#%+f4mES2NpQDP>tvd zgI_#c76dN0Y(z9TLe=zxfzslx7H=Ih1#n+KyaiagUfC*Mv z0Xb-ZC*}M(-}vOKguxBN7T>VO`n1@SergAB_kSKttQ84Y7#sQIcn+@5T<$lK?N7V_;_wB9jDb!$cqUbN{b;Ljbwvs;d#{ z@ffS_a5Ot&xQjw?=v`$gO=SS=sXH|E%dFjZVP;?Y2{}F9(!s*OWTl++GUm7D8sQy% zjpydHrHmX9dm;a@n{ii|mZHgDNVh)j{CfH3;|+6>8H9fTC#T-jl?zCu)Xxi+zuZ$q z+vk4;<^Q*_9T3GcYGYzNH6_3y*PI=-!9-SGc>Ckdy#!>8^>~DCv{JOSHq{Ftk zs;#G!+>p!N0t=Ue*kKRWyk`?`cZt=fMZoxEwCsag1D!^SYHkYQ`Koz-oC($lAB{9y z5MscJd`QYPbo77&Zhv-WLZQSgbiB`YF=0NuxfFdw<{&XT2fL}XslRN+$Vc~`04cR< z7KPn1^Un-eQ#@U!vmfgX#{-sBYfjtYOjp>$o;x62^lWWP--Yc(LcwOSiHTv~&bu=Q zdeEtB$r6mOSH2ElG@`4j3MTWgPd|luU6J*DSZ=&&&$hsGl64VbW6Y=av~{D3Ggx4% z2^ff`{Eb5_;GL3psJp)+U}pBC>Ql186t9!@1-khQPRjIzpcSdB z&MsFoLHo88M>pKJqFPWahzO|8h6jJ~_`XVtl zTsJ7}gI@ZZE|e~4!UXMRr|E{`1j_DblDvujq3&+AVxj)JxsQ3r1*=7^w^q#gnxh!Y zeCu^I>)P}vjn0tIvjuHZtjk;;j6xgDm#Njop!BRi|M=d>T=6lEWs@yYAF^pYy3Kr#K6pY`wT!xbtmX z#-vGu`RD}{4YP36Z^34axWd?!llBLNP!Pp>6w-vMQ}oyya`XQ9Gm)F~=>9cM{vaBk z;@NHSlKBDa%ATI!NfGg=+%7YbZ$sUBQ`@eD6=TgZzkGcFKzR~*rOV(qriN^VESwNd zebiu|mW0Yd!>$g3L|-=L@;3@`HVex@m^xFhqx4NMm*4s_n^?XzPTs!6fU1c8O)#T% zwn_|ENQ8^-$5J}2bLMIZZfXFux4*q5T7gDYB214NMB;_TRx%>3rF*Ji<4i5yTH> zHj>j#P#Ah>2r$0|MjAwwa{Kp2xO65FACT zRU6C+$$f7dgkbT$I^R57WOJFM>3a9P>-9~y#cRK-5W?0C?)zCQ#HR|o zH;Iu>^t7#`3#NyiS^Sfqw#0|*^_#m!p;(3zb>2veye(J+WKnCdz$jfG@gtM|DT*NVfi=Fw_NwwfIiN80$1VKkvNNFj}z)0n!w3CF_ z+B(+Fr^FS$Q-5(%F&yD}U3!M3uC6YE8mkz}lzADz?zmG$laTJA7CEXz>IT8o7Wbe@ zwrOTe1THN4%dezwai4g^tG>D<<_cUH&>tM4%Bl$i5egENpjo;FTN@+1H0nf0!xXPH zFfg&UNr56I;#Q_8uuxE#&KO)|hC_RS6j(0{fi9RP>Ub5`DtF&d>g9lgT*>XnQi3G) zz{RzhFKAkHmidg_2EGfP?r?@G4TNS;^VVJKOGx=E;hkhPx#u5!Mm1TNVKraOPBPg$ zA+^2OH}9S@g#||(dI5DD?%1V{EScnHWK&E}zIhlb&71n;(_EgGp7kzGOM}_==Nk1O z8r_`1?V09`G(o)#rwEI$;w*8IxMzqL2TMy)ms36Cx#Y&Phwcvco@?EdA1{zm*4#z4 zPksbuPA(pK=&K-vDn#IMNsMx{^hPO9VWj?86g~6QlG^(*FKjfDkvlcTYyyKu6<2t* z>=A!^^_kuxPRN<)O8+ORZ2%POd8JaUWTV>DT3BF+L9GDO!!8M zPschgC?x2&<*$9>Vdy4Cx$As$>U0xB9;HsG(w{w4iX+hwL)=Y`^i2m9%XE=An!X&@ zmR=Jy_duihBVZg&&M(IL9PZ<64G1=WXqhPxjf0Us@Vp!A@>1v1^31b&<~x})StJSF z!*Y*!8WvnzBNN}i@JLVT{8$YAcgr1gNK>T|_nUPG+Akkk)R(_X6rl?1Jr_~MuG=2$ ze$kST9hR-4A~n{i`igcXNuqCLV6V|_i9+!UA}Ds}EG#u~Rv|nSaemUTsnoFFbaCseoI9tI=P&G=w)MSka+!&w7K2b|d>=o+oDf8}ygkml@L zd9POb)dv(p!P$mOm+28RLlV}%djWtbfr(%NUq9InnH#^6^k!`( z$8L8=g+X*_f7f>I@CL!+FFv6s-$g^|-}YYV=MPP8skV=hUy|h*(M&YDT%>3so&IJg z(P0LKvk*reOzmXLIfpORHP(}J%TAIg`35%e{fo4!t((x9*)uc9t^wlsfAn!HtF#vHE+CPtEM zktPy-Bt8g)i!gE*HYr?6_`#<&NSzV1sfT?*AN2kW4P#BNc7QM+8a`0WTwBb-*3x%=d|666})?pMdIpb zpYx}7wi9$J(^i1RQmM(RX>^+2cK>!fw#}6syg__R?}!Q_@p=A7ZTn%sDk35X!F16PxycC=RN4!30*Pph6)u)(n~K{ zs1MmlWrTmIS_NwK5tUzgt%4xFO05pWe-vas@l!2#M8M<6_dMpq;xOc~eAm*C!=G)Xpv_ z*AIY|cfR~0Gx?nXJtB&hv(kW!3|q*V6t&HD+(dq;lI-{TX+ zmlj{Lg|MsqqXeZ?gQ)ZH11OC)9KR81G@A#xcB8$B*Hc5`j^5=Qk8SB3&mn$T| z-)sg$qAB3zk3`dCrG(gNj%Mvoj)97U!%w)Tt;fbqV<#)szXgBG@o~6SodORz?soTV zg!mLcMU+LbNHOn3Vec?cU9C>evs!XA_IY5F?5S!IqHXKS(#Up>@X>PM1Fa%5VXBJb8A((Za#A7DeIg>kN7hf;T`8p0Z!ZTDr8UeaW- zUaG%#o`5*fc`c4CuioO6JU8gCYrFL*5DfipJ02rbsp_jkMEEChzM(M~*gEqSkpd7b zI#w94-s*hS^0=38(JJ~Oi<9HsqPiyG@7G-y;GMzB8)uzI#8@G3vbQ6jf zo6Y?7p2dAN%J%5S8A+L;e+k^n;*+7Xa-bb}BK}+F*n#-)Xhs!&om~SzSoZRw9yIPl zSB1ip{u2KV`;2$I6#wlzqR<3+x+T#$M{8I59uj)M_1^5acD;L}S>ba{a9BO2;J;rf z2$v-(wgP#CR#sIvrKvTiTg-YY0JtF6gCEde!gf8k78(DFFVaC2N>fE4?xzi{dnCkw zXhjU@>{qDCbd5%2B4LZYk@@Yb=_*aA+ENT!G>_MO(oURFJci6An1Tq)WUi|XIWXzj zE1_4Ln=s_4D>QLJiA6@m6!BclNZ+`S*N3}ui4=o~uyi6{YN)*uvaGlY^(8l8{VH>% z)LTXMGKjaE_oM7WKNAO`pc92A<0a1U44TS)GZ&De68mVS91fpjeHA2&Pu8m`Jf#aA zt1C32gxhL^^cL3Z0b zeHfVOCNW(?Hg@cK#JcXhM#RJv+sl2DraD{R!kz)doyQ~6UAGtPZIKctR>!`PY)n}3 zdfb!5Bb(~=SbSv&&{%6Z;BuJM@_ZDEr+Nn@d+bt9?{nI!L`3o>37d0WzW&(m_$Y8| zxGBSY8;xXWup8#ccReL_-nFs+F32nCw4`CLZU-rAjtk-zMz(PuNIL4mxR1!CIyXIi|-3pi?9j>as|hl zop_85@12F6y}+X7S%2gn`89Nn&%*G*u2;}8AW}d>0I6AdR|qd*b3WWB`q^V#VI|vJ zSY)8m`5g)9`&nz1HEac(Joo~!6-@{HzOSk&F@DM*qgLf(8}^o###{wA0> zqFoKC9^3kK`FZHF$rJ~Hd5lE&eM0_Suy0~-*Xw}>cdvF`tkFOcCwQGY)N#`8mu0f# zavqeK?1^F~$yM9leQ%0LTt5}5Q6|Mq4tsKZI$Sw-Sx$j#$}nw|P7UxkeKBrzQ1`*N zgNT{L%i_S;t$Nvkp^ut|b;d9K8_Qz;pRwr|*0k=vGBN z7?Svbia8Oh1P86MyQ^?O1ky?g_1gz7qph8f2iKoh<_a|lUJB`$TDJZmVzHc2iPfmo zb=E~dV!LY4h~@k7NG7kYZm>K5U?|vN15DqNWvv#WG7Uozsv;34tp0Rv+z$-@0 z2Fyt{UH5oCzEcsmnH>wv=S&>Mt58!euP_WZJ@5TEI@vut?EjKp<{tmzXg0M3%EG12 zaRoeu)jdEJ$#p;Jy!VaajrsfftbfH>|H%CHjIjO ze4CmKe(NbNCnmtNeJ)u>4I^x<7>!s)3TI;97K<{<9VguHS(-(2;u{5aLZf`CNSHQs zrl;%Ct>%=FQ3o?Ea!vq6lP3JhWV3X;Yu+KH6NvRAAOY+9{F<2EhiD`uJS@m1HKL9o zrfQo!S44CrFnaU^e!>+>P_KqsVM=Z^7xcLWwm>biPu}MmEp(|0qY%x5F8GbK!lYsX zAqm?|YGI0CHRz1y>GxOI0B+%<^GaoAN{+%!x=4vj43eYEsxa9nAx*mMWBMZ0_49qr zdNX^Px0i4sxKoCAieHGIm_HdIWb-8y-Cl^9kZSWsKKunCOWiU4%g$^k&kEW6FIBTg z1B|8iAECC!3*FgCVIomYx!gGWIs(onCt#V^qHc+#f@t@qXu=M+b?=EE`)&d64sEQY z|BH!+=G-DTIL};ywK3ul#De(l3_Z2})={;Lx7=c{pj1c++@IB|7!RNrcg=of`YWV(_Ms zHKFiQDUOztM4}N5tESPfqfc-c5_F@z)n>IM@@5$oeF=5V>HSsR`6#SaS=@&~ODigJ zI5U~a6#O8YsS4a{1NE?X^5Y9yqS)^1*6ij z+=^BiHYz^(=fhp_GV|@OVS)$-VWEQe46g5*VwlrZ`E?E$6l-j(Z`9fZtU9K#Fhb}` znWICnrlGR&K7G#5^+$YOu`!4RzoYh$sk?}QT4Gdt!i_b<5TaMNpV$K{K-8|Q`kiOV zRZ&TR%6$HF;KDAt%eCy5Y7_chV7DMNx^hLO7$W*mMXsdNav>rqzr*h1(=YwT<)p<>KvFL!Velb>3d{5w1kx+f0^0M^a*9 z;#fS-Q|D?r;HA=HOHFpngI9HM=w5D&7v@VQfEDMXRHmkoaf6A|JL)UtD# z(v&q^pKQr{k~m1O|JbOuiY9ceUbSwp{Lx;70`KjgnTLUnNxgrVbe)>S#H9Yc;1@{l5PRg$2b8BO`);^0u-c^r)5NE*Hj zM6>NR9J=!52XU$a7M*Jv;78ggKkm)D+S$Y8OJ-S(!q$?>lq4jqd;HLTey>pDzGb0S z?~$9G3B2$L@#!66bJ6pd(QE=V7nKl02l#hx@ z`54}6w*w)Hc2vHbo>|?TUP&JQ9>qccI07) zyiYcK_ucUBX#h`PX)=!^5-zfZy!4S+Y%l!wa<;F3VZ=+pX!;*2Di!Us57eK8G}Nm_ z?XmbQYDng;5Bs!CB3|ho{qU7X^N=1+zMwr(lMbJ~e2AEL7BA;)8?k)8(*a-l|Q0`vWHv%MC1z+0>T0<|`NewVMM_d!|5L(aW@ zV=v2z&Vr8BPx_e~5f0^dWNefP8RNKj1nDC1#umR>j{&w&mOj!KYP)&7(grYk)xoQf zX{!Br*KcUCT_pEDUI*z|?J5OF^gcXt$DPa6sMe6M)a>OAVV)|tp%IB8Nk#lZP9g2} zgI!<-5&qu_;_~JjkB`Dt+hSh7fx?|ieKf}o!0b|>urE#DDw8N|6_Zwk^M^Nu?)2&G z4m5r#hOG5Eh#exhxWh-tDlZRP1PTM@bG()7<*C(Sm7JZ8rv327F5_Ee&#L-2i4$^9SEVbi>~&DE8gla z%r=EfCen$}x11(eBx{ldV#;A#etqk-V`881W;V6{!)*HAqGg*P8g;_#`7j99tWM{< zry(d*4gsO?<-y{s6!V3ach%Wi?V7kQo5imKY?SGGc`q4gRT`!$vdhoyFTXn9uI7&J z#A6s(l(*fK={(aa{=6l_=avfiwTh~lXU%3UvsO0grq&@tG8-%Sh0U4TA4SsR$}Ep& z#TE#T#vK_bu{U=m_oaOiEVxcKjJ7Ms5ljTD?<@85Euc{I5zt90z=X{K=ZCKpBs%P) z%G=S_^MFH~2HH@F>uU|GQYG5xwud#L&}=c_<)KEbWP#;6^aKxw4a&Bejh4U~-b9nO z%Di~7j_Pu{lStYbn3nrMU%jR`>AAT5sh$r}9S^nq(TD~nj=NKnisz5o^492n^ZDq`10K$(;KGW9skCvV!mwrqK-}}3d zpnoQ^|7RrNSOm1+sk>9SU+MGNFDY$)Qds9l$j;8x-MDz_8!64UwCkuArHUvYeBY`H zK~4-!+qO-ptZi_2;ewfw{o(9u(K;a(v-gBRo&k_Zg^ExEg&^)bQ&g+*%Yt8WM_zmSJ&a+tzS`yGw$*OM<(*ySrN;1PJc# z?!n#No#5^oG!Wd~-%a=Jd#wBP{r-Q?!;gY0s&=hC*IZM^9OJt+1@>OwQhmj3;z4Wl zFm{n7#v_18_(>`Ym7uZYN`H%a3sg^K8dC3IF{XWlCE<upV$>FBH??Slfq*yWRz<)Nwl(CTiifL z!gIH*S*0A++gC$@Gg#f`VEDuDdoKt=w=}!n4~ot%(rI5CSwESuY^=d>Zjd@*@(Nu*i?A0oni=g2;}N3_;W84p zUHT(+ebRHAy{oBh?;~rB`FfAp$`>On=u<{WjDZEi$I{8M;%oar4gxtNu-X$6VHO&Q zDMByTEKqQDI^>DtiRKw}EKW&+J$pQNo)ZY?ZsOjMaP?XPyMW~fnRaXW} z;e3QdrOFX?4U62zgWR5*N#I)yV=HjcTs@E}BY;$;3Q?zwj`zN~7Am&ey6^(s25Fy5 zR##Wm|JQmIQMVC9xMctq7=|g`cCa)P^3Mu;udNM%GZ$MWKTdt`-shhpC+I&!+lsHB z7;O}&;}$Obr|=t?;rqLt6G~+ag7mQYI1xhG=(A*kVKj1S)LXOewC-86wd6tc3u>iRKmcz%pPz=^?=Zm=+IvqDH6g4aK4T^;vFYbEx zerwm_rZNOqIlahtG-My@`ZL_p$I@H$QfQ%7bv%`I%?wl)2g5}BpECq$*=GhT2^|%- z7bFnsYLxB*Gd%3Xw6czOfOeyN`2I!E;u6DkxnAUJ%SZb=Ue)b=%^YOlh<*j;n)^j? zokakHBOtlW1fa9yf=e~p5_Gp`}ty%Ew~AjN9bz${>vK0z3C`^PdDAW ze+M_*VtH%!3r|grGZ<<6Ik{2c2xC>S(?;-i)E;#aInTmOswn+}Onw|l8lUQN-t^&C zNBad{Sf2Bc3a~c5>tlt>dij;Fn95(@W=iqta{@FQQS;|`b@_dt5J(Ug<2@s?0N?Iw@e#d3^8vE z&KKS}MB6`I`7jNlf|CbahUCt{3ykio4XuAhH$0hILo%ZQ!*JUQ-78y65Kk1^jqsCb zg76mJhT*@wrx-5(K^tzKf^SYV^?}xrf|jR%`YY+;F1*i$gy@j}p|)X`#ac|8`V4w^ za0(EA5*6xs`$JI)q}~jIKw6I*pdfrSh!hTWW!aCQXikWcEvYw}a>lJdcX|5!E*>HL zd*P=XxqD^5&+<}rG^DVmyo7JE{7h}Ow!~3)XZC6o1_Cv0@b{bvaq3U`S&e)+8pH*$8d8Gu3oHTx`UG<&Sp?^nADjql;WqzWz(_>GExgE6;Y0sYIE_r z?L>X=9n5ahpK8y?I~)X~D;@o3hj$dzl74QCIJFrM#pMHodkR8+hh-L1oQe*;=!a#1 zG~24ZT>(N2CL+Rcw+^n9pe>C~bIh-&_N3$1oQ!I;ma(x)RgTp*Q~a~mtFX3VuE)*~ z#VX9+3$eo)0#a2bM;#{0?0$p3H;8#a8SMQK+as)N*klb)TeEzmby|_z-Io~1uwF@3 z=;#S&X=NF^g;c~UYe5O2+NJ#XO+JFzQ4v38X_RrhEaZqWl+-0}(yvP;xH z(;p@Wzli7($2OZQEyAF7%tx0HB67{Y(@825R9A`&xf>%%Hb#=V{0=j#j07rJ*S6=y z8ivO-w4?pI_s}XwO~QLB+Zi=9^*gGLt7y?Lk;TiSWHQ5HktOJ;D=Nd#J#xQ+hW~ZN-`bH-$~Tl^EmB7 z(W=*nKp<4gPIUVC%Jy=btrRz@Q$!l}%kH~)ud~o>7NV#wez=}+g_v4>f-I7ZeEa%= zF{kGrKX;aG&P!dly=NcRd2chMWLH6<)khTH)Zrny?}=|H!c8D>NJD)6tZy;h@l8BK z%3X7446|xjC*{LnZAjenr#aS?d9pxr=e@gvLYo)Do*ThWYcFh?esS*wfmw`wE3JQr zn|jG2H9A<4(7t?YQya)4V=~={_nG$eNXw$G6nnON4-CmzpK5fiQszp#>{*qUL=A!V zvSoAKvmAiH2F!~D(bSL0l3$z1_jc{n@Vs3^joqr)^C!ZPwU7J*=?l!`0Qh%uY1Yf$ z4FM}?a;tdi>GgIl^R(1J@oaCl4fJj$y)L|&86p_Xi%KO{Whe-#Pir4PLBUObX{+V_ zq!k=j>EKy^iW5x=-*(<1)%-H2prq@Rrf1um%BTQm;Ad_(z1y^rSES_k+&5+(|G}>p z%c4Y;*6%A$46D`+EdTIPcBS!(TnMU9c#5(QseJ!E?h?GZD3@9 zvmVd)KL}f;nz^!lMS!w{jHa*IJ{)+oHJ5Mo#)Y#uiG4WQnaZO!Nqtr> z?va>nPkDX|wADz*lwb8ndGCt`3sGT zljr$?{oyU;prKE*+gK~_A&OchpTWkDlOh*~Mc;sC#P(;yL6~V8O74e(3MSypLkSi2 zxI2@iL<@-N-$5mpL3SnG@RY0{QNA9er0mgQlnUpzcDuWKz_RsM2n0tCMBwF*Eie;I zE`78sQ}746IiG$-`@k+@`=_4U^bF{#yQRM@0`o5@n=sch^0x=|KHX{VdoRY;@lDOh@%s7=)A(>*h*bBWYia8aqQL4KMUihLRiy3Y)@T7#{1CuwtJM}guO7JYHsz`3D%e`g? z8l$7p0*Aqm+ourG^?P6=2g?$$iJb7?B7}AkA-3h%WE`hJUx+Zw@0>z5H92o`e-zvK z@Iujjba^mzG(8=D(vVgqpuT?)RBQjBE2r*wnN9~ zGlpq{Vjzta`C)pgW}rd%F$;$Ke7!Ayv33{nOb!|1-)pVFaPY~`Xfy)n61_-t0a$q_ zTnZEgjmMpZYb}kZx--{3|KyBd7lSS>RSrAOXqBZ{1A*l?zwxJbj-;+ z-=MIgEm4^!&qidI-ZOGI8cD1t|?hc=IbF%`Wp z;V45}UEvm`yi35h{peFmA>x#7$TmNWq=#l2KYLT(-K9Auzj1b$Wn-Hi7-`yIo?W7G zB#~*a^L4R3){-c-Ow$)`-Ra08ihT1UBnu^C5mXN$?xrnl@PLnO9nykj^Ux03qUA43 z3F2;@uLjL0y}nh%$mIBpuf$%DVk-tw|C1=pcd3|^OiUcX^y9^Y?u@WXg=mCxp9MSS z@nX9Q*0PB(B()pog=q<`hi;``z=~bLEXOgG)k<~KY7qyKtGW{k!O=v?8AzSxs4L%B*leF4BZ?~2%@<7hzt2$kHA&p9=wC>VcM!6U zpUvZ18nrBrNi=-hqv{8-XPMC`5L)b2Pb*cB;AoZsYI#8L+uT7p>H-y4`q-_0{mkTE z&CXuLvJ2`9-$9|1&comiBMx4sO+~r=GA+0LvhmIqAjOAPM8i4%u3D#ofq}nM_7$2; z%9_xN)|%D+MKNClCySLELvAJvv7VxHodLjHRCcRJ2dp+m{1P5$MlM>65M*bv6vo=# z#&&IL5hbZw(0dxAp^R-~l}dB)6$v`Un(NMFQg-(HaGl@Q>_n?eS_zd|AO3@wWgCh6U#yijWOk zi{MK9hsbRh8_d4Q^ZcJB)2>3=b^Z{L6L_&aCD6!{dkT-EU5p83X+EY9{1rf*yjMbR zcrJ~}{9Ng^sLK=yum)g*VTeYiO^l2xG(t*~lB238pRXa1ryj6yem;e&de@TzWYp=u z=(PkuZ8~N(m>c5MFUC4SdsG-O5gQHHLow<1zD`#s%W(?Kv)HT7W)tl_eJ9{<)N< znk|6@Z&gA%?=t}ov!5_VUrm%%ki8UrKhQ`^IF!%vs)vmh4qYa!rszwA4uE*uK`xi& zroOo&^H^$nmEswcA<3=RyIiye+6BVj`5lNsPCw|0Hk-=xm-#7Rq>=!M7(u0z4`T0k z@1l3!!sxw}rE(ZiKp2{OZ7c~A62}x@Eugf!z$8WNA}54_Bj78T3P2jO(?4|=3a`jf zjzHWD8hDpM))9dj(@b9!!SV5Fwe(&$8Ca?jFIbNww)D7Rl+Rytmda9!%jt9A zUz^!>AVs*feVCQ&@rg}Q&W2y5V>x88#t`uXms2u&<6bkMwl*;_7Mh1WqP<&lgXjuI z#q8JU3yxJ^#s|d$vZwKd#~79#PgDl`1Tjz)fO9f~fKQEsl+N-cv!dwPfQ2;r?DVg` zxyEOcg5RId0u&?6^0h{N0>Pz8>BT=o#TKNyFP1Ncd42_PXfMK5E|gm>zaLWx39CGI z4_4kyue95tZ?0P&R=osA86uc8mY_~--jPhCJNU(!YA)+}tmk?B1H6Wk#X@e}APK!@ z2!rA9tW!U$$og7&gKSE9GSIsJKiae2L_`yo_F!fAF>Oj{&vMczny2f1Oq@k(jIjq$ z(Y?Oa1Dp4{!xZW`{&hyO7=0>4+I*7&T3vzP$+$J*U02(oKw!B&9<3%6C=}Z6U!lR3 zQeGL~+s_}_?vMM~YhTv&ms(DcMjvfi!reZ<MLG-RMBJrq51`biHryGoZI zHXZQugSpPuVB8Fd%g<$_6B74sH!L3tV3wl%zWjJ*jE3@f|s9Go6eK$P!4iag_Rlw5~y|+!A$IN5waR5eCzE4)BQzOa3)1}cJB5I`#_}U#ipd8=atMc-lgOpIIJXogM*nL( zW*F+YqIhbTZGGEt^K2OIbe6$B#M^ zjt5iGbXpTPA)To$j|Z-Gb(YgHL+VhhSu?PR4Rkzm`-`sWXHPT+Cpg#dsJnEXwrTHO z?v-4NAGFd+X>!)babp2?Kk)3a>-c!%XV$*O_C{^%bjyn>$3w0(TOz9<6RWEuX#JE` zu1}+&+yCCmTbBlN;l`q5*$;kx=g6%QGXD_}I)%G;{O28Rj{96*O;NXk?(BD`NJ zkklVV0)=;%x*UH{r%w)1uKvyHYLJDHvmEv0keu1;a2=irS!lwTO65Wj5Jst*oRkWfoP1Y!XG@N=vwG**>Xa7Uw96>|& z3~>D8+6knaYt?!wf_NB2l4~vl3)+MCy?pt`Vn6`eD3ZD6rEjm}=CrN*oJfr==z5K7 zUnV%Yv>L9tO3-ezt9NfZhx0w@wfZ_MoCK*@qT5g(H+zzx(-WBfY2g7}^A3$>T zD+XSFVscKGs$|<1CQO>B0qQxEE8p>=e;gDId+IjEK_E2FZZ^+2xF0ne2)mEtAaZ{g z`K3^QvLH>H!a6xqqFInYc4U^_`q{S76xYT=@(M=R&}I8uBF!ehPRrM{nGK7OAx&RA zQPr2Sb9&bWyCLH<(%G4BO}yUO9lH$1QY`{lf5iEJ;CtL6_{JvU0`5S(MIqx~Clbth zO+GC^VyJ#rx7Ab5 z?JxJ14b1D((t_qY`s%meSNEwyPU+f~Tnaw)&1uLTvx8h)uew@8JkQ!mN!2FLLr$Z! zn9p=F4O=WBCx-c+99AH!Rf-1o_i+C8l!)z~P zcHg3Tp62qi!$rDo>NwX+vGq$VxpXFw{Z72Bx+V98dk&V4F^+eTl|>8FJsVZUQ>D-r zsEqDk9>0*KY-0+G2*YNniSRIevu;t=ZB#_NsAdjGkMi|Lp7-A+CR=ir7tFq$^)Bi( znEC!aZ@#FEJqsAR{KH;;fAE;hH!j~D>+i&KH{Z}L2@d&)l>KEe)_*7#tN>Iv1ssF- zNteL^4^LbspRLD?O187@()^RBC7VcI*Xakv`1&{dU}h`1UyA4Jts*@w@Ypg?LTXAjow za3nOW`fTjb>4Epd>3E`@Svj~w%5*%ffh|z~=7)HfKs3wUlPoTRZnfSdL+Dd6Y_om) zps{x^eT`=Uv*nAjBeiye^{kW|;d>%PxgCl*J-GR+Qo#YomBUFWM$Dc5I4((M)G zD4TWGpw8UlqLGRk8XWNqws)#%mc~^G_%vJc%g5Xgtr~h+{n0y~rxVDVAJ`IT1RFJt zDn8uZ-9d~qv_4NphGi0DUGF5N98wB>TK;;dK>k{p`Pbnt(geOp(XG#fp6;o1q0~Kg zn&MW3K?C}N@4^>)Fkg(Ld&kE~73)k0XU=rg&5&t@FeXQE$MkP#fL0vD{D3R26zD@jb)=|KoMa@{Nx$t80ppIS+E$)Y4^qnN7Az zOu2Ab6z`n`d(%1MCqS)4v}WgtQ2FNv66KPf5vRdXAA!R0(R9C}((HLz|<{S&(AFILA76OUh0s zk*C)z40|oxRH^-<1;?R3?S;14liB%B8Jjz5MYp9Z<92$CJ@YGn)x=>@u%EALPKUcC zCH~F=0@msB@J3j2ZdNOV)h2j(bPtJofEm*E6)w66U5$ zUnAY=5P1|+c)A@Chf4qu@0d4guviSUP^u^%yb#K^OpIt=sirVJ^U0uO64JH08?3F^ z(jl;O z7;%^=w8Ioe7Graf*0nAQD8^5AGimb`TyITP(38RffA2RFB&Mhee!gM046A=mms!T7 z`ibr`hUEL6%M1u1!8gF!>LY%-Yhq*mHoz;#?>BFp_5IsgqSyv&zS|{uIIcS{v(1ti zN0QuE83cC@cQRb=uL>Z2I;0k}r6N0H+L_`fgX39hac8%?kQ78e++IF@BpiFgLCih< zc~r-y>981TvP8{mt6fxl0}kX_oM3@B(O>z0sO!95P}ZeqB6*U8Sgms<@Kb7P>?k-n zHP5L&nyr882G=rAIa{I0!js7yOIyNb`Qq!-g9^^NnC5tyBU|7Ou*~5z9aQ;1?M>t> z+ts+)3_ESavyQeBo#Lox8Z^s~VxL#YIz{NXUuf4DbDdod*V?WqBzE&8IcS8vKP1c& zz92KvK}knDC{014Z_B^TznLu1oy>L2PeI4!t(`4~>Mr$`fq?z$@r_wjv0H~3&)J^q z>4PDbzh9DPZky8?!;0=w-AFi^7?FP^nzHvHOJL`$1=(ME;{{cjywq*F+c6+8tY zjJ@LRN@|0;;M&Wd?CdWy=~gLE5-iHb-u?q8A4pIUy({?3#KXwPV?pBeD~me4cul3-KO)$Q0j%v zO@qB?$bgA?M$fFD*JzeEYgo|_-8r#^`!r8b}3L^p9zcX--bh&H+V_R(eQiRtEs_jM2|xlhV0cacGoaqnu$ zQN^g~R6TJVy%;uc{4vZi8yB&le7jFsI z>oEN$PZd?Y@yR5=VtmVrbd z6^Bm^+9_&SnetUounsyM>xzl;;nu!9nvG8=(;^P0A#1&B-$KsOPW=^N+PW`4JMID? zBJ;A{hE!F%n4xaD0r3ChGlrq}>lw683U*3bIA9I-qVd!dtT^S=+VyxO#g6k9mL`S8 ze7D}&YPnsqf&Q9lBF_$z(n@6fnDt;zo`KbhMl?2FXrn3iR@og0yQRLYB(?&xaO3Sj zo*J0p&Xg=5DCjT@FqZBT6K3zuJ;>C2(P4B|-!&nm@lyw-AAiS|nE@qO6^22tU6f4O zkxtfn9YHp(;M2=ZjE`qsw29P`3%3x<*0_XNFtDzyd0n)`6= zHN3QHwP6W#FVm$RKy!n{gZngAz1!<^KJ@;6J=BPBOpb<2$|FJ}%I2{j``x=-G~j&pkmP%I zE?T2Qg#Cj8>V+^WHP&+800ED~N~J&ZPJjKpHBVB3 zJXUTxR%#R6(bzPc0?o|rDr7qcad{$H(f(LkGn!`2x{bTTetvK^kKe)YQ1>SA?frdd zk&G*(z70XgD;07co-rEP zv@U(T!@BotquC#2xlC*B6UsREvgBnIW|wHqI@gh32f!vD!9#;D=(QS!-V+H4fFhe1 zbbUL4@81OHRa~B<9l8I7xYv4JTWjC^l{=YEr&S7=2SvqW3EwxX5#f06S@Dp#uGEY;)dqoIqUEim~HdfDfSs4Nk%C1W<9|iedx+p z>&=pZrRIHHZ(}-m*h7GY72&bRmws?!ur*7~X1PwP`Z8!U1I%;>2H<_U!JX;|4cQzn zPCx5FMY#Lkiu>`Jit%Rty#gD$=u#Hlzn4RAVZ0AsXtO{z!hI0id3a9@&G(g4{U>0D zhZOr=iD44E$I*J^cbf#hW)SQIw}~`K6&G>2eqgs;3J&?EGP`4np_6% zzwm{eK|3U_@@KAorRKhK&SxAz0r7i+L^yCUEmUo&&m~D`a7ZDs3JMF$90-CE_y3k0 zFHDp2fq5Eng}vLi4@iUffxh%|ZE@VMD5S$Sf$`@%N5(V#!NyNcPHe^|CKknN`?5us zKS#`G8XO%-%`YDBH5Kg-+)h`qzuGKaV1`EKw+cJ{aOhCi9XSIalnMY4IIsFc+GpY{ z|MH6wtW=|gVoZY#N%^1%NFlfI%)BE)!QF%vh!_-Kwd_R9T!*UVk#zcTV?>r@ZMf<{afXF45R9Maxidd>V z({(q||JCdE>}Xo`<>D^f`Y(~e@oy}R}&9CRu9rI zHI=&P%obyjFH+~d(DkP7eTB?=0hEYVOV!%@RC^21Sd12ID?h4mKdn>+u6mp>T2OQc zhPjv}nG~-5qR$UcZJ74*XnTeB*ssf>;Bwlgkab)Ufge1|?LGEN23#{2lfEmzU?vn8 z$*{zx+5`c61SqjEJiZWD!M#RR6fhtWF0d3K>~{z62noB0+O?gvt9nxEUN~B-kg90#`1~6^) zII!>(uG)d*`S#Ic#r|MQo}-++ zU;+?!1(>QY;-ZsA~TaY>V-X5Dm`8#K?C z{j`nU_1Ncf*mrk#N=dphr~<`Q{?edG_dichdX&a?hu|%7A8t`FsAyj|gv?66B{VvO zUX)b732y1^y*MN_t4Cod+Us9pN4a0P$I+(dC6A?(vb1D|(&^bcPNPJaw-hqShbho) zuaVw9>wFq#e1ixdY0ckB4OQ8ohap*6hpcaU*Mz$a5Zy|cB7DNJI{#j}wS^r=>fzR) zKdo|RwTA!@w5G6g-b_eD8-FEQaoz=wxT#f$dgE zLBVu_ZcG+xYa*lBl)s|z57WBk5z&PiIfrwcLJR@mpu)G2gkSPsX=EW@qlSO^0ZyP8 zM;cNT-gJZk%utl`@1aNcESgm_NvJXYmOrw{8B7MD^u{>DUhBAyTjAT*GGI+`JJTwcp(cu zgJ~|yoqDOtl0J>YPUc)qzD7iu!{v~q?7_4K01)|0S^6bv)_)*e4GIlqzKl<`ghK|Y zihT=9_n7qNu~x0IiNEv@N^~_+uN#RHCoMX+?cOt`GL3A>r_T?j{k$f-If4&!0hX4S z6>aC9Jvty#;osptCmXz0+BUu)E5uACf0qbz+vQ2&btS{)Z1nedw&vib7eM{Ny!+K+ zvNd=mi;XF5pJahL&>V*^&0^qB$;c?)f|c9o65W+X5&1&#gT2seKvB~KQ_qN@VEIpE z&JRyCu2D^-RU~BH1`rD1XFY-`^3;gJ4Z>V7o163U)9pwP#?bi|9Ogp}TlY%g z25=LmdR^+MS0g*wU%igF-p_`bU8OkK@1Dv7!qGLBT+alW>OsUPeyJ?iKFByYMrQLs z1;#?VP>;n3Lq|wRNG@%OdGFf8JxEYbt-8OTOXwBC3=o_z!*O;Qgs!$VO9VD|0C^EW zi7*sR$>UeU$>Gq-HKqW3d6@RMUCEdPx*pHCiT zeWIyf_FF>0Aw{b~0AtJTk zT7^iu3nfwt{*HP0*Glbcy|z;sWZn_x1l>QKR82{~=_I!ND`j296oKEZ+%h|1wO(v` z2s#f|N!SMh3!%nWZbJi*)R$%bcU8*6`BPUuyejX`?=KGx8f7FUS+y>Na{{hNyoCWK z=QIun6&Uzx_Z=M#>TbWwgJtbUC;Fu4mB#pM+^AkGX{`AIg*+kvf%y2no%;b1*|k-# zM6IgTUgah5VH>ECKEE<+!#XN%FZKz*8~vW`thb?5{L^jx-~0Lcz2A;Hms{m0<^< zbcr_FKC~i|5IX!ZyV-N_n9}fE&CYSFt;4KGLyjT4ah8@ zmqpCsU`4)0|0I8ZO_$OPI6amx^GE4cJb>V4BoJoqe4llGR!e1eH*$WKp`3^ z<jtu$vD(<9@cG~+CqwfR;RM&5gjJhq~liWB?81xfuM-#YliKc?OvnyQ2> zk9VZi=*(S@d9)HE6?Z=S27a0jxH|A_9{yOjT;rGlA5$2ef>oPA-*K-}ZTE$Qhi*&? zc1j$J8SdK{@oq#u12wf~yLO`x1pMJomc=T=hFtu?D5Kb?Z1w3ro)`FEzkW%G^dT>+ z_|kvV3*X#dkr~$4zgY#XpKq?SOb|TDax9uIQrcm7IRHQ2TYji`k#0M+ZS}0%(K&i3 zU3GJ5boV#zddVcTmw*UGiMs64kGXRmCkhpYmaqc!ruHc^-+MgzQWxid?))Z<2CUmZ z+k~7mNHXWVM67*MzZAJ!SXKtgwK^IpLIM?d4*~}8%XG27(bLOp1~lc6 zyV2*FZi|A!)cyE`fqxD9^0p0=fmtk0*GL9X{`)VAe*}eq(jC7#2sJaQ2mFOzbm^@( z%M2hkbP3CL!{=V`x;OrLz5eM}zB7=i6;w4iT-wGSu{>A&e0YN82LjCDnnvtO^r{BND*A5cVW=HGRP#w@Y324yS zqJ0%73m2A-9TmXmPJH*BLna&EZ(oBpgH84~7>GgxYBD)tgWEli7-;fYpg zCB7FzNg1D89ruMHm+h!Yv=ttQ1xskFTl5b=N|5!$nG7jht=6!|*bxk=5bbT+Gx6Nx z7^V%8<&|Sc_nT#j_pp$!>T`d-qM5}ofN_Y+Sf?B9Y>62d2q<%hQ*m3@&1cX1`0%lT zLshbgNV~r&deQEKT|4Cwk@zb5##;jELzVt-Ci53>3zfs?)(W9)Bx#}&k(-Z}M9+GT z^D(S7?`z()Z($q8cpZfUCKdrw^&3S$KPh3szqq!4Ty&<9Hyf&4bkh1=x@%uFF)6Ew z>3#YB4^1wUkTF& zfd0qXMYV#dDm)Hr8h3jTz=%N z(#%}8GJhuK_m8ji`iO$wyQc??^23WG3F9))?fB?u>`WY5u<=qzWz5bFW{?bQb~1Hs zE(v+ak|GiXB5-;P3C8pMsrK1F$q{BQ!IF@YrR-4|NJaUZ?!&-?h{;Gq_4t7WM@Et_ zY*Kg%FclK9vZ_|w*~4JS>iPfnDNyBX5=viT#et=aN;3OX(Xv05@wW{Yr_G0d+n=l# zLc{1U{8gqL9_q6X$j~abD;TPnR8V3rNj^BViJ-OT`OiQbD5Cp|khs?|0Ty;o5Y}pWy7TtVz zx`6%{cKDA~fl}rJeY*+N+r)#jj`=wOe?#KEAch5J_&2`lj~hlL53Gz>D^(aE@b>3B z>5o_b_AHSE@baK$95{&o+sps|cmJ2=`^O6Ze_8iT3)r3>rK2;(`!~Q=0TC z|NZ~*p`!(bZUgE{tyXym$t4<8$ngKT4gcj)W;1Z&$**{VuzCLd&hV~;?nyaB|2Tq( z!ZI$l9|SCUGZtx>@li88sa2UvDnfy&n2wIeEH)eQZ`zH2ecu3kkw7A71Vq)T+9}&Z zB6*7gd7okztG_;pT%5y~zvDJ24C66Oay`a!`}j1?9y6C!Cfl}5(2W5Ed|TI?FLVu0 z$5`eo#r}`2wl%E}U2w^6XUSmM`Qr+$k~-S;C;C z5>zaovp-WIpYXw)+{awgM}F#eIHpjQNk>zde>q7HG+Kyy%iyG!5w(;!L&5zs2$~EU zuTcJnM<60Toi*y!KY6o153~7-fG-#)f#v}Jdaot-rJkz3KJ}r9=JTV=oQ)}LNPWy+ zZaqk=k3QEakfDVZjS>_P@Ht%eo@L0;C`(};YDMJy^EdxTxWLDZ8H5}h8XlkY zu7nBY)^?ZFNi5XRw3Anh>>^FMnUT z(qS-AFshW_1EJiVtnb0ZEc5{VZ(m7L=kvwoU7-< zpnojspFfhP&_hw)RU*(xLLh;uIFbK`@)kqXTV=Jy?6V()441+?rVU)B4IEA158nu2 zPG~S!uVj`(Ty*ch!!%t*xccB7UuJ`!Afq(b20So2A@nqhzlTppMeCNR>dZODZld{>5#6jOUZ2#(N&CC=wE! zmB=axM>8+*KhstIBcK^lB>-n=T&XfF zfUlf*+8NO#czU>uxxGDoR6Aa5kp=wKrvjWJ8qA14Q0V6yXQ*9!{H(a45seM*_7&G_2>|5LR`O%J_0cUj z5;kmmFK$jprBsX97veJ7aW#X^7xFsZL>jB*0r#N$)riPUs8Igi+X7#;pNh|`w+be} z_GK?)Bw>$nC$6c#7U9U~av+OG(T~j?d|Bles9)uGpq#fI7wZc4DLFo#7CP8_k=6yr z^BdRG<1oY1^x_v1M#;W!A8B@f4?(w_cHRP@K)NBRt$dY>t&Y`UNW z*p`q>wLAFnoWJ;*edvulpl?_Key=)Rsrhsnl*Pa?@)mB)=|VangxmE-{K@6&8&$^5 zt>Tix+4T#RE z?jZH=3xpMYq?(S;qd=*@>01&6KuEKxe`p1zWEv0v{QxXe@u%|TA>vMhwRX_@7rnO_ znhn+$RwXCcZqaG6h6RweK8zk3bus`_ck5e%`#dF2yJoEvSp@!1McpL{K>gN@>&edi zQ?)99y>F@gW%Ya6>c1~hrjqx9g62f#4=|?NuE7sk$YWGCAV%mlK>*q zqHo+^{T`pX$hVgLCG%`fV!v5$?ZBtz8V1%MH{bY}6gb^>b%?oXeRHfkTxt_tb@psF zTNkQGUpcB*wwxRGvB6y4es-5lP;^Pxdzk&W^5v&8&qIyRpC38@VnKW#09)}XQj$wE zsr)R#214_#{c&x$=EO!y&JRNU@?d2ekL4pFzkPA6-gZBF$@WzPZya&5y7g!O`e`^! zz->e)>Uz$bbWf~9!S~#H>_vV#%RTFAu6~QX+J&BiZa8_vJFraFC{RGvbQp6%$}$(5 zxt)D6o&dcQmJrWd7Jq_4U#FNqQ>rc)fUr+qaD75sR7A-i$xr(C`P3Vb2p}_v5JGD$ zNVY;8iz7z2I|FAV3xW;Wcyhe>U*_74rZBIiigHJ;_ zD84F9Iw%F2-9M-5A`&o~@NqiqDQ7QZ@eGltu41N4aoH4K<^SZfzyO~`lD*rO?Lz~| z!jnEA;>7WTx?q+xIIelz4YwhR``%w|1*|=bG}|A)tNjGjWu)3~xtwplzdzqk#b03g z*$0O>KzFt|Y%_>}>Q-fUqk}D%1>fRuM6#34Ez#_FKvXAX5;7$bKBM%WrLkjF=j`<; z0b(Qu%WJOpEJj|?&ulrhD~l0-Zz6K=uW5x>X!=~1wwfq%X~!gSz|du~9m%(8e|x6) zjAPiI`GX?z4GU-Utjpz1n+B+All+@MC7*4l2||N+!lJ9lDh*zqeW!A5zk7Kw2MLT< zln;7M&RESq8v=-hC4YtJC?tB&hYX^p`^%VKNslNsJGcMS+0zAPaNKlr0^Y?Emsg(I z*%A5Que4MceUe=_PKJd^zA*kt@!aOS$N1$pO0?>rzg^lk`)Ggh?1s(ds5& zUkw#JtggM*X|T({JM!?TlGE5(Sqs|u?tZQR7a>QssNSzq_2a=^Oma|wJ|Z0xdeYJr-h=(}Z0r2=>7?VC^{QJA^u=PlZHX13>z^7FL#o$; zd{pCP8zU51KWU0OV(Q~}*r!!1U3|az(c`~HyxtO+1EZsM{h4bL2m?g? z5U}%s;6NRlTqZyP0sO3+C(RSxy@?{inALbtc5w$X`+HvPz!AhaRXno7?T@A$WYz*3(5 z80YtzqqhP$0xky)dsCA_iwZ4JVTCoTkgw(Zddx6cV>?wS51MiGx_d@Oq>5AHDAR8ZVbVt?*ah#M(~TmoCKC{jLthY? zLeO=8ca!rHElJm+!;{R;ske4sN-=yFzr(hR{a2x1<`=Nblba@I=XS+lrT<6STgOHH zwOhl2fP@H&poD^;bST{*2ug!=E1l9HF@#760@Bjm-93OvcQbSi4TE&^?9tzS&Uv46 zd(Qj(<3}AG<~w`uE7rBv+I{leX3sDIYIbjc=jopfWO&z~jv8O(=F>n17!OaYXT_vdJ8{-o>+WRptwAFwTyGxb<2$jggpX7SDbb8Lj!^;@#Y8y6%Q&)j1!=CqGt z4hWY5Zl@tQCtK^bN413SoyYS#!?}hyPU;WX>%=tLk17jIhue1Q6tnC)c`_hVV}ipU zxewd-+}r%$v;LYcNo@@I943)w)Jwu@wro07qj(FnB`G7Ksz;}hyuu`Z#VC}VQ9d!| zUOc;finge)m*DUB?HJ1<`{h0NXVerMsm~Wvi%nlH9DRDoCh^9m`Jp?c^q)mw$H~`3 zb_W;4t^s{Y8nw1_LXRUs+YCFayE@<+B_uYxd=;PDV=pUXBTiVd5Jo-0h0=__}LQjQK`2Me*~An9K?K`|2-j3UhlHhR=Evqo;QErJgkql2zLMiD)Wd2@7jZGS@7t4!g2 zBN=$DIdhq}>tT@7%u5}?QxTpNS5+qgKALvs%@GP25)S)KZzh1LB~QHjIz3(Q`3@rs z57C63B|DdJ7XRbLM=Kn|(|#wkq#sl?G^&r?w}dJ^1d485UtKSk=6xvDcCXuY9kfe) ze#J=4W>yW_0&pdRgkikyo11BO|CacLN3fUU7>QoIuvsCj$Wkki`+6hL406ijjNBAn zZwNnbCg9DaG+fTU7DOTH^HtSxd+TzzwUCbixwbfpW?f&*^ZvkB(*A4jlm;TW;lPi^g99-G<; zBNQsABxV)-=o-7GIH6d=&CP*+Vn4UX6b%YJUX>MRQH>)SyxjJGAuqgX3kOvu@oXLH{F$?a9i@&CX7V4-K+x zsjz49d)-A-@vW?ln=H}$D;`uKf>-%HmqT_Ixar@|wx~eZd98jrOj6bVAQ7UpFpZ^0U{hP;2KEwK0s`m4q z!Cjx|#2bOM_1W?FewTxzL??3x>L@hUqYhPF7d7Nkfh%Ie`WIH>@9I1siRPbd?Xy@) z;o4e%VvwBTc=k+;nfsdTONiVz7l+NLvFQ(U2CRDZ#|yWe*GH>fMJJ@Hvt>?OM?QT< zh;Gk4b4b(w8fZN;asPsZgi4O>rg zWA>;+=$2ls2C;6I)v&_P7!r&G5rod}g13zRQmuo}KZ(5@V_Lbiw3xen`g-#V!MnQ! zs!{tQ$T*iBhB~a{-hy%a;@2@(*Uf9Up2U_4Rk^4nm4b?4J^qt{r5)$iwl>LlENoQqSv73+m)OIAg zx}2Y3y{hlt8hNa$?OsxDI@|?Y-Ul==8ONPz!_7Nn+{@uLq6~-A{bT}KeGgwq1B+u1 zc7@l@&f4CuX-$CMzP;EoX-Yb#sE&g840id-Gk=F zHzf+V!V?>X4B7=w!z4a$nN8BBkcf+Nq%s>60rRZaKny4}km_@Y7A$L&Tr zsZAB8u949qfRs>$2oV(bdjdw=hlEL#LvM_uaHh@ZNzbpIK{|x74thJ$iFTGl)hXjd zb7H+FFF;C%M(UUoLfm!yjPCxMqw%Mvo5cB!iSy(KWvu&c&{$^Xb=LgO^x)i|%d}zR zI#!W?E4RB9G`Uvfs(#&%uFQ+m0+>X^*E8D05xx8 ztbl>Uejy@&UugAvJ(6+vUzg)Q|Co$wQ0xy)^8S8~@DOe>3eLIj@A2Tp(R+WVtg{yV zKo8GgSTde${7{*(T;2Du_+_PEzdLbxS=8C%Ir%#Z8m>iCHF8dlLXGE*W6sB1dMDON zkld-Rqep!3lW+{!uulEEsB+errzBNdX5*_pZmPB92zBN@_&(46@Ob7Se&ZtE)fo2n(=Xu}6noMfx2 zE8eC5hMp!LgqGQV@>5{yVK|Gfu3#%aJ$5#O#r)ieu~6CD{|jWN&GVeK0?=J|t)@~% z(zV0j$l}q}riUWSZNlB6qv^cBMR3Quo6G0%HbBxl*f~EO!*@)^`sb_steYH)^>rX0 zA0jR8&uOV4;qLC<^is!lNdQ)3t}5K1!c9zmkKo=XS=hqX2vl2>$GivAcobR$O1siB zJGYGsRaVyEYGt7U9frfrbE64sE+S?VqA#+hk1q1MA=>?Ojjx7JPm@SD|NY8=H4J5# zCw+`xvt(e90w(WwYYCbDl)$uHUe(pyr`wUyQQwIdgTZC+ivAJo{<>9`(Ty|5@+C6t z)jx!q^e&vEzwA6Pt6Azlj+kUuqYmmbeI(UeSYk7*!n2gWfAg$yuQYJkiA!5iFWpMo zBA*@e-id<3>VxMd^nb;Ua;t=GavJlhBqIe~Q5ka&;QG72KmOoT>&uGh=#RYN)QL>D z@E+YUVY>|PM@41vO%i+sdIo|4N8#&P6bCRrrC0P66TbD5{*z(YKt2)4M+L3Uo}P%Q zs{>W1{Vln<3K+NG$r^$4j*qIrQYO>h&on$V+*E5P>{5B~*_H>C{s&Y*_&$Fvrl!z(eHpwAwr_B}BT(wRuTpcJC?pj(#&q-OSJ8Kl zOQ!2w&qDyjn}&}Kk2PM8mEU8tMB5@0X>|&W>Oi=AWaaCW^;JTTcZ52zX6m{WLoP)R zCX#xuhM*X?c@%Zr7bJLUR73Dx4%xqa-p6q+{(C74iY0$S|DcJ`7tjA`i)d9SgoJcK zl|~%%v6-+^W>xkPJmuCkHJhMYuhEwQVV~x#5CU8+|K(DvYn0Q=ZI94FeHLqPhQ7yq zl7+PwdVE_yf4(|JjI_fi^Ju0Va32rSz1o2u>wY+YBPi77`?mlAw%H(+Fh<~7O;r5v zoiEIR4`7;b=l#aetgHP`J&EU$0!xBf2NB4!#9kxV(NeR!9nrvOoM>0l3&sXGB*oUl zhpMoz__ME^otQjcrf#Q@e<^fNPo>;#WzaIx_{nXLcx`a?- zSiF6Hi{5V{i`QGUdY~p}uZWW}N6c{Ib8m!=rE8xBYIc)`ETJ=6Htf8|p5*G~?Kn=$P@x%zBsX63;9(_STw+>ey8yJT zmYX{6+mh6o!`vZ#U%K zjN;&4g01R?-MTteYpPVBk_#s-YJBUG zp0z&&>Y;&4M+*(|)9$mD_0ipvv#z>PAe4BOA}OJf4$;|WLP!ep>BMge^Yi?s#!0&( z=4rJF@6!Y8bL{R<*DpoaBIra!gj4^Cj-eF0KZg}(=COMvfAx_n!jq1#pHWncv`=Qr zFgP1dD(=n3u}FegKDA)p7I^;n;{>inXID?|!0^!5d$MD=0>!fCd+l>`v~xXiPp+L8 zlQJ+=*95_^DN1PCwjuv~ov&u8lJ1DVLZ|zyoKUp)Ej(wHo<@++cf8d9KJmKYm3^PW zTTDquMHX2yV#+k}dB37{jTlS3O_W0GJ0@Ph(r|$>b+U7P{5rfjwSu4T(6FQ2uAV@VDQQRS4_2dx5(xzpEz3%<|~6+u1iO9IFU8 zeSMMjh=72~#-=0{_GEF1;3>0;IUQz1BbbQ5t$ObMHIe-DLNW=C$vZI>l_)?;R?+v_ zjjzQ>H+d%s#d_QMf+f^K0?9BC5sp@9X9CbTe}W1)9P(v6n>ZspFoX zgsZ_|1+$9w@_)#z|DcCNU=Do`QeQOu313UWMhM%iV8yKH|LrXZ2YRqzJT*rbii&S@ z3q%bQs`U%UJsF829SBp8t3ls@sDStUZ%dr~3EkGtZtE^bxQK;?G&*@gOHR$CxlJKD zJ?3XABuQLi2=6Zy@qg@aKoSR)w)wa3^UfW*+b`)+no67s#h%^v5*5z@yH~;GJrC3W zwhOX9dI8Rn>%hs$Nq`Xd>jUr(=#2kQ9|5pr%$p7`P>DQ(#Cw*7!Zj1Mdpt$v@`k1pZCXgVPot_z{LBBv3WT2@VQK z%E}g!@!NN3<2gKe#-sjLQSlS!9MA~yA06O0PFXR>Hbpmj(<=~fBUx!3gkw$ExVe}- z$*G@h$7ErgVCgBd=#Q{Z$>Lr2$XyOKQ|e8zq2 z&f^FW8znedkWV(xBL7w%{TI=d#_v>-f_2ia_;Jo0s}i&Ko63LdFo3iO5D%gs=(r5u zdAC_K1>a(O{l>x~f32*hJ#i0BG{Vr{_%`|5t+K+ccrs+sw$04PRmC_nMC6ry7tGaA znc)X=Mu7zP7jDxOcpKM-3{nZeXv?A1bi(oZ5BlnybZhgA=*G2mwyyeKgsedJqja}= zPgy+964))&giIc|lqK+P;}Q=b?kC*{)TVweD%DWy@M%X?BLl-?8NPapU9;LGCV#5< zMKZKJpw?ldBm1tXq`Au05!wFBFuOJDk||Nx?u6p7^ic?Z&Bh*1hI)9LU>%Z@AQ-nF zDW|QjdUm}2%u=I@aq~ktQpa+C*L@(XWf$%3eZwMJ?0nkiQvn+cHXUR@l!u?qib^7^ zAV1%!wm@ofm#TILqWf$jwM$C{;}2f{pQW>K#XS??E`($W+##2Z20qf=!3lK+d#$U= z3^!$oyaMi1G^J8j1oA{V34g{=F_9OuX};fFYdy(}h%FKEX0lNE`on_gUyR&lN1@Ex zKGW?C#rC7UUo=T^afd&XA-$)I=AFIW;qOJy&Mn40y~YYOz6B5uVr#v};@pbxSznR;k znUa{l*s5f>XpLje6jc@DU>m-(+VD+})mrEIP_uEgbt0(xU~?5egD2zrmbo=o=YsR8WjLJzta)D z-^gSWz0e}dt#-#2pqCm(cYFvB+I|!D=AB@Ye26%Zs0b&!;H+6G(-}Ed`c{Z( zaNJuTVHI~N4~ZOEuKyuC|B%2Uvzi^9sk7B}&uT(-zv+HIqzx1A zR0AQ|EB9G))z!pzwR0Z)2M>&|sy@b_-pO1P%T#WlQ} ztxl!v?g|&9?pS0xx3K;JOM5nZaWz?MXqWWipe7*-2u3B^k1V%sO|RGXi*D~$}@u@HZLv+plNvszo83=M< z!SI6%CAUP&RQ@YP>vk=do~8v&}PQksNkFbz`9=tP;cIbtvy@HplifJ_pM0A`bG2oyKX4NV3m}3P#Ur_J2ZL(`<}%|J=lAtFR$YulsbP~)uJmxzL_ z$U~Xmc@s!C#m1zJH1t#dY?lxbsXwp5&q03Y={m%1KA^j%kdG81Pi;=J11q$M&Yrwj zD%HoP=wtBR)1T@HXJEGA5J61ZGg5x{UtR!QmKDlpPH2~(>}@h+;-YLts+b-375kx2 z`JA?~Omh2+FOXScLoq`!dab>h0x;zrKh3JuRk^4QK)=*HF!#e}>w3;rAM)WH2@p z$#Q;OPwFcj|%toMoUq>k~j3CfBn8C^)IIdbH z?8sl#AEexseu$iC$XB4VuPrhV3}iMFwa9gaV`UMcE1M*XrK71E7_Yvm#45Cwj3vkGtxY$ic1~AktX17d_vxKx* zBqXS5?B2sFA%4^6A?@&xDlOv;$SP#A+BC5K;>=Z6k*iCF&Ddn!wM&B4=Hg*peVYN| z`{`-4BTs|}Ct+|_dgYYNWLk|b^YeD@vI#ogfGJ%nRC3odu^Mn6YWICK*5YeA*5tX) zm&5Y?y2bKP^=hS!80iPHre>dLje6oY^CyoX=Xh_s&4S84mhD7;EMIke;nWr2SiDq&`$( zn>fpz;OVSeTC8WlP!SgsG8UF%u0fP_6Q7FP9623o5om>)XX+2#;JOsGd;9j&r53GJ zBYZTf-)}E0Ll?x~Ce`Fzh~P?V%{tTGs)62psIj2p%Jf69D-=p8HaGUh;|?e8!OQ4$Q|7@Jj*JZ zIbx*1|6L^x5GLBrz!zi2^!1xY9?v!t%yzfp>fU4(4)Z7?qHo0$9pQ<`h9P0QifPrreHH+?fO89z2+*pR>~A$ z7WP8z)w|18XDj#FWcb?&looOhFXAE-m|D?gK;C^jS!K>tVK#O!D<-?PBjK>DxL%Q} z=;UVji6pklgx>q?_-OL&+2U7qyY%2M4=pB=-AjV(?CcB^VyH8vR)YmCCC@~i=M-hj zQOE`PP6~Qhy{OH{ED!Pw0ZEEdqDiQla*lsnvy4M7k+m_)42;eXOvoA1T-v3gRT9|` zWd#tw%ZAr1$tY$O=07K*XLb8G7)s5gpgkwds!^9)SnmdetNdo^zM1#;gv&B<6d1zM z)z@11d(|>+oMUq+p9t;^kC0!EzLx{5Mp24(;E}X>Kyh~LMYi9#z%d{*OhyyO5^SPv zNF+#3ub8AyC&C25aPfd;1?5&VsK zX5DlOM6<4!PZ4r^4$U5-=C6Yc!dsIra#(uaOw!aCOy<4bPFdV2M6>El_HjVmJXS(G z<-=y_UuNapuzCIM8j*+S$hmp;rh7$3tW0w4KGE^@K32`a4=hfNKAuX!Hy0m%v62#J z?Tc-a)5P2E?!E_#p~Nzm5yVWLV^kn@0)hQC>4%zOHQaFK1YWcm9jy#mAy&7EoW17P zY=G-7`|W5z28U3$Vt>QaOPAw_eVTYfsE^w6+qc&i zqM(pq!e^}3>+Wls59p-2XYdBEU&}%H{422lh_?8rs3azX8Oo_W+EFwteFtu$qBm#R zcekT}s4mwAj;6lG6`nAU4f&Qe+^O)pQKZ)q&eE5ZlrHzkc0?sTz#G)iCHCp}>ylsn zjs&`rh0`H|Xu;n|z|m9jdg6TsV>uB7eEg&PA<2uBR!8rA(k!&Rc*7zX{~r=K#8rR;}{Kxr&9EXC#Z9M|89Vd zs=(4jFpUZQNkY5ui1BWY=x9e3*X-4sp<%Y=tupbB)K;O1(}+T@ySBVC{^H#UFWmf{ zaVuZmj8*?3!B^hggELB3(BNuz>ke5Im>$<|hE^t{5b5?^_kr(vr|0eyT)EZiAyawT z9cPg*+erTFxDF&oI^oFoLhq9zl8rj(Tq;Qqki}Enc?<)80mlQ@3nP z!(ikX_cJp-T@qeO$oD9)q`0WouH%UhZ{~BD!-qHn-`_ zFwWRsCs5zQdtBPT2|8hmC>)6IG2roXb}aTaUUnP(T+?UGzJ){4*TKDLn~|A033m%$ z5C0_Wa1XRZyxQBg+~byjO#x8zC*VFW+Xk#V16J1DVv1=Ag7(SfrlTT27e+{`^ulye zP2Hqabxnr9pa0cJT&^kxD`sy7ErIEtul z24=0b%5EGA>=gXvQE&?fgnFQdhF%u4%J)~KtU+pi?h6XIQE}li^|gkRtMRgJo=pJ_ z#3;cnxT*>7$(F_bjeY|puGXQYVCjCrtpIvSR8$waKKhQzW0M9|8fp~_$0vg1xlMRj zqOWdvnbFC^lJG|FC4>W==b*8)$H@%G&<}&yYsZCSB%QjNzM!|Pwli8f0*DKwqOPkY zx1!yP>VJM9w`+SI=d9P(Q{xyLsyh2oZpFPuxUhB$6;L6eQ+~{3&zp&5C1m&b2^gO% z>hL||jAHBwMn?=XfCA${N7K~^&Lh?lplmKiwQ#U-xhz%qjl?I^>?%FhXQ^+>s&sCb z&y8CYda^+pV>@FAoPr*ob{Tj-Glt&GiiOA-p_4Dt(FT@^uK6(a8NO1BfS@WE0+Md0ZV%UPZ+f zMgTD#J)Xc>{N?j}^%es3wK*_x-Fs;DiA$0`sXqph=}w>XXbN;hdgvErswNt91$BGN z`-WnsJ`NiAkmA?B3x#YgU?H+80E=M;FTEH-a3|2eQq*mmu(DsP+dns{ew~+e7iMH8 zz=S1D_27ff8B#72tS*e`RR;(hdo$Ykc zk{X$X;%(o55S#JRs)@l}gMr3gsLyzANQKk#RLsbI139b{PHGhTv0TQqyL*)dPzohK7dSsR)naobN;| zr_GkD4b~!6OVZd1{bb6shZ`v{1|xUQ{>!64cLuRQck@BIlytvfu6;Yj--a;ky{^zN zJEOX8hmcBDk=V!E`eKU1;H5wD2+=7u2Jb9Q67cXJa6NjK_2BzlFaNRi<)eaK1onG! zYuWeKVs2Wk0_8a85(7_=nWh!>jO%X0>&=TXmws9L3u;~GB-SK6IFs}>$Qw;1SW087 zEhRy*4R?i;>5-I51*DS5-ekN@7Jcv6*M9h?FH7l`Q&Fu(EhENuB(xWl33=~=p+`zF+dr3sl9%2(`Y-9v;Pw*FZx`_kwjSsr?XP-6i8aJM> zNERFwLiATb8@;dX38BTe0htVN&=I}e?nlx9j^Dz{T9LvfP`mpe3#Hgajs^QO?gKB3 zCHS(-m3+!vK<00#_Eo0mY?`H~PmS^BRt}udel=bo2`?B8S?LKI;3yreiW-;L5;^^{R9ziM$vRG^1hI4YOXJpa9GL7Q1m>%_8*)~Z^8vgs@6W=h z;|*5@dJWvHB{lpisXywam2E&>Zz)wr8ol&ph}Zu(9^JlIk3J_+wKNoPS=sjO<)z!5 zyqEZ8jVF^B+Kjp-lI&~Ee1V^x^n1BWhNPuR0VG^%0K=}!DX8(%WnU*VH9{qhF9k{$ zQ@~6_mSMfI9HBzjzuMj&du+~x7=n9Jhnmp7n!p;2i?D*TpM|p|s%=0sICf>(3P2GE z5GsBZJLtuvbj_Zn>CA+wmZasYn;>&MN4d0`q64^T#%8bUdqdDGN#icw_|3<8_03Zj zEXUg6)ZVs$CSvBY&Ut64?|FlVhuuanT%c%IlOk3xTzE=9)BxGdB)g#>JT5ri;3wcO zINsU26R@?z3jDqKkJ}4cku5vu-4evaia!oC5s-1SwJ!K#TwJ|;R7ix zD^5yt?LR=VbMJC-AE>&z(N~k0aG>Y2g)G-Lk?|vGaWnnrNfqW?)p-S)mDbUAJ<+KN zKSnaavJW!OpXqxkc2?Xj&*Rt5@j+Uq-!)^a$DkosQg z*llt=0k}4t*~5x*@uJ&#HPJK;UcY$`zzFgytEh{UVUj^e-$TKlu4Vmu3|944-VE%T zt%lL@#;m%Gq2kmxeF3T=yO~0c5k$hd3#+OJQE=AmCfQX#nqJSTo|iz6&uUP3dFZPh3L#DtY)9T zzxg00++6}jSK|}I<&ePQ$4y`V8adaw>*&zHCbJen-d9dOOUw{|0|${$eOW?d`o3b%}dwD5I^dh=1ovx)1tYzvUy7y@Q%_s;1)ES18AIv?u; z+ey+XgYiOvp=5XMsv2HyB3Z+5aa>E1EF+mf!HIZFz;+W($KpPvXYWWAuH0cJ(bd2K zZhP!%zq7bh(>d7s9~%7HTh|ehyb!z8|rp935^PassIz(0h$qV#uOA_(oxeKTwBkk6d@jNKuu@~?yUTO}jViMzpyhMSb({6OytBX7UV0D;~5gRjFz{eTM##+wxF=YRRAfY9#P&qs~n7aM5j^iX!Cj zE&*^2a>bU`IcPA?y4n)WjNs6+p!$BjJy}#`Sazm_oc-E@s&kuj~8Y~>-!Cq>_ znBQlN3?KGInv$uZ1RXd8FuSPi8zP=Sp0CpNZtmmvmi#6u`?I4UTn}bNT^jq?j`ym6 zoYxqaFT_(nZEPaA-@_$LxlQBs3z<;7{;bWdN=)4#yW6YZ7dB%Of7cK{nC7SR5?r0C z&$EQU@p%{5`4-7#!+FtUkTEdtr<^z5DVHapD>BByo0XY~CT!}P0A1k$sUKR}-?LMH zNnI5r=t1JNf>~GSvHJZK7QMrHZTuEUjmxGs&`$5|HsVoyy<>0Xg{s@t=cQnkU^Y&^a#)a;aO0?QG3N4`9oQs7N<9_-=K>%nY8I&6U(A zqBb}p*4r?2Z=IRsU_pray6{8eru%iS-Z+4TXk_NafkEGLZZ`{X>iV2)NfR$HiJ`~8 z_fsjxcH2;JJ?zuYymHOes!XYKH;ZlaPn!?hNFn@t@AJA%9&o?TnCq}%e2hqXb!zid ztNQ&=`_0Lm2^*p+7U0<-e-{;f#Y@Yp<~XLQAU`!ZAT~#=oA3qp?~acGzt-~@v@bI* zc*kh{gy;i4vzqHn0$v}}Cm6K@h>+C;N5bx^b9NfXXyH&l)`oIhz_%`e4#nCTe4RoA_~LLBY1ZUiL*dG1VXnMu>VqW=uI z7(daCH5CD}dg;ZbCHSTlQFqsmufJwO5PDf<;}%4q8sA%zGk?XVV}g{vIE_21pN~F3 z_cFQ1eMSQtOmC_o3vZ%loAEUodm}3w1@OjNh-vw>zFY$Z(k+0Qx>vq96@~B`a`*@c z1}brB2jWf!i7s0zW%CB})tAz?Q%5e@J(m)1J&x3YFR_;j`CHkfXgp4Zu9 zgYXXNUVPRqwLNFL3!5&oI0{o?ii6S1v_=30wM^XogsD<;)ydj$93jR8NWEx#s35&S3H*b9KPctJ;w~> zyz&1UnPcV&>*zq^N*YRlU#UK8m-%4M)}d6jcds2KuJy^jE;)OD)nhw0ZFUR!xyoS{ zT0TWrMvXo11OAdah*6gAPSBSx6|hNN*Ug(#D%U(t(Y>{XykfaTPsnL!p3KkR2O`L$47ahL3Rolhn_h`sGw7wOzt7PB_J9OPiuoX?)FRVUrBdt)?DEeh% zW83Mol;ONjz4t@PYO*G;kC?CYCxqj>-f2kQR>U*eYqK=+&wXyHyooTo#wcim9Cqwj zCrmEHWA<~I!A3b(y2lH;td&y&6$fjvhgFVY zqS)b@dwiUwPNUXLt{58(BxZor*=4-6nr?gV2s(jwBiUVm2fEuLL&~IG z`Bruv){>usajIP7W$F9x_IH$u6dW94%v`$Ik3q~OeYzkZIkOV6-@QVz!0 zKK+Yo1zEQ9oy*(OH<>>`7bH&|7w)Y?_P#i}ZclQ+`-Ege52%R4;v=4dE9f1QbWOd1 zNpm;Eqo%iVbRH+mp0l6Ldz5d=`E_zGAc);XQI#a0-^@#2OsLW8a)_QflASq_lldk_ zcWyqH)~h7}S63;}p|OIVmk4~G?Xa`kU%p8ooU$~cf6SMW_4MgS_hoyH(Do$vK_l;n z0|;b&Bq ztyo7(vAnx*JI0vRE8b}*X2&lySpU*jX6F#xRh-;=@B+ZO$T~^>4il9e`BUWeI7pMD z1Md*96roS$?O6iT6ovue&N8ePs9i0sD^x`rRUUfWM%(H@fhZqGIEmW@r1RBm=-Q2qUAd%4{YCh*GGc%VF^p5u{Pe z*48#-$|eDIQr7iw+p|`lUI(%S;jup{=1F3BS|B-D*;Xf5m6YR|9OvF;|oCMng6XpUKfa(x}RB8Zf!#Cz2`l=$IhPiS0y?e)>sn*d6LuQu0KJ%X!gz=G1Cdse#B0*evNi>^L`ia5* zwzjmVn0g=v?*YQq_D$Vrq4uz`EDbK>+SGY^^0^~o?3u~sPrEj>g5#pw6$#hwvzx6! z&2=&o-6Zb@tDNMLs9pLpoYw^*kRv^$!VEJ+*u}`pt%O~QsHiV=G|p~$eCE`amTJyX zhk-*Smg{R|Ujae7>M9zsn_k$+n_#Sdul3)`GzvZRz5QNeug<5@iJ^2rf9V(tt*__j zn+1Xr(vJxitka`!WTG@U%+Y$)9+c3uH~p?wf;Bw*{n9e^U;k6Q_`gNDlrnu8k*KN6 z$8=?;JzA)?91?vau%l6Y2a}^iB<>R~bq_-PdP&`L{74f~-tORt-n z_59Mjh|aotdTggg3lud;M)LGCo1_>Wg*bmNijp=80NZye^q2oxBzY+68TurW9LVMI zE@)IJ@h!cU@OH&oMKT4xEJ~v1XT>tqGs^5#)Ned!$dVxl{jvLIas7d*cntc1lQM4k z!1yyNMfl1W!#Jayo?EwVZb9&?trw12FAknO*}0o&#OJib*-Uo|+$3)f=ESE~Y)zC% zaazLYIn2_3v)_kDYC-?|MRr^4qR8IKX(hfe08K(>2K{adWJ1UY9wMdn(*FztP^v9yX9i(eQu|4@Xi# zKEPkC_?_yOntaKFSH=1*pB}j)vL{@deNpKHJlWS9!mn21^toK4>z2c&wV9)@>z!(t z%Gl}}TN{>-rmX&afb4YC*owtv8$cG&y8E$%Bss;>G>%Xmcr`GY&Kc_RER5_gCpZax zg%$S1oEL4Cs>;C%%&kt1`|AA<&wHzcI!4IdqbfGBt!#ONl3bhNL&c`XHEToW@nSEW z8;`GE8w`9SSnkTRD!shCYG;HMjSKStC6#)WyQuV$mn_&HE~sZpdyP2PU^ED1Moin2 z7pEUvpuqp{5s% zddAfI0D`zEDYbw}!s?v^FHA~(ZKPM9oEh<|QBM0cd`rctOp0#_2~6Fp`Dj4tOdoS22`_T%b4lI5&cS=s zoE#{kx6&9s@m_@9;&rY`J1@UKU`+Rk5oXL6ry(hJ_vKCR{myz`U6je{)ZzlwFF7qb zLr$2@<0c)&3`P9_o+Ivo!X78_gS^*l{^rdz|+{YKw+r zlxs4cU_Rv~Qh6x7tT2fA84;v)wpxo!^-@&YK~vpo}Zjwj=k`{f%=Q#TSpR}VMAXl0!BZTXvy~S zTi@zPwI+gDlga|j_I`gqDipvDH8w5mT$ly*52Bev;VAF6fe{=Osn zl=pl-t$*6cg{-)UPkWF+) zt2)(rxNX5oO+?0WvoV8`l*^K6ThyDGQu_!2$!659q}(!+BjZ%sc_hTVbp|h}wppSk zcwRNeNw7H|JhXFOd$>6R86^#okjS_eFXLA8yUC)H#y-zY80P-- zm0{6-OH5>r-Bw|vOV2ct@QulOrTflRC5mIbQSXpU9^bzlAdsO%83KP22K|>NNRegm zS}{e+9Zf%+K`1j%rBOD0h7Qzv&|`0*J;|>^J8(s2#wA);K>X~H?KeK%>7UII7m0I6 zOAEq=lJuPI7P_;8az%TMBvgi+HG#aZ;3Boras=0$%UmJ2b*zHjyw+~OB~-1tbNXP* zX5gKT?*(m+3{76!mdA?Zs({Oh@_7D}GZO2Q=XDob3pCx?QhmpHEw{m;Pig{#(dqXu zDT%cf$Z{%A5kJVL{EkMe6>b_Ux7J6$U=wZKvRFK*^*Dx1dlOGIvNRJGVh~{dED}!j z5L`nPC<>e>)k@B9C?VcueztbCScGgwjCZ&+Mmf+8@nO8s|o zhUc-{HMTo-?rI})CwWnbln;CCrhjinWes7!?T z#dHhqSaPRsu-J69iGJt8+)A1VLa3?{ll3c8=JKftAugrO8Ma2scOThev3G>Vr%_BL z=f_0Vcs-z<{)ieIWYjx(SZN!cK;m1aKx}3@g+s(t0vRpP?&g3DPIXF#@p)agEdn-ha>z~;8 zF0}rz@0?Ru&70{g&cZ@*)>)$V(n_wew&y=8tPH|z6($+%p6g1>{z|SZ(5&@5kB^}q zs6S9gm6~!=&IQ*pQ1CMP8{rIw#VE*(Bz_WRN29chrXq0_kNYR~;d6LJ3kF?QdrYq4 zW{n1Pyio4%!N8xG_34xeGceUmc;q6+Ai%fYtT)_V^+Ea6*NE+Izic6@&JxZMIiD#wSer)KD~DN0Z+|2;mEyr$eiv+Zr5( zC#EzQV&sq+T4HCI181O*%Et|LP)LfD;Q!N$NHTrXDTsX0$BmX8P#M@Q#Mo4UMn_zU69*}XDgK`Q$Le2r5z_*uLlNG{r0v<7 zb2C@!8X}DmutGv&qBtBX7` zo@IIG#ZcE4-uIAOWK6M7M5bKUVXs9#q_uoq3$Hvqb$+txwg(%x-M2hGDHx|&KEUF} zRRNteMWcfk8!(2!(|Q`qv$S*mxU&0r0`dUsko^c02v3DBP6F?%wwv;+Mcp<|IKK4U zag{p0JUU{=<`r09f3UrG^Jq>L&{7|j-@yUf=?avw`-YdkL5{CC$MZ-JAB_nD5AJh| zQbq4-VoJ*H^n0W4f%{BYhIr7%*vs~`Yr8RveX$A0bpupXpq*b@O%vr|(k*#=#i8yC ztb&;mWUMt}Pd_nEr%thG0)JX5g~N@Qp;ALT(~;*FDT-tH>gGZ$;94NsASbh@Bm+^z zMxI|EBTxyFt1V8khmA!|r!ti%rg+<{Y=)`Nmyc^5HZu$ZSB@fMH_jo=IRsmf;gw!+ z4nyIG*p7F~Q+HM=*@Hhl~zaRg0DZj9fV(ERLq>AV{h=759ijDo6ayAPV2R{v!c=0 z2`is^e9_K}nQ|ZP-aV*N+Su9#dRnl`!7JUmhposiHrXT*VziVyd{~wqUBk&U8b>)y zJ$4{h^vwJ0B=&@VN-yv+RweY1&19ybwB<~@ztID4W9W%v4Bhl>w`S$oCUjJ_d*In# zn0jPFBo!}Q@281+@i}gV?4_LL_^(d+9IoJfn+I2&HQ+M+#F)l)rqneX44cKWpB4^@ ztjd(EMj0MZl#7vlhmJKG;$}pQd8wINdejK2j;_2Vd?^%Nx!yYjQ)DV;GfyO}g^EHQ z&6X}1#9~Lj0((bpt-HXo(X@M(!bwY~4rjY+b1@&W*xic%1rGd_`1!+fai)Be58z9GT#A|}ns_lMf8 zID~g@EoI0Q0fgK0zRixvlZMXKMC({p9q-8*Q-?#Fu@b3vIW8TeBG*B$&Zc+?Xa!opwoJkb}MU3lJjkK}0Ifdx@8 zKxSnirAyr$%GY6216gjkjCoOYeyb!zQpYhASx3KVzQv{-AB{r&Gwj zE_4FVqv~F7@GUVY3{CW|l5h@3#}t2wEW9=Sw(W&DQxHVVletg#r;D*4cWJV7+n!C| zY9mm9{sgl+IQ2f~h((8bxfj`ltjNW77T(|Bs-MklrD=N;c<3;Bb{xfG*wwudqSZ>u zna?8i!0QVZ45m2HrBpPNk&1icf&E+UOho_IPt@*;O`4N-v<&wU45?X%6*amw8=5Un z-U9E>rtSdUBSAY$?2D||!b!2;49cYbgWnc$?vrr+cj*23p=)ait`A%5cOSE`%wjmb zJ8`m7HpF)8B1OrYt;o)NJUJ_&s3dm{ySWgpLjTPo`q@^DILuSyj}A>hZ4<}ISO1X8 zvC1D^)#a8HDoc4!`@$J$tXX%-sn{|_SHoEyRZZBO0--fcW&jZM!Bkk(ZfR{rr!>!r zJ3{cr;gdr1_v`!EW3$HPUw z-aZ!rRq)z+G+u5NLM|xHMvKOvA%%L)j}R9xj_pYIWE-j~p85rkQ6`o-_H&pJ^m(7J zl7c@im?uq`pd+Nu61(=|M+t;?_ya0P}rUr@o(f-BX!dxWtK+ z!7n9RDV#%06k}a`8=*%+RHrvp3TrNmU)_NE-cw_@T`Ff|!O^l@89t2F&FPeG4z(L~ zM+^nHJ=%S%)XQ|oW8O0D5(9Q|=@|{=7f&0q>=KcW zY;`$EbyUrsuWG~P*DE}KbgAz~)Og)6{=BVY@#6#87O~bGXR(2(|CXO8#+o@U_u$V^ z9?%XJ6*f%^%Bvl@jt)zX8>m?w$4Eqk7+|TVQI1<2FOAQ3v)Ik*qv|~E2uQWkMFQk) z12D|^xz--~F$9Yl{K5J05RDz5vgNEG?!j4a3%+K_z3XW6$i7Y&9s>A{-}d*jJ8{q_ z_1Cb;i5@?7kSNw3;?>|vniFl>;8PQ#W*U4wx039T;5u=PhY0v=9BF!ws^v6S*$9{*${_y1UB1V}m)Xgzg z!+2^XvSZ41Z|T?ikD6iY>dMvbM-Zb=G}#y2-Q>*QCIdpT$B)ud4&Nnn3GC2Ek&e^3 zDai#g_OrUYoKte?aYG)C7ig7<=~U&033H_S%1N*!n8C13_4yRr+?(-H{uYnJZ6)!* zcYT(GZc4f#*&jz^W14xblvyu2G&;UxVqR~Y_`)b4ka5lKO2AQvpS$j z#Ur+#9*#YADrmeh^D~$r5I5Q%v0qBJ(r~hQ!+ilBn3{cJMZoRQ`&McSz#FL9{rF1T z<_pCuZq5V96LT1|4gu0?Y#N$d?|G*Z@H z&Ri4nbVGDGV6*9)a)yiDc)*y<=N!Jqc3DjFV*khd+Zz|I4Paf+S?U+ueXcg6^MI|>#GvB&Wk@SmeyUm@MlDl@Q?mDiP@Em<906GMH% zj@e1a%NU`=?lSix3fhQO*f)0~&wXJ!ZV@)Um|C-Esdu0{>tPd!%X6L|vRDzDQr=9! zGwD9d;UZd;nv>*`1chZK)-kwmmgE(%E*uG+`!TSG=O@-+aI%W|pB)_Cukk!Xb593E)#M}1^*DV}BPh3Zf8t$&Ehv`L+zxmRC8zWeKk9i%$-A>>2DEMCa`bd}y9QoSkU z6Fu?ZbPDmkIkthBH%DkX|8R>yJq-Sre3SslHTw12#j}6d0x#)(v-0+ub7#F%w{1$z zrfpFefkNk#<0wkol$!bgYLjbrOJA`4hweBSEaO5>c3K~-$;TRrrrKWY6^(Poi4CT+ zurD;bcZN5jF1?p#8Jm+SWECMG*gD!st&;z_j|>wq*B%wxId{20L_W}1*PPFU3!Lvv zjvh*Lor@~_tEm*}7^J2-Z~{anD9p=@zxcATM<1!aq3G;tvNDYCZrI&JDsS>qgv)5+ zY|FGT&RzlIcdTEnRkvyog##I`Y&_#=T-*|YgWJY0S8GQ%o$E(| z3#Xsgb?m9qb`85AEmi3=GszsgN#IxRi|R*9KYgVp6LsAl=?20`-G=n4LcMyrSyTBm z9KQr3HeuuBelL_ND69ttDePBY)}5c<5@43XBY9l6InX{bN@h6o)4zXIo0&L?YpQUN zg7l^R)HA6zk@HYTRGg`3KAMXZfmTCvw>Is3$(cD0b~UlUJYJ(?bnE0&`Bi*68|dK& z_L;8{lW#W-U?dRPvgfA#WXUt&%zt3#_h`tZNdV@M?sPQaz{PX&)?jJBj2VY_KX9bq zn@^tJlkas1$>w)DH%r7;Rn>i`DKl2jcUU7XTc{)^?`~po$i6QFC3>%o|I}W_-`cB$ z=(`rYvwXUYBR%9fYwe}}cE)90fa_bHL7?;Q4+e#rbq-9+)mMR?(m5%#q{nH#=DiL3 zwKzgd$1mTydAe_8(fZ+0XPle%_vXAmnx#WMJt)~b1;m(@rByW}km;f#?d;KSFRv46 zyy?PKZ97>>db9p3j5(Qa!YB0Hox(g;&g=5?9aP4lrBgDSsGO7oqib-R!t@-1_@KywNo3|$lr~Z9H@luNI;Ryb%eM^Ay-&rBO$B7uvu_~xU%qOU9ic5}H!2i7 zm^ez)DTP!CoSU0n`_9c~Tpz8HYnl$7>|4V zDdO$82=rDlS5dtaN)BYUAcWjrJZ$Oio}r_}UEHKE%Or#9fjV}&pQ)X#cJg)@5wyU8c`Y# zb^V_|aMi0%@{oM^wRfVZL z-f=chwU_v4K`Kedc2qB{G!<)|%M#~=D*{9H!vpug#8(DESo4Ev zx3*?c@}YXlI$MG0h;K11TyB-XIw3>5O!PzbF@55uRw^r= zvNgGT!BCMZ08%>B*Oc(*uC&^txE!zB*z!L1BionJqILORH^b*S^-L@iw4cvNY&u$} zY+*TTZ|@Tvg$P!x+?C@K(%ar~R%zgO4$fvBxfU|p6;-b>q@y%w!E+I}YjmUJ<2EWd z$AdD2b4}E8zFA+Gl)geEf?lkJ6S^f2k{KMVk7DOjF=?_xVvS5p>IQuYc*_GF2%gy{ zhgxji?e?3yyW&=~mmvLc2)7XGoEbmr#%6LsN1Jf>P~+khL4ha~VsCz_s2It&m5hn= zc&8B2Ou(MOwJI8Y-PkmA7efsWn`9PAw3(A0Zx~<`ANC6fnZ6b|13kQT;^+j}g|Y&l zd6`K4pthj}XpXifPI1w{;z!%J+2lV7biv;4+H>iFNcp^uFxwROb<>s&_(s(fCDiQ{ zg}OZDf_x4KcewbCr`cewQC~NQi>CL{rAZGzLH-=}@lBNygjYa=$qh~2snVn=EjvNo zqz?j(zqaZWM)g8AN-LMkWM#o+h0BOd=;*aj&9`CuTtmS>XRB_Vs9%`lPhS~%z9_4w zZBOpx=6@@>sXC_3YfkVurCOToI;33@EYv|T*rr=8O9`%YJ)- ziNbGttuDm^abqE~9`d5o+#PLg+cG?tKbsY5xjqE1KesuK{&c4S0lUN|(JOH@dlPfh zhSzyx@Z^qaj?)Lz9C)RL313a$)DWnm`iQk%J=tS{FI9E9G4~otXfv-RY_q=})pUB( zX;5e&@F#!#{f`%pNhiAAXZ7E;nK$3L7XQSu@oy|8T08ypXd&R02Dl_r78<-9h~ zUIbIJ`w{BHn|y@VJ(mxe;RC(GOHAq3dLy~YgH{@;h0Yxm-WMCP!Fwl*%$RXw7QNH{ z7;38}VRbv^HFMdM8$`eIS*LW&90LQnhZGrl_Ock)-Swq^Q{@` z1nMTukwVRt9`nOQ?K&c3276U_oNeg>qgJ^Z2naoPG9SH}PxVo@vU)yPXVNh)Vk_t3 zvCzeS9_Y3db)-nD44E8JA5GU@UAHavw6AoSN&X<+Q*X>fRo&G87w}fl!SEI`nn{ms z2mfsjJ&qF|d^d-#=5e;?OrZD51@$vGMgf_8Kg{5j^K@1jzy5%CDrAoEaCyi`H8e&^ zfp-LtEaN;?;A~N!P5y3GSfZ@P8*of3W(t`Mpsw4JQWKI}FVsjA05e0(O790UA4MVS zHU7b0>p0W$3+KLM=pk9|8NXF+;=ozo^&5e?#<5(cFNGFQFJl-TMmeyL6?rBz<(*%+ zxz)1KP1h0`&vRd9FigP-xf0W&=B>4Aai4;|GP{T2F)`-;45g(GT?B{{S5{@jK!&H( z=kExPFY0Byujb*5UKU>_|%|1YO6nOHA$USqNyOdyi9f_5 z{CdYAr>Di;h%G28A0S=5FMVx;RPm(CX1#OHEEhRBls=AT#q&3;L4KhSq4_h&bGZ{IWd^XQ3Tz8OKvXKP%q%v_S6*wU=1%f5qIm>z9Y7tIG?MpXw%y$Tzt#k81|&M zP%nM-O*HWJSCxXNG69uUPV1bL`CfDzOsDAkUdz%A6x{?Px?bl{ZH_cuiZutlTQ(fo zz~t>gXO;aqbhx2jYJ~s_7;!raIt(i}omDPj)TtL2H8On}21=i*8TmNOO>@@chW6z9 z--t8wU&@+Pw$&Lw7_e|PmP{|gvzh;Nd7=4>E9rgNl1Nzck(yl8!kyKOX}cX9^nyqt`*bIdL4cCxNPnf>!TyIj@%Xx}8g9(zNM4WgjiX8yC{ZZ%fSH8vG}peoE&C7e1sTqK?d3ba75DmL zVFB|GPlr-CO*4nx6+t%Jz3ZbaVO->8=s)G9lN>HW1Uw|CCZ-7pW9P?F-8VL8eAjPJ zRADfz;@{6=YPjH3k{4RnPxKV6f^Pnj*}$hLyP&_{s}sBalw5vCp09J?Y|13XL-^eT zp(#8>dOYK1t>=d%uU}7dUIai{G|o~t>I?!i*4}bdi_84%dsqwKZbZ#5leefb+mD7} zc~SWYXWjntN1Z&*?e733u2gPypH} zvCrhbM76Ivm%HKNgd${^CAO+;@IWQw-gsKl%-i{f{C+=O3lrrzR(x7^!>< zm-(NPo7K4up%rbFOfqLQ!vD$=n42rSwri!LOda~0^uibI4!3QQBNmq?+&gb~5fnTa zHCZGs6`|KeMrrEl@)mSDCXZm{n5H7v(_Kk0lE>llV10L9rig>9;LM-uj206)ZL zUcPCqX(Wu*@hDu)Gf8b6K>U2@Rm9c<8?zG0K^g z`+2KL(Odml>hIZQT0j~jSuRs1Hgt1;@kp?=xv3hPpMb*atj46rdmxdzDY3t?A*PLw zrc5zE1S7*#g__B2VW&x^ka>4JH&y2JBwyrSV{k&DZbKu6(p><l!L0(kIzC>KatINvc(1KWq*;rmC{POCx`ubEq;`#5hdIqiC`j{ZUb{^rdy+ zUtKwT6HwdueH`lz*b+aG$h6Qp# zJ2V{FqyxLT5ofP9aX&q{ahx)6KA3qvoECV4R*+=(M+}7GU~M$fx~5Zm{~NSw98S3D z$11zHLOhl(A&A(iG?IH?oK#tTmdlGNH5)WOIS*531yTIn z=CFpg#{`BM)K9Z^Q3+DRLG0debnTWW)tGy;O`&V8dTZ|Jw@7er)f~iuLeJwnePn}f zduky%VxP5>1j`j@`NYZ*AF2+$S;ZKpJv960sM!<4Lmjq0dDc2P(NGWK%-Wxoz;c2( ze16trv)npWKa6ZI-EP>tAp=7GJ9Uy4xo)1j8l51Z7FoCDK4vu;-}mjigX9N$B(G<*+Ym?lYf|W&h$~Ne4SIQ`jusql?W|B*n(_KOOzHf+ z_WnQ8P%+=I0uA2@$UaxYefLTl@j>Ebf3K*UAa68|`CcsNBCF| zhH{eaFV`OCx-Eb>wkp3a_r(1N4UP|Jsaul8m__Ouna4Dr9% z`eGdX8p}BxEtoa^;Zj6B4AIJ?f*NRccR$<^3pphDn1PIZ7QOU@mmru~({xi`=6!|% zLgJK2*FECrm;Sfzl3y&W=z2xP>biI0;_;uhHG)vCP{a8>u4U5x^rHPWvu<{~A-kOR zbX2?;XqQPxLdE)$KorTsM!V9K)w@)$488&`{8X>A>m(jS($4W1_dy_0C_};8qBsQO zLH*81o%OjMK|+=(GBOiqx3KZcK!IOLo@b4F%_s#vq6h7T9ZTxC8`u}i8IuQ<-L%GGI#=db*b}MAyXdRzH zj&$qp7}_4@<0N4!Pror)w+l@kh*2s=6Mc1BxeZJhi*!e{!8g zYx`^mcX(lhYzj48C+2wb2WTS}OCL-3fE$?u1EsFa{6OHU}gMot!~U zYbJ{UD{V`$p~VVkzgWp(83-@;PKj8ZN3))nQD>$wv}uIT1?&72+hYD4txP9Tcaq%s zi~|3K>6(eP7y67i7H44{B!3x{dDbfQpK`3m_me4~wirsp22yx-Y>EyhC$@<|-Q!r^ zJ4oatr(r`&9Or`AjCA)}%4RHU&kGSo^N#{1L6S_p#_H^R`Nx#|;d}Td!ii0n6B1?9 zfatHR1JyF16h15CR@4RP8@T7Yzi*E0=W~zpI_Vf5s8rt5H;?sdzoAxT97xUy9v@+qC7k`Ob@3*J}-Z z{tVGCD|dR6@wCQeear%tm@Q0L^ycG<6)`*7wTV1tyQ25V#9@+|-_X8WzC=+gh>%IL z+p78Fuw&{sp@y}ix#mBQI)wyrEl5P%Y+=0TzM#xmEShIZsr|E9j&v$Z1XtH{&+}P< zxK}h@)b2JaWiC*ZKm<$n+kMpa-qVhHcSfjQ4S*8^1S9z8HWQ5?y=n(N zFK{zl_M&q@)Ty^)wFENOLLg`B3w!r~V+=#n)ELXA|C`UZpMXI6xdMw?aog0;>0`L@ z<*9pU0bl17d>~z+iH;ngSK843Is~yDqHg|l-IQMz%jGN^rJ$&oi(B+)9au5dB%=Em zgOf!uvUbm8iY7VnHU~Rldmx%LubBn_#!6ezURlkiP+u}> zKLqrep{sbv5=2SKT2oy@?MPIVhR7(TrWtCee9UW<$vupO1D@g>kEbU_HC|m-tQ(^w zUUIm=;;&iGjj;9sY;$abp!8;UGm)iDfZ&0%|NVNW{p^&PsdYdb11aJ2oG(s~G(>r_ z$=+yxdbhM{N>6by)yfZ#rDlrQXAVRj*JH&NY{(wTC2~g4rn)t<1c|2B!*)(Pmx{#@ zFGBkampb$&647s)2yG z$rGsbI7kBZspF(w)r|Jy)auRwUv-t1`*4<_u|0bHmzyZ>cTd7&R6O)jb+{DxmRwwa zoOz~8uubq^Y(qEcuBWODTq}AN`ym{}z93b0G$=v@NlU(4 zsEDIzx~=-hqoJ^@=ao0fvlr&Edn?7Z(XDg#W-M|cyv2o(WXOa&R9=R`&9eYdM)6ExC*=aVx85-{Ty+U(|~76 z^~Tt zOY*x12(}z_?G<&E<|F`N~z?prIjl2Lhhvbp7w ziDNvma^f0px4eW%8n>?%O^~ciH)(u}GQ$zCAVr6q+9rm*!=7PQyK#x}1}_=@Xojm2 zS(ozWI|1BP4liG<$D_SBpxI_VH-JLW=LqKuQp{_cbljk*t@|QUHuFuru4zh&zSOdn zh27Z`rj^aM%#UH8Zahli9=#{tFG3Z)0EfeU$$_}_qNnO)Jr!TODOkmHOeq_=W-lnmmoW@L zDPB@#=iZD<&bxuH28dnuSM!y;XMsv81sVplpysAL4PZDarkp03n(RKxGqN^c~b;A|wkxWsfE3*)Wuq+2~7{lp&|5%CZSZd;=tPI{VIYTVqX zsva@}XRx?5+9<2ua8!#_)p4Ixsj*xgqY;*4g$4$LV2tI1}A{o-6;$1^Dk=F{(%t}mu} zD!sg+=*+NCS4c74!ed%&A#4eQ8N;BqTiU(Y< z=HkaB$czUDOx4G@25Yx!UF@vrE}c2Z--@r)A?tgn;&pOiaAI}31z?+*5b=qsz>EK) zg+eNuu1*sZ*M&)XuSBo^4xr?$9OC-gIG7i3v~Q{69Gb6Pr2}YnO>gN0qMMBv()uyb zj|v66sH#Gy^U3tm#lD;gxQD{aAgW~5)nQ<-k;bPNfGyuNA5cFca z;LO1I)3pE*pitpp*J9sD7ZjY^kuEEr9w8wKnuxaX`fvA0QV*L1g+~%DSOtZD6M$`1 zbni|{X`&LxnqPdnKD-oZAf2jB&rT1@ZnN$@-+w5IqVVHp)Dp7%c5|MTX2Y|abeMXO zUpcRlsK@p$?n+!6}r`JtBl0B0R4`~yBjsC#nSegW~K23dS zH`zVfxjiV)Sn_}n|H((;e8q37-|-*%;&}+71wRb~j?r79py3sX;~d;heaLxMf3RZxbn(TlP?Z zZRxET>*tsXD#Mlhs|^^G3mzZh5PJd&WJLgWu_-ETeIr4OQz3~TXxNtM0Dhx1 z5n@aGg}GXsgr@p<`NTLV{5yZqeZt=>Vnk#*%cyM4T@a(?>>+nz?iM4PnEa@Wc&Tk* z_@va#0%ldE;>2d`-fRfqcI9fFUPPfDy@(|xq3Q6uYH$q?M_F4N2EaO!1BN*mrDl6X zEY!DCQaC_OlQG6xO%jfXUj1y* zo3B-&kpPnGV-E}I^@Y|5d^?!I6pV!Y;h^6Ck2JGzIT~hRP>f|um5h%>ODgIivO)r^ zLiOD~pZ6luY;XoE7yF*@t;de$7DV( zReju4!?<}5uo}n*I{nLd{_WelkA!U{o^fMU244$7<4eDf_ph_~f4?PciInt5hx&a{ zABp-TtV3iD^Iuy#|7Eag1DK$c+Du;gyYr`OAwjPO@c&RS;_c9m7~PkL&SO`dGnByC zu_!;9AfMJT`**Z_l!a(r#A0Zw--k)wkN1FHWj1r(t2Gs-jPuz#)G~H^^SWiZAYibu zQjY}r6T&T6E=BzTqnoP%8jm+jYtwvs-v#rn?|(qC|&L;G&}>bIY3T|ZA|pliQ; zRPEK>L}ns0N+eHPh$4jq9egc+eT!qlePiTvzlD3|Wug4cj_aHUohHxlm<9MLfE)h% zm%2YYpW2>&FxD;qh*YGPg4b6xHni$GM@coWGMle6L1S!Y+_&Jx#Z>`IrHEFk**UD^ z*=Z~rGV?!3Cxp?^>bd3;INBhZNoAR~6WVRajyxY>ElePR8H=i;ga_^9P8WK;N1<|e z+VM=Wk;_J@kJ3{yQR6I?KSznnA7LG@m`0xNQ7E+LXx+t$VMZd0Njo5@ybY1OLjm znrK_+^TEVMr^cMSd1Fnb)b6$Yub50HCmx*Q1JTI-AYaozgz-@@i8EO+Niccoch3KN zs*gi~bK&$(RUrI|SJQ_VM5Vw0j`6p35)TpzMy+q- z7TXrCz|Qw|);sG~CY&ciSssGsj*iLulUMp8F8!@Wa-C!Hh|lX2??gXi>aG5REGOiG zE25NKcmxf34gR^+-y(t9%5g2OmJk#p_{@A7oF}W2OZC~Tka_d6<%9R$QYXpO^F5Ls z0-!QqT&_cD)+<1XV)uwemCb#lm0tRliOr$lE_aOtmK{B7od;q8;_p7GAO{?D)ASbv zDhYT+#&2&>iW|u5dI-L~rvD77;*BVT2PZEZ1Lcl>uh`>ypFKIfCWQ44sd7QuCL#`8CeT$GPP0W9l1B|9%#G;Y_m>mLZj^!z8 zBb9j@n5Av)OgKLr4ncbsaxlS!wuRlu=bZCF3Zx`5n*#iIn&s6nDxpSKTOUjiQcO8T8!)!v?^K|2C=6$O$OP_ChbmK)@~t(ppcd!MA0-U^N+o)4(eT=gpf}D5p2DEU5 zd6lx;d}7Htn5OCVNTJg#JIGER8`^IGbZKHoa?xOGhJENJ$sh)5mWynQ^-r}o`q6O0 z?28m`jq4*_y(E^sa2D(co#q38j+oJSpJZ*VRDhv#rmvH}Ev9^EEci>W%Cw| zJ)mAEc3-;zf7oo#+X0slF#|IzOk&f^t3T9!$Edzu2V46HBe^t|sY{MBDmrW*NLMdq zFGzjNk(bD+Z)?CJ8Ys_({7YI`aqb4Ct`@$G+M1LOF^ir1H7Be#>TmP{HuLeY?ep8U z`kT;HGcz+MV)1TgjM{kefJDnyKfmy~HXqv~?b6aNrNL#$Bxr_4APS6!gCs8d`8g*5 zVk+9awy?KJFOc>W|7^GyvhORbuh%b_8BiMRlcaeAv3H>CoSnRk?!|8X9b`q?RmH%r zLYu%P(wbw>6Com>s48JGQI}|F;9F8suQC6a+aUEMnkpc_ znjrIZD^sx9Kk*avi07a;zoc?QF-2NjUxcsg;ye0wzfQjeSv|Td|1~CtemGZos+h9J z%VqY&W83^G^3oMoNGtUe6j$&=61yS4mVnH3?YKf%RRJQVBnIxPfj?cn2r;SJ(*ZQ~ z@9zD&++oWfQwoM)kjp+#WG3|!xx&p0O7WpV|24uXx%^pQA%BFfe!{HFX zqWM;*E}0Sov~Id9mm5J#auA;k{21-K2!>A)l}F3E{Y+|CM*2$~Uv`ISIkrZ^Qxg0l z04wbxzy>Eq!F*gQY;^4L+|WA|7f1t~pK73s`?k!6;y#DTrbm@?pmV{rEFmvrTaqhw zSDr0Y^BEqk;0fFU&`Br+PnS8eshBmN9%Op?*8<~TvUHL;IrCNX>7nuf@$jJ6`E;gK z5AaDKd+UGzeKuA{;E?)TlR^LqFwdo~Lwj8P8O8U%azpKM-18JDL(X?AT^~iIz(gwV zS3+XCG?J+HR5LqDo#;xfWr8g`Ijz1gG?pkJ_YO6FU#FFE1;C zyP0n9SdN)Eh9Lyq#Cqu@RP5iHs(Mbehh2+;nup0efS1yxG&GNG>8IVmuXfhBCYD zNm!z=4_H1h9|!xKYx7GNohMky`)?%84lZ0`o_GXZI(&ZsssNS&BhfBEMF~6)#XJ=%)KPWD8+#G7;SC6&SS^g?? zE~?*(0FB_Ewz4m%Mz!oj;+++2o_ifGK}s*ZrZb%j60FGpY0Wc|XPhr4ISZzYu2Ql) zjQEUgRCLmTSp8%!Zq1CdXd0&PLioSNC$S(LyJD?3%DUa6NZO3)m~o1p#!6ll&I#W` zfeCZwWNf)vh)%$rRErvUEu&cS!fm+mO*D$e36pD@#x1t$9)uZ}^%4IOHJ9?~y&>gy zb%|bS0`eb9R+Jf-)yObQp3_!t16htJ;U_ofYok+bJ5rS42p`f6G!~B zb%leJWlgGc5rnwKl=B%c1L?r7oIqLj`$gyG=@h)Bk*DtO!GfFC7rVs^cefwM$_4V) zEKs=G8~=|-;+y*jxMAq27Py6&IFXoBSyUv9>`E=Ix@an8drDpQw}i?%Cn~I%lKCnfMy1^5JBzYk0A81@@gBLegbR>)9;j7z zkY+Kt7O4CAYDUlSF>y-!n4Q$^OQjD|w%qk6vOT%J&L+bN(P_~6ujHC`hCxBc*nhsFPXtZjgdRWG9Riei%Pk)HW9-`*Z@BB( z1~3m9J(4=U@zONy9fk_GSfK35_TCqhQnfN`D5#?AnAVkhlNz18!E7<$LVE%M5S?Xi z$(pTv&ulJ4$BWIPK(i3emVU$9FOIyPMO=TD=HSE$V4zIfgQz)*Amncv@U; zT=wM2b2~P*W9Lt@goO4@t9P%X-4Xs48z5TL0E>MP$uibNEdB~e;C{YM4O=TECZ(69QBi`fXW+AJy`tz>|%>2fn_CPoRg|rEf|&hL35Xd; z_V9Yfm9^n`?$(^`uFHb@#&o9&k8SBU;qM_(`6~1A`sY3^BL$O&18hZOxlEdzdz0d2 zrlz(nuf*`8yKpWaKKDxH(7cFe(>;Aq-7w49Bb!|7lq3zh8Ls)>B1EcGDW|$QD><4t1VWjE%#rqikhAqL$_QFTdkbDEca7l@8gVKY`53M}0 zK0MhxUaRN=Tri%@IE7Da{aNQX)W(qZFr^N<$Ee2a!7x_-_J+lqw1Bd$9;fm6m4ETZ z(;QK1YFsxlPPtSf8p^uL#3f=j#+oi|IIBn?R71|z0Y2}6qcfJIj)wpIHSn?G4F)aV zZq+W-)Bd{dpk)O3jWQU`Uyad_HsZ@)lajchE*>f4&+3(Y5qN(fC&%NorOHcAA(cla zrbvx$CB-#k*4`L-1$8j~JU3u@Iwl&{d9qNN=A*k$zvpT^3lgOBXi4j)#RA(p{{96f zSFS(OkN!(kMi>Pvj+T_erabD+#Dnl-?cxsspSMOqb# z)p61+h@}P@ru4lL71z!F z!FpiN$MYP8v3vuP<3R=O+Lgp5`&4ViiQL8X{ys;sTn)P0T|LAQ|6=TpViBX5T%zYkLh6#N@u>KvOBD*KAM`+l9rQ;2IY!vVGRKM+iTC?a;9E_x6+0o zk;`gla(~5)ZK<(Qvl7Y^T%)E(J$ki^)zADeU(| zQR$L0#N4=j@s$KmMI{psjOcF7u*oOQiJVQaH{PEx`9k@jbf3^N5df05B}NIA$3J%4 zYzX@R-4Oh$UiNNv-wOd>GLAd8S+j@i{fZ}1EYcx|TvU;GKL=#9pq;Y-5Z-qTTP&S2 zu#}6CS+|fA0AjJ2GI=2pWczz{upCDQc?kmjtCo8&DFk#{5T}H2VM+gB2V7h z`bIon0h1Y&y#yQGrE!iNZO_qIHZ*D12^^D2l7`+po*T6?QfvL5Htn6gWr-hQZkex( zQb-AGRpZz-5C5_n}A z{$>08b#;G_^Db@L5jj@r6Zt1QBe(T#x2_D4-FuZXowPf*vNX=WSjl#ItU2T29-I%5 ztc+9p&I?lF*DzGyPCdQJfV7N9Kax)w)f;R*Lp_u{rXo!GIP&46Neo&P?_E54O zTeH$~Iho0smxDuDgVRRCGC4~|H3;M7r`JVel4?i`_1~Qz|NK8ROnk%J*CNx(fIHLw ze)wB3K!1$osiqoBH%-(4Zq1_zQv1iquGhNdUF`3P6;zL=h5TDe#BOxHlEQIJbln-Q zx=PplX@&f!`MMCdb@s*%efs1wD|c9QAJ~VVP+1}q1zmcp?<5Yw?`B~|-s91C`T8ri z8g=PQ-#46xkr62mN&~uZv{kDGyA3nb{yFCV^eS31P57};gK%V26n!^y+2<^RpUlo{ z_4Jv}KVMQgV@BNUdH(7lzT>~rPX50#`d%UoYu5(>6n!1L`Y5R`?hN9i1i`o%$g+^Tv3s?eYV`UHBYmORe{R-E|U{-xgjDJUhd}Z~IAbnzuslimPG||Abdmugn zxaLQd9vvG3cMwn*;1SL<&CH3H7pQ%b?uF@>$$|3p&WskB#p=&DZfqPDP8P;~*xyfE z_hTAzru(}o`Uo4{{X?ZXs@LHT&%lr0^(P2XCdHh?LCx6eajf>#bs5JM$#e7P@>2=w zYqHjc#C9iEO^k$+vc8Fh3AgxU~K@hMN?OECZ2SdEo(icE+$vo zs!TP^6WVBG-$tNPz{OguNjtMmQ2{#nf*GoY6I9#fAf7G;UV;!>`i_+ zM6V#BMmO3^;2!|!+TWZLU^bvahg3%dl#rq@-dXpsXaC#Z0Tu5r7Yg zN-?q=fRp_X-~0c`y@7#KQ3Lxb^Rd~DzsmeSPU(&c$mhQmsJMUce;Ep<_zUp(bLoLU zBk%uRx{vdLvxntw&hzwd%@w4?gFb=BBO)`&|EH&)t0857IApK)KgbYx4s_sgDr#xI z|LN)QVwA5~?uP$^Q~-F5>j`>fVxobLYh-LJ3$R=i@ud)|B+k_Ba6{0#2z*ExbMs)= zEa?}e%vqH9o?sjqHnO+=ZXAH3g}s=HfkCF!!9K58)DTY_RgwKE^(@-g#kkn#m9J%F zqWAY5QJAR2WEolj=Ww&1#OV+;5i&4qu9LMyv1?DfN)O+5)8@P`!9XcS?m9v$o zoPL4IkZWRm=k==>&3eNePbE<4_LT+7ge$EljZv8@Cxl~Aq>EIT#*7)o(xy<_m#lp; z{v7Fl(cUvOk9Y=;MnVn?!{?wP0;_PFAh_!ThUS+7Qj2~cqKX_qW`-1U`inpDzy9Mr z^wJ1@w+o*LxV%Kroz#dF9U~@vr=R9p_Y5Pb-SD1ZiX+Qio`(#8{K$V-_19*;`T30f zlL%f$%tpczL{8d>dEw#z-=9#w-klVl9Ay#K;AU(b?eygfDg=NKhPPy7^oY(Mf}()= z6wT1TO~C6jh681EACkQZ0L!0kCT|XDZB%ftpb-A=Cqr(%u(fpx$CsGpx^~?q5&CQ2 zN+3NF=x`KmnlH#5?WWJe_;-6W2#Q3s-(ct9;5bYk{G1APVAc}>sXGvIdlBjUcdMM% zfem^^)KK5u_*(6KGZcK8-dbUmryr=`^K!P1D~^c&wj#m-SjSBa^!e_*hnTj^{P$!f z#lb=bJsm*ilm4SN{NsbirC^%xJ)C(s|8AOBS4B_|c>MO8%-jFd)6Wf&2iFPQ-tvDL z3MKw+-=Uy#(MSKo!-Q1DwmWymE3N7rH^-PwFF}JPh>M-2tvkN4w=V_U0W6EB(Rm|p zwNw)bLCfCDgkM7vl$%Jewoa)D?~f$>f4P(9u(}YC^eMzog38_ZtrT=O;G+d??;vt6 zuQ_dI6b)<2(nQYP><15GdCDYaT4$VfBaDpHw|@#Obd-`zQsQ$VEwApcdQyM4`qz91 zuuP4E!SUAK(GdxF@H9ECryips$Y1A5f)lklx5)S2>A-I9l=nv6ArgY04+ zPTm}Y6?p2FIVZ1%_z6>iV=)5LB*NhDE_7ddy1sNzvYOq2y~TdoRATau1?VloC|okG zPCF?I8fj^1FPgfo5p|6F9%*%O5KE11Fz$`0ukULPtd8#f{?OB1x(M~KBxJMx=(f7y zI~o$(=QeBJwl;O}gpAI)^%O$?7$RZK>mSaJr;He9V2t6 zdDV6k@E%xX44wJ~`2SP516G}%zuU01&yEPE!y>RM0{YdRMgs7~(o32d2u9L`>^J}P4E={m{OWO%7oilj^>_Fz za?IA;7e_0pj(|0GYcqGiKVtBn00JPE0$QS(Zw_0rUjYxqVo(fQvGGv=;6>I(mhr%7 z#phBr-`Ky3>l6c?l}PsNn>c*9E^EvHFrIx3Xad6kyLsR2hTE9)YQRqLL0Dju;(u-h zgc3Rn%>YGD#!^~LtY&8X@_2(`s&bp{8l83BJLTD~q{H7}1b8*uRq}RVGS+J|BsQ0; zC7vx8pR>co^bPi+0g)@gDZJi%T|j1{OiKc zKP)`pRZ$M~v9Ynekxf2H6T9`%ZZ?^zu=r<2;35gYu9YMce!d-_rV?@qbWU@(8%aw` zs~7;9VTDdPZftgA!%i+k5K~R`LqlK28E$&DKLH+!fspOy&wfY$^s{f&yw-oE8L@Xr zg}njaqJyG8z)5*PN+h!)zhwXzeqIdofyfL;LjkM}su%p=QAI_CgSE8udnm5Cwzj(6 zhmcQzb1?)Na-Hzwb0`)W;(-KTP#~W;Rw@S=Zs*fo8+HKlKY)Z&AKuLN%LC!>heC9S zwD(2;NP%Mcp(cDQ!6dp4Kt2jrKX~xJX6qYeukQHqfEp;PHWR8_ z`_awQQ=7p#jLjWd6M$YI<9=pPxdZ`fCxA<(f$AI6;auDwRFQ-)z}JV-R-ldG|Gach zsU)D@o6=y<@~%*}2K7GnVAR7Bezes_T_NuGU>rrmvr}IL>20(6e5=C#!bhykh!#Ck z8Myp*=jBn$|9Ldh_+1AKRmXK1HYZ8qkNEi1946Aji%kclm%FqdvsWJl4+3`Yq?cuH zYU0m|rYeELm42R(175%4jwGf$rsI3aHT@y%WF?yJCk_C(lEb*op@2I_CSZ>FEj|4n z-xt6I9%zgc`R!50HO_i5oHvvIy3}%p?QnfK3^;8N_PfHV=R)U+PW|Q9bdbS(uW#S( z2^|?zHF)0)(BCNC>aQB>R)jm`7&}N#{LOfYHUhzK;Zi}yyHF9MOFFINNoDm?V?&Tq zoP9HnP0JLqBtpdf#D8Vd{ZL@hwp*~z^*(HqjTtguu7}T_`9%u_2pN5qL!@uLA@$lp zF=w(!HO{{p()-H`08njTFWAr<=%kS@+5PFt6syaLrt{r+Vl1p9Rj}-EO29;2Tm?43OUQP;9E3+ctpZ%Hj!)*WCsM zK6&!Nyp>^GBi*&HOGMqwai!mVL=A{&d?XF{H8z>AC$nhey%*BA-A7}m9=>VoZo?^Q zw>-U8yk5&rE!AFv@W=_?rItNUGBY1;h@7l~ zoa!asn)2Dz*7YMpeKzgX+>Jxq0jE0&A0|@NaUzTANg@!o)VN-D5dfTGo3b{bcXRuY z3g8dOW!jpN=wY7WayZ(P)9{mP_?z93um}BzN*-&WOKf$@c@5CT*o}8Pya!Q`SNo-> z`Xc=Sm%iubggHBG?V*)hzLSx0n9hc(ujs?;$=GTXrT3G9_2Mxao^ZK)=hild`B>qQ zeBQh2=apMCJzBZMm9iaxN*PMdl{Epmv~HaB;sluU+f}6nPWz;eHDF@E&pjf!g`f=9 z8F!Qh=F&Te%y2a!AF>1${JGTPDhR{GtafOXnLx3fLh}E&Gk#zy`eh|6& z$VTb(R`@d^HB~dyY>A}#>yYEb_EnZpG{~_=10jCSy$vQNVN6ucHHuU|bhI$oCn%A!zI!MGWpgz2sX5*nkjY9*yQ$T3-KX()Are4z zj^{6pJQlg=G$a>BZHA}GZ>iffjKU?5dQJht`3{S77}r}=!Pt}Zo;Z4Tgz%WVBS0m= z9EP|BJ+St*GoE?NhBzWc)ZWQpWwfe~VPCrh;zd$7NFUO~lG?@oB`lu`pnuN>TWKah zQv}kyNUp0*%bm=3@_W^N&W?c&s^_Frp7aK!+Xs^>BiA7?T# zZM)|Cb1}SH3@V#0`VqT1_*j($GLswRdbl=GScLPvSM`kYAdme8j9|)1jv=Afnhz>3?Cp^OTePpH zuEUx#V>~nD));mB1hN~i(qL<3giiY;qC)KX^MIrm%B$d)UsBYWIXGT3%shR&%&KYV z&C+~+@#!7}0|3&anmSY@$w?3Ql9$;VX?<)`sd_{K?9l}sE=xG2ggD6&dTNcL1eCQ1kOyCI?-3x*6B$$+rp!uYUr~Nj|6}Y0D*z0?mf49M%K_RTz@5m;Tt?&p zL*~Magj4bO+l7x=o_y{9s;Ds{97`_koSKuA|DwHJ+9y3hK}OELhi~wsg?OT{Jf|4J zNRejCYM76Cds+~zEY?oiO`P(bC@Q9CMSmgPoO2RZm0HzNiQOPs?$i;ywljKG46P|JtXlW&1|&1NeGiwX)nH|#^Q>qJeH$DYn5kr z0CgTYE9&_B}~1=FX_&ctcqKbjz^y3Q!j!poTqK~$2dT@244d=xW4+y8egVJg_rdW zA0cDwGZ{ZYMeIV=O&+cdkBcJhi3azbE0@&U`~d4F{9>1#oiBrYCs%>!_~}Zz0uMp! z7ONaN&+R`J%uf^p=4qD2CK--L478Zo1S zGDPU@vm0KO^u0jN=yXwFIqJi2e-^kJA z%kwy#nKHNnPTAbBo48SxQ$OThl9*!?P!^z3b$f?QyL`S_WkgsRE@97HLIo(zbmb-A zW)`GMVQR?!DS`W4V05c76UYecFY5X0h~=%C=IB}$M|y-UO&ASgiiJT2XL8}2WUIWE zh5_S(IV8*@6>#lJUAZ(Ld@x>E83`%rc*GXDXmXxoXNCKcOkx<6glVr@&(!8@HaT@K zZ)66iG5`ZR#E)vFC&7Jq`o?id;1H zv;vDhi}GyM!%b57v;a;&q|X$Ku=4+w{%!QiQs@Z7X{n@aQAGiz$`7mb8b5H{^bonC zf+mj_cv#VD-D^ZJ^ z6}CrdWYkob%?;eGDe$)Kq`YwpdV2aovnBQ8r`(ofv@7hIPtj<-$QUQZ68v?>7>f^D zD)#RiwB%0hwQtyNcGWT_Ud3OuAUxF7mx_0?wNBdrVi{+T;8gbcPAR>NO2-i;+*~ER zImn>%dolwh%An8GW+IRYt&Yge+5R|p8x$WAaqsE0(T$%pXhO8E!P#Punmu0y&rbuW z3)`MnU`LJ{OF|+&%0;&?PmFxn2KvdVJJ|?W9zR9912S|U3_d?oK|GG7d^S;YiM=$C zcupX2xfkH^l3^SjUzjv=iTPgQSA_ z(mxBkAV7~+&+8>SCa-9H>L=hBzqrf44SoPQ3?Wd(dd`+gEcp7(o2$V>s-p_q`;)1T zP&j1?xS$R5&8kv!tX9e=Cx~a$dI?Ow|6`xB@WZZ)B(9X_cPJlKEHjKQC3om#F2IPvm3nCZ0e0SXL%i^FI3#1A{X>OVDle1m|>4Au)6B ztosY2d_2x~|0R}d??GJ!Q4gLEmWv|$8zSwzO2NAAi~hol`}^9$%sCn*2W$I@W95(^ z9#H~$F%(z_NYfQA4WEJSntg5_TM6g0q3mbsh7v`FyjweAsonZ%(?4I<#YSx%wDVXOm4=QFfV&5p+af)&sg zh&#wC$2Q}rRKKHJ%SATy#mOe^TxkR6?+;BsZ7)n)lbVpDW@N?iA!xu8EY$R;M%6{)SV*_)h+|HPfI8H^$P7a z`r9%P42Fe;R6-l*mGxAFB(iD0biyfn7$^CKs`*TdC9H-v$%~=*!Ue66L ze)gT=3FOTO3r00N%4&j0IRIQJfp6wGqPJAwGV|?TVdHL+Yjq)oGgzd)$?5Csx4KbR zOkE^Mb9h3Dpeu!+17!K;~>(C0zqz7)-WGpyfR(9d@A4Hej#;ks0$3#1)pW?CnZoHy`0 zx6@p``PO^b_6!lL`3=xTu1Q|aM)0i zm+{jt9s&**-&jBI#VEMF;S^K|5m>N0by}b_(=xh)1PYQc^rj~Iy@?rAlc7>;2V7r4 zci5b#<3}EXF41lE$KH5{hqc0Hl!6d#wacN^ijN$NcE$6njXvZ}jm4uJ)snaFhZ8HC z7v^QMw&!$mVAFAlgsU;s*`0wIIc11}fkC1En)K;_w)V9QykqJkR*KNoD9>W*@iZBD zwv+M?!)Ktom%9q=CxIu~II4YxZ;vOr2DY@T^gjt6q~&$wCu1y?^;`^YIWiQRwIM|I z@c7*4zim!dX(}q77)9tc{0+rbljMc8bOh&q^*49ynAlhwdo-mHIiTI%T1gV^t9bND zGGI!Rj}9L|cv_WuHcK>)u#{vIoqh0fdv5|VJ@QevH13LV!MoWxQe|xYgOhdx+$N2L zYL%`fEZQZx<8Ak;BEusbA_8l@U= z07XV7a+ci%Mvu9KWGJ(=w7d;4ipI0g$^pm+tzspWV4Rk`bH_hqirC^qEvHbr;|z>Z zeVT4SzZd|LGX>J^Hy$|hkA8$k^qUdk-e-?Opca%hh#Ea6qo}Xq9>sdxKwVPjov`o4 zj159}i!tL8PAOcXS9g(KuCg!Mh~@>LC`KTo=E zfcIrK$3<)y6%rh{+3_f#Mr~FCy2dGj9IsE&v}-hJnn)0Beql5geTAJ1mr zz8!E)S+c7rTkCpVJtr*=Xv6REx}62OB0Um|u{}_X+x8eN2h_77XLWj<`?(737oTJ% z!w1Z8N5+SvZ(uVjHRF0**{Au;fvx5;wJ*45pEd_lu>ry4-r9&P5s>MW#5@77&NVm% zxI(dEr`rVWLf)+VM^i36*-2=mir`Ppnt)1gcyeHQW(4Y+ii~>nY~t( z8hUyL(^iDrJl)=~0CK{bM=pMGs#stO2sp`j9KHcAt)2S!V&ptOG5;ON^WkQ3cHxV# zd-q6T*aMG==SwQ62Gpq}wn`$k12zg&j$UhH=3ZJjx`kym?GF;s%I`zbciSvDjB*8a z(b-LYQ1aR}Lecv!fe3kv>Umji{3Q4oR<`}L?}h-mE+&`(}eA zn!*WUnbB(M}B=;mg+uv8uzu6p;pjP{3D?x@Lqo9#)b#3gd8 zB}|siawlog(=SI`-cgWBz2Q5=V)0L+*tMZR>)TT^B_qJq4ANns$@Vz3MK93W08I|<9ZxBL{pR?J=hKTQNdk} z<;-e)>rDm_3=UL<6hw->@P{_V43=1)9;nD=O;z!W+9X0sax=m#E{nkz;9KGz>?Wny=726@3@_h!>WNgsMc+d5 z+vBy7SwesWWr<9%87O|qjr;LRz4|a%VYI;Gv;owgB3atwJZDIC(%G=hec5t-fLtNh z_BP20+>X%t?$Ayj0g=8MG)I+n{WN)ZAid7T2m9z4e?O+CV&1U?tDZ%*{0)NwIkG4Y z5IFrdA{4!OAXa$2UB%q{8jed}O+)~tRkcTR`=H)+Xz|W~YWc?*m0D5Fb!PuLuN-;_ z)F>;q@0zF)qLx`nbX%O8H&pT3ZVul=K1^Xt6Eh<)+fOVCIu%FICY zfDr0+>Ryt!&iP~+Dm2p+(QXt5ZRjne(J0-D7%=J{H&9{-3a4leIU;k+0jbD-`Qr0A zWW4KQJB0m(Ls7cTNcI}1*^ASlQW>Yyqh<~Jo2tfYp;&mFn^)JJ)U@SoA8wFyS*#8q zSrkADunnbwIhOc>>%Q`4w~C*7-Hj#D{n&P^uRza*iJ2Lfcjx{)S4sIeL9*UF%7Ccs z{L3>qt!f*-Q5w)>pe+ZK&R=xo7KPK#{N#y3)N)ZRC_GV{YWIhhMj0}~aM@A6Jl%fF z*$%l5k^pH`UCdK|GYYZjf0B^~H#>~oqY)}gC1*Hr%vO3PDdTNS*p_&;m(l}0q#h}n zIe~ATZ{Dvlnc>oTs3}+Di&xah#0A(0BMEip=-U5=_Qs1I%O{4sT)clTwtwelDU1eJ zD#LJ^tGvG8txKWGV4$tr%o8A^nk7#1x;NcP+)WEX41ixf1R$T=yVR*-_s_E0L<3kg z9KQ^gk5DLAJ3H))U_z$&rN(1nzMn&=!nroM+Z%UN(VIUATpjhBPOBwCrdkFmT4FXy zh+WTq@TZEAoD)2N9#WLGwhwJAO`hU9D1nS*!ze8O*a)YkL-x}{#>-zTJCfulTz$3F zQQR81o@mzr1yPqKheM}dC=Kdx3;Bulb22j(2f`F_?!zcD{{)^ zXsqctmWU>7yqJk+RFN(=NW4g!@FwNUz@^_F&4UVFgM;@fbg$r=ukW$D?|3-$%tXfp0RzaF7AaN=UwA)WAEXbgJMGQcJgC9-kOEbmpDG2Rj!<(;BKnmmJ zCE>VDk}z+&s^c83Q|9JB%!%2!4(AQ;{_adfQ~Q) z_vQFMbUl8pCD-BkC>r%PZ|aZywxX5piJFoP9)H6hJC;%*@!&DRZcHtnGHrI)RJ{k~ zGN4M?#=STfPrU6CdOR)E0ma^P11XAVq*oi* zun}4Ek}>sCGdA_^RND;vuM>F+qvEHh?CwTUj$8dN8RQ;NIApWMJ=kwOaTxsn~j6E>0I)7Dii70JF7D-hNFJPsXm{B zU!0fLg$VRlUUhKUHYZeLfAeqi+wF$$Pd3HUi(8IDYCosLH0||^bbwHkE27V|W8|%_<%&&8(R*{P#1Q?KCIsWF1Wd-{7h08wuG|kTZgv}a* zb)8F^E78?Ny0D8|RxQ-T*T5DHn8TQ%nlt!}Cv{9{HYcqdr>A*Wt5hEcka&$&_pxGl zoz4QQKt>5*iwOGTDbN6x*5+ijg2VQ-skYeLiu7DAh6eA_O+&dQkD9}!{4sj>Q}f!mzyyskJcJ=Z@E*5Z7>C+oCk+0(w0?4!Hv3RtH; z=3}U38F!d(%nTJj2e5FP7=YsBBMbQcv?#Z<+xr5u1nNlvmnNG^?ke}qQEplG9s#Gs;Td2ruAQ5-<0toA(}l*g4MYjH}o#@?zs~%m&qhJW?~-I_ugy7rE4Gj+rGuXK^)- z7NcRic|bh{WJTfHm&+jx9bFGyf@9_8LJ;Q_a!eWY9Mi5#c{Uwe)MYU&=`fuDwX!J| z0@S(DH-t9fTL>gG^W2f?2_tewE^2Ro>EaXa>zt^t1dINHn^6ASC3|j(7p~tWx%y{- zuJ>v?b*6yr5_j8ElmNg!^Qq=ywf5Z-^9dqpq@ZT|s*BNV)*NvCU z0t3aX2=1I2oj4(Mds!NXZL1NE@(pX=Ai2|9=9WEI`OW&JbAl+B?daF6EzQpXiK)eHwOg&!gOurFtqdZ?pUgL)pu?Xj+BULle~HQZPxPCsbV^&?7w)L+J7b5hS| z_T7hF?NWB;A70?B6M->@Xlj#-rOzTOoqPYpemN8c*c<9qt<$nZ>doR4E7TX!#-sto z#rK?MiH}Exks(qwJ4{Hk+G3@L5+fr-v{zkl1(ZQ#y!Lx(H{*@|3LkKf1-bPhnD`5i z?kG*TqFWA&Rsp{u;7Jque!4!5^ZjP4-^w3dH{=eBou{71SPumC-EnNO1CN2Wt2ek$ z&qcJOT=x%7NQn=udn3@xFH0eukoxHFK3tP{oIr_?gK-B?Gm3$GpTP{%P8wUbp*RcA zv4ui{DJMGiJ@?WUOo=5(J4 zf}M*ReDCA~(Jha8IPU#@z+oW!`$$fX7fiMkr#hL>S^r>#Y+IG!fuLF0NRdmFRf9Jh zg@AK(m*(!*HudB!4ej%kVhc*Gn6M;@Z-BeadbX!wZiSP&6V2PVZ=uGdpDZs|qSDI0 zt8P~=yoHPvI=yKT?#BG=cCw9Fq$iF$96?-)Cjf5jJTw~sHpB$7tY*sEdSm~}2Mh0E zZWb9V&enW5+aH|WID^+9H~UUkH|1bsLSWu~yv$VZICYHiSqUs%?yFLU!bH$1oJqP3%yZXBLMD3)oykp}r6~he)OCl2>$0II(Y9is(>plHoV7W?a z@>M8WcV)LrDD7Q182+&gLer*!bfp238343zmnr=arDEVbU2l*s_`(q~_D~>e=sIhY zk$rld^XiX=UU9l_-g+tj5{jrk$-3Ap^MIajCos94ANZMQ+yUxXqjv!U8scI%N*MbB z0QM%0gvVn5ihl)?u52R8F+MQRu203B2k*m<$cS;sOPK_TiR(z|?Doj>I}~PxtjY?a zTIHAd|HR_fRk&VtHHPE|Nip3vlJK0q# zv^jJ>onJu`R0efK;hLd#!!ryTmbU;MHtAS_X-KiJyafA?4{% zkw@6H5WK^04kOJnkIH`DFw-UJZ|Efdx?HRu0#GYud8j(6g!Z%zRrh;{8enR9*WBu3 zcy&tFhQP*91`vn$M0II6jw($T9eaW~GS?@H+rx+%TIX`cs9PyPmQd2J)DG^uZCjU4 z0pFD6WpI?KB#JbQ&ztfR!vPdl8s%QSMCUm5gilNw4QxIxA-rGf2J|dpyFSMRkE1+B z?QPEwzbZ^SGJ27|^`ZWU1^sKKiyq#3B)Hp1;AZ;p&3m{?qX#k0ytYYQPF3h^8Aeg9 zfWyx0UhyQ^8pIuAqbHmA(OvUbl26H|noF{&&0xFZWk5BV&-vh^P<3{;?3#JIBTzN- zgJQ=Bc=%89T6c0qvh7xGjTb50#{KdPdeFOc9n=A+(Kj)}bjKmTf!Isk@HS`b#Zl`R zr;b6t1u{?1+aEko&*_LDRp;<&OFai6#0+YDWomkmN$Fx(m-KhXA7H~GIlJX(2+v4a zvOZiI=*&s?toed=CGIv10woXcQ?GQYI8+r;0Z_E3s4^xXN6aczpUf`$(585-xrx+9 zX|IzcV_i)xA%3_m5O&T`ZI@;l)O%#SHj4nnQZ!mf+M&CDyr!HepI*_@IN#{c^jYmd z&@22KUB`j{pnxl(BD6> z#xwxkrdsl ze_R*5u-J$`|8n>G$Lk~1K*!O~_shX0ZR#V^xqFjH{C{2@@KfIdZ|ssp_I(K|0GRU4 z_VbVbVF&#y=Fye*d`Qb-Uy;JA3TR$Lb^lKahu65#qqhO!Tn-5c11H}vMyAQ@7qRef z54!6Y8ofmoAC}GQ&G9QA{ZWw>NuqacL-mgz`T6Rgz%}(?ihkwe|KWa2q-px}^Ijwg zzmVGTJ`kb`MgOus`)ibbKl(}1=s?vhtN00X zlP55ZIIO~WCH@c7h@UN#=4-UEd++>PU>22E7sve{)N{3mL@%TlkIoMZb8{}OKAFlS zT>f8%{FA7nV#=dOq16xB*bJk?ff;xQKj8Ow^Y8b1ixKd}{bJQC+K{&kB)L5|@UL+! zOiVAo8e0_i{ru{AZBdUIMg}6)yC@a@6Ea|pPJVj#>)EddioSncfRjs;pVF;Kh>!W- z1GjT)6@%kU8gMKv9=z8>#PIXC0oN-G=5`QS(}0>aTOfF2mQSG1v_H0mb{CeYg; z1$shZDt92MKk<4R@N+j4HOGxG%B3k!Db2$;|BW^z3oym#T=fBwt!3&L?%ZHtWUR1V z;$gqsR*zi?d?rLmwT5N%{cw`sV_;Z-n(>BE^ zs<{6$>HXI`APf?93YD}5is`DipqeOmp^Lm@oxCd{SjJW9>A%n(yhTCe!WU_1XSVu zE=^m>er{xq{-V~_khnOUjNdU7XK(GqhtSpiSs5*02QGV4TfF<{N~6KY%+g&R{}8~-vlFRM zk<5J_?mssE=t+RnFm~F%_4sRD{QIOhd{*m9kTKJ=<`{N%X$T-o5z$dcL(PtMhxLM3 z3SaM#a(w9H)&KRxP1Hp(h@d%|mZ-ky$VjGvjKI7wt#!N|AqKyO%@@vxs}KGCYYGxI zN|DfkiKZ($V3$c)Fx;r_dNzS2LST|O4uJ7EijW<$>KO<(JVG|7EKv6fF8gB!c! z6V&6h3sNw*CCqJ^yxl!z3h#FdJS&}LC=l50mG)YHrdzp|@dQ|uBcr0ETGp+#!u=(-pQSD7`fs1{iP#)`T;-w+o#E;`Wbk-8wYV+&qo$VcTxDJ{EjS1iW_gI~6+<(-4JzvhW6vdC~bxd3k7Qjz*9ntwi!UD1{-zXnr4R^BLBI+|6Kpy8PU5P z!^n-3X9)V&Gk`%m+^@79{l8jauA>dUT_O5#~QxGn>I1U5+VlLDE$Md97*{Unk zgBDX+pW=u6iEyK`p%jdP2TqkD^UY7NWD|I$NmO;;w#pQlpO2S?0gil^@aXv4Fy-H zO$gKCbLIk39|CT(>FwOP2#-Ed?=W0|Dt?@05dWk6F(4h6T=_% zWE&38qph-PG@sVpCQnLcg{$fH?N>ufjb_9W^@dQ8bE~>NKiDA6Lq^ehfe$d8xKY)+ zjf-2y@}{r+YT}mhxFJT~&;L5L|6_x-QWixN4P*_zBq3wXWqY3`eC>8Xoy)oMD-ZLO z678_ry*7_^-1RRUh=bI+t3K^$sbOxf> z?_+M1x}AyOS1OpxlMx9l9MiG0&vec9rrYBp(nh#{cs-Gq7DlmtS|6G9?^nFM7s1oj zJ8Ry6br_y7B`}fOlz!pQHd06hbn+>6TLIdvX?W8M3CR(4mE>%f4n;KX68N_AjEHN8 z#X|dVhA(7_R*e*e!0qoaYz}(7+#Ra6Jlk{@z+KMCWbE}NZ1LkUoWd#?;QAng-xxMq)v~PPN+G#yfS(Iq;PJUW$Ug;6sa7MU2KeV&=v4 z9V3-uH@dI5%Eh#&)5qvzV>?L$x=!dw4A>9SE5pY`L15WGKyQ}W-x-~9mb$2${a{5gnjJsg}oIs1U3}B|R zz>gkSwVWz^i!NlK6e!+B;`7Z^h-*N`PfVVF1xT^&+to{G;(C{yma8SoiGUbWMrz+i zozB+0nw5)7GM0x~#33H&e13;mV(uU&sJsE7XVQQg;`4%WN(N1rF}c^cHKGfj;N0!M zLrO{tr#CMal+A~lcgvtyxJ^cD-Is?gMxUIK$#SFJ)e>;Lr79^CYaIG@08H~((R6}e ziW3B+-;0elzllk_PKaq*A714`zJYFE$(emCQT;0m=3owujV0$y8Lm)6#%U_kS^7+E z4(Z&$+g=?`ORXrFm>z55sy03Hm19uVk7a@<^E8*xzm>S-(nCM$WL)csC8ycR?7^?? zL&y+BJ!dPZF0FeKs@bRG+sOnXKNH-|^{bPaqR=u=H$N0~8rSeAuW`=Pkoy@(?PH zRy-R(i6IhNE}E}Yc&v4DG!b(-%Jv_)4ie$5|e#8(uucy%bJD%YU`M^ zQsy&3w|z!6@%q;R&WQ;U2lc!XTb4?T@5_OGOH7#;?2?Zee~U@aYkSfJoFJN~L5mZw z@~{Se?Z<-{UJMN4byG#@H-(h&Xn3901GUn53K}B4ZdH0cjc1g5i_5BDbFm0Xgq^>C zi&OCeRaa@Z#Lu(e_)uxM4kx(WX^$5^4X0+~pyy+I!AvasbXq@Uhnv-D=jtMjo7ra| zFOB`3)4rP#_pFvB_lBGSD@n4#2?50y7}n=l1sMYP{?x$&A!@I5*h8Vl=i4%dQV-L&Hx zBd(IO%t{1;G6(zZ(cx)ujon&daXZ8!L<55p?DL6RC?qft%xAUpJu}9!)LGoE{vZ=%{e3l&NH*dmWR`lOJDdy^jwYs*edd(z5KdoT?F6-N9ur7Ka%U z#^_2X8H{6Z2Mv9rsKm{wP6XGxHAlv`N(Zw zu`t{wERO$x-vso9H(uV*w;|oEQ{-K|9z0$+-pRDvl?Oxi+#%i?efj4j$yUIBZ61L(hk;KR;<+PZ|Ku z%K~!!R5gRR;S3cXwF<*jkSV&*GfO3G2!zy}94zrSKpMvkd(;^6>_9+P{o*zzs#(Cj zCkpiEd%bATB_*4dm)9BK5~y7s=MOuuAM*bG3_%oagD+Ek%re)tGUOBYE68S^{_yH) z2J1Xaj2+p$D{ngfEi4eEM&3W57|(W|OV4?bUY;NW*hMb_jO>Z*+Wa|- zSc*q;$r=Zr*L@&h&reeW<9^v!J9Whu&uK&S zYs1ro)qY(voucSB8*O?nHrJ5RdUKpJ?gDqIaPTz5`cww4oDAl_#*iTp&tHeMj zAo1K6%3#{mrPSJ9Y^7)P$2wlg;wuisT$|kmG_Bna)Ud3!z)CKz>7zaKk~OQ7)<@4GJPt5N%bVf zDH=xVYVUrZ4xTxAV1JRzO1|mZ$%`%xEJ?sNZ>n?TD>={#?GBuY)uq5>YdJ~J4i+4P z^^{kQ66^yQRV>ATyAu(QDV`LX-=T;z?amWSVq^vlMC|*yQe6oQ(|o09#*1PJcdLip zlq||AHigkK%h7~kA&f=4&iA3?J$}&W7{rG(n3JRr8*pTjfg3(KdXRs#KGIB|Sc)L& z!|!sRD<Re&@Gq#rEBgLZ?fcM}&z1ECw}Ve@ zvsV}aa6$-m!|LqUvoMB)5;J@d)p+|h{a|&tO@5?_T3uSqa&q(ZlJqYx0Mk;Oc>Lkw zgF654rIm$A8Uk*%A;AJs*=HvSn9^he(l^V z>BWWtkDsbBbjKrf3hzC5O3&l`cAa`6c1jFGCKhb*nB0yqd+jJb@R2%cs%74Goo}08 z^5xR#gsq(6M3th!DS9w2O);3!)QbD^!g}sgB>&_8?5Ry8{Av^%ki;h&$aUF;I&; zHNGY_st&jXaf{K#&PiY-f^%jogC<;kmy0_~V_3mS42m13(jKcaFsy!K93u0x(?kyY zs(1(Yh}b*qv271E0d>*DsO!<-bCj!b=&j${Rik0%-Zti*`uOqDe&u+}n~PhTm!-V! z-Tt6I?F+pg-#kb!flgVCmxnEl5yYKtoC+`b_LXLTD(Wd7e8WTMeRRmd<9~jzsSAiH z^}MT*oMMFX4w!D_sGUCI(qix6A&O4^JRl7sc8uHFx#+#6XC5mgDnl!qIPXF?4gzPr zWVM`+c(ZEIEr#-{*4cgHvhTXG?RK|F=maU@!(3nTN#%`HK zB&uk=#!&T*-Q%VlF9~)QnD&B|A#IETWSwGJCmdlSN?KDjKrR6vm}3CHc;OTL(7+9+ z3=jHS11Z2g;q1}~Oz@(qNUSP>0*!Rn+2Aio@Zl}Ms=S{5#}vgb(7uz<4FYtn#>6x| z16P;Ur#w`0n0=N{d$i`xYVdjtIz-bTEJST4botpBF}mGk?{mmF%!J9E=?s9k4vXl zxkhZ3c;ogSk>BI^?qO-dw2a%H%|k zE_KB*GsbDXUXhW{_|(S^ZRxB6TZxXyL{>b`z8+JI{gvHxik7p$k(dUlkdDGg3bW8& z$})MJn3oJXlKj8N9;DEh)L4jc3jG>P0q5Zd@7~Dc{0g*Qptl6WX1}DV_%PeKdY>w2 zTrG86gH#Uug9Kd;geIN9A@zexhh~uDyz#>>R$KU{uN@56uHT*n=Z!(V3 zsopzwRyqPmx&Ti>8lOtouxuJ(NuXA0QZ4j3jX}V7pgqs%TL4)Nl|R-u-ig7j!wt%+ zZ(oWOP>w>)m)Q6+U)!FBnH6e3;2D{OklPXEMn*)aGvn>>>Uyih(J&zSf&MV1iDPma zj?Qq5bgqr(aBe=d%kctaEiJ{pIfPX}n|jK~Gq;j05v#`T`_ZQNH^+@%5>|7Mga7zu z+fsV%?Dg*;siLeWOR$s}=8_#Zbe#vyJYcb`CA=S=U%uHYzrc94T{A3!?)`7Td zJBep>Bzs2J2M4^fpgytapQ_)sXU`oj@ue-=dU?`mtYpQr$50@a;%FM0kN{X&+2t=i zK8sakNsfgs_XggdPOoE`zCo&$zLG3PZiy)Coz5PSeHbg(Kw)OZRa+0|BAyNR7{d!) zv-GwH^0>xt5@ekDd8?)J3Bnt>#ZDVTh4(Ft=ot@sKHB2b`5r~RpjW-%NpxhIF}$&h zR8etrf*F40wJ7`^VbsZMAQCmbRF}@RN*kNrT!+qb04fC#6}W$M5c&8Wu2?zAA`fGP zVQ23K$J4T?0wKK?h|^acxsjfyE8prsE}bQewGX3--2NKbe=BMDMZ_W}54>2^U9VZd z+9Tmbzjk$+Y|rVcBPhSRWdHtF0%zdf)~ z9T0Zv2R+ym4j&T55Gb|>(AxffSCEwP(}pv-2}zJRfc7f;;;(CDF7$HZ$G)z0A=b9^ zoDnpLgU2s3xO|+J9pPYNXn3|FZ=Pl^kLy?gO7>!u-|-(ZkmrE=qb z1|h4AVte{V!+Z?tzAAHncz-pooX*mLX;e^=jMR_Z=@KsJ#C&C(FAnr@e>4y=v%}rB zSKziHY#UdtzpzYch*0SWUo9t6%5B^o?lIQ&wygc)Ug}dKECe4%_4RGcCPKtpjAB^s zU7p#M)jb7Bpj>*y2CK8~DF!mdB_K7W;AqZA)9GH1r#;Dy#^MM_+nTCmy3KkWSqqe6 zbSUshZO|pW&8^1T506Ie;p^)WH`d>yyMxOmWKuWB9D5Vs;bL^rK1EyC=2M0>y|a{E zhGC7UNBrPlo$vm0C0Hrfy52OYS+BAfi^$uK&WybWEQ$Y*v8#@>pigh=eGLz|e?< zbhng*bazXGba#q~C`h-2q;z+efOK={?(Y1aqjSHxchqtIncrcAbKdjh-fOSD7OCAB z%lC4Roj9+B^E%3j%`)xiyIVfD4}vR@BhGPC4HA3BZ#AffR1{=Wvy-!fdun}1K4I#~ zaBL^Fn%?soH%UN4ne|8?+3{gDGn!XKaYkU87iHFnS!rxPPA|w65JydWw&L~E z+i8;cD4l((Tt@C5ox$C?)0ecI*zizt7!_r>T(3JxLDFVb$K3nn(e0XwT(i9l8J0hq zJmg%C2?`tVoQt?d;O@&CQt|}aFHPJXvl*Npc)&28Ymi?K*tcG{(Yy^?#4rO#Ec)y^YK2do2m6C*3ficFMVbzwn*y*@Od0&0kQKxCKP$fjthE8|fUhWP z+6x|sPs6#egZ!R$;`r!K8tF;qnuq#+!JT}R#bS*WQH z+W<;=3{9bnk0xji4X)K}mLQL?pL|+Q#A0NcNoj@)UXie%UYyZmah%5O0T+8r>{%`O zDvQ~eaS^KNWG%#i4gp;qZA-b4dY};BPs08-7Y!FTS^cwtd1Ep0>N6fp(8H0(&mZdc ztW$%nZENzXm6~&c3Vq5wp4$nu75Ty=P-)dPcoc9>H^X5WTLzTHgIYt{#{SW%)XUe% z>G2$Hk_;CNrfi*Y#{xFET}82=k*DT;I}C%Xa4wU@@SuUW6~>uiz2#ZnqDGLz>F2#0 zI9)o)Sj=&4z&j@7k*cD}!^PGA!+>enI1OH&fx(m*-U9V8NP$&K! zJ~!Qv=Wrf<*=eZQQaqR@9S?a|HDH^q3bLko+gsHZkb;qA=i7Ja@M7?E zue}~OV2sr=l6uzj)a!tSb1U&sXGEHGN1 zczzgB2$F*hihJ4rNv-XT-$<NTjehXSfcBUe2JrJ%X!?cxzcE<2PtHZU!EOfUhTeq z7>ffn4T*Hwt|eKs!}`;~(F|b$WHhxterCo2cU4&)Jsa?ZGdr`b9ROj++LXNgS)2oP zRigfb@zVBPk&dNn`8vj0ifwS_Saqr!AtH4B@gURqHFDRUw5%+VpSVg8O00wMbuCw8 zZfPEH-s<5}A4a5hv$|+DW_P|w!R53PWmVLbYVae2RoK@n z?ckahHC=ixUl>?h`1P^->K~vy=}lXHnE9Sz3|Z5~C3n*PE#g^k8VNa*?(?ZsF_xxR z5$|Qty7UrLiM_+?MoSFqaG7pmPbr#uKl|Lz^yxLrvSadOE6V)q+joXc@hL}43oo(1 zCSi#mwAL9|{=sN1Eq{F#ZP)q!CggCtSJ(Y8=bq^?$gGjy1?e{w`y16(Uv#;%A(g^eMIlZ+_xMd)epgME|*$l|ugnC_`kFIL-#4wxT9+3BF z>#lwW@e)77kcqSSI;G(@si2)KcL^lw!^frR=M-e z2l_!qx~bihAbnNCvOYBUB!>F&9i zcPPi221mmqe-NWi5CfGF*S251)=s8_z4Et)-iG+bJgr=454_$^t&ehNEWDX4O}g| z+sJ8?Nr-J-u=SEj1TSD!Z3$vZcr8A=#2Re$u0J#zD8P|0MkLdXjJU`!9k78|h_;(f zRd%8j_SN4d3*fuQG%kpcs6#f9#qXG(ds zPoYYP(R&d%tM`>!dXy$l3hB45E}JhMh%6P^kHbOEuD1DJyf{rb)rPE_6l#QNkcX;= z2LCEHLKwV#bJ{|By$NfipS_fYH5!9jgF1Doy=Qg_vG1U`_|0?ci3zrt%Pm)jiLY3p zlQ@#*;1K;5^vZwogh$?BX>bdUnZmutDw$w3as#=>X)|nh$6U_m*1_vKkX|JsEtA$c zypt?5i`Rw4u;s*7&UahZyne(&$!|ZtA<4#H%TSzIUn9Uf?yJyz=?v|LM2yhnYLh3Id%8>WrDkc`w zhm51B32OBfZ)tFL0Q~(Emg-AgrgJ|zbA$CGE>gC&>O=XLW@pi2d0*V;a8s}i@70v@ zxOA8P$K`9H)_#QdZ59tD2^kqDBUH+V(MDt+jx$+4voK;cAC!(^ zen~|weuJ0MMo!UUNoatncNSjGM7J&1&3fXg#`Z+ih@G6AT&j~j8WuyM1yd{%Cn+tD z*09;cQlPOya4H`oyHNz3#mnj3aLz0hM2=z$GX@d7`_wIT_6ynxu@_tQqLN11?e8+@ z7hyxO=Jyt1RtK`Os&1jES2%UBja6Oe17Rr|I<(0u+Ye3X4Yy4esjKGjmc`dm?7Bo} znW*I$)+2D@RgN|CUH~kI&xhuK&5ma_HxpjzA$w~=| zRHO_HOC^~Ce9BN5J)W;~bfdUz7ru-huJ^-|Bl{ApOO~R6Gfmh@Wz)sP)=*bsG$p=O zZT7jV_5RlDDosoj?%v%#CX^L~AzL-j&I5qJa3D7Txp5n8znPmTTfO@TBTO&78A+K* zS~O{?Bh#?hoR2r=XpqZLG=>ZKac~fd+Tl8lM9@&QEme8uM*(BWuvBO&zDK3}eR^to zyi|S3_F}?LTqM~Nd45%_GDRU8Ek<(FZc`FoD;_TCRGPN zWoLo0KZGuQ4##j{6gwsv>pai4I<&tgmnF#qZS??CBOmEGbup~;w9G%v^_7PhJ?n;H zS9wg%i;0y-SNsBu9O4x(Yey&zG>|1u}MV$cAd zuw?paYHxQwec|FXVCw`eZR))KswQ=G5Ru2E-`hp4(5aC@o?{ze%%G{M$Hnh?91V*X z?$4OnKhii6ptN4cg!RIMmdNRYu^t=?j~T}8`1Y^Dk+%uTM<66hDPQBsv>Q{kzfov^ zJed9moAok*KJNIFmlphydo+-EGGbDp>`c6^C$UWC56$U2|L2!~d>GytY&g-*TwxKB z0;h0MK+06Fux@!OE|K+)*BdJiFug@Y{6pg6^fbvs!ZK=@;7l~Z%7|etb$LQfZFj79 zRefHpQYIOWS9R18k;7?rT}(c|c(7VjXwoIb>9IeE7!nhslg`M%;5s}MjqK!9ny1jY z#NakAz}a>0e~10YN6Wy0!B&=0R(>MbLSi@XsdMs#iBJ$$u0#BTXn~XeLlF>ErKG0b zH@e?6!mM~V9(q=%?(OYOa-H550m8UKfcmgbV=-M?(j`#fJZVrNt4Io1reeqD@so>F zh$s`59o~9|tcNr@P`CI0=GOl3qlrplMusvC=EKTp3Q$-U=jDCdUj2c}Ur#fme~7#& znpaf3UbFy6|3K=)bzyP-W zniBtJF^s3`=ge^v1LP7#8*CV6Xt#qz`vZ@JWxljC&DJ z7GEU_JYh!i84t3_XD`iOF;0!JluNc%s0bz*@Mbjd&bbBIFE(zeeu`yaRa>Qab zF`9M*ySJ{)wr)^Fwe;&3`z-f`q7RHt}L+~d#M4q)sl#OlEen9>hLKs`N z5%vI0FUf4BUVWQ1>RV&`8+xx4;w#a>0m_Vw-xN*LS(Q`pLMp!2`(fm<{H_DW+Da&60+i-(9nXh**mc8*gRt*ICWO zC1T4U`asvVaQkfDH&PRHWZI89p_E23y$Fl1)$w7TLPHtvW!j@u%qhCW_DmY1`*YwZ z6Bko{dZzD^gSLcSCk=K=xMzq!7!x8GBl0Nt0eVuDNw!wkkY=n6wTY>rY55db12X)P z)%An&XY12i;YEOy!llRZB@Ao89QE?~;3i|bKj@P@n!QL3vG9Ikzd2^H98ZAFM4#wk z0o9m3*qqb==GJDI#`&%7AT8;cf(`Mgs6Rqkn|=81e_BI>p7U@PJRsDUktM#I4YliCzN zE7*VXTYx!)=>6oHBqlnTX22ZR@c8MA3552qUq4)3I**^)7MV^&z1ZvZAqYp*?P6fe z+e{;$CmDk?e2330U~L^E29W(peSUa3{AQwESa5*7vGNLgSg725Lb9;)lfI@KH*5D1 z36o&G$}zz~UbGzbmEW$!J_pZbS#@ySD}TNmZANAcyE+A}?S0rVJ2lA&oD}hiZ@T^J z1RuBcGBu0xvwJT_Kiurkm7wf%zb>d$_o3rf-{GP1K$Yg;-uHYsf zqDvs_%`3&c2FjR-2#QM^ms==Ny^|)SNa@CDZ*w&`UgKuXgKlR-Je2H}fvl=;I%nYQ zz(q)jejs}x<$$X${FBa?wHi7yn|1bvT#K3KNQu#3s<@=u_tPG&6 zk|N9Ihsao2I>*8M5abe~j=(H7pl9llzhBS#+WxH`KBxrxNY>WFm#e>bKNUm!`Seou zdFJfSh^#FC9w!redM0>svwfOgU2)+2Pek&D=jPSxpt_eGSl`~V+28Z(o|A!44Cp&= zYsa#}{_#zCFOXIo*v>Qp-P6Hx{&ppLod6+u30F;}ZY1c>gn@AVvU0>E>>_Ui#8S&s zhy%Od$^FEP5t>RBsjvj-o;VlS(n|;8T>{_C;-~REk z`QLjxBNAq^$}S?$<8*&7cKTYOIIJxeNE$5DF0^wwe9c#JsD4s$>D*q7LsqU~(RC$V zYCdgTL|PU9f)oh0kpb47j_vATX##hC+}f=S+}Sm!xOJ>Y0`46 zKcZGIl-0agUVl;AK?JAs94!6O3PoJ<9~b%GvO82lDBsuh6bMx(a9wiZ=mkecE(e0p z#9IwNMJ+P0KNE1`RRA!tWy^$y{bY9y$7VwsI8#Z$QY=;1)u&6r0ylWA>HXCr#tW#J zC$!%T`KSG6e_Q52+j9q#n}ncSjf&ACddVP9x$H$&z(CGRLx)NRv~NP?AA>&{oQsNz zP2RXmjX8I~Ny)aqy1FWfxsA{awWg$ZS!J{Q-RSeqeD;KYHbK8{P^Jqx(f|;!z0nT; zd1=n1K1Xm^oZhk&5Gn?y)Zs$!3yGY_s%%& z`yR?YOK};IY)6-7m>kvw% z4%di4YZc-5V<1&hxmAbn7u6mapAk zQQA1z3tykOfdaWd3)WIK!HZX4%$C^Frr5aW*x(7JF`?SS0ZP8Q&;EBM{v^(BM0kjf z-VQ{ZV1*7g`i723l=RYjvAU?}-8bHN7F1N!=uJf3FSOjXc2E%&eaktS>{PmNw7v*T zsi4|h5Ip1TyMt!^sBaXY?RJM9Rug}lR~V>oeV=c{SsgHMWZN93}iu@lA@I{i_gmPXw6iobgvtl?8ndNzXZ=G zlL!ILN{g=Xnrojhti$~Y4u9Yn4U$2p;vRO1VX zaLSH+QGqJZ&phh4n)%&ldh#0GL~-s_D$YqMe;Oc|vYT*z&_;w((CM*%zWfQhjSx;p z5Nu%O3RzsT->xQU#GBM8vdZ$r0g<_dFYV}3IV$Zo7m9ojaY{!Y5R69$2X7To4St zv-dc4+e|Glya{IEfau zs)^IuP|joO7*@o9wfYAB7bTzW!-PwKk?F7(WGDFT&aZz=XS%}Mx5@#SStQ^9kGOJk z8PD04__!qbh{5_q(0s<%a3nEcR`ULyAI~+{%9T8kvRbd;g3B800$5Et^0kb@$9WcoR~9cR(7@ReFjXDGD@|AZaQC zLb4?$=$!Ar6zH}swkS;n8xQa>o*zVhzN(Cm;3G{oWZl(b22aO!doi0uKriQo6a1oEBkg#zR2L&Pn_( zLHog&N;(Xd!x+NqctuT@Vu9@`h0|tEdomw}%Q0SlXC?RQ6ePti3j>&ST-8WvzaEH2 zW1G#!Jicg8PYB>HPKn7By*h~{z$`79_k z0;}JcE{wiJcMaggqraclUuPk;+Lu!VT}1IFi7mIx)H)_^1?Z2z>oH68g?|CW%A3v5 zR(<2*CWIYCJP^WS!DYNC)aCVKQ2s4+{p&woh4;&yFO14`Bjiia)lnO%^5emesab~hANnf_lx#u&;C;Li zFPXn{lm2$HgCFms3FifqM&7|fwfETF-R%b*5(!7D&-re6F~S(`)x~A%ey33W^@P8A zRS&w)eQ)~+{&r};-MEBNU_=_^GyWCkc@yp_Xg+tOQ~upKa1*Xuh;Yf&-eCF-k|i}! zhlrt&=M7{{_`i;%U%cXd1*fBfWU;=yj46=uyplZAQ+HcEAY@C*1?2T$FhyTG|Mq(& zzJjr`@-5T0v9S>=C+m7lGBe6KLL__He-AwT4)#w%i@&V~B4Shm0z;RM?sk9aAX_=v zn{87S7$lPkM+mQ=LI7cR4uA3Ve^n!R5pT`Sqix`)7g{$*e@=d-(R)D!_st)9WR9v* zR|(D)sR# z$iI7=ZWrJ%NZhA(9wRao^qdPDEU6qebpP`>=s!-3Mw1Oew%8%y zbbm`2Q263#Rl$dHP&HsYw5HL|FB}2!;ZmBvw4tq>Jq)%qz5Y0rN?8MIezINXQh%jB zBbME2Zz1Q2%0V_CN|pU~csM9xfx|$!6GOtxqV94Yjlpz-cl-0|=y*Sj@*tDahu~aSR5;0!kO=BT zml7q9JccURs{-kxB!(4+Qt#iRGyvlBO&W*QMI-$W1w zWM`X>73;@lGw`8fvss-8r%5!}ERKR?BH)$dK5S|tU1!tPr5|WwPV1N>p$0? zYkWOJbqr)=N&UYYZ!X29i^q~{QL9tHo}s&4y|5F0{Er*V>kTq732DqKg^u_K z!P!A{Y`70`VcI;rSp-i>N0V(Au;r7mri*NBks=!Fk>W7~7?n`fWjN%Hkf2&~J8|7A zl-DyzUCQ)8`qSYpD%nn>6U{bZfQHLaM|Zr7UYXOBc_E7_q;vpY{z()({vA3Ma-jc5 z@*aozL)0H`YLrxn{<|SB)B1apRjA#s=0*0t+R(OdUmc`OxSgW_>%jfS*MhE-)pk{r z;~jcCPd;aB7B!fgj~C0xba|UYs;)h|YWwmWf)XKRX452EvGpH9;FMPe@vmKv1dXS% ztukS$`ukr)%Wp({YWa~lan9f4sYQ{p z(FfmQPDr&|_|RUgg$MD=5sR=^lyxy5_(Q#&dhSBQZSJv(mw6W}=MH{PO?93nKaBdq zkUp3b`sf~)W6;HRxWj^f`p3xaD8c|qym9teA8x2pK7N_1$xtFBzC$5vL^M5clbD;P zJCmZ%=F-`2VXv>rduzXC!nehsd}k{vHj=F8^4PA=s3kTRZ+mi~C@wP6!rO>Nv^H9Y zL_SlEm7n`ur#7BKAYFv~GZX7T5!=2i`cRn7ugH3bG;JbZEcO2R-dO=exC} zDZNs~7fkS6e?^9;KZjr=Fd!iO{V7BHlr zQDIblx;sd!#sf}?=!Zf_ex8z|DoI|b*mlcZlDlJ=$fllxG8*G|Fi6dU%78Op`(~N? z8p0?ea zqI{1%1QZTM|9R`|-S6e~WZUPPNNL_<$A{kSUw}36eU+X!N2TWcFM69ZX8GT4j2Pt5 z$@v+%7$y)oyqRfDk*z%^ixMu(&ff40Zfk9e+5AE!>n zU;eNM{{Oc6|I1YpIqa&N`jtLKzPZ~AOTqHny9NclZk;wUAApygS!C1)!gR^_J;jR7 z7!=)&sM)Pbhe!KgWcupgs^!*osF_6`Z$M}Xq!P?z$>YaT2GT$v2Xs?5oYNS-^herY zj=B4oa;$$wBa`gt%&Q{J<3X`0SPAPZC1g31Qp3J*v)%F_gFsk164^phxOY!#dmbHQ>s-HUT zR5^$S7)W>*Cw~J+e;l~~9$Uh3=*`?uxq$Myca@s7#039HxaRzIjT||4?cvr##>JM4 z4=$2tv$Xu?HY{J&qz{VmfS(6|0q$Bqb){!uC>HKMc{2qqz74L>ufnkq6Q_YAlg0k7|Ih4wuY!@EJ0LE19D=D<0(6vdX{vZ*<}5~oppV|(!SDFQ0cDCKLSU%jny=?DYm37}9rqabN+H>^bE;?A`|(gR+3WpxeYh!DioDtok13@3RmDIOu_7 z^gw}Gvi=q0tvCBr>FID*@M^){l(t{H2MKN;5rmx`u5+Y6D&mf%gxGI{aM&?Sa9sNW z#in-mq680i>xB%pplafa%8AE#-^d0uO+^ zMsNLrf3^5&%R0k%o$Ot>yPGq-X?HjIop-`n1`L*Q`ZH}L{P4nEG0ZT5e7AlJFM0Bw zfKC0%!meADgAZe!#q_9MISJNevR|))PR`#*D8Pss1h$2Ckh?0Un;gdag`z<5p;7Wt z^893hOAwB>6)~e6cN7DE>$_@!&*`bYPvdpl0z31rb{wIneThk0=cU6Wk;lcl+}>Jj zs30u1xBllZMId|fp{>3|FS)2WH^;Ikx0R}!v1O)Xq#hnibvF;e^%Z%(U8`_)b>LyD-S%Y`^HyVvhETS32+~<-)`rT~F zOM4Z4mDl3ysg%7Jdm{kJ(diR)KG6sq*~VCvl>NpiHKgkRA6ve_Vfv=iSw-Z6cWI7b zI&DRw24!ZvOy_g89^KVF1%|D{`wZ7vBzr@1sn+MzBceKQZonHZH#ftHh)7y1PA-GS zF^4e^K3E*{@+Bmv(?f?;rvafk@kevTon2jwVKytPsQP8dp27*@12&M~MZQ zSY*x}&ic6%e>B-o@W5&!BKtQGo3sp^4n+qqmfdj%TL8GQOnxV4M>~>jVHh zW7pNBJlDGpZVUnoV>bfk7?p9EquBuHC^uJNyE>sc<$C-W*dFN> z+$0{afX~t!r1<(qX3nPo_IDEzXLGY_O)66zJh{N^Vceg|iENPg@lf`CeLSP!C{)^b zj$*WU>rXO1&ZP6t?D|;!duo*lkuhk<8OonY zh84!hrf4O6Y`BYOKiRKZMf+_aplFEj@wwLz61^ZG6y`@+DZk!Q z6G+F}jmX|G<;1Xvrx^jrA!sDD0)BL(C6K`(%UFdRRl&8vcFPfdYLHKN_BLGVoqc8R z&y%|6{~m=4uf0H*YysfEEI_!yWAHbt#*?=g`9lNIc%HOK+jCDrk(Q498iQO*(e;Qf z{lPbnQjr384C?>%vEJq>UQhA(0S5XZL(hPR?`3671&=zBmScW@RRk#d+Q-x0lHp(hUZ95 z)FZ#EKiud!3)u@{rVQhS~z$@v%&s1BQr60Z+iyxq|5DJ&baS?Zj=I`uB5 z=Ogwwn{2Xo{yX-h)wdFuNB-a!I%i(v5qQv|^XM^eyS^P`z!s;Fpimy&DI zIK2D@#6kM%d@q*aFmz0?^msbM3lH;Xg~+0YQ&2L-20@`Vu1xssc$0kB$|@;FI_NKW z=I5_us8E#zgwCJfv+z?20pd@~jqs5r5av|*_22L>q8|4<%Djou7t*4$AF;VNJBU^2 zya|uxw%P&(7D!Dfg<1o#Xgu8ZRB#aH#0_Por3t6ARVhc**-1U06-itiD0$q?pfzjx zSaZIStB@s+IMz~mh@%gj{J>qI1j*zyuePs-?*3#xEYOa4w<4j4&xhy1+`Qg|ev`ye zfIfp$$GgPkQp0WcbvkhHb>}|)-2ww&Q@{)0doq^ytND8~`6!&*duYER%*Oe_D~d<6 z$SLM%l7aGNn8@*zK^nL5tSL}Gj}-=q;zGrB4b45bYjY8)yhpL6rbPpO#b!2X@kRV5 zCJIg`e#e|yTl>Q~oC~YVn${yt8YdS4w@QsS$C-V$tT#px8B1AHG-4zAYs8*I(g1L7 zeXt_qv>$(ugmjt6p1OxfCny}IF4x7Blxo*$aK!#Vi)c)#e7?QC zJv^LrU!cT9VUq@6)*sqFItm5ivQtb~T*9Bu9@oyJsm~s`>>nzLI3F@O?X6yaDMZg& zsJWLm=J|Tf*jN(nxIz?g2^GXp9F1lK0ow3M5uL+?bBAI=sTvi7OgYUk|XFcMHCi(RGp8tts4PGFYQvoiX<%D8OORMlFAr!NG-+?r% zPiNga!9AIo3JV`O%|rP87)c&I;x^*-fhUNQxX~CM85DLO^Iqvq6f3I&*&>5V5cSY{ zqj!Z^jb(`GjrMwRM^jdtG3&NYk}|f5ph+hQPScU8vyY4K_6dWldDCpIDEoZUW<`Rg88I4XnJVBc(ujv)VbG6&&UBp!bv%5_sgoNa}KA~_u z5;`_x2B~pz;Kc1RvaCgqCu-z@G>l<~2__hFk{&(3;k9yqr}p|a-1VGB4fa8vEf`vKdt-<)ICMMd#SL3vUk|bsXp9$8W z$f(E7OHbBq`dSLD?)YUzWt?l)!+{sKu7-RAYw2$7WY#zB! zZ(=O>lq`w=Gk*K+uZf-j<#ITBN?s*6WqHqlX4HT2H_2k2GjCTa@fYc1PWIRk+Wy6p z@g~Y5G{0$XX%!WJr;-?_1eR%>QY1(r{yFZ4{_B$*z-i5X^PVI>x!4`ujR%H`;tXBy z;&+<@Py>`W?83@gToGjC^zRC$!GLzgJjQ{c0xfLAm&C*{w4fC*FTs zsk`B2jn=#Ia7o*+Q2s+s^27#ZuNNpA^I`@h{o|*i%MfyNcjMX5c8S%zUnRel#Dk}) z&T$pu{<%73lk1b-HwBBCr#}aegD; z{ru5+PZ2F0(k`;uAAJ8dFOY$2Q`O%3G!N@`>kyvs#+`@ba@wmKe7#TyuIEdox<8Rr z|F0EDxChOoro!;QFQq5p_2YIp6PJJ@4vjC&>5zv8`WMo9u4>;^UT~;gwyK2&+{tHi zzjy^&5G9VbCTRggCKgKiKE=Qp#UhfslT0%1Vf+=*qeq&+IAxP%Q+Kt{f|!R#>+E35 z_~Sk?Cuf_vm_!Z*Jv}|Qo1+pFf z8QKfZa$F#H-Z&9zq^f7V=;#CPeuH7GdJw3wIbPD? zqKpfrdj8jh{%4zizcC2G;(!%G+edi(ko7&Y-DYO;jC5Q~F7(~fGnfWC3 z3|?9M@Nc(zNV|yayM0}{c)X} z@@oy(;1wp^Sqv23lUe>cYadDu2NPOWb(d7z&3>8TT*woPI^Y#|<(g|sKVUrZbl2Uf zO)pVcT>Q?8dqZQMS$v@J#Ny*rvGlviUN}Q9Q=*NZIlnQ8?&W<`Hg#J6Zr{*e^!~Nv zl#|C>wNw?xxWY9^iS?mf8E2ODaeP@6;rm*9(GVw6wN|;FYV9JUrw0SmF{{t8JMzV7 ztRc>rK+B0$(;@s|ZAfEnu=_)$9YeK5ce}NW{jwe}?sqBApHE%4Dkwe{k*53!D-eEYS;2nh+POEyr#}W3)fny98 ze`g6=>nWTPo7)}yQK?#&>3I9Y(SmtksixJ!2lj>XB46QDU3zt z)PDP6C&%bJWsSE25!CBM>uNE_+n_h{aJB|n00u1(b)eT>THa#Ya+-N=zxHl-UIQWcf(i*T5hN;zRWm~~H>rkYTt5#;utHC7yN*A0H zdOtZQc%2X(dE(H}&>o-dkl%Ws>`>|DA(wFZRg_oj3gu#0Y<$YXK@oWL79y>k?lcfaW?2Hm$+6D|N50&26ebpF-X4dD~evGw% zsd2i2wi~#f58o=fLxi3o(Aft3KYIz-_{N@#$q`Ah+9~c;0O8^}_Se6tU&I(5Yi=1qG5b3y$?e z01%c1g;Oe3bqZ1&mEKF2&y;OZS-|?$`blI!iYiA!DeX;;PVdnx%XmlD$?gkv5@8c9 z{+=-<;=)ZiFT?h&GV09hEOI(bn5JJL*eRpPt_+uiem&KxHA|MW2GB%T%<2|&Gc1H8 z`Ai%Zs+KRU6|mGl^Wj>eRVL0+Wsam>cGth>u#bWUb5||BJKC5Fn>kyHM)%GiuQGQe zY3ER65r>B})vEO>ik?USd5kA5o1V9Rwd%c)0LzEE*eFrZl*-0WJ_;Fg){8+<_!Zyk z-EZI2s~_p>wMoInnuA9Ki05RY>G$`Es`}ocOIrG}WGj^gn&pi>2GNmjq}|$k=9eRo zMQyF4O-@2c>2?^n=iwl@_LG7>X94gwMyI{vXbhX7DqFSld4Fuua3X_ZrzJ+0Q@2&* zB_@-LZ=e(bh?WkfU5D5G5fS?<zN zgI-=_!?h0`b_Q`0Y#@?*^+^Ed&y~$LGoF=a83gT&5ClQ>LaqxY)6UEYDQXjWO#8F+ z+g2T~DBCy!tS_mL-@wbBM1uG?7`gHw(!Az%M|qa{?zDQc9m-nyRm`q^0wB=$NTMfG zf|G~vaQ&9#W?#{BRzO+$XqG1z@bvI2vi;b(;I+@0kDt47Te>b4%}jHZg-C=&)v0RP zDeUyMKWB-*nv*l159L!9@`* zR98;KQ9;LomkIe`WlCkd)L6er(rC$_8}@lejl(o4SEXISfCny*ySk!Vm7=EmrkP5u zPPhvNkMj<JDJe}lH_qC{|+oi@>p?<8o^JxiiwS2NVP?f&0yx4w`R~tcgbw%rt zK;XKK?@&>UtQ1J_*yFsdD0$%3iY<<6Zq<@_{hKl^lrN5lE$R9@B=e>v1{){i;x+V+ z(mcjP$46J43~D?#;{ zpNU`4U`d9vbTz?!DeyVY)!8zSb&4FDnO<|AF*Bawn zv;pjZ{M7z?hE5m)b4HNgZgDLIJ+c{WjxbBt(?8z*x;2B#$XxRI>G9X=_7;A4;8mw` ze%hJf{@CScqv+D;L(9B}ORMRCpT+4EtV8>&tZu#Pok{E~RX^nMx7Ma-`aTxEUb?wjG4SqPKiP{ zUn~_nz<5)}m4aQva`FuJ2@`5loe>K%isi|e6#l{VLT+1IBvr<28tD5upR#R&OE4Cj zK>6s~0DIofX;BTxXNP-(7&L299!3l{UoSTv$JR7K2x(}#HQv8xKfmr+VImz!uO;U> z?Moplk_F8hf|{4p1AAY#-4#ultgY;)HnAtge+dd@l*3@_HP}8vKRyfBTcG#2nEzzR zMKA7k;&2~r2gW)*VYeqxhs9CLbh65!Y+g!4lA_(I_nT2=;zO9J<7rEK+T5CR&Fzus z$)4>?t@Y|l6atp77PN{AGdscJDm4c2rsKs5fcBMI)Bd2ZSmZINp4wT^&P(y2~^#^L#So zIQgOCP?J6FiEK}PJZz*9;nKt}N4b!@>ajmaPEVL1WpLI!w~9XI-(1G;wMa_YytFrj z0(1!E^|p_^n7(0b78A&sDaSxTqx#ta1Yk+E##XvRmAGfu-v&SJ7LN50MG9yC12(qQ5+N z$Aas%!*hi#Y4PWM^go%=4?)q4>q9z3%)NcLBAU3MJK?8sz+myKQl!nlU>ZMR^>Phu zAsP%@TVz4FD{lbBb+M@N{zbV1a9(Z+D9ewF0o~cue`lA->xBH!pg{Zchdq1Rp!xgS zGWZ?1{}kl?^+hR)L2U*7Y>(dYR|ELhUGHUt_NB8YsPqy2M+7J<{H8$VucFAmJ^+=s z9-WTy)k8Y5H=Qyn7ywP9QVQAs7k~GL=q3?W@`{dt-V*>uyb6~5@Av%Gw6Q(rg%4oF z2-gR_tR2=Q?7vlB04GYu{U?NtL?*d{De%uc1yhJuf7Vj@_Yd?MfDuQPJ=H`4sB7(i zdQJZ;S^bfDgb?@gCWCU^Iwk7vJU}w7WI6s!Z2lVHpS|+xAVt+0lgwa9=&3x66(Eo_ zp8Q;)Vf!)O`QyI=`Unf5Sk5rxT#YerOPFzy-sEdWFZ3-w?&w9DKy2=2nxIPC3^j*b z3~Os^P&n904Ucw`|12*Od~#B6rq4C3+a@I}YF*ge5?q{`owG{tvuFID|CjL&X2LA6 zKkKs>`Z3RmjniD_vd?X$llLxf5leghwip-ZqutdyEfip}+%E!nIzsd*-R{PO$5Y?C zUXz0I3%Cy#3qs;UA47?H(6h*;9Q4F==t@0(e0~p?BB;sc&c9fxqJ+{{nQ#<0RaFwa z;f0OzU&*PKQNOvmp&O8kCI^5|Oqoc5qd$L6zCDW|w<;=% zNjIF=(?pwD=teg;d(Xi5ft9;C{E3HfWp${Rv<^HMKvJfLcDT6j824wO)qiYL$_Mc8uWt@0;X* zhKl$zK;`y{JbSTuhGQ=PC$-HCtTG(6u&|E)&PuUPYWfXeW$Dy@h@W#DARn!BKZ%xt zZHHRqY7fD=3VrsX?Y>V}0Q=?XY7adDrFlN^&yEo8d5FRAmSiyDF4JPEz6mCU-;Nc6 zF+Z(-|9|Yo{vPkZ_F##3DQ#xe7MLcC{U)j$E+rX^d!n53TEtlfDJ; z1>GInGYFN90V7s`?FX8Ej*x{=37G>hu)$!0nc5g3d&Hy;+y0-#*7K74h{D$UbdGLG z++o6ZY5pHsg;|n;qgCA5#igH71t>Z5GBd^T!z_pL#r+-In9T=&zjdF9p<`#2 zm4Yy7i!a^#45uhoO?|iF;JL@Cwoi;F#75x=ZS~m8p!I##(}Jf8tFl?#1{!Qn#2HPy zI&BLL5y`-iyY#D;?Z1>j{?AtoFy5wq_XQw5P75}C*}D+NsNH$X!})9xyJ+oe$m?2w zr=sxoEr}n+=CsYSekwb#i<1=j@d`1`)U4!>tNoE$0B}pm(_w9vAmdxgHglmdez3(# zK@Sy{0;b#RS6=q!1!|EP8Q9}|BD-?r(yrWTvsJ?c94`C)6h*rdeK##|gf6j+le?fWxx%8^5o7CtZ1;2K@T1B41dsX7N`Z*`N_OAwUOiPB?;$z%Y_d zgdLBFE`e><4co!knzWxN7WW7fJZG2;ns{5%J7Usx$ISp?AK)zvtv%fHdf&$P61cYg zEXg<}8lDT3azgYCueJ`Ibl?0_sm$zhTj`C%*l5?!nejs>Cns}6KT+RpU{{4j08n@U z&3#)hTtFu2*|VU`u)F#yAjJ`Je5l5o7tsK6b4EOzlrK83?b7ihHlYmV-BX}IS?*A4 zVWEDcrtgHs$y5Jf6iRw0*=8~vVm<>(!+NUmjfzcKg=r;Y`N_EG@;I_A_QCLWO`X`? z!UHW!bPiQjOSGU1rDq8=frhV;g-Fz~KxKX!u2GWk5-_$QM46AQnB8FB-jqM^)Q$B` zE4(>xgMQ*I_wu`YMqDj&qlEA-9{mIDq5>;9W+6Wh**cMW#T@~&iaE>!V=7t;h z>j87WzS=|mFS@)sF#}{2aoq%#6*ATnFP}5MPIx2)>&319>L|daBb}a17jXlM+2^{; zC-MnUvy`b)4SbnAu@YJ^SPhzh3!~}1mm;4O-1Yc(YJ8CN#B^&vem@}Nf*oaLz-8+D z`4Q|(6I+f}<2L6s@1a-=wL$56+`Z#>mRPL2^+&pOv^Pz-I1j#Q(xFrG%xC}qI*I@J ziWrJ@=Jyc#^y!Zj6~)*D(K9$PqvX=u-L<3_tN?k44O=^Ih}RM89ex07<g1|zPW1%G+Kh-(H12C97YkT$B%p8XM~Ts~xE z1Fpqf8HM2tcXTAlHpPTe!B8rrl{q+4)eN}{tu&pMpKiT2E_`l(=?gcFk)!}1hFApatnaj_ z)HcO~$y2uT^*@!lVI;#uy{W?ScJ}$t<@Fz_2a}m~@?pQM@-;8-7EOz<2jmCML?;S` z`FW8<9BAIx@-G>}Wi$-)!i1P%jdDFT2j(S?!D|^F--|fV9`nHiGIs1$c0h!2xjKr` z_!S(y3G&pRLR5{Hr9E(Nau);n-_-;n-F|_XG{An4Tg?hg=P*)RW1Vaz>9+H2LLsp{ zrfk;jb!&@A^T4~CWScFS$?1aAPOj$k>LK?3zTs6RZ(q*V!W@H3bDss~2x7&*X5i=V zN>_R>OOs#7yr)28laHh3#Wq;Oy!XnnxtFC)?u4bgCXLH3LD70b7-d=DEj(6c1Qa7p z+pHh!gdksFk)feqew*AW!2F`Km=4o){-HSX%2M3H0aBE3LNCA!NK+FuI;4jZM-q`x z%>nL_IsmodK6+7a0p<_}7%s`7k!7USL@$B6&d=pV=9d~7oTRkAZ%(%VbbMBDfb;9D zXnmHuLNf*|&0ePYDywBZdzJO;xlf)@JHDf&b-9lX(&oVm_bmAm>%$K`z`XW}JUk3n z^^|hJ_?7i3v0j-U>6ITz^3e@bibGxW_q_J3pj}Ob`+1>4$Qgb9+~f_HYY}ch|7CTD zTmAERsg8)HkM8{4cn4ZyDl3wOHf?Rx*7zq%<&~7zH{UR`J@;_ zuoI=<6c|Sr%I4uXGp2vjYUpSjHr|8^RRVM`L;bfX(iFZG62Y?EQ3X zH`b8xBQ61`r7-#Jv&=>R3#=kF5hd5_E2t;7ZNbehRRD=4C>Uh5I(Q@?@DG06zT#`x zhYq4|-T>{Ga$Gv`Czzi03kp&Y1^O+}z;%B+Q)GeSJLxY~>wJkNzo4oJodT{{R)o z>-+Sk986Axe>u{NPv5AOxz)ZunFH9c?#az%p{coqTp zKY%qKhO6w2SfdH$p5}4&nA|+beShgWnrUcIN2An!T)t@tycxG%m8(VQ{sLY`cN4v0 z>@j6E-*uMAf!K_aUTq(L$s|xK^nL^Sko6atR6flA)wccDt9W(s>tj-feEk1F6^nOL zFfeJ7>5HYLeOEll0?Jn`RHMPDk{fZtI`(69p8P|`GFvu?5aE;6<5m2yw(Q(Pc(5~~c z4%|Cb5leh8g$LZh&Yu>fkLd-TIk@$-qtPWt=)i zBTE=3N(yjh*VGF}oB+hLDg6SdOqVYfU}b)_CF5%I zbvh!?S$8IV=XC0r4^Wwx$0%9;D{1`WE9oX*^ZeKH@-YB&&SU-Pxl1!Q zGbJd_hHWV@7nSPRLV5f5OMlU|K2`U-F+yyz2~AWl(fT5B$CI+jsEg1p+-}t_I_~Rh z#~{i;vwuk9|24?FS&35p@;SV0T2I`>Zinh*?7~>z;M!RDZq+C31!_0~T;a0Q0tcUW=`!xarZLKdyZ9NjuN1 z2DiE8h;JB>w@!-2+QR8C*4#xQH-AgR@dmmvj5F?w&s3O=1xoq9LqCE4O`Hno(tfz_ zr7?3`_`ERb`s`Sx&1!H}=C!89Mj+OLsQmZsIzkTpueVssgG2 z>&b16-=D>UZm^!;4Z*no;adHFe#@x<1N##|ylW%}z%F%E&#NT^=&AWj_w2`l>Q4ol zU`tDnlddO700QXE9Ky zJP>05S`PwnB_33QXox?V-;ALCpgH;euczeZyfAzcH{W>W@q5BOHam3C1dR6ni;$M$M+OMsawh3U1Y8Mx1Rz2ucdAJ%FgfOs zdg2-T;0a}Jhg@Xxrvrd-7pi!M0MHkJ@o@#z;At{UiX1WrC@uhmiTGm8x01X$bbf`D z0-lHUKr=2`{1xd0Ik-po4?rMdHmvFC>HVE#$n6fR+5Eh+aIv&BNo}60ADOoicV^ts zkKmNm&G>=nxo!V_J4g6 zy{Zg=G{**qzWI|3f~UMSVV|kB_(pydD4X^CWSi}L^m?Yo>H)C`!g>JrfZ)3<4G*XN zVdpc^RW_gqBH?Z|R#g0~FJdqK?Hu!G5bzSrMF?5X9`BCm0Alv4+psKZZ-i}p>0>c~AK@l#W5w0e??l!}X#QbE5S+!7rFvj|>S)Y&&{3LXPh+y?+ z+y#_*W+>|S{{jq)cajei1-fEJC9a#TZ;*#jW%Kb*SuyePYyht!1~8vLw{_8q8K`~? z@RMZW)cn1$C1nhPer5FPC`+c`B@Bp@L-_eX_W#$~4@f4!vpFipvup3uhucf&9m+#;4EtQ6{^e*nDg!>0?V zy_I7ki>5PHA=<}q9e(P`^3~nrc&b+12urK6;v}f(EC(=Dsxhz}mQQ=~WT0+mYQ46v zFdkMLV&;QAm?WOxW1RJ@On1VC50~(-|Mb7rLD;0jx58%~8YbsmuLH!~6Co1U!37dT z*Iw&(8E4sA_K4rs2k+~@-+y?ub3>OJGxy#w+JVyph+qK5SlqO0@X)Ep7>l);fmQq6 zP4e5$Gx@iwUR+umm^5@dF2U@f@$#db(sUxa!D<6!=@+f-#~ZL3jpH3L%h`!??OFEC zjg6kvhVb>3cBnkBQtva$qxhjs_81W_qStQ9k^RXB@%QOxTBp9n<+jJHmnad-zrQ74 zH@LISq14!lo473k6>V%W9&v`gtj*O@`f>$r46&^Ph_g*f1@g?+Gw~wl_-v3c;!7EL zX&7!aiig>A=deNc$eK@2*7G5odwJYyZYnq-A2H9U}#?#{QSn-K^V+V9l5W zz<}b^{;<@Fc!Kf(pS$T}@g0hiq|f5XW*hnmmHP6jL*vWN=0M*njj1v^4}yxh-tB? z4>O}273`Il7bU}NuG3~6nBk%GSAfn-$V8;N!3AIIL%(pRvChC8+PQmu=E%>o2S%snM4jF+a$^BOA2GK z%`Gbq>wN7x08FaP&2xWuQC;-h`Ab^%+A5XN)X@wTFJeBhX0BYD|Bo@6HKp^zd@$SS zq$2XcF2_q5TI2Md#^pKYo#`@1I%&J>pnE6*C$Oqy7mc!U_ePnJtnF>!=y&7Vbbc2XSb`pC#po{w3z-&QV^Ya} zEs#*su=M0F=p8~k?=3i^d5skcgo{~a?sW3mzPxDJU2JL^FG+cDIG#V0jKOV=XtjM8 zMCdl!Mh{XQe=EZF3tCyfBjHyl%^;8Cw#%MMs2GB*fXy@0KI{0qHmaW^4xm5&>}<;H zs=RgNLe$7uR2tH0#yGb)rebE+qaB_wvkOqq{v}0Og1Li7P+&n=@+yM3P#P!&G_u2c zivFtoh$8|0s_gR&i|gaT0nY+Mu{^zG5(oR_b4n2I@KU4}Te5uLM#Lq5qP0O4zqg6M zjO4Nr~-+*Qg}zlRC2Vjp*h-7%5G?^frU(U zZ`Zxig;c4Jt)>ELok_1is96IZ(%?vi0?B$yV-5?KqI6!FM*gwamyJP5PbfGHo(ZYo z#iC9*s2!gVn{^h~0JKr3B0Z*#O=@5X9eXcFrixpKGu4k=St@3;bjIR3GnB&oB^6WK zhp~GLaXsz_uC+#Wd0mbl#>2D?6^1U1&%t`TRT4Mo7Sx7ZxkW7ZeLaJ7N< zg4wDIE^?fof6ziP2wR_)y#q?!rUb`}7p~Y@v_J+1w^kMNL0)xd zQ+W|mMb2*1>|YchuCMRkPFY0*ZKKA1_xqOW8bQiQ z+3NQnaZ-^V?}}L#FpRKL<-Gs|`$-1`mW{MhK9dX<^p+(1ev!EJ9`ZVv1Q;m3D5f%m z1iSKce333d+v?mMQw48-vam8|3PHQRq+s$2R;TRQjClUB;~sqYo4&J}SVqe9*_r)l zeM>c;pRlrc-|tnXwXYt5fJ;z}NnUvv6ETfmmSOIs)oMAxCkcF)+f?x^C@lmGsKz;JXvKyehkFq0yx}V~QfYJ(jeS;E&&}kRE%3NjTk9qz zV&d1E|D6)Ed>0bxDWC2DwA?8d&l-9(A)g$JY`md%U-)4=C7O#&&E} z#T8;j+Vyy517Rcv3J3WhnSy$N6_l%FQocIrd!MOZ5~=<+#`b4Gv+%(K48t=3T{qXg z$Cv@*TLeCsrC{Y9Jv>Dav8i^(OGM9J961Yerj%mPvwkR9aAlU4#-9(l4btTa=;q|c zk+0_u=XI1GQY*JnI|pPtB*E^Mt1jCoq8H=J&=n1Ko#Qftaif`^k<`?jn*L+Evo$m_ zAbH(mS^?DV=x{N@o%TPp0Gef`aEdHB!*O+OUr9W>>AY>LqUwdt*-#jnOV_YZ))U3a zMw}(nM#6vxU&){=k@XIC8tQ0bFKaFLX^~NRtw+TA)YT!cuEuk$BV;L#?#k-wc8w#E ztEWVrOy=MD;m5lJ( z0X`Q5xIE&i%JA!7$k%>Q1ZMmtwQ=?s{8XiL<;r%<_j z;XU5M)qfuD^TNKV7Aa;+JJ?^H6nH42?O+(NIjMXb{H)oFq zPsuUE{9~~}iFx-`by>@CT1M2uo&|_?{G>;DT8N=p{4p zPqqHKCk3UClootHcv(SJ=V`t4zVXj$)-{=Eo=7AvtM#Ji?wQ;hXj-t|_%g5&w5YU9 zI*FkM#>GRi#+L$O>di>qR;DAU~Vao+U(Mi{s+#`J2z6}f|?{NoimdwTdiu!Q}f9KZ4E zryf0aSpkV(7~ps)BZF}MO!Kd7>US3(MgM~mb{VmwcFnv{u8qt@2&pTgA_AAC&T;J2 zXLeqE(o${9~{`5nAYQ91ypy%i}*Bnmq zr&b-w(Zfr!tAtomgwPoXP0?^rR$Fbk()>y_wWr1mAVxayQu-+m@`@Uq=ywN13%Jwh zIvaE{Ax!}Cmp_k`1iygKX@>l$6#0hR(sIOQ%u1s9AlF6n;)s;E|-flcB zgEnt4HAYMXfhWA#)A3p7-zqslZb`a z+8kY(1%HzA#yI~_f^N-xvnel|q*)F)&qyMXcL?Utt$fS#adfL|+GsK7aMP&bgdfV` zYOc{-gRlwC)K0epyFc#f^gvy zaz+akD>Cutl~+H6>RubIf^1#S)L2C70vX1V1nt5+7aX{pp+`0MD!C*cFJv$BYnJ4` zz80PLzv`c;^<+D2w9L%EhKyVuiHH|a;Ztm%F3553h~5Y}tXiV9l6)2SoDLsG^uIbYoa$^!8LH%`QRo1!MOXCGuD3pLbhX8mOaZma6-QCK z=qg8(bOQX{jxu?^*c}dj^_N=@n`O=c@9Ib)s z!t7lFe&sSg<%plyvM7zm3tYeR?>#@?O=^XHtP9dzuo7jKlmM{83~ejk&FIvBQ!`*# zp0Ph0JVx_OW~2H2FWtgX=v0qz=o!ZL=Q0^WkaK5@BTe)Co)xVHFk6@h@A-2p$(9jQ zQK7%4bTWp3s#{f2urs-9RsPmA*pYbzuDvTi|975>tX#8>d^^!5lwg@Y;xP)uo)Mf* zeyLHg&^Dg((DRO*!|umBtcUrX;x1zf;S0^vu(QNeKo>p9H>B-Wc-h9!6qvOEy{GzI z!?$fg)2VPzQR>nxnE;iUT6@E_mIWG78NS2W!i>*u_E#v55YI=K8nzYLhMrX~HPqyo zj5}K+U}6IN!`fgz%@O9eD##sHAhN+jdR#Sw|m$^ z^+bxUat8}WKUq4NWNQu}OwvFTsN!i|*iU@rb3V@LPS3)WA!|E6k85j!n zzqfH7__Xi}^Qs^5NRYwMC9JCYp*%LOftcUbY2Q>BfWAPg5DaV6pLm zGGDBxr=m<}?Y95)KF`V?il^7l^1v5W+p?|_F|9&X^8%F%!R7PCk0c%(-VMbl6+Oq@s-ah1@*x3n^~ZBh!E^Bo+KLk5=2D)E*b4XDdoz znX;<_S%$OV%X@ithD(Si6Wmo_N^7yvUkcMdn>Ci0AJbPGjwgVt8h6LVT4OPi^Hl!w zlk4#1uZ#4WK*#`593m~$3zdRK)t{H?X_wDznnj7)DJKcOxU1#bDkhC##*EFWlY}_! z&uVYr%!jOo?RrHCfBT-wXW@BjK&#DRpYAV82fLWgMFtEe!t=(I(+o?h#NZ-6>HZ?* z#+&pZdTK4&+@UiJ_J6)CAhD0;(2Ul8X^p}F+_n>>y z;k6IEIi`^>-uPP?h-dihjX|;(JE~h?d(m0fI;z10?g;A?{NT50z6GKgjnBz%75fql z9Q2V+O79di{~9vmuUZfI3_JK{ea8X4h9`si{Fz2GDkS(U()d_v4I87OBX)1~W2FY$F9RZdG>wnS6X zq}%z>-EZ-%8fB2p+@oMm`7;y@p{Vx$%TDzig{yvXha44q+~o(>h~FGVtwejQD+<0`2-XGcfGpqM&9XbQ3%r2a=ZHY=pVZM4CggUx*+o z35AUVI0viK20dSjz?lVJpXV;crPsQ=Q~7*SdpI`yF1CNP3Br`GC#hHqOjBGML^Nf22;dMd!ioIU)Hrq@-7CssCMkUGC zL8h^=n0i2MiTDeoM2Yg^tDMMJy}p^3K1UId_F*l2EY(f;q7!1H45TZ%*{POtF9gIydl|Un`oglkB7+5H-Y42D zl8JM<%mik_BPT#cIaIwLm+*|36$1O7=kCeZrcdR zv@J8{1t%|+Hw{l{%dVqAnW|36(w0P5;qs{^2{t6G^#g_p)9v0)PP;u!OFjDLw!#ck z0}8B8LnL|;CpZSTW{UejeRlktA;HatIx)U!u)uo~`vGO$QaGQKV=dT*^c*BX#LCU!mo zY2VklK3_?^W-H8gyQrOO6?|JZ)|gBLXqQ6d^2EV0g7I_YG&fy3t@Jr6LWg~p^f?!CF?!GA1^VRg+#hAw7eqk;=Nd%O3=q@d@A^6z9is2WK4 z^#k=LNem%MEK6-e!&w-Gv<`M=#k5-Xw(k?_cyLVNj{RO9bldR>%xN!trv`(?dUl9T zIy77>7n*TfIXig_A;x$!Gx~)c%SotSMy68woIZn$pSM$<%~>z*>FK8pNJt5x4)#R* ztUG*q#dI?L2qecnolX)=`^<8U?`SL6-X?(&$75OXGzWv5YK-t*4_Y0@|M(1Lo3*6)`AaWeFr$W1^qX*d7?^MjX9A!?P znFTu29$70amY+qctBJoekV%T&=ualm8aXOB|BVQzzt z)r)hV+44IMvDNhW8>T>H#RA4lZgZ+(i#vY_bfWW?wiQ)XxTmk}3PMd8+zQ#S@wq5{ zea5=w)d8~Ghl6C}>x>R{Tx4!Pfl*nty1UxFV354c3Fe?~kqu^NRzL;>Gye`B{04_m zlAWE#4yE86nB#WjZk=n&AZL2`h#7U!=Tx6+JN_n*bS-SO0i~+b{R@`P<|-%@)YVFPkVM;JG)2d&w99rrG(i;mx!7L6VZHV*^k(p zX|1X=-)@S?P0?^5L5pUsL_vWpELHiC?e-6A{r z(TYajQ*A|kJ<4(3xX@`YxjOmcvVr4AgTC*gq7zI|#ktJ&jM*kQm`P1-^K*KzrnqGB zG7r}@jtDJ!Fan$i#B&&yP_|535IR$R^Ai{gj|kq9HE$tM=H ztrpYQpNZ}rvCOm5Jhb)bCd^=2N)P7i2;al+Za-FrsLIdKu=xhwvY1(hlU)>4t|CrsCPGuP+VCk))oF({Nqv{ln3G31 zbi#G$b>m6C^AM>-yY`7tFHQ0A^ZD0ZZp9#7mv(X()Qrlbn=m8W#C3F56I4btEGk6O2ns)<{!x>MJcHq0@crdO z`xlrsllR_zDxS7fVQ6I%zeO+zj(T{p3&)ew$DPayz)QALfByCN>d)t6!6 zm#I3<8h+E~w!ge7Qe0Lyl4^_$nz(@s?eiYAWj$_dPCeX>+9V0g21wq-4ZD7Knu&a1R{QDj%d_#J4H_9} z6Jf^V+Jg>qoX^a^Bvm%iZ#2b-zfTkmVZu3D=}I@i@VT*tQ3mKI_$Ne)lr zH6M56UisQ?Ubf?_Ycf=)V72=alZxrFh49H_P$S3o;D@{Mr)!9ma%~X=23&HO>!9x| zGbh{{v3y))fJ~}^b!-IQBJlZ0jI=6(kgpvse^Gw(r*KRVUda7!Q-@0Tcv}f($n#W! zJx+f*%9WL;I#6DS=5C%bvo@J(?s-Aj+w;1^m{Jp+qBDb{l7P91o=KazUytB+khMmp zrOV7Hx&=o@^zfwnnA7o9^scHhS~?Q*+HDpx{o+vcOuBVrqDDXM1pWh3*knd+Eqh`; zUl(`THuM>=ZQ_`vT0D5omyhq|Jt4(rId2uKtmuWhU&n?xTl689!(zHzw{L`FI&j`G zGc&-eYa~cuZ6^9@fUo7uWz=lb0b$jW-mTv>_s{K=}nAJ`2=tc9@+;R|Ph z;j$)1bd+y$y+bt>Se*R0Pf=WOo>i$;-=^jxK!FqT^t@%Xu_`~!bl2$YonxBKLzX7Z+1i{(#RG4S2wyOgre=DRJ`n1cP(vuh+LE zzxRT?SA&yOx#g91!_pbF-0&cgECaRaPgQs?LiOt4+IhN0SBJ$6v4QDxS{pdmJvwDV z!;z3ZO2|0sjW*8HA0i^VXgBm!mX@bCPL*`|GujY;9*P-_E&N>7=MT(Oe+hwXuvqmm zNv-Ijtp|LJfiSYt%t1xk4hovN6Knj7@;zLnpnHD2&0Z zAm5Yxw>mMl^zdRZ)kM}&J6uGft4K`oQwHL@&*(b|Ww<5%mFMP%liJCWSviO#T?Mkd zmXZyBp*tvs<@`(UiZAHC^9kOcvrWZmwbX|6Ift!hI!vXJ3K=FRf_BBL8k|~{*~;(c z3)c67ja3aG**46?B1+}itbDzLxs8pRhgG-nyQaKaX+AI&#yNbw&1y#WUr&nymo1IQ010^eaUKJHu9#NCj_YgCi$=2%%d zxG0R?GPV>BJ~vUI_IZJcdHS)x1_>`^6+q5d;Ed$W&?j0+V=&-C%)(afZ;6=C2KEhJ z^P6TXDBvgCCnCH=P$Eg1+Om(KLyapIpPUa4lUZjIjpW*2IOVm$;IZU5Ck<^>I$dVj z!;NQ`GCHuh%O3_R#e2b0Zzfx?)Z=L7q;wmO4h1!dO^e@V42c)>>HG~BCvNj#xK&~! zl}lu&LzCfREp(Lx%|sBXQuu=gTlWiqmMTM;x)lD^ZWIQ5CxWXMw17z`JhsEqTT*?j zFF59VJ+t@@<(x$XX|%yJhqG*D^{hD43wuX2{2@tq$EhfGmee@|z|_TETt2)XU-HC4 zRU17ay5}mDZs?D3g3#U16-m3b9dDnQ^eaEZ{A=caX7D*C_P)H|DIN70%XrZK{% z>$lPv2tveE#?ZEE=u5f8^1I--WNVHa7qK1I=SRd@5z;h9@$ahXSO|$}j<@F$e$Xak zUQTp9Rym4qc?Fc<4&htlSV6TN?=l!cc$rMU;^)m`XwNoKSI})pH6RM>cdsYiSOpO@ z&g*J)EQTyX78-#Jyc#sfq}8<5ID>-yR4aUq4y6%*F*WV4fD*U8X&blJ79?S3HQ&Q@ zsSagR>hAo068I0Zib9(P`plT)&5w(sbq&8XyF^CKw{P0R<>toVfwz;nA@n|c75LiS z?V*u;xANiHi*ov$1Mqn-b=-I3n+sHkgqWQude$5#e{Qy2-QlA|)~eu{af`Qo@di)k zZJv)o-)Q-C4JP}B8j2K$a4TqOXEdDK5$Cbse~>R=?wD36Z=oU_%iA*(Q&6qsBU?j$4gFtLH@ zNjQFZ%;awj|8X@~{K(Z~uKu{EwoI7-C+<|m9#f=1TlW)3Th5y=4E%cj;U41J_qZ|Mh2MeF6?ax0zSdCv-$ePj=?6>U~8ME z(x*}E#X)%Rx@y6ye*?Fjh0Q+GAWdbNzP|R!p{dk;U{v84BtI9+-OHR!51RhAwd3yp zna}G4_K+?=SoGsk*7`=Z;P2ak?`0e$XU9+c!@%b`?NmhxB`q|;5{tKccHP;R6shj; z*veL;>Ww@u%L1pP#A-(SQ$%>*r!Sz=mxyLF&$T&(K8N_z&(64(=ed0mW1ErK#XTXWN{9!k1 zw3`o6YOs0kiD&>v95o8c)Kd5~^(w9>PLwx4Fs*gm3b-`8I_xWl-wR;jG0IRd`|fx< z8fM@FjvexCr5pR7=ZPm6NG=gs@+p6PW39qBSpK6|-((=uhRa8H%v?=GE$NQt>{f1M zT$PGgo|XYbV|$^B1Jyq)35s9FVm6e1QCPRk7bsuUitu@h19T3nIye~BMA@`~9c7)$x3EmIj5Yj0`zeT{^f zOF@^b(!faX0iPbx2W5kgM3&2p15Dd_i}fezZ&bZ>yz)$jEm?TqI~sD2pEK4b#TV0}vU&5`jzM9DNC)rP`?SIgvnf5($8Ier5~H^WD5G=J%30Lks+>zV z!!*cH$1yv@kdBZ-P?{W$Law^ol(DcoQ3U3ea!vD(NX@22L!p-jO|FE&^70lcv^g;xxr>Rps-orjcvb61cu|7D}iNo}GozIJ`_%PPl{f zr7?T2)pkKB{#9@M$xbv&5`~%0Ol#|z?(nd|E_Fn0PL`7X8}LO~N%T7Mhb`-TY%P2- zZ?3@wnFHI?x$2h}n(KP=2B}UFqrJUH{SyDt82DwB*ikb=*#b$Wn9yUvU>SG)Kk=-? zpRmd{oeE{8bN@ZU){XJK>dqKY4%rZPgUpoF6LKnv9n7JZUaYq+Ht{_4f&LDN3Xytu zdSu%Qc{i?;NsnV|3W(qSsCv&-#1(vBc*b1)^;;goY(03S801N=`UT-xeWX~;$h^6% zqe*s&y3e0678_oX0m9Yg-W<19C?kLC<8t&=iXg;)%OQYv@T)er^d~RLx?#7}JiYOr z+TS~D5}7}(Qem|CwJ3Ho{)P$_JrGQtq zB-YxOELh+q{Lx{FBY#a5eNoxjC&~GJ!F|qRsyw!A58OI%rifYCz-#?uXNh7@++v>h z{kp)X9>Rj|cc-$L6Ta5$Guy>VzA$qdv!2!Iy$E;;a4ifQ-A6h$cFz-&n40(W#WFvX zT|yak0Cjrx`q$GD$;GIqpP;U6hE~%g7ask|UP>d7o5G?d^_fCPbP=;S<1_?*JRkkO zSO~g*%HXRs23DaSWzp$5hN?E*@z8!&ir^Xb)^)eMa6k zfp%2puGZ`~6nN6Le@0v6F(y5F(ILZHE5x8S9L!&!Ed2PfS1>_^rOBQ9TrcEKO5QM) zf2MazE4Efi)3^Ubf!JH9$AwD<~EJ+Gd3K;cF^28aaXVr(fV!XtBYV|Ocj^Y#Y3T7 zsk$XzYR8e`3*-?wf^#;g%M*U>%Zqzt}5yB4w~$aodDQ zWIxX1qGyw2x2Z9qiT>X9m-EW5bU4Il_yVZFCYc$O7L{ms8k{>@DO_opbJVLXTj-nD z%Sp>pc24l)#;!^Z3r(g<1*^AkjDVVG^u-F@vu@Qnr*8RqQ_$2qY zt<`@>M!vnPp{aHK%4lL-V#EEzB=1u{n@N$HMF8MYA)Tm9Yxgq@f2?RsEx%-^bUbzt zT{uF&75~B*wUiSuaY=aV@|)3@{&l`SDXH|IM&I_=D@avU+y;frtcU+B9rI6@fbIW3 z&b|UHs;zq;5EN;pMAD?CyG6RYhLVp#UVH7e-t{hUZ}e-N2{d9`i`t`BwRA5Dp?IpH_g$)qN68cDi{*5B zi`S>0a(~l)U%z-OMU3WI7lIG$y}XQ{)q!&xJiF*CexD#UNs9Z5e+X2B=W zgvnv&ioq$7)5}&?)5=Y~lh4Q1!C@jSA!Jl2%XPtG$lK63)^~fspnNxmfMGpyKFc=E zUy1bFV|<(8E^fY=F^!yRL9ESrgH_?tfH{4a^yTVP;b_-1%Q0!j^-U@onu;cIHPfIr zr$Zy=CEJ59l?fgKaTy^dD}O4f_zt)@ltC zKr#m9p6}$b#sO}~ig)fzxQJs9u7()kgj^gw8LQ!F4+!ctOQ1sH$ zn(9Q4iN}iaTohrvDbKpAsL<9;>>>G(i8Is~FQ?G@J4~C+nL56=b5Kk9x$5>*ZQYYS zWkbVwJwQm@7I?R(kEZRS_Z56PkNOLw3K}(5F5*m?CMa!shl3N$qO=*FQ7Ahhmi zlD){yKN(-U=$FrM?4%@w-@fFUFzC9@kkyx+ucTg);n-4m+ETD){{%rX6Ke`aCQc0( zCCjPS2kE=7;MGDin~S>6S6`Q?SR=0Ph=d)@JwLP0rbm*C*5Ey&=qNs5Iwvan>I^iT zf2`U2wC$Qph8OwXxS5yNBR%Zw{F@P`XNIPNP}4&|g;;b1N$oBSqc(O&4(M8B?V*$$KI|Nf&jFC3jmn*2 zk>qm)gToZ5TtxP>&}$Lt(50xQl0nu{$;`Dom{#_XJ@n~Zgl`eZ59fBJMOT@2{Kr$;ul-2p`o!rR*5J&_k z4rqm4(YEoc?QjmmW@T^-s6@pj?(O1(l72e^iW9-nNw>=frF3aUZeycH-HYsNGt9%V zQ{#I398)IJLOnrTLt7T?>#sIva%xa}Rqok5rGOD6bk`9g#^07KHm1uBx`T-n)Q`EA zzpp8Ez9HoCsIzgw4^DP^7pK}a)cY0BmZ@|s+%q82(24Xmlc+w)4(CZ8%gAiOIY0zt zfG<5jX{fY6(vg#3DC7R+TS$OZ=iI^*Kq8L5&IZ5GV=VOhhOX@5Fui(C;FASk*l18X3G)xbA5?GRNxcMND|dt~qSW$9un{CFy27_R&iLazPY;$N*7q5DT`!+rR0$j#>5)xnF-RM! zRr?%ZQlFVrF4QG=J=~~E_>>k{*CyZ(M+wrWBb+-RXKL}!%J+PZ)t=?#nXiU$OeQQZ z5OnG%Rp(OyzP$v1_aN-bl>mKIwxLhYG=IhX=`_ty{bOgL6S zyk6`iAcK5am*WaT;cSS9>6p~BJW}2g8H%kV=pxUDEuYfe2xC<~Y*EYY$H^n#X-)Sk zX~W)@x~lISVy%jl^Ft<{=mc+M8R#LhI-I>LOc9iNUDM0IrD+gT)PilQrWY|oYWs#v z&Wq9&!;d|B`g>*XtF#6sCnS0qAu+~8fvwUK{@lt1K~T;5s{semip>1Mp&V;b0jUxL z)zSikZjK2~eYwgm9y3U>G&9i2)CjD5`QBCqj`69#)B5<9|ELma1mF&grKpX9--G(7$9K13Y2T2U$?PsTVMh2r10ZQU`@a})e8v zQy<9f-lx*z2PLB|ty&sg`k)i)0#$ks^*j2{2hdZ8>V^8kOw>udq1Y&lm%~O)`hc$q ziO$z{b9%$i=YWKGwh`y=Rn0Vaf5d6iye%PLV|{(NR!oC*(yE8uOjO0;_V`D9sVQyg z(6m+>B%CN#ZSmc@rM|IJH;GnLO%E?53i-svJIs*`st0uri+5ArxJczt>mkVm1*QiA znWXABg*-^}%HN?~Wu8w;1IqK}V4>Zru1`W*1IwBMOpW94M+V9g6W__>aV!R?Y(QR^ z);eW}S1lH7NbOA40Ts(IV}@GqamX+uA+Z||{fH8ZWys#X1XDcp6rYo*Mw>B^jWS=P zw6>}9P6$brC=^- z_|&P9b_Oc-YXOtQ2q(__;G2nG(kzPDWp}=cAw#_UzNL8ga+~O!!(V;=A}PO`G+z1q z>AU#(YvK&Xe0R4?2?MdLb>HJNVVx$&U`PY_bgL(ZgQZw;!?9*epTaP|xcl~l02}? zOXqgBooU)nl3;Gzx(ji*$ueqRHre(Qx!J;a#D9)(c!3?qfa+F!*$A7WtuQPzLNIWQ z6g`MJE2NSFB$njfk|zogzFjBdIsPbzuNQ&@9<;1ZMuC3_exy2Vl-Uh`hLT-BDi>w$ zMU#BYh+CNCGzVmCKQNCZ&uvT&7m6QwsT&(&T{dPj%*(?cKaUU*>KZ9phckG=zEf&} zjR1Bjbg~b!%@~qd)7+_UAlqFJLIn@4e3?@mf2;fWMPhdYT5rh;6??2q_HwYqJ9F%( z6@3~kIL$W2b2V$5nw3^u&u#k78|~tAxnG~m1+vC*|LA<$#Dqr?3$Dxd-HPdbLUR15)t+6vL z88q5}dp6RMsi`%hGFWDup+)AAea7xRf;!@2J`8AP{)5^0y!Wg8`C+u8c1g*R}9rEfGTUprn-CAPg7 zR=}EGv_~Z8Z>%yH-7XzvgIZi(s7krzm~Lg1%Z;%qSa`#!ChK))E})|2OUF!_pEIVC z6+(xK%2@HN|q zJtj$}6vF%huowgLNIg19-SF*k(_J=*)bq`(HC#F7mgMh zaifmUQcD+FI7RkFT?!t z4CTp!d8wC6acQ%D%&u>T^sdi_{KjmVb^ixd6$Lz-U6djn&m4l83VV#gL?gYT2PVEY z((Kj^v5aQog}sd_kh2ne&b-Dus9s3{=*RC0jOFj`AW6Q*v0c=#II1PV7utPiQ`#JS z%uciJE)Ucy!LfhFE(^@ zUp&!5#}L-k%IO(i_VbDRg9+*SSaJl9v^7V3NHXcmTr)rGa>vu>ZgG3OIY_*wyx^9prbmgv*j<9obQE#Br(dxX}>Xp$Gh;-^4R^2#bz{(8X0`MXAlf9*KiW{%I zctK7*huJ7F9mOz=W86hE;m>w^NA11l{q94m;`B-;E9Wy?S&>jx6nBtTZ?J! zbJ+3>B)oAjBSdxb;+=uJJE6SE75+Ci!wKyEIa-r^6^_k~ro<)9D6Ofmu!%dut0#Tw zr9J|Rw>9*TWyp=y1~c7wI>Gfrid~oh9Ai0-;poACB5^Ptiq8_0ZatS`oi!9D5_P>B zI*B$@clcnVm=4=+Vu3Ha?SxUVg2ffdFh{T^9qi)SgNv9T9SsKY|gIhzR@ zaXt!tmjYAIi*G%Op;<2{PC^NbGBf)x#$j-L)H`DKV>a#p1l{=wkBpt&Gz? zSf}gD@iN1&_H_6>HocZsKd(pg%lgmw!tLw7ij%h&31AqD_8?NBS;AHt?qPc$_uB8&*B91V9RnsEDN6oz$hXHx+et^7Yob`59XpjkYZ zy*v)>YVB+;);-l)QZqRXQ^=ArTMc>-ujHQSjl8<%Q=aw=%zMfwr=T#>9OyH1!`M_K zcx^2N?y88YjTx;N0@LU56pm9o3O@a7@AJmTZBjGM0_agJWBC<4Bd|D@H8r~)dQF=M z`;v6eV+;hZi zX)qEO@6QKXhGQ=GM8P?=N5#+zPFa!nzACWwwv?p3j^5c@yEF*{FHvUGwK*M>5r3e#O58JRBkFl}^K z)bH!;&q9db`R31m=;C&*ci5>r_DZ}<-Oq`jPWC>z8cY6;oP3l;{G_sB32UZ?4tmHy zfw}omiGjU!oBDB#?FLH}UE8NpLXVLZ^Ggq_5}h`ubWnVkbG?es49~_Relej zE8W%+U~YeTEC|%gY8s!egg*JjYk z3;Tv`fWsn1nRbThR(#luD)^i}k60zH;z|WYVO^9;jY5)jp6XGa%4iRzSG3%^yWJpS zwmvXndHe>lb+0@u?8B4)v)P;lwzS6#uEE^8%V9=Go?B<+h`mj_Yfi1J&zY8%YaN&L5~`9(Gk z2T0?Y(WzV*^Qv9b!f0y0HRM@4PKJhbukGuPsQ7x{w0{5g8E>q6vC{Apx|bli?k+yS zI$O=BlDWTBA(KyT!j4{ndsf;_(`hEdCI7xh%krtech}n_5@enh@A3sW8Ypuy*G!EP z`hFS>)XSk|TdKG{x5byT2WU^)y{RI{Fj)&n|AN(?PjEw5wMwWWKPvyF%P;2f@9Jax zEmiFgE&!lcNq~{=gNIN13IMJrt9keu7^E#xM#SbmjRj9qsQOWf^2xqeITQsKohoO< zuGF=Bma_dNPd=5~8CyxOey+tbuY&_^)+tRLv^pkk_iO;U53RLu-{5FtXlm8>Z$9VW zMDjFNd6(nu)8%9(|>18G;@fLz>_)3)oZUBeQ{7Kj_wM`rv(!AVtG2p4A+lU#o*~GeVAcqzp3( z3;Thlnrzy$<4b}M&MRdBca(d)ZVK(#vN{I4zFv<(J=gL3@~(?Ird$0vQHd?2##TY3 zD%UzA0K6y0{kfP3`(`cvi(lZu3bce@l;5)^g2iF8((pwmlsA@|1(oEF)>h(O&^}Tn zeKF)cDhDH|-nBHDZ#>&`dG1tVIRt)3F;i_vH+KPMy1Sojj`hlNv2rQ7gu#`kJNj^K zigM`?ijAekcM}c$-}Pht3D95LWmt8a9bTDnsj5r2JwTTab8WMp*gaVC+-3u)n&{e; zCh~1rr>Q@vRypg}S%6f200=RodUi~hR$`fG%q#`>TLbY!WLzlzlHdQ2oA!7AbK*Wp zh+Rxnb}ZrjpC;7WCto0@r<~B2>2}HI~+1dxnD3N!zAxK~1;1nKWw$dnDhg7wrj@ z7>{kgbBL-Zr=(<%BmLB8!?gapr{$%@0#E~#lckReld2~-u%8dH^C&6W<$9j+T04+646Ilo?0mfgy7R}S$?@5cNA^StEqQ>nQo@pVQswrCp z>B{?K0S(h5OaIWl{;w-y{=Iv#+;~Sc6}E29{AA`89ldFYepHGHwvrINa%#2A;b4LP4gxF=P-tg>^TM(H6@xfR!4sI4t+VDQ zEgz|Xxdn_)wnSOp{R+<9abZlrt&3b-T<^+tyX4jfABd1w$OHbzXnoluJuZb4r~$Y3 zWJFk+0*k8A+{HP#;R13zFj0Gh+xyoU`C*w7!-0iUCLC?&14jXJf2ZM%Y+P2Bc4P6cueQ|P>ePAB7A+9x2qxRrliPJHW^}?wEv*12o%&oIyCLy50 zso>~%b4ID)Q2D9@#NUn2FK-hM@0J?5_{KU3Z5o94OOtenPSf)=K7PoXw27U6GG(#< zaH9;E^=Zqob8x)3SdTmG`9oudQvY8J!LR!dZwSEDJ)k?>!IIjDh>DQB=!0Zg=JgS0 zmYyf^r7g(E&k)AjlX9BAj1aZnV#@$h2A${I_no>hBXZPrpmlCby}MLgcX#q%v_|7i z>_7leAsrn{EN>5S^Jy>N_LMcf#C@>f`FI*I=$TYFO2l|CpFZOEMje!vf$>=~W;_Wd?Ihb}h6MAH3@7(!G*jI}F&RqEUUNx|8AoDRv;IGP%)~cRbf{sxxi! zFDx@6Nsz)-Y_W5qu$7UZz0b|Lu4)ap0hK|T?EPX-*~S*eW3T(*Q?)@WH$v*6LK2Cv zs^+b(DuX_TnHoE$w6tetip?9LA4sDaaATusX|}J5GGg@X-Ur)s9*((w1+YdCo9xl- z5`#W&>#akcT&4uPeX_4R>+Q5@#zrg1&PY-0FGw_}Hyye=3H>-b|Nz_GcN&e@#Dc zJtAkJuE<+51Ox;@p1WV|C`Y30Q(u<8X?8r4C_YR+oQ8;R)EvJ3(ND7-H2+~)%7+r+ zcy3kVTv(aU#_q6=oetdS>h#W3Pw#Q{0OP~Fxs*#IM)9ysL^7>r1PQVnfl@MtcYaQv zUaWBa8gaS79{|b#VEldXnN^Tq^e1;PxBL7Hi2;1P7=~WA{?+vXIn4kHKfKi<`&+~F zV{RHTw3-BN80g7c-42LcNO6T4IQ-TW<*nOT7ygTzPGGP)SAG`9nQ%AZGX-`B zPf47bzd$nnYX$x1Z4d_>yvn!UDnmc6M~%tkp%XtVF#gvac{PwQbb})@n2r(wi=(G~ zQ}2TR+~5x#T)PYWCSFdgK)_uv{(Z{u=|Bc@GGgeiEn`zo!*UWAJ(DHy+iMVQa_ppmB{AA(y2NnEdpt|b7*K=`RNxo_@phB9)_wT;V;*spkLGmx8(k|f95~h%zy!bPR#i*d5kWPVX&**!c)wrz& z>}=#zCOUq8NLx}x0TDeADDsV>m<-36 za(90nS$FdGl#MCFi{IiP= z|2}X+k447ww*R&g0m`VXmOxE0J-}ei_%SEH5eq?fHA7t`>=b|$kn}IMkOZZ5hK7fJ z$!q0=e@NAz-bCI1Kc({1YP%&f@dt z;E>VKu&mFH6_UU!G0vos7)Nz!{NJvCEe(F>P zvAbK?K4;w({M&(GBFoDYAeCo|BbiQOF@7JNi0#%ZSnNfhK-3ijl+Q$D5bbwstkluN z&#%RUPsGxe)Gg#>jZaNM@#gGP$yazM z`xjb#Bx-=F>WY1~`<*#Y96^>>P(Zkif)r#0+G4&Dc7MSa z;+1fg`eKOP%%yAuv|Qj;GxhNS2la`=!qNZ5>{EXr7-PKYwBTQuz_R zBM^8+62Sf03i2^+%8^WcweYs}cga*6GJD;kJ!RFcu`ea>rL zbBvw+|0{UvF^xsm*)#LH;ChM;0oPU3UpW{^>zD;1ka=amh$%=a~B@vi@ltJ z7~}dg>V3C{Qxy{z{hQB-Kc<;?|0+=8d}Q)7ksb;sM(jA8tJW}pLd2NZTWwxjHOg58 zjbh*ULOor(OjzR3rC4G<=bNt7N2EXRev$41v9nY#?#ao`WyEKxH)XD?`x;tK&D4fM z{D$3rT{%Z18;`@3*%!4u4oHx}LlSkUDUD;^l+`NFChm&Q@@%9X^G<0Efs;6wOGF>uU?c!w2 zqN>4+m4UwKZ3n&e61tnd2IyQ=_XAyox>fM$Sb@4}wY}7Ze?A;fz3mKo_l9|{i1KAF z131e3d_(Cz_?grG{mDHFB2?t^f-4`cK=%zp#5h{&e{RIQWsRlem_4Z>k|?u=tAV%f zpgm1H-EBDshpXk>q8gUrv23nPwJn?Xc!nM$_Ttjn1VGTu?X@*T`ty9RBFXWd73u=Y zvZDjl&!SYDu(R#^T+ky`_InxdO-pS6hAD>Ly4}6;d89o1{Q_5mi|=BZ1q{j6`A2Gt z#eJ@4mEPgbJB!~uX%|&)^*^mwGdikpcHD(l4J{?!nD^{+|vISgdX^9)4)(e|rG#BqrdkQXwsH#``z*A$cBOi$zObShwvq!)~2@S0T z?DF?UN4@|t-zX{2_rQhPiIVh~1n92gexr*U9gFMnBnjCo2xLeTDy2-yUS!axvH!h` z+rVx!QhiS@8Xi_|^3vHh>cj`XfNlO4^<#x(xUu&7mBt0N zXM(IfH#8Uh;SL0eYn>D}fpadQY)d|KmB$x%dJ-<}3bL}Zv&%XomjojVK<5*?2H%QW zyx4uwNr}0?nCtd#;(-(L&*XO*fV^i@@~O4H&%VB70#^4By^_?1G)6lvtPf(xe~v2O z;u<}Hv=YBJ6m0LtxRT%`xJh@>XKsEMY%Cl&B*ifT{Vr9}Hky;kYVzWISYe@fqyCU7 zC5aUjMSJnqdhHo8UCD~37PHB?Za9_8c&2GXZ`q|Q|APl~_Ptd5ZI^Wuh9=x$m~EfR zTKqwVz4i0-DZBx?0;aXl^^K-Qe;&v9)3v|`10*CQ&*Mp7>t6PrSo$5P76;2Yv|-DV zhBig$lkg}=qld?}+i<_4VGf^QC?IYEZ-}OS-AK z86)D(Dg4o?GAH|!pxMGlw|Wvi&Rg_^CKj?8$bl4|ne_^nzn4O1SC`jV&5qh@92yr! z?f7B>w^Pmf5L%kIEteRE3s+~H1=65v07<(hY?^{K9(la$(J3ui6VGeC=(m=Klufym zFB--}v2Y>&nCS_Qjsgz|=-jr+!TSlPy3JN~Yr{n;r!6~XXe)EBZCneb&xtfI#CJ8C z2rdsxh~G6~=gR_U*wdLy=x|-$VRn+_g~d}C-TB6poeV@s^2%oh0kb6f%^`zy4+Kmq@HbStI;5d8anY+NRXWn*Ix^dhBYc3TF{ z&)#)0$|%b}sGU2Vg@CgR6r%$OIU;2tBs0=6YcB;RXn^K>bI$9<{+`E{mud}`C8v;? z<2(M~$pv1ZoB~UUIjF5LNZ+}qE}?25XJ)pgc%-30Gwh!f*Jkm2)EHz&9P9f zHjaH`4)OK|*YkuP86jIl5chZ)NB3<@z&1!U)PHDrAoRW5+>IsHwwsQ>WvvpR;R`^6 zqbfB4deHE>f6XbV!&4+=+8eiEH+$W^8e9a6GgUsa*^+PpwJ-VBv|egyX{NH)x&DpxK>AegNaRr4mo+L-g+z=Nl{yYg z*zthl*ctS=C`f|*uW}5FD%uIxp3E*hx1R5Xw7Ylb+>WXtx3*kO)=7&!=zE91mxJy3 z&N;fBw||&ka=w+#W&Oh9j8k^o`qar0*do4xGj>e^rO@M%<+7JL6C`mNX;xTaR<+`h zLfFR$C1C7jK&_19;ouvpPogxr`FLn)27yFIxqR3+8w10)g}a(TZTY|%MD2Y_*X*R}XuvEe|Kl|w;Y zK|TfxUiGSH8TDaz(cBMjBu{WYG$g~D-wwJy)_@N5m!PEVxrBnl8?}HXyG7F&BIii9 zJ|5N+Ge_4M5S*JkRY(pR?nn=n!!xN^>U1=(idkvM*K8PY!;RB19 zdr$IFIIA^TH-)FIz>oT60SC*F+`#YrIl?QZDWbvIx!`hTJ?}MkOC&}u#>VJK(vK4`#yX{sr~`l44pRgyI@0Vj?LOMd3Wmf z`Qo;2Ah9go7eAAdgRfWc_Vp=48Q8e>yH56#IHvRV6sd=Lro!QPi>5Jp{QM53=5xIO zVnro~F_XYJrBvNveGMDj!(<+{(Bf06Ql`P=+yP)1efHZv65&Z2I4useB`~r;+oW3b zX&Mev78tLfp*IC5cwG>%4iXb=#J(o2Q{iN-yB!g~5K(xhe33H;rOR`af0>wuIQ-4t zh}@~$_4;VhkAzTdyYU7~H)$_oYKPu_R-F*RdEyyvTHKme^M|%Ou=n3A0rs7q=dAB? zQ1|6T)jk*CH}MenLRN*cE7tVID|EhRt>W%c7F4&Uf8^N2K`20qtb`JOHt>qp~da)mpaM2m-HjvSLghk6>zMZ?v^*_YH z$k<*=&V)WkUb4HXGIZY+x{wyag@HSBQ7jbCxCVVSw!{TTtq$OgsLxS~FaNL;Ara~( zW{|G9yKRt;uE9Q=L=)_a^M!^UPwN+Ga8F%#Qxak#G#zHphUi-KlUKDBIV(ET&D9>< z#wL5PuqO-|bK5WmTXNL;cs>bDTRbS?(eLHju(9smeFXj*p}}N7?Vw5fQKVT?aMt;_ z+B`BWM+0enrs0Iox)&iL3L&G1j+DR+L@PPRbIA~FwggJLI8PUD? ztjlvrDX87Zz-2hrZ5t0LhFt0vhM<44#b=IQiFaz6cG+2+e=u`&vDE5jAA;A=DLR@@ zM%h-w%+cEV$vnoF$aT60O(Tr%8)WWOznOFPW@UnbJH&pV%>Y?niOZYnxx(IMcZyoi zs_c6$@DGDM_N+Ll4zm+aXX_lJA9>wLEz64AsY_7gqIeW!ezmtu;Vr-YvA66g^2Y6S zrtSh~+dFNn+wbVWCq+#BW~tWmDS|{gX6Fr>(IuwRUY#z-G~%%cj+^= zq;Wm5RTJui^`DV5_@`|(8zc%ekzw3xsnoCF90im{4qKDpZPd2=wyDEuz{>i32s zqG(*I*jt(mE|S=VANwtNKK3(d^{b?}?GyoTH_W_tbslny@+DS}mrmpNJp%gyHMaQ| zi(%=hsZX_|oERL35v_OP+!)8)j)ltLp1i0#d6Ld9_Ju#cn%qRri3jUl(Qdw454xL` zHSISM<o@>(0v}4Q50h4r`o|B%so^T0W6}*kH9^l9i&ki#A>=`g|H+QP&x;# zuXq`KWo=evY)6sWba@X_K9b~P;XJ>cTzP6+hek->PByPI+HC}e?)TFAq5Z%!-(;75 zWbX#2oi1gsV8=~fHf^8SM*;e!{#^BnCg@s$Jp#X!xwDf$z#n4Z)pJB47(-~rH zd#g(Fr*JZ9ixyNCMJAiJ9P!c*a4834u6O5fPOP^|)#t~!zJI~=LIPajC)f@@&?P`+ z)a6Bt;@V1;<~Cl z#q8$ygvocCw(I!fw$)>f%s>7BrPxWuaCGTskH7YP@7m&gn80Ijn_z|}vY`zv#(62w zjG1_IGR^`eEiDag;@F zlVZfE+91R;WEe^DbGm~daD*I0dIxfI$(ogN=Yw(CTjk7|d`);gGMxavx+CN{V_W*f zKgGZM$>99MISI-kqh}J@?q6k?y}k>XYP|!9C%UjB)I-0|IaHX;uCp{T!sQ79L1RVh zY&PkCSegHUZnO(|5OU1s*eWjt5Su=iBEapw1fnKajxz$!fM3P5oiE$Y=BfVT3Fc@0 z&mZ!ad~iPW0&s}atF%&mf_A{~D|=Mu0g4}S*82Tk?w)RLhisJ!?LWBV(`#!va~TY^ z+=l91YP22%*V(NG{gpbN)-~72*cVmiyYJaihTBofMsxXk^vM->yfOhHb8M31x~z`M z3!uOzxXZxT<@dM(SwsS-ZVcP|uZx}5$=;Uh5R+=6njUtH=FU0S7z1aA+$;a^Xt-$UCX@EJ&v6tuD6_v0w9HZxQuveLy^bI*;4_&|^cp9SpFHqn~yxQ8fxF*0c#>ZHeOAW4fHqPIq zOpF64B4}c|(dYe}Fm*+)z@vt1^GWh9*Q*5s;X~>%L>Gi?sRsk)d|VLwyT0~ zDrNmAHx7}cKN}&GEh*R3_Rb|x6N^}E=$XUUgk7u{%r!w74O540mdwr~LCI->_6I-K zpR4%4Ul~$)eE$4ePVaKcZC8V6dsPF{_5Rl zKX?6XM0mLSDBvh6625&MIAvwKXV@>-?f3yrHJi62|A!dUpZedg0*eZd7_G{r8-3xqBQ+EKX;QgfO zXQU7BdE(qVLqn7a^H;jcW2-mizk$*uSH9=K^e!gD6{~!iD7?l^yKBO?D=kMv=@R@9 zPW}CU&DIV4XvF;8kzD#;W9z#$?ylN`JVOJ0c7G&r=6Qf=g0f_MM7%QFBwYx}QIye} zj`1qoSk*W9tKVXI#l-wOL_2W}tjreiXn&ksBmqktj5*Xz@?#q*7U7LR7|{H3<9tn! zW1toz>KUx8G5wsSr$7gR$K z&bMx8aR3W{FJ;OPyV?E{_B@RP3&}vcZ;k#2)#H5(d!8%AvwzJI@GTu?0sA7x{;x6+ zLwK1a9DWpQc2jM|Q_OjW1YLT9D=J)Y`D!Rk0zjzn z!?d}hBPc)pVjd#u?med;`u&Hk*G=Pn+N>I@!e(jnqk#55?(dJidiCAaui-vidFk$R z@Zu&2cvMj4tjF-mC+3oP_kPSK4$`Y=0L;;Bl=25%)t^Seoi_viWHUofT>fgP&d)Hq zmt<-So*(RKMJn+0Qs`Vi0J{ec&SG~pTT08XOP1FF0k}Yk?C`aNLQ--CVA(uPvj6c_ zw#SV$6KwPYD42hBB7Ufbm8^T8tCiIx9wUISD&&pt(731Nf}vS|3$bu_xs;ylckVxc zR;$M#sZ6FBc3O#tS3U?}u>U<$2xGU3m(j_)?@m-p%ag{(zN86BHz~z`?XsyNBqX0B z3@#E+6x{*Ukfgr#FTAl;Y7Z@VoYSl;beY;~#MD|LtP!dh^mol6t^QP&(5^ zZJ~I=_}f{qdjGja{NpQLCpdVe(u}17Yhdb)x$#i1SBeoPtKDKL0lO9G1t?ckEv9Zp zT24;5H=cv0r-wSO-uCGrRhYtvteo1|n(<_gvYL|l_T02aSZQnytD8*e&iAz9_0Pj& zEMJJX_v!G78??ldaMjIpk^{=0*SsbqXGE=%2;zL`SjU=)0IR&6e{e*$VuPyls%<-MRWh{SFG{3*+4 z*Rrq2X0?Ncx{wvzisiHf1%BF(?WitvS>c|jH*V;t$1t)^Xl;G&#EFGe!8R6kD|&=a z45(=|$nLT>*0l4YJn;M!Q_-^kcE%GAb2d`4rY zeoC>cuP=fw#iu5sR`;xc?}=)l68~pr`^_)p6|?=1ygrXE z{bOd^77n&Z6J@whguLn*?`{9-Xmi2qZ3(avM}(YBDbhx&reIDF5dvsGd$w#QDht4maOtql3|=)g0j zQj}FqGx~t1m&*&?RjYN7V@i8SXz2i!)Z1GMNJz4MNXdNUy~WfH4tx-QX+L#gvBSB` zHF6I7H#Qp*YfhR3Ot7-O&V6-Of}%}K3l`(Im*L?YJ|_@w4zsFp-IA5YfNXf*J~50@ zWrS&mAZHwdrtr&5b4&>X{>u_vzwL*>MSrNO>l1Wp5SlJxL3qsryVb*CP4m-atzyEe zhc7iQ=ydFu-X{W8D(m%awDi~AK_e}Q`PThjn?JRn7zf12hdmPU#-wju9E?aE-NF}v z{5^c&iEjl9m#_6&iZjKC5HTz7(1Khjlq_o0O(33i2*q<7@v$!(TjEO4z5eVpod4UIlN3Kdz}4BVEx zb{YQ5LarjiqUQGZIg6+c;GRWUO94)(!Hde}C6o-nx+ zSfF8ieCXkvz5M?u`^tbQx2|nzDN#}y6+tAWn?X`35$RH?LAnK`L?i_SiJ?2Bqy&Z* z>6$@$=p4Gcz8m8_o}=gWeE%2*?$~>;zSgy%Qd_f1AtGkQ8p}e~lE%@qg4}c|l}6sJ z&`A~)+nwVfOFJ_Kkj&^T2z$lK1Xbb^t;iDu1xHdMBDmNFhCYMjWQ}4I18NTV44^07 zH4=&>4S~CsEFxiGb}l zp0Iw&Y)%;!^}uF$x)c|+P=yfTxu;WAx~H?Ox|(3zT>ePK3@scGei=Z6-5_d8-*mT$ zefnFhl2E;eGLw$Nb-jLXcy`gbcMu7R!zDXY>A`vmAqx+ z{#Q}FwWQ&uhP{5^48Klor?%GidivqBh;}VBT|GpFI-BnMh${`gO&wmCa{Leu1|{>( zmyXIFc(xbNo`LZ!I1`(AWU}@aW_sVKo zQ#58)2FI@QC1h#AWUL-MI8Ewb0$sX#Z{2or+athF*x^h0l-!3Cc zr{WXhZLW53B+fhu1-cf-u;>X`q?z>FDZIG<;BgYbxdoUf1%6I^I(~JV`h-f$jbq0s zFWEtm5LP!NmlP`8N**o_`f}owiR{5~it$$d!3996H@0Y;pm8+CY-#;@g^#^VLF}Nz zrI=o(G>E$%eIQnzcq~O6GbeF;LbSZk?lAN;Sw_t@r_p@LlGptQST16E+ita)%`VLAzq96l>P-m@ zWn^e3cv8qu%6`OOGx4>DG$uK-;+Xj zN-~Pcw1A4=ek8#$->e|tSQ`$aVMMlQtJm734v`uw#?FXP73Xbzu$(5%=74BdNv(;v zEq>=g{8D7DyZTcVnT#vJk)w8}mBvR|8s{lzvt&6x zBsTYQU1sld+B0`tPkJOCgaF&M))fdpN<$H|GuCr(;cpk-NfMbew!zKZ&d zCT+EsEKwcr@kgD~M0vSUWNGDLGQ}vfex;V2-vR$Cm^p z!ToENY&4v_EJyNV9mx~j0QH(Oo&dzqYbknWx0)!}I+&=)b$b6)v6SFXg20|&02@o}K zj-m5W-;o;y4|%L#L?QJbqoVqKNajUz&Y;%}HvLAR=e2wpmM-2EGl9Z7YkYuq;^>wU z`Iw_yw0yvv2pXUDs-t0}cs^ws!fj)oa(g%0_`-f?gA%=P!FZ%cgD#((g>E%BH*XtV zc%cL>Y)#1evoSXkFxZ#^V7U1ee1=LdHlbuMI`z&3_-YZ=26ZnB357&W@NnL&gQxi}Ws| zRxjz|mM=OxJMW5rG^T>3Kq%s{5E32U&Y5~IggT=VPYY^}lTLzF2b$IInd;LX3$=T9 za2vqPud(MM3}0S}QGhi~UMd#(JTC9eJj`s?Z)My3(RwU54TTDDNF&)b>AGWPA1*FY z8aA(-W2(2q4A z5%`vHBeSL;ifa^p$x>YLV2)sw$mm%|ZiCk<_T)tPz2^a~T~_CkPoQei7ehXqwsT)} z{fR`JE&>aG3I@V;=mik>H8P`3TrVWhp2bZW(%bAh|62yZMGH_J38Wd{8HhL^fxMZ4 zD%Y~8-vCzkVQM`F#pTc}Fvv5qDT86p%i%?JQBPPj$bWyUbr%S-I&85pu$7h?j+hnt zy}F1`Ue~JFV`Ess?<|Ex0yf1|eWKhNL=ebuP;|L2BKw z27!EXiMSNGd=@Q!8D5K_Nj9YvFZ8Vf=%?(dnhfT;yTuY&Ic$*RdM=9IWUm`zMRrPN zA3m{!pQKf%kn6@3^`e_S_hycphUT^}aFIN8%WgV7fzN7WDAHIdB?>r=Wit>phxlXX zTPuPw!&FSVhIr!)#sGAG^C~d|kquCDcw97S(qAjCwrlQseDiwK4*C zz?lBOPqdJqw|Pm+qs4!-pGi&_+mq`t_<8jtE_)dT8gWoxm_d7 zCTTp6=`?A>s0YkZu`W>u3aq$A7t~W6diGas-xn~WDL0#V9oZV*RI%6Vc?G`;R(HWI zYYqREHM7GL14OXM?(NR354J-*wsXtQM9(jGW=$BvZ*dQaSIsUSsJA83cywXnnm>2W z87Z04S~GP5(xh^#Sxuu#9}6F^lJH)Z@Ydy?kSO*Eq7}NG0VP^a-hz*ES}dZ+R^Od> zPQ+G>7O)6Iq|w;DWZ+(Z3gl5%EB5hM!VS_+D2z~*!Xj)8WQJWAollNt&^7qwoRAxU zFNKu*#A*6z5Kus4bog#!jLhDcJ7rrr233z#m?)7g zleyzz{DN`_bI)k2y)eg`Vxlx2?RBe~`Tl4InmhGq4h**>xv=<$h~uypw3MSXVD5Z8 zyp1|h!PL5UKQVD-DBAOOHtZETLN4mo%tWX=wUytwVKL#Wog7HXSDqXMSiZsbLBSgj8Cv`-nBBaqKis@wEJAf~DCVJ}6j zDTvD-%!_Yz0c7UsQ&yU9k&|OQ4s%}~_26d!jwTR}TpGrG4ij2({mj%TRVaR|%Vyr8 zC8aVE)^d-pJ;+d7z$mgA>7PQd7w@|Z)sa*lk#NN~e(rEY1Mj$8 zPz9CCPY@GIw|oaN0-*NyrO@c(IFmfYD%|>PP1Lu=YuJ*L9dI_cjCwxb;d~~E=)j|R zK82oGPS$J@L?NJ)A0r|%hlCqsnv(`cdmCr=R)s|BR)X|2YFjZyg=;OtH6x9Dv=yRK z{3BOQKD6*nQy4|!z>O5QPjsG)bxZi-6?Lu+yTYpHbIDiO33?{5KfrJ53^POHf>rEK zp&rvJ&P76hyNj=B z%lqKo54Hn12a%V~jOZS@V!@~_`)IJyr%lD7plC)E=x_Fu;D&*Cw-YWfl% zfDte@E5=2V>AUyoxow5rMCR;)#7%>sNz<~QXPV>u@l9QR(a7dxCdGiV|v=RC$ zh?LQYZsY0w9;jkkUg5}UFgbjI!L^>Vt8?(OQf9iUxZ%vTR}ef zTaJA~#Vevc|LruR5&~;|Nc3}Rvr=om2-b@C*1)*CMC>HedumaxbXUAAp*Nj#mzw|S z(kM9cgLPYP>-0(yDH)qu#eBS-u5+L77bTIhdfk}2Tq^Xg>s9s}FV2x2kLu$4DZiLX0$i=`#% zupNceyYnGEwCcVU34Rvr!WC9%9PE`%Dx$g+v+%jMQ5eR%5`-+H(&EVCxO1lboi*nl z?6Ut%M8aZ$5yB&}!y~v~*WL^7G7Ky0aD9}Z(fQ&k`W0lf-tcjZFpx)ZM!xOajP*?n zWlYdpVtXGr)t+%tL}-Dwnk(j{KJ#= zvcy5!UIM(~@Mg7wx73wd#RT+b{hG$fACesWUmGwg4&8ikH`zM2eH-P2@uTM_)O`1y zO|xCym&__VImw06xd`5k28NlNk3m&#cj=dyY;A2xalN*G#NFW2-)B3*vQ%i%JQmD$ zXNWF^;LKsUHPRj~ZNES;FTM>V_bo5`2WuX^C&$W=4=_f};)CGYlZV)o;a{(?gjYo` zlPh{k0hQaEx>9Jq_RjdMf>DYeTDF;p1lAprgjWdL&%7#Ls#)b8x*NCW=L##n)A>c8+LvvIP#>7K*sUaLh3eNolsF6pdrU2* zLp>UUT32t9G3U%yl(JVGZBS}|TnW1oHSX)8IBq7QY=0agd>Ut#zo#*WjkoCU`mc@d zp|YMg?-FKP|A=v2q>|Dr8mqe#JJ5Qq`KMO-eD0o$jhyiT0qd+$UOnL<{xJiQ&Yi~~ zF>-U7Wdd^H2mO!^{6mly7rs1%xN~gMuoku?^M;Bc29^Rhbmm{vJQVLve+amY*RTj= zib(;*t8H1X2#_8Kncs~f&@8p}7q%?206(Z(t;Xt%z8_{x0k0$M)JkfM6}dL_CXywF z+2i@PCc?3P78D$AJQPepME^iV+l)&dS#;-FidHTc3sj9VYn!}{=%H#IS#c2R>nSsJ%Dei`r_JS(oaUww_!LnwhxIz?=;u&G}qU&zJ zLby?sGcDM8V^-a$780eL@g^+gdUm-ryu^9rW9Gbr^E>xhBk0gdil!1b!WVNI44+jy zT}N^`W)q1d&~|XmceM`%;NC$YFq&IS+H%Taj)#HXwY^8WY84C^IW8_zE+46sRU3WO zTy2be_cpFwJ`V z``hD&z91eAm{Tf3L`$ca@nWc_8KxgT)rt2c=u}Hi_o%gv^?9<|ttvcP}vt`DJQV#-7K5G_KFy zfa&Dy?26+A-MUUc>!(3vxcM1tbc!eL!`P33LI8e|+pOwPVb_A)@nf4EnNJq53$#Lx97$hR5p-Ro6ok{{pGWqr0{n0BvU z8iww~atHE2I>cNz?53vfcPsf%b#(z%ZSc9t zbb1E zAb(bLIZf*|0p;}pmgpNXJD@;g4j3LPbENJHCN@$6>^4y#b;JiVT(%C_@~mvz&K@$v zf(Knv+`vpmdb^njcWqvei3cIR(1mj`Tx1YUl0q=?$mR04zu{kb)&%gKqAF6a<@CE* z%!|e*gF?)jUOl=`2ZMf!cY2{n3B?Nv?rVLWghIAL(%hllY)o6RA?Umiv62E2oodcR z0ofTV6hm%T4LFxsmAk(^&lz<^?783*Mst}#JKs+lR)FJgTHyj&k^3DF*?@4&ZoWM` zc||^!Q`xvXjxbzYqPzL3!!@KhqGl%}y{MhVsK|9ZIU00p{`Bzu4nq*-4cxh|0Otbd zC3N96ajV*r;Jvs#3KQ*lX|A|;5OX%Dc^hEMD@i_2mbCw)=?^r1@p-LIYVB6(LHx)I ztmMwZcHI@^0K#zKCZ*~7o$%5I8}!TT5YJPD;p?(PB3fP+F|iMFc>|MqBX{aO$nEac@*}rktLIJ1k;2hF!RcB^GkUX zbFVizd*+YN3@l@fYUV%bd3A?DMWzNY{(T>F++aqxSF0R+s`G{yJtme!cIrQtfj@k* z;lk`x+^7^2HKp}*?OytbsRvfW%mhF7Q6Xiox}V^FDpk?X!Zg)MIu$oC%al*b@AagCEtZMQjZ*w)4Wra<)6W^qxj0G_< zv7a%oGP|y{ly~5^8@$`__}1#@B6f;q{GvhmSiVX=0K)P74^zeY0NW=|OR>xzdFTYNh)dP} z8~VX3feh%O=YRPJ4m3`;4q#KU?bhWJs9eqIdZO#*_9+Q>C37sZTA*(N zN7wjz4Vd&|=gZJmsc~-#`6Mx2rbG92gh5`M3)QvS6lzf%!6(v~JhA+;i1lRLFx=H9e9 z(X@0(2?{PWRky{ z!|O|td^M+}v2Rk&bF5rxaK1Bu?1tigI_JMY=P(A_;^>zh_c(>_+fd|vV+ap>Czo5*RtIE;}R$3 z2$M!j1q(N|?Qq54p;YV@lB#m8IEftd$`)!D;mtHolyBV8gip~Vbji6^2WOeXsDe$q zM;AUz7aiG8J$!%uz{)6F-4X6gAs$MC?h*jyNPV7R_0*YiEtSk=5B_IM=;Zv%{%0Pa zs>R<^Zt$&q-hsr2R}i5k{ad{9H?vmF@HpXUspr}ld)V%9b%k;w0F;Jj9*lu?XlITGC6z{(wxNRsRq8t(mL~;-nQk8$ice{|hdyzg_RQ`6@5&bA~|a z4m_)X3xRlvbkLusqW(TJ2ym(n0ZfF3yT@rcPUsqG|ABJo922ER*E14VY|tIy1+cU+ zb9M6nR$TtR5H5;IQfU=K$Ko4#K$)~AUJLW9Y6?|A^#akwLgn;-ca4XaCyEn^KV+R# zDjHURid&z;?*>Ulc0b)*1@1iZn(JGzv>kJy)^Lx4hwodW&rg`! zZP?vcQi+FqGo8RT{N*Sng`?IBpR!kkKYH|=dQ6cbaD^hvHsm?X6vpZWCdKX1u*bb4 z!9HiI_*xNf^c&m|0M3P;$#vzCo0R+#!iCDAzSs?yEAG_z`j*Embzi8&y4eiA-mQhg z>oG?(k6U``FaHNTayIt*kkT`~-gALErn_uvx!b^TRpM)V1>Lv0GaZZ!)yq>>+}&M& zB*yi3g0!9F={}d5_1>(s12PsPN1SO6JD`1wlCH3aL&8RmunBVJPjp2e+;#TMx?IF!+JC z{z?|5&JV#aS+BqKJ1~uu zwJR=(-a%~t^#VViUDy^N)SzufuxJk42=uZ(X=v@>$QJ9n25Z3soI_3n8YITI^l&@? z7a!bG;Ke=p+rYG6HP&{*1zfM41Rkf~i22Va<(vF_v7AG?iJS1j@T(DOs^%f){AT{UCqxNF&z;vfGMvAV8pp1hm!VYXxH>KT=rTyC}``KW$JOSWnU~}_# z84vhBD?HQ)We572(6!Qhqt%9i5!or!RKJeT`Q@cXuYjN7 z+;fvijm=Ie&#FU5CywsGJqVMT^I?PYBIN0w5PcvFd)*X@uY(PlhzOgy4uD&c4`X#+ zflPqHD>8`C;n!XJhbi?U#-NCek++r#ufC7Qd|&SV#rLvs4rK>?c{m8?go)MkEl|9o z1S^}MX=G)TH69HIb>W;JW;CVbk5*lPISE1( zbN7kRX-`;KoC7k)OK46I%dilcqq)B3258K+5l+sZPsi_$4bsTVn%{6O3qriPiCX=_ z$ue{9SwSNN_t9n9sPk6|OQZhNC0d+;OFYHL!A&7Ju%k&MxIpDiHyl$qW}Kue;l`C7NB(}(Il$ycl|@ZEGI|{ee+8X z&YNGp_(i5KZbR$(Erimtndk%a6XwUxYqLA)a;xqpd_cP(fEA#BvcqL^dq!yd%wJ?r zeeHR;e5@eTIn4aOzwQg>568Xc+n~Dr%@uWNSpw zdA+&z-QjC#{on}kYoXWx3s=i4%_M?9?ju+c2>F-LoOJJmk;tVq)v@|!A*Njz+AFsG zy&j7Yp0Ngtj|I9VfDz-pP1SOKu8WK^kc@w{WuYJt+U>Eet}d5m%VsQWBHp!*W) z(7CI_+eP8_J!$-};^!4dng#Ti)NHKKYIu8`^?S7Wj-OdVz(q>L#B+fFobbMaLS8X| zJJCroP)zR$dupOGcC=b{?-o_jQFBOQnEu)6mCC&K&>&hdX~#7Ix)9NgAnina;6MYS zJZ=)b@h4ICTtf%=|L7@@>j+=k(srxkBM;$Qw9vHz!!gz3m5@v2OvGVFwPz{<&yiWw zwZ;+AoX#)f1RbPLXU+By1{Ci$>W{`)0ZL@Y@uTMQ!qkz`Lnx+`Yuk)}$~@3#V;_RC z`IJkeoN@5nb#xY`!2LE&@OQ@Huix29Kc3x@n>q2=iducX=CN*KyWe6{NgFIV<12PR z%UAxivB=EA+9!dE=h>@P#SJxWr6sZT2_+!mKDNqul=?BhOGH(RqSF>mXlfIJ;a6bI zI=Ef{n3vOdZRy{pu6`C0QL=K|UA%+3>b0U&9gem8;hIlNbhsfmYtFnA+q*`OmNx{7 zjoWm~#_9X|qDDYCo%S|}_Yu2Qn3^wk<@BTQDfQpZ%^AD0oDx`FkTx~V78zynUux2C z7e#fy^4R}!Z3q1yxzPbSyJ!)+nB$Y?adrdgkaus`YpanZ=0zQgwYo8<5f$LZ$zP#QdLyuoX{APnQCa5$Q!mf(+$Phr-Ne z*jFobad2@6jWaGrBLjd(xR(#LjXYkCs*U(!(sjSlC-9;1Z~(hZaKZ;)AW#1EC1=W# z*Yjds^eqkHYhyufXMlS=Su0>9`9}ejKs_R$iagt24KZpiVvm?IO%gnC&F+#gO$@`m8G>k_D22Vgm>?Vn{`r9sDLVINppsL0x|(t=K88* zLPhXP&v8IMtRLZViT=sySTFU&lm!b|BmG`Xj5)RK$dR?r6kXebYkJFaNTtSzp1=#dY=9T2!Kdd~Y&FoIFD@f#il3O`SRV)dh zHQDeeYe*(^R}tApkN>m=6HZ=-hnSQs$g6rxm@PrN=mYvi~$+gZ@Ixs;;G0;G#?nRKb~_bunJ3EA$B zI9Iwg?8wnyHFx{0GY)$kvQ*7K%jMG~xc^DPGW7w;5*hznCefZhaUY zao3yHaczjba(B77Ig&*W9EIyo`wZ|gfce-jG*6bbhSqzSH;0tmKId(w4AmaxYcJOr z7d*G!!CkGm8Aw)+fZ81IGgXQuc{q32j2_ILh`j?8x@*;YXhx^=CA!czsCko8LxZk| zK|d>`-yVyZtU5V%D+iex>v~nkBZ3g&1)9p8Ig(XJ3yb*5Jj!B|&JRaCJf*hFgZCg} zP9H}->RE^gFZr(Q9+L3^S`7a9jevUByP~3pv7J+&8t*#;oogxN>?L@Vyq{}a&R~u_ zTRx}Uyu2<-g>OLp=h+Ce%;0}5%KdmkP(BV2UGg5i3ic?+wvOho3`&3gg>v;B?eVsU zVX`KLq&HQK5$8Q(GO~8l@RhsmKnvc=(v1JKRMakP!dF52W9v(2yBXM|fo;3OfE-;Y zw;A8&jMpNZlD7DaDo09vf*=0t)G7>0q1|qasr~6gEE~$2 zs;ntuQSEW<5N3OUE4pfF<>j)GfLnPO5!E^(pzg;;nXSDN90sAp zq{KRwHxD0L=EKvaGa&AtqqNoYLV@<3;RYdjQan_aDWAb^X|Fon*V^CaD$+OL<=-QTV6O9u{)=!TR5Fv+0<0oaJ`FC7(M{3VpIuK*5 zIckl%jcmVnu=b@U&ePkzgcs2}%BH8qINp71X4Ai--)GSxLOY=L!l(R%cQ06|n+grZY&(?j2a^EhOBftda0Cc-3 z@6xjV7+RiFP3!UT<9X(z4|C8R$`R!u50B~dbh;XlPCvAAFz@g>@E|D5D>45(UPAOH ztUiTE_R*i>Dim~1<;N`(;wz5rlvHrbMs&u&033st%q4ApSKztC*ze`s-D|ej#tS;L z4Vw{LrI zx-kJBqyD1ZZ^m2yE%khPJpxpxp4z%v&p|dM;sF_S{f|@>K=(Y|Ww+e%k&&>7lRb${ zY9MYt=3r1z%lFn2+iy7T#Y?BgDQr2D-{{4e$~3C=oTuF;7SSECb_Bc@p|42(Bhub{t`Tg1GwrwDZ9kzF6XBcAlXy+)20LYmXuF- zi*=;x&^0Ba@?mtI08IxI%dODk;}^l|D5^oy8@8Ah~JkBP`K=AtatH8)-xs0@4>?>FnN-1InQtmda%3v1>E-hw7asS zmH^QQ9AB92-}(yv_q2k{NnRJki_K+$7WIN88lZ?22GV12c>ROK{x_E9Z(9X9j|~~R zaNPb|#yJRB#_OpW-d+DYy#L2fAO=85z4b4?`<>|cub}FG&;RpJbgAdvDp?Cj--}-V z$Cb}tE7bp7S4`!?82@x^zTUUx*10%76GY|Uel}Aczh>J620-yd>DS7FUfB_1J*RL5&&2fO3JW`haI%_9uPft$ugz@1; z`cSga#~imX4gdxu4Z6lgAHtRozBfwoSvQaSrH85*>0N_g&l*Zevh|h97CX}_yC~9i zR}maWU>a`2k~*#?vg4W%w*)v#Jyh#xC{9qxGy+zgt`cM4!*4mER*Zf7(=uE4Sk-bopksx4?920DYkToFaRe#iEdoDh2Xv( z4YP11e|0&0WQ1kX`ZJ~mH@9?g0659#^j%bh?3ue!^-h%2ZR z_Q_tPhg)6R_h${Pn7aMk;n-iFrsHzU2YL;NSJ-SdAUW-fp{u{Ii-R1o!TkY&syw?W zHM>>0C40sjDS}QdtjVYbN z6{|7iV%S$15lij45UpT56v=+?EL+C`KU{9f!>Pcg(3%i+0~P> z`!AVYCe*R6#{jgS`3~Xc(ouYLk&=t#>=enI4jw%~1Q)U2#{aj$hZ!uWVj)d+3|;y8#!Ly!UOVfHK9u^Xaq@qv%SfWPBWRGy&BX z$uE$eATvCEu35dh~cX%GEkK2l_9tcD1M zw3@Z+gkijKKSD}`w}Ti$hEpS z|3x9-LO8ze5; zV`IC$Ts$>e68VD*K-79Vfy8>QAF{MMVzHF~yn&E~wS_SbM;Szbefzu|Et~m+BB;kS z#@_-MJ^1g2uZ|2As8s&y>?|>Bm79tY6gu85zX1% zzH;gOvhlS=7M*eq@I5&M7gehx@T>@v(9?6>=8thPUv_zFK6NI2YipU4_goGBDm+LP z6q=|#CXa$KVQRb4k+5Jj7l*--+;G$m3&!NE^Ta6N6ocr&s9-ddV)0fcKn!i4LCIPE zY`a0~7~HO^dNf`nx;yQ=OMpkQOP=bSE|+(z)nIqAfU_1XW!F$nXSM?Y}}4e)Hx(twz!SI=7Hdg{D~~PLb+% zj2r1FFiVUjh6Vb64p=oz1-ZM!Rvf?aHN1{n@EG*{QW+I=tfa}L#U!8d_yuumZ!|-X4@E)y?OvNVYQA`W6d^EM&1_}rM0Q#5GmiNaM}yMBjLSSGZ4|{M4KPc7fWG; z^sIWfR(|r}*6f8h;wPIkqh(Rhy~EEZs+%MI(oR5K;`_-=;=|Z4=pO5x)BE;PGXR*V z_$K>Kw~pC&krY0kpTWn|K)G`?~~&Tag1Zu42a( zBWG{e{f*0$w&Ut%;0L_rch??-(uPi=ZOzDee9@m%V3s||AP|1UEOu&VaettCWv+jM z#!dUO&Ras-i0`MAt`pdeH1cQAW6kG*j6@ya1e z8OB?M0`p>^EBDs8VZl{_r{(i3K=;QwoUhcF+SKyOK|XfaepfT50v;|S?lYGZ#YGBj!-imnBm8&P^e(lJVo${QA_teI3%*ztarpNJ|}LWyE^5YVQ#vA#-2gQsX5y;{&%>GzRKm2mzc_E$+*Q4+Xa$|W1ObSvT?b@nv{_v{ zil_nhrX^BrpSzn#qHgHC&~O$&%YeqhtuZxC>t4_SeF{DPQtSSudYdTmFzk41{!H)m zP}l0MNrJXX)J6E|@kX~&t4Kj*AAl<6ug&8Hm=kI}mc-{jyg=5}a*C6RYzY;+hEt8- z;;uDet0;o04w}0Y^mG$er8A98xxbziSv4nM=?<^DO zt?JRhb)~A_#{{PDJ44(m+gt_HS1O}SJ$+i{M6Nl%Fep*TL^(0`&saM?T2pOBUMr+^ zaht$5vfX`ZS4sPcJ0{6%pRh)FMoOJHrh^7fpz274)9w}i1ae1I-#94nbpFk*44P_2 z>nBGTYud2VQ zYwB(Glj)OHorv+VlLDrK2rpx5u;@cuXzYfEOHC3fAKMJ9@U{BjMlfM$@UJ}Ax2isH zfdKRlRcEPZN{fq&E7?{ad(E~dJ&ot-d`Z!DxJOCzDo}|mT85e#B17xP+DBKgJoe=T zT#nO5^?@>>X5q2c9u=j?jdk_1sNr#=`2O01oR6AxG@~#&{llS~nfGIX&Q8-0@)V7! ze!Lu1-%XNNF7ZU|vm)DJBtZRm&RdD$sCx-DTHxBjnQDP^={+```!x&}MZGc4_SR#N zSX$8SmMKl26wO@I<2HaEOq1-!?Xqe*cT=aQh?8o$8}?hBtV%p_ZrCJr-ol3$9&Z;49(1ie zlSBiJxu|Rp-dLiX*%DJ^txM@Q9Dm-e&Nh69j1+yQ`;1pI%uRbv$EbMZWMPXpOVZ+O zKjIo8Hww_-FbqD*u4jmGyFraY8WXugtzUI3lf49Uh&cKb2yj7M55d-reOx#zJ%YF< zCA7CvQHaOr0_-tNwnr|S{zMAu^%0d~ef+1!5ODb;v@ zyO}e7(mGaFQt}zo7+rRfjEg{|aLaKk9WN6}hOy8{~?7TxL{D7&my?)^P&P= z{<5#4)C{G~_@xR9o5GHm!9$Np{hESWJZe}B_0vr3uKl?JgnQ~|Ay0i_O%tb`*jq8H zTj5q~^CF|GyLQ?XvmK436>B?2wtesVu{zodZij7#fu;Qo@xdIPm^1*B43&Qya2os+B z(&=V4R*re7k8#lRiD`c^j^mQ-RB>Xh)9D*%BUT4hrbhm`qy1n}VZ4^5$d8zjl4g!8 z-4dCfEUgS+*iG#`}4+Pd+a3NV&AK z`>#>1Fm=>CC6mMVSR-WL9nai4b_JNk3+aH9<_2T(&_OF-c^c{<><=(HEJm<6<+J`? z%$0Y|@)zz(sTRBKp9w0F`a7~M^fKvQGRV^8+9{qxBfvL&zOSz>b`g&GY)O0Obx1<^ zW{>be{5^vAUby}NM06X6^j;UkJNPtqX}iNip9r`gP~M8UdAIm}+#R`Kq>}T!I?P~x z2t%BUi$LA_bUA3&jatCX*YQN0tiWL>fl>-?X(oJ7_sOsj6Nbt97qD?h_iL2j(IA zWp7pb(uM=Yp~B__{jC(&)xH5oDB^6RE+oyPwx1nS`0#VF<4~B=>F5!Waiul8VQM>S zjI(mFq)U5OdppD>WMTK%)Mqo>siaHk*m=JF$_Bv7bk?qY-{_r5wz4jR>Jhi4#uU3* z$w0rIHf4Hp_DT2s;mQQNAI_7fI*|V;z(joeyq#P(9^<>ITFPpBtn*nkcOd_{%9GQ% zA(z;~x+4^c_oE_KsLVCbopm3bJWffw1x=HT*TkHT6Nt*)=0d@ZZw-dkvOM<2v&Vj` zP}%tqhfBNGc^fgmWZ~#5KI@fMPtsS~S+oM;mKEo6puZhsaYM7|P1(s8j3Kk02={pd zG=n0u=ZJ$E&AG{}*5@`&hF@f7B@()8Gi$3#ithrx(M+Pd?LE@i*?VzxSMV42H`&>S z5a78tdn3wo)kbXys63#-K=;|Zl~OS67YJFQ=J!nIGKOCuWW22~BN;E5-?JwWA?5W4 zb?F5SPLNltzkmZ%FCwlGl@PwG>W%pr`DRXnt@%#PisaFuUSDA*A^n$lQHDi{+s#wn z%nvIfp+6SwJo*c(3t9-bNP4xquv~$|=l!_`&3DxD87Yp{!R7$1b-_V5?O7QB$kI98 zpcQ4oILM=wTLVyQ{g%&%Ts31?ED%m*0vTAMI%f&5ngs?$P)41!lM6AkpRE4)CmYk3s(vnL2|-A_HR z<+49!As9Yzla}A4%p`{8y4b-sU1grftDrtKg1&&R0u}w@c$P%%TUOQCj~HSI>Z_!p zvYppMN>eoNp3b_R%;GYH8dff(MxGt@+d3mU^VXWx_k3rT&a`g`b%b$i*)QXEZj$k_ z-3sG#Jv38_5sUO#k9+k7u0B3Kj&9D4*gLqxh~{6hOS9FP6;hjC)mgQ$Lou^aC&9C+ zO=Osc!luPNC~+1%0{GmTf|)#_%?VYb(tBJ;%)Zg-!O$U^2gfFAhG%Cp01sGuWp#CFhCK|jnM>4H1d&oYq+1%44(aZa7(yfj=^i?yq+{q7>23xXKw{`_2L30y_rBcM zeeIw7dH=j)6waABYOUkD*7~fvcE?&Z>$ic7DQU2mwq4=F;q2WpoGTIa4f=0BVi&8H z6)5DHfepPnm!E~5s$BU)t_ppLeRB2P8h^kmagyGLQkM*SsKEr#=&0_?4p#T1Vzckhvkim|xCTlS{bGr_`zxO*0 zJ;7ECe2V@)O3!;{#`eY4xB8=sdMC}pshMD3Bz9HwXKx4JJFIM#l3g9U823-IRPbm# z=Q(#f?dkh?T(`mtE9=?lap~I!3m`LS>`Tp9Cc4U!fKOXA3PI1ham*LAAX5fg0~SwH zq&q8HV>p{1s?*T)k00HR6kV!w&#bL!fh=6t zhVg0OEiRc+n}&l10x`l9ScgQZF7C-zSsS7DFbu%3Gs#dhNI zUoJ3i57vL-DZnpKuLx^@Sfgdtc&>YTmO$$?!vT|v*4Tb0Z`FAEtg(jM{`B)o*PJ&2 z2qW(;p{Dab@7{5QVVZJDJ`Rf0b~M#I^DEbc@qQ-_X~I|5gSx_O8d-#7^#`@E3;y=i zJ>q?k;2AerULrHM*3pK>E~rCg7<1$Jk-BrlX#-WefpLDmgK6yCas9`m$Hghpmi^p+ z-l$Wbd9UfH<>ZNu{pm^uqaLJ}%sLjek-SZ7^)p$0?NcJ=$V1j5#57{QI?XPa*n{Nc zz2a{)IFR-#aj&Er-E)6mt{80xeYY*ugg#$KOucg)ky-3cW(xhc1LP{#iDdDUv&Z>4Q-Z-Q8Z!^sIjO z@=-=Px!_35ke;YA+`cf|$o<7|XM2l*-ps3t5*g}udOQv@i}4LiW#u~U@%eWPz~7)0 zAw1fBrSTN3!|l*3Q`%f3&b!vTN$t#$OO(9^nFG?~yIV6-V31Jda4I@>@@XZpv2#bC zWa$LT4PNuJTyGk%VFncEqus>LoB%O)AzYu?Qq$26f4@y59ide>A5YIy9)y->67~&` z(e{G zIPE`8$F&~pcP)aA3^6`qCnjL>EO(8kIbc~NpixFyU%#w}Ny3fwK_8y&F6(bv>L-AE z0tp()AHeyKwq)3Y^l~g!|H^nB5dj1na8Ah|Jp4ndonca)A3>*qFAJ1qXc#7|letFL zGcLzipukHe|K%{tLgL{{0esrxljrPC(MMK5yY52q1^e#Dan;t>ctxH=+mazHvh0l=qtg5s%F#Tja+wkZc0hh0V!Jg-YVx9t)61A{6&0&RPofZGed5$~pp=M5n{PEEe2X zmGQi+Qm=Aqla64I+pr65Am)b1e5|zrXzp-%PTBafs)Zme#0)wXa94eI&R?*B_k5OV z&kQPVe8mQ$u1e3c`{fKM=0z!NJZuPXcKfQUZ{g)@wR5Th>&o2I<#)JD9n3hOwUB&H zdu(-fY^XcF9!T7HX-XcgtY=t%8@|GcxIi>uc*GeY@7?U?^3RewSJoX;fA`E>45K&j zeqsKU`~tvb5kPn{6{6?hR?1W@$?1%?)ME&)n}l^>_bVGDVpWoL|0zXygMlO#1$;Kh z&t$pOWE7zJ%JvKeU9afMs!Cf|xqb`f=v-2S97F!r|_nix&7Nb@tGE?`N1*D4}rWTg|&VgB+GK0j=FnjX)L+`aIIfb0!XePoBm znG+_TtX7dA;e)pdnu>#2!yJeiT#lC^XqBA8KAh*dz$xuywny5zLk?aNlA>}nAG-C& zAp8gir8c05Va>GS+>A&_pxF*ifp2A96@orXt%Pj3SABF`=jiy}Uq^Yq$*bwN#Wmp0 zcyLvRXT=j^t>GPMg<;z+#Os&_#Ej-U=`;e+3`WY~#$wqok1my6E+Yq812`Su&3YiW zJ!n?aUg*)fceX>fOy`|yXa#Bn3$n~ZF!SILqT+2E0<{pJRiBiAPD0O}bM1_Sfy0VC z+dDSpZ!mX=EJnpfetM*D6ENUm1g`!aO5|1qy)CJ&taV4eOmDszVeM$!z>JdezJ**K z%ZAjkR8zJYSILmql*eyWXzUc)R%hwD~wLc zz?=9?Ke^hA-%>e^o}T={*U6NS^E@>aBPK@nP*UR((3ZuwNL~-vj`QHj8=Mzq9UJR| z#~08MGZo$jQSIt9!{DnqSqw5~y!2adm=sB};bC;q1ih?xRO@tPUB(Mw3z%ItF>gkX zhANto2F26K49j;re%}C@M6(vuCtgWA}?!EBAD9&Nho%&CnRZd7#@uQoc zEumvu)G!bJUL2$a#vN&(;cNFD|4%U-u+xFE*NYvkV&}d73B<>jD5o1i)*IieGH}AX znRv&G);2kGo@9t^tya?87;T8Ebl0HdB9 zB?DvfbO^Fk4&BXgQ}2fAtOVjXrniDkp7)sQuU)hv={HGYW(+>A8$OZ$dtH(i!WnIz7y)9(gLb7^FJVd_`$Tbsh3IHP|M2=H>kyM5%8&u{fm|NVtiJW#d72tjI`yU<5OmdK2y*=J;8&sd9^;$g_EbBQ*O(R_U zesX|`XB>cSeg1Lu!ec!>gJmh-$LA*Yx7jjDBgUOZ^R`-YV|5l66Se9z$_M(3R-No8 zA`EAaa#0Y5yBl%E+*iv$6Cs&azV%|_@t3QZA?O;pfdpW+ubn40fY+R<74)upCrA3! zLERhhj0Gvc-WgwrA$o2xGp?^QTPNDE2GhhG69?Q>c=O!oor=vVO(t`p)CsUuSD`rBx$i{tv__*U=GQdcQfR4SwMP4NH zlAm#zMK_pv|ID~_am*h4%F&06a&@ZnR%W+uYm4WjG1d$##sTpmq5_sttjP4Xgpaz# zl3Zsq`j#7_4qDmXINP2HG+fo6sZCl{bYlJKz`E(|qAYqZ&r7UH#|1$Vi#he~lOzZ3 zFKr}QxR#V*_bX!gYt+6WvZLzA9JctsV=q&PaLq54?H1|tuBxmc_3)mpBVgf9wVTKM_bjy`Zu{j6gOr?5AJ;DufT9?qS|G*j+Fzb#HriEMHi=q|E~wvL#Ss zdnpl9l|Wto_PL{G_EmlkR+(o&DjeG`(i? zO=W{euWgqn`xKMDvh~vv=K>QFx5MN9uv7#VJ$PoWCL5&ey}o9q-5zv^gKanNa=H5I zJWf}fRo_2La+JxA74Li9W1vIz@u-!!&x0efILb()dlUKOlFLjD8OO4bAqDsbh!{ne z33sl;kHl7@-Gw~LC@(}xGw_e4B11ho@Oz!?mdO%ChMRHV(FT0dUpsLUC3J@NhRY~kg;+ea3^*Dlb0_W%Wc0_Ek3T7OIDKWTFtO4JlZih5 zCWoc&xt@gExi2@%L`!>_GM^ZznA+%mIsuWP8_Xw43oJCB%G~R}+^X9@em@r+T_s-6Yw0WH?VY{6U- z*LsAvA64okQeVAuBsM#Il2$>xKF!ztNXZ+!YG9rxq{hm}-Nu;iI83&_HqRCd356 z$#|DqQKzg{ZK(ZUM^sb2h8@N|tjny{JZtP_SN+~{WrdTMbza?{7+E#*|Cjb(Y)qw{j9z$y|niy_@xUd zaAC87zc8-$#_4Og(k|_Z$8rT|r14k}s=i>6b4$c9=sc-}?H6l^`#egtz@Pf2)L%~> ztL*@(osA#hd8y-q;5Am3#Au zEK}i#%@l#%YGAgv%`KhhWZR#M{9YMo#5ppQp`U(j2}wPN(T*bHoUkt`6ANf9oVl+YNw%x^#Mu2w zO8YSacObgZ+y$F*emSklmJn~ArG6ov%s_n2a8r8L)^DBX*+I^YyLZGyK25NxUAoz_ z5OdJV8TByjor9DflSLS`ua5s^_EER47}-ni4dmEJy&HNL=0 zQ?Z&+Nw!^%_9oF2W|YvTH!x8#`(_k8CVhW}T+&OlVj$wqtpRm zeVZ<~{cvw3-^B{=TfMU;P3^|C1|g#UHDlEQEFwYU&d$XSqH;>WKf&6w339y5L*9k6 zIn{yD{&0Ex5?9G{w>!jkCk})dQ*jpaCrPs)_vY=DCnq9}o*;1DH8nP=$_5Cm>vE?ie3aV?a>77q;|3Og3qE5s`_xRV0o2E-QGFG*0vRfmDKYbdp;2DrioZ!Q-nrmT(%bUNdRA- z25a0R7Rzf<5J7#d(OH@0V>9i>=eOUe3}^bO^A$TQb}Y>%Ar2hiAo50QFg&9)UGRxY z%7Ro;f+4cL!wCWQtAV*B`6Wtw@o+v$+H)+4J0SSXE2;4~hbPMp>)Rr4DFTL-bG|!| zt?rkNm+$nQ$*gZD8{t&Xf1;+2l*nBqijiD_V5E1`(T#5E8(MO$vvw@d+8>V&lfPl| zlBgVSy3N4W>7mQ9_(!v@9|o!7-hIy06zP;x|7&mNU~N##bz?%b#&MHgtJWc+2Mbp0 z2L7zqVjMZm+Z%rzt7+FJ`tOGS4Jpl571BM#6_>|q`arMgtze=1x zpad4?Nbl((6(>hfawI}%e=_e+FJX@6d55<9%rjwUNrzFeh3{<4pY z%dwJWSXV(yJYX*`z?rTfN=>R^18LXQs5XGL@ye@>Xa@vBn-)+!G#wnW5*>@TySc2B zN{%fe;xX5btR9n5E|$M&KHu9?$&O8Uq04{8Rm$XnTp8?pnzd>>#9F^~buer{JNhZ3 zmLV`Jb+jpPEv9BG!zhK~rl?~-@_#ZPP|^kGPV{ji@k4oEE)*IVu5=KL&n~Oz_#;i4 zpWr!9M#m=6U=lHgHxe>Hx~=N>1tSmBV*!s;+v8k6qzB9kD{$Kgzg0UNF01J>vsr34 z3Qk>{k$D?nG?Qw<<4rE@^p&^QnGoSd)ft3uzcku!|E0p7?Au;OtwutA^sZg^zFMi1 z(qWkjJQ%wOHqE@~A=Qh0gZt&vKirwkCjwz@M-ec{dBM5$WToZRB*{-t`NiX6w<0Q+ z{g{}cVH62Y@7vL2iL*)MdF2*&nfyNxe-m1FwoZB5&bK)YS>wOi$Hm%;6Er{5FRK;N zqAkzAy@N%}!pXKFy5pHLD*{=JlI)jl0O`0W;`7i;@W`RzY&Seh7^BDHW)(4K$?Wi@ z7Bziicv^E;lEq-&uQ12V?N53CkDOR+EeacnhP;q$(@5!S4S8>|tY|84T+6Y9)M@9^ zBf&5;yZh$x^izGN8uhV4D24b#_HSq^CZF>+EKC7gL}q8usL=XZ_}s|@D#pyB2Y<-C zg^+un_mI^nCDo8V$J4u;d{VEBmQTqi-=mHCbY8O3Dk|gnRWybXdIvblh1yHVzY!~% z*S2|2ZxAH^Q`MVDMt(xwrl==$Xnhx&3^Ia^;^&-oy*ixAVE^N?F(QV2Q)jUOJhh67DgV-PZg*c^>!_*|tteC)iB9Y2>})K_L?P&H_<8&D44A>c?FWgS}k zqet$a$+{0ddl@Ceht0*^Hgaic?Y_~rzk+lclt{t&O8d2ig~8=kqlypfR;XPNo4e`6 z42kLSj7RSF&hZDm_f`ipzLZUDS8w;#KjCTTK+)+Y>_v>N_8^@HM@`qgbapu>o9b-H z@cU>~oI27dU@IAqVYFj7LC;1@$-WZxs%q|beOh`>mGpuP!bbRm)eETkQIO5&4yx1v z>3ixe6}$V1LX!o?5k(PWER(#A7oP>R0J>z*uPHGL5->u`2at=6AF|ZwIRkLkB(fE-yjB@*(j_1vB zUvN=g!?`+namfth+@Vga3zUS#)uBnb=f5+=2X;`E3$?9r(_3*0NG_XgP;^n()?VU1 zsAb>aQyttv&Ekhb=E+4ELf*D~m76K}E`6p9Iw7!Eqcu@;@omcUfX2t$KvBo+m%nJh zCPRz?Uynl38zr(WJ#V~F0%`_7^oZNba6>;j*xKC@egIoYRG820;EzUje+Iq&TmgTm$v<9F(l zFlqv5j753WM=Sob-PnIU^6xK;c~QH*V&Rwa@-#K93j-0Avwu-U3HZ}ui~=7-^AX5+ zZ||5JNBsjP$~)&yKoB^2ApDFlC&-?RT!pA$KnEUr$SfRPB;w z*;@gq+DolMWGnBmSNs?B4wXcP!Lq?z-uPe>$7!O60o<>j^|Ns7|CalP` zosvhY`8Z>+F$_hZ)oEX+!+0rI$Ct9&~9 zPm;pleiHbAoSWNu*LaQa9^prpD3K2diUmmMaIR7NFDt73j+WgyNVIp!vkMgTBnA3& zlYdsg0859}9jK#taLKiZ{(qL=&$a#vb)m<1w7mQ^w6>tyMPW*+=8Y24O5z&Vn7tM> zBK!f3_@@Iz!w7@{68aE{H`~?f&q68SBtS1VIavvN{6{jwe=T_Lr-U6nHpujQp9c(h zQos6nC3hE8pB}wUwAUmOciKZTbsWU96bZkv!Wt-xru9e2AO;wPk-K5Ne})wP`)>7B4Kc@DE1d5f3JDQ%R}9@sz0=z$2B$ChwRFDnE12O2smTd zflV9`yx*blR|?4ANB15kt7XI^Ny(3juV2$Aup5ip%B`wXA-m!nYOVoetTDcnLSxX? zrxm$Yqc#Qo_TxYPdbrc|!;|bg!~gN&8M-p}^ zU0sE!qZljH+wkfk`(+drcy)evnjrFVcI&H`HnuG9>5+9GG}WO2|E_NmhGtXxXE&o! zKr~JVzkU5z&eMM_7PDtziHTGN8tnG|m}-r-K^C1|;R_29VD?U@L;BPevHtL6I%8El zUVm4{^;#@=G?kprOya!I8c7WfbX-^PAaS)i znWmASET#Y+VwP^mR*R=*r9*moDj==wOmW0jiZ$H`hYP z`+bWLzNUf+&<~Pe)zsAJ{&VCqGdJat=aX1PS)m8z%$>d6W!{9<>|vjKP$a3xoro+# z1tgHrb}>6!g#KZw|1iEB4 zWd^IKrltxU>^0zE3fZ(l&fomNZSvl_hq89vov`L#5IarEk-=PP#W@oXJzdq+GaM(h zZ*9e`b5R7f9dTop)aoG9>m5(q8T{=K{kx;`MD%{a$1%6TK9}EhPhxn=6_Y{*1-l_pfQVVsiBU zKVn_~bvFL(B~&{#Gjn8%^~DQY=yoFS*(A>jG9k*^bqRmXzo34a93UGS8jg9*3qoK; z0C*%ZXmcF@h}0agR$r%xmww`+{~urEZ7s%03mpWyH3?8R?dO}{mdajxmH1*0e$SNw z9_y`ojY*B^XKeo~o92I;nm?w%@fy3C^n3{Ym&nXNPu15_ATUzm?!@&Eg`59%*yZ4R zf8{py75jgD-9IM}Qr%S4*dNi-LP=cOe#03iD%utiowV5}&K({Dezv-5`shoMTAB7k zmp#O@D@E1UR0)*^la9V^ZI9NA@w`=u-UfuJ^0$(b z|8A(hW(uq;Gf=Uxv_6E6M|y^3Cwt25?P2Qesd7C9ga|Jz!a*BC_|4xvhC*J(yHJa} z6MJ0q8B_gPNKbsFfKhKU6H~((B+>1N*6whL=&a&sk_gse?b!qAzI55;r{dz8w65<1 zJQU*cd~Qs3%PC6O<@16e8L{n3l?h#mXER~Z=px?0EGpRi8}6n&zIV?j2QJTzTcN0$t;BNNuAy7IyJK2!# zxas^* zRyko9PWyq~oQQaj?8mr!Q!h&pK#&Yl$dM%d7Jm7SvDjo+(ZzgP+-{7xUFzJ)^0Vo% zbUcS=n)}fbhpwq|ks93@x{BB)+X!~?Q(9VV6$TTR#IL6e=i_bqpS7T4PLui7n2R5? zb&fRpv5>J9uD|-bG5D_$(f_%0Sc#hEnF*bI;ZccC!MrWB zTwEGSlIQvBLJaoLK4*XCj$}T{kNlid)zQbK)9`B6V_To6pf${4bx2{kPva>*-}zKQ zKMHaQ0NoVLPtgs*{mZ8OxfB0uv+BvFI({Kc2SS(b&zo@?ZLhA9BZB(*mC!aIK_EC1 z5g|cmhFrtbpi6DDP@S?@2NNY<^3^NEwyL%jYNdUcFSjqs03D&N&&|_eYh*U5x4YLl z+ihb)#$o20Ou~$z?d~Np{WO6z1)ebS?h6Nc(%)a!2TV|oCf(vRcS47!zk?2PV-0pA zmYf=@GDVU>^Vk4%bEG2;D3DcwSi|}ZvN=1)hR;HI57QTymnTnc_#w;7HD88?L28U$ zuezoUZNI8i5lVk8ygt|5thj#5g$}Uw_NYAw$L!~6eYn(wgpfA+h*MgTgMIH$7C@&# z%)32c^&L&4I-M4RoFiIsW_3zSOJ{xnI0xXM4Oi^as$IxcSLm>j1c2StB0Om&@d{5) zwo-|G1+VdjGKUJ}qdyms$WN3E`5)x_zn`%WJ~uJ)97cS(bo6%EDE5&)ck!#cJbE{t z24oD`jjgtK2NeZJ@?R#>)gEjz5QXs06!IUWB1Bl^q%Yh{$&3~&DraLmVV6%=l0l{A zRx?AE!ZeV$Ijro=EGlSc=};gR;qER*pIP#q05!k}>m7~Q{g(bXa37_stBZ=|u@<{^ z-yUY+eJ+b)Qu~qSh{ULS%q*gb9}78wk6>8o0b2SzJ{t}k6zizWW!IMq;M@Gy4+$ui|9V?}GBOznV74K!;0ep58 z2X$}vYtiE3_s}S-8ScrsK>2$QNYlLUHcDSY>TYR_?eJYmVss0gP|CVG z%bs?G=?{%#k=xC3WuK5;RUPw(=koVPJ;o(t=#D9h0-C~h!K7lkc`L6p?&81qKU1xD zN~#4N7bZ=wDmFe6LNB(_jRU5@qg(&qOabt5U0hDtRfR(BcTz7Sz!8*y1>Ng)R@zK6 zHeQj?OTPjgYd{R71O>e>I4x^7*zDvqP~NyGCTxVIJtf;Za$hk*Pj>=baI*u ze8aBa7BW*iak}ks(y7rayS_sa*?2xEn;kb7e6$Izb_@)n^Ftq;7ki5SwdQ(rdiZhu z7JeF_=1rb;X%PxAIL{8&bhBwZ*e7{sy>FgX5BGiTt9O$yGV)78xSUy+j@#%E2Az5` zM6j8U^Nenk;qa6?&7kqrZ_l<;JxW`+Yp?CmpjCLDMFA}43;SkPzsJk>f9I6n@V<|y zYFPT3Ub}SqVJG?z2ka+itwJVldqd{=uK&NIclw`zvDOWaSN@xby;vtV{fE<1z9Pk0$d8DSC8_~pr{ z=s@f~a3|gGUqLuo6}?kDE=PIy1aCT*4S(mqoYr!hL&}(5Ezq!^F-@BIY)#IS0;K!K zFlf-$9>d!DJk9+uXc-wH!vK&VcG4cy{Gj#-D#^?|WXT>BoWR~)spN6D_g&g&nFM8E zQiY+TJw#_upFZ6w)O+BKMhEC#Xl`q8&fj@BtjgXI5sC>Qq0q(x3>)}Vb;f7TEx7Pf zJP|BR(8vT2+wNV5K?H!>jvl3k`o@eEs86ks)jHlGn-caR_lQ5O#Ud^Ka#8;-{?nwV z^~a!~*z5dAlJDQ3HD})^9^;0qm92CG?%pP&JvGjRaVE3LmC=?MH z(WSA=YY4V<$O+fZBKKZu8qCPNb{Tf*BT;>B1PMG^CP^2Vo^ggLDMLeICEw!Prhngk z7sNxdqGguVd)s=(`zGk>bhyXU;3w}0ajn!po>Y2sI5U3c^=-omspzT(fN;hQX5576 z{X^ehJ<%UJyf841EY9t>JanHncXw5y=dZCZ7UG)F6$>@F4>a5yV+j(lL}fQ#nxIy% zv`~g-QBHZZev@IBpnS~BYk;tBOriD_6^)1sd+Ud~%3!H}3wPy?op4!kYN~t|;-%8j z!|H4e!(Q-VDI1fDqRdgy5l{KydE*!1plkD7^T2%`#r%WD7P~qwKF|uhq=4NTtcF`4 zK5L+-C$t!?U%ZT~Zk%F<$S*@|WJYVpLpOAZVf@g)2 zFyy0t01oip>EHhahcf`l`vVH>$;86K!hKG#p9^|Oy*=ZDWrcxtlk42m2mQZC3woo` zhQ-CbJ=aowJwkC#j+x}Nkz{`dv;FG}5r4og4BGzHNavrY`>U6Lo?ZlR;Kh4NTfe*j z?^d(t{y$ipTtwJe`wpyYfIJ;=_AhuoscGm`U)v?^6w^1YthBX-SDkvUVVw4Lq}>}LE|P`Z^tq94Wrnk zJUjX3wQQR%Me&42pqjU#bD^2hQpzr~n zA=Sr^A5k<$b~r?Qo&h@0=VA%}x4Wpy@L~#ZHBbcpAiY#g1lUg^Uq$|jCG?-qpL&Mu z+RWdKqI3_?$ZiJ({-AWqBgQ5uqXdAd%Zz)ZdY_Og3a5Kb5!iDzBbf}p)$S+&!S(_W zY{3*adh6s39zFrY)Mti<+yUsHxsk!Sl#GnBckeF95Qd@bKS!d+s#PQtS{u9G(du(0 zw7=cZc)4C^$W(Ex?oD9d=kLjjJn zsJ;igA*T_owVbakDT4ajJ$pT@CL6$DvmU{2g#6!TcK_!aruLyM$Q%W7e!G>+(r;o0 z;MPpINe=!Du>GH3|If4mLyo68m@23985r?&#fQq2|FmL$SwML4UQBMH29CL-JAlAz z@^0nR!+()CAd(=EDB)grD){kV-^YQ^0)0jCD7kwZ+U4rD>CaMEzc|VC`?(U3MyI0s zDy2y}NvF49plpW+M4jKaKKuK)dkO-(W}02qR-$@s}OO-LEmx#z6~ zRLt2CZ>BIwB2sM6#|^26J6jMuwoumX+a&2f&S9~qGHE687z3!={-x+`9) zmy4j`z*oe!`S(tOm`|bPpOxB=hU#2~A8b|@=SW!9e}5FeyG~EcYNG;3ljvqE>zzuA zzOk^F^o6^tCHmv3hiqy)BV|RMRmap4uIhOXtJ7)JG685$F1AaE4EpLCGoAalA`%l{ zXE^0eJ#9=CuxdaEo-RXkOSg081xGzkc8OfJK35tn<*}7&G$y?tQ>l2{IYeMLjtGUC z@zywFs38iR`C`Ul7fs14IGa%4{8k_ww1Y9d52IP!XKIXAaKm=gVG^FSm;uur^X+6_ zT()ZxGrtqac{Zx;H0tQwMWh3w(LaK9Th+c>wR72OtIWKvy1n*Kp&Io@N%>lpg6ica zgpgSx9Zx1X{^Od=&slNfeIXESTicRpJ?)Gv>4Zmaq4)1PX>12ZMBH_;zQ~R+o65ED zoVCnj0bL2p1vNNhkul5P6{;(Gon0WYKGt^rDhFg}ywY*4yVELVclYre<{vme8##Kz zFnnZPb?9+ASk7+&ZwlLPeN!>7b=96?zXXV>uWBKdnQCR!vu^~4?N=D0+LK+?fOj;% z&?X&uiuq%gK;oOyMShf${Z{31Tt`=h9Vz>gV&F=vyS^H@m`SxdR-q)HB)H=FT^M#R_C|&4QZd6RvD-REErI~G)RzI1&X7LmNXPCRVwA2mU z3q95y#rX65%r(Y6&k?NVFFczmiR)n!TK8VtL9C+Rx3m-^(5IT{2ZD@TJl0#6xe(C&jZDjU!i)ZwP?)O8YUdbiU-7J+=YOc^LPCt$m7sCf(T88&`cVkZgHS ztn^kY$Z`4h5=_HSs%b&XL0q;c20#Zej~`4B7Urwhj}_*1Z@fH2G^Bt005LG^JUb*YOmcz0 z=t(YXx(wlLpY-sDl-6Q6jCj>_y{dBwU7yU*Fa+wPSv(UgJ#`Lw`KHf{HMrh?%;oRD zhu~SN`yglD!*wW)NfP5g&>5X_A!OZYYna6#Ngk8j zJ)2eY}P7)voLTVOO88X z=vl?!Xej5?MJLuNnkJjR_gDAhyG0!zNUy120hIuXv$yaPT|`wbk5o!}v^DAJ#$qMZ zUnOF}z>I{nkf|Ub@Jh@^`Bf z)TNe`rsEpM%kTB#YcnT~Dx|wIh1XG5XSf89;7awM7=x5!Iu=Hs9NmXPa_Wi_?7Iwl z#q94=KGo_uRd!?W6*Xk`RQq^gcr;p;amu@s$X+m9Z7CG(tRz<4os(w=Z$6I?IwV;P zRReGfG}=+?MVI{NQ}-r1m#Le~J5n#HG20wmHVbe?eDVP0!1!$icl`KV`F8e&1(X{R zr9ZvFUsSj2@wJFpruV$Fcl@eM_6*Owju&O|(qHWGqxds@BQUAU4VE^v!r9Zepqs zsDG>x_W(3ENi-W(Vw|$rP-H6L>w<~Da;OD!w(IhtKj78zAwP6p4ACguGfyG@b>Yf1 zxm!p-_jL2`DwVfB&z#S}%UPsL&w-OC{m1nJg0|}Da8!a`)xlv+6WdZ{xG0fUp&;Rz zs==7fyV=!>38}U$a0_VKEoB1ru|Fx0y`&2TR63P4A1++~=5&z6PQ;A!}Dz&hEfiB)@c#zR-%l)E~8cqb19O_kH} zAYHxa;Y_01@hOaB#{Ae32w%pfVS;PBlRn1kc*+5x&FvoL_0vy0kk{&sD%mJp?(YIg z8?9}Wqezl_!>gr&3?~>R_nO+8S2qu*bLz?(Sl%XW-3oov~#WUtcA{=uXiMje?Z}KqS1L_*>as zb4htb6bs|>0mSuQg@WR<3kF$(;~Cd7cx{)g`Hu3jRgzDPigWX7HYpwM%ZpDYsZrWH z4HrZ9z2-rlctH&wq;3bz)ERP8#8PiW#c8Orz1o4~!pA+Aph|xmH38 zdHgNX_-0kk4yFbhWQ~n76Vl?qAB|E6=oNV4DEzva$Vm3;DOgy9Yy3M3Gn-kx4JQLf zm+Q~ho7z0~!7noMxM#5sK#b3ugjlTLFjQ9$R-bR8D=Cbj{D*;Xw|X4ILtt$&#|Otu zVu{Z}s69E}9Hg~%rYnpiWy~U8{34WjsR%T(dm4C-Stm@H-U4r;3Wn|AtDY>-xD(C;-Y)iR+8IX)3e-VfNR)Lz8KLe7FZR)Ve)PI=xt!nC+a`OKSLCL23#w{Q z(s68WVH$bMC`GXYv!z?mEO8>d#CB{|l-1f)k1=jFsSa#)m9U-F>2Z#ATnHHki(OSz zeuJ_iYUVP(hZg6K)Ds*cj|+$LS%-=;4fv7}&L!+&s-JK0_Mj!gXSOu#FO(|GYlo$~ zoHrvcwH-I72W#ng3Rd_uuf1TG4OxVQW#XCRZay*N!*@I?2Q8Ov?I*XyUyokaTU!8x z2^`W0(tDb_qX$BR6xi8RG;NZhI2aKc^3+n&erOE#(%<~QTTju4UY314SXJEh6dO0C z_Afa$l1$Dm60L-lfZj*NVYbxtefmZ*ynL?{gTp$U=yK+@I>&yaeBpW3Maf3$91c{K z3^6e0m(7=tZZoifI|EQAPq*Km=YU{QT4rQ~1=5lBxVXF@LWYxYJ#+3O8y9fVNR3AY z7$xTfrGo;?b2bl7MUQN+L}d)V7KYzagU=lGJ->B!B^QL&xO%layS#8Jj^D6{nYDK$ zxqUXC=7B8rHdT4+Go`b0&B&&)`a0IlCKH+mRqBNMyzXMitWOXXBhZ$-s47U;D6Mvt3!7 zY{YK;;>xq@CU46oW$~44{nDYlaO<6M75YPg_%gY+yzv zyRFHwSP&qZa*4iQvQDDRx_G|V$|8z$+N_LB8Vd?^o;mia5Af*U9Ixs1upWvtCoVrP z++t1|(B)6m6;CAS!@I|w5*=+xw4T&A-0eRAdZMD6Tg^2G^PZqZK zdRdklhm1tKfONP`Dh66>cDMkK&wfZ(8^`nIa3V4i!BlF7=5x$-A{S-cHIOT;28t<@oMB`U^GOh+QQ6$kP>=p;&~Jw8l?lMw-7h#zIh!e2~o zh~=cWHgY;A;6m+F3x3pTyonw)ACIrCCz9~A7e5Zrb^DNyAB@Sw|7eCuN;HI+@q52$ zn;k4c!}3Opxy~XK^Hltk#K)X-rnsi}xDETLH%*j3B=qf(9~m8r>BxHzPwB8RG$cBELflUUFdH99X*vW03}xZ_G4Od%G4c>oG=mZn?tEvp7hCAzxZaO(ylUhmV-myb|g0 zoYGFyCMgmr7<=#DQ{{Fx+h!cvAX?|Juv-z9dB=jvpamK7P=FUW?_r3?S-fDFc4D=c zuIbdqSM}TGk_ZfXmHAn%2uXd)v;Ibrs7~x5o2HIu;`q=|sp4Q07lSM>cH$l(DEK;C zXS2<{H_m<+)p%-L*^=pX7Uni2T-sbfEyv>3_>6lIYmGTkAnXqQ6UcYR&cn-VcAA-|Ws>d<+>z=A0c%Gu= z_!N)9J6^yX$_nL>M!J$K!rs3RXZux1`G7YdaDxF#X$Zj~*)?@x$FNr6J6n|E`QprZ zOZ#M;pic|h$L=a96EurNmIoJwVD|C$YE>9YL0+COBy3%@<2OQvU7NRnN;TV9HyUl1 z?}*eOaLwf`Paz9`_XgR7Ypi2Y@{HvTKkbXbH10m$m3xu%^IUHJE*z2Xb&9b05f>y! z_gV2WQiw}MnCe`%pN=_I07Zwj^>I0-J>tF?5<8>$C%N8O3GAr28fj_(xJDM_DzPF< zw$aw|F-0cE2Y0%?Al-Da!n1R^a9KZ;dm`)NYKSyNJ9D&$RQ=d7a;Bwv;ES4A9g%60 z-vQ3NiVU7#3|XzS?Y2=r)6+1f#yzcjPHSuXfzjq08%0K{QS$9+t>_=RD2qaqU!$gvp_Oq|5UNmu}JlC z3$+!nzxr!+3ec@M9Ubx#t5UA;sJNfCii&4TWkWFRVspO`?KPf!7Qi81!`uOCJ0)3y zEbPy*r3uMAzG~c+dFh1~Z1Vr{^_5X|WJ}vvkl^mF!GpVNu;4BU?he5r1a}Ya?(Q61 zgS#DqySslUnR&lEGwaS>tABK#UZ=bE?yA~VPwlFLMnJ^Cebw#2H>kF8EXLOXWq*Nk zu|o(YL|H9nR4028I0e*W(l-lOS+_{y`g*k4CN_Neoiq!yZ|ATbx27MpoBXJN54G6X ztv%s;c4SNXqh{I3;tkunViqnlgBbhPfl>DZ{2iaf*nV!r!?K!#0@6&Jo8zYXo`%Tw zSKj25KqI^TrTpBz*VWFPt~U=p>Zfyp55bU->W`!2oZz2$=s)Id`$X~U8>^1uytljc zCOvekB{W-m^_mrL0D|};R(kD;1||}yeS^=w!Z=;Vi#M3n1uxHUo9mti+TB#6o7aPD z4wlvNV#^Fp?wY!ukdllbiLL*o0KaN|m1AnZb6vgL8v9Nl|50GNP2OFm2mz&~Ltssv zEv)C_VqSAKDQ?x=(x&1EEwa|cb_!|6y{7{= z+h3fc5;VKOH}@lXSe53JpZI~^w}!3FCU)PAYt*>5(^wUvhp3F(fSM15xSP66cjnLm zombl+A3d>>TH*aQQZz@oHngsp{gRA~%q6~_C=&~h_~W^ZTgiEx0$;Zr^+}m)hO-nR zA{A`gVW;aS4*3o`0I%yQs_2iV(-Cndq?2_1Uf@tEHXHlhMrA-c9su`2=Z9^+j zV!JGl^DabAEdVX<>(8DjKviw)(UkTJ4RLp`<|&`0kfe!33#+9+(MmEb6Xu790{c-+>_PrluPJCFzUP%gmvp_onM=%roC>ust2!FfPA^(* zy9!K72UTF84ChfUKUN@50;s|}KmqswG~ewmb=CAzFwN^<%nF?c{772EdLW=(-bMDo z8X~N``%2xCQU(*HGbhHj(52n5leC!4>$E6uvVe@32dj=moIy736-&)S&~HhV8h~z4 z&0RPHt-I&q#RGHHIlX{bXu9AHb0b{3S4BCWkwRKL^uVsPibW_O%4%yi`GWu^2`2E{ zQW&)654~7gTO7kxlcggvTp$xySl{E?LyfwVMH~q>VS)`LeO!1&AFJKdV{gx%)eI;!|8QeD19e^KfIYuBFzL zOO&1NNWBneHV}^{8Y){(kXm*8x2_1x<{}PlTILx`_oE)%^(S9sU#}JJ^Z1G;mIX%< zE}mp;o5NC)Widn{f#@p|32&>#to3esq;hn9w>ci=r`p%EX5yT1{9;2+$_;rNzZ0y#noE zY+|HOUrX#^Rpq&Em=~uVQDq^C*l;MStdwy|NfkZsKj5;y*6m6a_JGp&b%JQ^hp98=S$f`D#dlHRORuH7 zbBC?2e^_}#N_9n+%(>88yrCmI4IgNp@e|Lq9VzH1lpoi~!9}w*{91tRC|OLpHkI5I zWCsf4TC#~o(+=EP6?@C*w0g=*0o43+ml3bsaEOTpvgsg#Y4WWm;G_4=K|zitH5xkY zlEoy2UfL)A{#1G-$LMEl>W!R32t$oMpft?%mpGyt!Lh>Tp-jI| zUN7gGQKk6kUPMK=&Sg;}s!Y9F%HQCPLS$9?;O*@|tp%(!d1iMo($kOgY|R`l<@*m& z{3QOFVEMJG?hE8b+QWLr_3d-_rTnD_!|PM2?eQDu_D-K0(9X_6eeX?93YDgYwvp@l z=do9vV_l=d%ZrPn+J-PLwEp~@94kBXzL>M~|6m|J@u~~Y47jf zO6S`K7k406Rrke%5v)@VI1y$drIT9~tl?N5KC-}kMPt^v~qIKAGc3cF|jwDzlW53z;{H|C-yV$ zt&j$F2nnch@G1t?XEEVkL_l|#h{iCVXKAW%)kb_M8vprrHOa(zB1bg6jt==_HGKDz z*(@6n!4%!J)`4CJH|^cl^4iPq$wee|#lYn0{+6e4*n7(Jr^ zp*Ax{$M;%{kAT|anqluJzURg)1T@XXMPYC(9;1Lnr#R;vo<4kg*u}gK6JA`NEB3*6i#DMAD5%wqgzSNIBC}CWnqKXddoS$E(VUwT z&&d4NtqB%s3Op~cao4#UuK-pfjlAcZIXVtSxiLEpycG_IJCxQI;}I-}fpT@X(g!)q zpIeqD8tqs1UeMc%M9GqEKHT~%$Nu=J@w?^W{}+#}jyDH8aXwzRv46HB`U3|hAp;*3QH=Zfw$Of64U(+{ z-)Md19G{lAcP7(hft2j3yT!yDR@LcISwY|fk=F(#?@BKb)uErB6&(FN75+D9crN2w zI2*fR(7w@>tiZ^cPMvtq&0}P14p>`>H-#C~15Lg^;YXXYx*+kuvPHeWjJOIu9gSJI z-^Y#i(Uli&PcwB?Co?+XAR%f@$P$cqoCWH*j`qc-Y~Xi`o%YYcD?dyKMbDPb4}nS+ z8ZBh_=Vk878@f`nr`e5({@lUq&E~-8v~KVsm;h8Zv$oah3xz`$?h$sUa#)|kS53gD zx`WG(2k8weE%$4%*=>-=k&2A`Oc&Mf)du9NjI}b2Ac8HJ3V}ADtJ%k}PRJedQZl!G ztL2N`JyaNv$2_dv0(s*##+J!^+ydhQY@*_?86hrQsEB!IMn;i&i#)}(q$953?L6Wy z*P=}5RxF37pBKMlriuJ~KK0zqYQ#nAgX5Sal?Ul=f31!T#XY1o5i0F72aTrsRn0I}*D+iGu=E?$-y5FbVobu2=__CD0Y-<^WM` zM?ascbfD>Jc23ksaY)6Aju*z(gG}s9^pemu7HS$f&{pf&3@4RUqr{P+rj(zCTaDs*){HCH)%Q26%Mt)G1b_IL%#}Pep z=w;G|HsJ_)HAJr{@7-s(8-9s3<$)tW(yGbo1EM!LR%;@Rso4auC#v1x(Cj8*`Sej z!B;Fo=&jyOK)xz376BI35-gc3?ngwmb1t4VTb3B)wB58{-ovm(b}3R$}Y%+8<{` ztqkRZX;t1b$xk=xF>QJRRslD+IIIYgMKuO);u3FlU(ci7IP9Ha;2L}-5Wyrg5etJB z`C`L?tXe;1yaPz!d#NnA05;z5C>qq7?Ll+98Rzsw?}^rm=s2mQq&4a9b*Q)aYnA;N zy(Kh!R9lCXgGilkseQp()aod&^>&m`eqx#DEE`^@vJ6nxN-4u`byBDk#&v%RLMYeR zojDhlAhOq+Olot z`o`!&marYqQ_mx@Hx(1$rb5T(n9`*w#Ljw&o_{;peM)u?in>!oU*_u7%V2VTwlfQ* ze6ze1)MpPY4CLR%C)HYa@)|)1 z{YmlDV33ylA-X`S(am*r^Ja58Zb44Fwm1{B1zg)l6%`e$ zQ#|5ZAa#!6B21!~%|B z7gk7#Y|M}Wp(Z9z1Nu#26|D!ad8(yY-iur-3E=2>PdTn~Fo_`J9m7jZDeZ`?c9$`_ zt?j8Q6|@&XC?&-=u`>sk%R<|!2+`{s-L3wNn9C{VOKu(XhZRqkf|aW@rq<5wD3g)a zM)o)?xv!82T1F|_opO^xu%l&J!YJ$vct+V>=non-En)>@jIp?vVaFglHY%up`-6>e za+;Uj*{Q1Bt@f-74__!zChIo|)obT=R;LefPCii~RLb9!`;S1yWbD7eHBFRZKWH>I z&y3ov&$wn>KdpFqqnmQ9U&ZdCRvSipXJOVI|MEaI;*_ zYr1?#q&}1KsYDjM0$XhN5yUqr@Y5y_M2xlVNTpKtNih!+x@aiG^#V1%*Z03c8f@Dv z-<4H(@{A=6mA>6Hg0}-z4GaF{e{3gLI$y>rO6Fw|lyX)PWz+)N&SMfB<7hP-Kt(qN z!gJ6@IkGgp(?B&OD56rWq?b-*82Lst>5Gdocdw6K1zl=MnlKN{yWuX(0HWeMc}nO) zg?(_cr9%~mmtKAQp57l=!4GS<}Gxi@(~~u{d!#v)g?nmR~fhVL~-nH#_@JE;rgYR<7#p?yA@OlSgXTu6q0Bf-G?K(97@P)+=s>lZuduJ=fR^MsVAZTdm?q zN!zIvGK8=*0NwO)=y2mNqdt$Ov9?OQ`EZ*9#ZqruWJM-N8jX+lyBo{12wQK8U}VvN zPxw!AR>!O{2y(WonR?yhGC1y*DI+5B%88j-BDsRE8W&Pa-laCA>VcB-_n0Wk|@(c76aOc_FBn*2rHzNaB z$M+b(2b)0gJpbw9x^X$ZyYuit$s`Jm%($|>LxMjVUHHdB!TTj0!uqv3*^UYA4!+A3 zItr#ninS^xwOE*o+Cu1kHXznkj&h0lB>nDf!O6HafwTyj+SN7ZM5FhjC@B`3=Xrxo z%OzdzPy-#JoY{vE#)i`<wuuq=E z8z%KtIFiM_`)9-X#POhq;eK@j#qw93E=5+V2)WbB21lOI;1ToI^vX+9Rp#$Ry1Q+) z3hEH>A;XS~eY>@1w3q>6Va4*>)(KYt?ZO)hg>N{o8rM0rFJO{bRgL6t9DMSQJQ`Po zD=YLwYvh7GkshQkG)wy&75JaqO578dy7zIWBO6yK zccYfWHv+Yfi_N8)#-qvn)AUoIlDaNR1;UAOBCmQeJDk>w6=adHj*B(hCy(Bh?#?|h zK=HK~;4$q(ukP(jBn4S$=1kvL?nR(XtW^)0qMF!X_H@H>t5h%fjS(hT!+9e6MNI(- zJm&7@DU{%TR}hf`y87TL2CCt~%rIH5N35cz6Hz1y7{uR>k5UlH>Va<;Sv>v0V155` z#1n$M2{9t?QYoZl^;03EP&#xwpSFB}L$G<=Z$DV>BnF&&K@vfKpt&2!_!3Bg;9X$b z?jAtZ9r=1!NnC6#zUB8+Kd+NrasVSV7ZldQzgVy}qb#3?uApIb!r+6auW!Hj<^Vj- zDw7mBVe3-@A5K|u1xo;a@`E#Q(e#Y7N8!EPNyn-8^-md%kUL@OMH5jUnJ-K4l?jM_ zm&jfAS{`p2CkS1iwIbAa_cswweF}LJY|T%XU75?7mUfoc9`);YpV{e{@(p+Mh*%CB z8gs8265=O{8n&!mdec zhqs79QKa_({_@KwBEa2=d&`Y=G0cMZfdUaWb0cI71HB#)C!1P5$Q<34`(jh@ddd{2 znYm-L{$nQI^Ah+3M_(3kwj58<($l}H=>AjFzf#PND@L58Pio+Dj_}Y$ETPG8Z5>B0 z^E)F^)va(Z&JYQ!6JHp%Q*05|XpoY&xyBpJB?)3;&oNztY-Pe8B<}(X;ehXvl8$6ZBs3~MQZ(G27dQa$efIeS~k|QGSy6}A^XYUw0 zExlU%;`nvh=SIeuJky+v#8HN$N`=a7rsl)@s_ z_O|-kmG){<8xnI7xL9PC4 z1Zl$6sjNB-4H!7e-wr=Mbg6Q+t;P2%et+Rk5T}OoK>&6cxEs#AkJWYD!FBCkG-*Tq zK1K%}ZE{cfTLszIwTp^t-aQox<`AJ-d{9#8#ksIv_PO5}r8gM7P<7Rv+@^`a+|q6n ze7>)g>WC_UG|V#1G=sb*WCuXkRWUJRl@4^M_5lS?kG0An?!aR z{Xv#NFcQO9wtqO!IiZDxT}Ko#U75L;yM0+|);C2vEtd0H3IhriW95DxMc2WCeRj49 za4L9~Nm%ElA1*d{-pLtpE;ziL{A906%x>3y>f?^_+WLOAFkPprZ3#o;N;8K)4;S<2 z#%H2w+f23wkNdGokA2BZ!oJ(NX3`M7bL+^u?6{d(=`hi)2gXQfC1w7JA@s_?{S!%% z&_>t~HRAY_9YGs3rK=UnPB`yTLKUMAk7I8Ll+^lw@N1D{9Qpzbp+m`cW|0sD{Iwwu)8=55= zYPY1@NXcSt!E}kKUJaHT&p1Z2*f+liR)JC6SNz1aH?dfZ-&55nu)7y*OV-!6P1D@b zoSmGDOP0Ca`yr9}OpYhIfyv}p-2T?08;;ho6H zf|BF>>k26)Dy>5&2$2K!HczFnsrL(fhL@9qBPc?gNGF*n`tGyKtsY7D)6iEM&F`6Q z;T_ftNhvk%StSo?jQs-e!lQ^Un#r{JhIR|BG%1**FSsd}{hKIJkypG2n53Z8$Q^rDl zU6o+p<5$7C@6U4Xm9E_(wfC*XV}bkT0R->qL+m;;M6ywaK$Y^{pCg=I;=t$C>;o`9 z#3M-04WiQKLV36K8^&#!hDDxl4wM9C)QJ0vy)ieFS_|yvp?5=}Teg*og~__eq9vB@ z7=1dA>d3 z)w4Kc6BAR)4#+k2yzC%~boc~|kyz~&IUeSm}M?CGtQ`!@KCaL$`%%6reY?FQb2CK0k_&z&9oALUUi2O5{g`1uMoMWO=Z z+g{s3c4FF@Aj2U7Iam)@!xevU15rJO^A;KluQ`I5`-cvPH3IJoqcQS4=Qr z=Id3&j!eVSaJ=_GOuhGWD~u`|R`WdIVHMKpHS;C0`0cO|7}Vbm5^`{{s=;8GfWKLb z%ongaPdAQ^cvn49&!V%u#`ja<*9wBpxB>Ut#Ut`MIXU=dqya~+J{<6186NAZnCl`$ zzI#D8OWCjxTf)*6#qGJustGW02?_hl+*=2ZRo4_-ujEnua>`T+d7!*_I#)-NM|4~% zBb3Vh7FS*AW$t)SnA(NzNgnx)$LJ`ENs+Bj2MS+gu+p`V2-OsY(;6-(I2?|S4mcX) zl354V7RtRJ9((3SVz~hp;)v~S??n_Ukd!a8al9VY`L81I5hle&t=~Lvz1kW1z3!`! zeI0O1g?9^y7U>03i|JfY1aAo-|Ge-9jZEThC8|Z`SziDsc@bMKNuhbDfQO$|sh^3) zh=UdI%I&HZ-EkQr@vua=S|V6u9NcCU9^0MBAd|PolHoS3{3`JJ6n>tieWXtBjI9w| zSvppdU8sjlmpR^nS#YWFp%I&C-e)?PhU`XOxCr zN?Htg?P;w1%I{WJ^x||pMU0CCIhIl`okdz4P-SarRPV}*-8F|T;E}uZS&U4R?`|I8 z*c*;M8VpAk+qzi+>ckNnwBESlPHn60fO&m;J-tGoBiQH?mheg-DS{G7hER9wOl%uN zWR>fckIT0|wEP2nL1K zb2FE{`L`R)(Z1iJlkCnOgnNa;esH6KZ^jZpVOerLe1oMFYqA4!n|_ZNtmkX9*ghsU zBL0#-rfHwTlK`@$D)NTVsf%*Os_$N+ybBueW85#wtz^m+Bea=#!RPd(m4#|_I(WF^GtNuVLR^FuHsXex^Uw((vFM$C&#iOWy0n2W2dF z1|8-kM!*(?BX6EoWXoBD1~4dJuLy_qR6kh z{8jVM?iA6S#d6I91L@-B`Po%5nzzwGp}c`#uIJhgA+dsWAnQPivk^6xX4 z9^HFadbqnktKk|DBT1T#w#d=?Y<#o7Dpt@L|2jBxvHkMA56Bc3TTi^2QyxiT$gVaV zpkd{Hg^`?yWr3w%86+kChnht9n?azA!~CcI3`7J+)#U5f_8W8|(Tf9Wi5=|3x`S*m zY14Fj7EJ&pzt@t=QmrfTs{14V!FuzmYvwlTLRs>(O10HKs3~D#oQ`%MrV>qeu%()w zMN3RD#sx|LvEPGosoqbRxXNOfGs(z?ROI(ety`YcLDM0@;&nJZ#VgXD>qfir)RW{_ za|9PzHsE3J#MmxI)&#N|m}1HW%GL%GY!yc!RoqD4#Ry(uQAgl5K-gQ%r%)r8>#d`l z*UWx!cE*%U-utLX(QC5OL{(aR*bJ+du@zLzlA+t0I_f>-KWACl{52189q?&W?V9Kg z6m!b+9P(#ru2e3NRYCS0h{Bg)>Ue6Q^%%g!zIR8V*B;{H--~CEc`CrRCEjfeHTRiU zmq^{Up(YFea;AM^MeXPm^4e4#^RnFzx75KT0izkLILdkxz)H9O9c~}FFn!6)qAW{W zkTaAk3+&iDWRe32E&2@t-b7`pE@bi0GIJ{EwMkJQR}z)3&UHECsW6pGPdfU$eXg=~ znB9g)CSJbhF%L>Er1kp}gW0d)UP!k2WQJALZj7%hQSEE^xR)iw5%sOvsTC#4mEdnQ zS&ZNlyV^Kenv=%AA1}L3E4~oK%`7Y!yhU|6U6l0kxO%7jg>oBycU(1wRDu%V{F=|? z`IYGTary%S6$G_gh}i-EX-lZlVHO<|;A6=)<J4OR;>SS`b0W#f=wL3v z)<@1de?jb*tDm_h%NNZg(!~CrgfegEwj4j|wLAd;_p{*L@T%X}2l_{s<6*^nNrb&O zkAdw9n#%!pOXA_=RL@T*Yb7%G{LXW@h0#{Z6aY?fp&W6$sWnW~^qkUU0j7;(?y~*L zD5EH|U*_gqhY^q9ectBKyAQJlP*@k-^deaXBvRUqrz)di8jSNQeWdOOO&w2T=%1pV zi9W=~%K|dXj}MN`^>o|&SBPZBDEC=P`NxF+VIy?HgTvCdef)~}H{b|?6#^}QEJ+HZ z8R;^7I`PuKUDL>C&u71HcQW&ybQ$Ef4AK@vONX7fkDz7Vpv zb_k0?FmdWWho5a6-TsvKZ3$aAG?>cB3axr`r8Z)ZXlbAqZZm;IG>nDfwN5!P|9q;! zI}$Yboc7Vj=@O}6&rb)GL!49Ufa=JtPe~N0@=}8jm!N#Wmx8m1+Dhf_2-ADgqZY)} z#hW6$QqMSTz1<2fT{u2@O>YS6o#PMkyxYCua$zWxXdx+I@lci89$L<;WNF9J?8DLS z7l;RqowPm&?3?^Tk?#C3^g)Mom#p`0W}s62a}?)%S)mt!wNL}+{%#SJ z8fgGls5GHv?6GVaDodGh@>uRuGs#zE51!W*)+5b>mfG6Xyw~pMeth7lb!@0fR^uB~ zQ1a>q2&l$6^95uG)^58dvKymmU~Jxo)-%cP#!<E;7}Eoy zH&Zpg&g(uRUAvuNobP=vjvI2<7bm$hV?hgPmqR5VI#of1cpgXgF~O2?!bW8y{RhzF zE5ryz+wpvSdiRPKTAmNJsH^g^8*6^5tZ=zLeO|IIo4+_dQdbA$cQ7!1Mkt(VJ0*eFU;07j_T4NB=Q9?Sf*4p$B6*-}Du=jN zs-76-g^esuf z2J4mA7xjq@Hbx^P(qtcMULW1=XF1J=Hr`~s>PlgNhs2VIyc3k}m}KgP6_%4hEqR9$z_i56 zBXtsHMMx_0)-v`2)<-_9O8gt{97X1uU+B}Tz~!Zf_ya@cz?5w(&cM>V&Y0`Tl1wW% z-fUa5;}2`Iy3EA6o)65_^JJ0Hd&-7x+BlWcKp+10_78T{R;#g3L@PgTHFC%`8wujd zyD1(9zrh&}b)<*g)Z0&UV9H5XB+(~Vdi(6N5d5KQLM2JMP)pVLYlp2qdfOtPs9l^c zgS!pkxdRP&s?wWgYm<|cypKmi|Abu5}2V{X}m$w>;FEgheenkTVnc{PVJ`VL#I}jRZ_Cyxja@J zH3<0^#Kw*aB?9&Oruiv*S4VvNg4M>NsQwxCdl!-TU#;0~8ItbqTh-Ex_b@cbqtL+w zcM4;_XkN_lX`W@puS>tHZ&V$ig!L6_MvES&B1-j?z{%u&(5Qf99BUeuGl*eqWDX>hc-p#5Sbg*tGCbLBmdx~JC-SROb8Y$br&-R8Z)eJ5gX-2R|noO)tt{Y%W> z{J_7f{NX4e7~9c+-;&l73r5HIxw^@H=%&#orY&fDVPZcGR>&X*okIoU_pA+a`X$cI zdA5EN{5z!kmnjAFLnE=P%Fl;#_hbAOedzoN+q+3t7aRl25G}z7T3#eh+6M-IF+)}! ztdkv7SY4GipK86iE$Dt9FXL5Q>ITufND|a{NA{bGcthRi}!1{O#X=4-dT|vD=c;25;*Iu`3vC zJP~*>{fE;Gj1ox{4UACwr$1n)Eir@3ol)*tQOf@@&?&O2Pn!i(2Pp2px%3SMSd2LM zdSbLCEd7T)lPtR5lvY5|dRoOV%+E`vrxR|6LTvuQNB;n^eqH*Z3TNf!ray>~jvHRr z27+4ZA^c{NL8#B?Ut+Yy9Q&xi{KTt6a{2!lZ5h;DxpvYlH1z)%Y`)CTa&}9ZM#nwe zU!%nc75Yj;v`Z{yECU7+R?}XOgY)ZW1bqm!zKyeN(H>W@J?{UA;xF#~%Moq>5u~N< zC0>csRmAKBBAA5&WtRpE&WwKeYz$BIYiEv?Edw<4FNw{UBhelNfNDH`q{m`77S3}0 zA2gan8OfbSaIVRY)an~)3;{B3@Kzu(a46L5xM7I+rD69;{E|lN$L9zz5LZ!sx zGhHBeZ-e&oKLq=&l)-Z#;&7|pW-aG-qHSXj`|q)uLP%i1f2+nByPYUMMbT)A5f=D8 z%FbwXGjsFoxkP_aAhOe6rSBi-B)oq9Fvt^KkQ(D7@^7*J7+FukFBITj4Rmj zA5qNjP%GV#9Zwe{GxPJNOY2bot}U{hU?b`2->G5?!Uxmf`cl9V*sM$;ah{_^*$bAOhE{)vmWtUFtlx z^e_*&fhV>Myeat|2k@Js_C;=<0!Y5N!}Q(&;4)~!ebk1)LLNW-x80CP{{IQ6N zy8oT}w+cdY2N|9R@5T}DHQ}!|*w$X!>frEy(Rn5chD!f5fk)JUqJa?g_)YKku!LbT z93f&W8#263G!ITrXkn4L-u$Tu|I56eNo19l#=TS@nA`SimR;1le9y;K)%nrTb_K%$ zn*MD$67T@=#Z?24jyg&FEcE+h)b-+okrs`C|b z9riB2pgbUfr+1nOo2_h@&z0ny3K>Z}x38-RCI8o#_;!-OIXO8^999B^*X;lodGlNJ zy#C)A7IIY=xG6 z6qo>4XtQ&3Q?ZmXW=Bn+y-H~z*wnwi?SDOfCi@1{MAp?$g0`0zRWjwxEr!=~Xoj)I z^1F){)N{?(#Pc3wN&d-PH0BT}8u3@h!n?oCFE+TrZ-Xi&k9(g54$bw^BCB%9K83z} zcz7_oKA0_0XEhzmGAxiQ?C+PTU6FMC^!8s!0D~Y8^eww}c6OfZ*lDk`yd`p6sX?M< zf_rCagP2nxhwl7&A#c~Mnbw7$$?N|+t4*lBo!t=#7k?U>B0^LjwaPy6H*%h+6j1B; z_l`xkY9HSmEy}dU#dQ@FP^GXq(?;O31|A$(6qnS5(n5fHu>Lci_n(EzuZB80c;Fpc z#7A4(;H72Wz0Q|sr@H4Nyhc68xUda&1xR(YDEAb=xjx_(3IvO(Y(&#LlQ68umJ1ezo8F9aaH}xvRb89NiW$n1m%5IC# zT#E%3#)6kT9Y+}k;D65hJvnGeebMMf|ILX)M6iRsbgADvfwC4gi7L9$NZMu^?&pPG zDo}0pmDFGQ(6shm76&RSA^Q%3ZBYJ?1^zQdf9M@QQqrvX{5H0Vx%jO|vHhjD^Ch1% zJCG7e(lSmD$jV8Qc+8RP{*MAm39#x(nWMlg-anXeEF0q5NS63G&8cG#q3rA1p?Tvu z9(lJGKIP1Yx=$8$Q7Wk>xQk3MnwaG4v| zFG~(|_^r!>mj88cm`R^s*Ekg5g+o!;NzHT09pZO3(%xRlHyqpTTZb&7MxWU7E}fHY zmOxS*=jBL0jcF~Jyzr2`vQ)D9@7j+A?X301sKR=2ZaojRmVPlu=P)UTvODz3xjxT) zsq2Kgd}95xZ@cIkK?%1B;@P$tJ2UINJ0w3T%x{HikcZZZMM{1>X1Ma3am1jE&~mBG zl3P$W%cTElxD|LsLGLqs)qp*pq!2Q-GNU#(>vf-_(hl3mf zEhg9dU7nRdwYryKSElVtavX$@#>Ub5su)|xfEb}H94L}18ffFX-(QtH+E?}knvN}6 zFuDu^cEA!Uv(6wcX=9cX?1(j}ZnA~U&?TU4(@y#sbmoIB|3)|cU@ev?T%RAAa zLS=-$VLm@UV?>IDE!*KXP!X%wR1llSeE6m>_NkCo&a#v|?RD==u}=%JjpS+|W23x^ zvz|cYT=h;5vFBeGioc#?1o?VB(9q+9u`N$R5c#pFaxI>1c21?5BrBAhNks`P2HTU-`B?sdfK@{{M&N;3YLp@gS^6IAE z4B;ACkQGxW7UP;N^>QRLtj<19bUYEUB680X1N2k}9E!ft{`a*345E)Dx?#@9mbsq7 znCaY!;n1MXjJL=I0hgE#p=RnZoy8nD$_~F9-*XM+b=aNR9vc>q zN^@4MiXp%5D=u|7^9HrEQX%h)Mt40N*F$d=>XpdL$+3^7bArMeJ4#x0PM_Re?cV(N z#Yg~_A}N?}{;W1-eYRL{SE+-QE$g_q@_J=|3E1Fqq|uuHYU!c>%|m; z<=qekTXaZ)DpYhiDYXxe0&!o^^~#mGn_lI}^<w?L+~n(Xo9-Egi{*+33i zo==Hi<@-G5STbPzl}Z0|E>w&*ai~<{X%FKMYP z?>6V~*z7zyqDM9F{mgw2T01cOg~A)FqOFA~gmn}VHjLhCAjF6ZzHx(3Bxx{shxCLx z^T7Zr0#o!4^OS}i!X09-TA#)3g2!*F!O!=yvPIV0Nq62x8W86--;n&&{P3>lYo+If zWnD}kCd+Q3bdx>X7@`@EB~+t@(`~)U*E2rywzN8{Dv07LVQ3&&b~B#czjNNc*-GBN4<9I zCkE7TtdNfnKuaRgS{S7{=S!o82E5~aQH+N#6vRE%&!?GpYvrB9f7)>iuuXCsRwf5i zHita+(+b>`Ktou9~rICH5*-^XXvI8LUbml90q&I-qT)rE^;LWV>}2d z=f?iP^<|Av`noYJuWb$pj9FJslFqOZkCf3lwNxeF08s8Igu>wtB@v2^d9EAluAT?M@0pFKU*4fcJc^+ zw1*#PG$3&k7cl<=ib61i2uPVRN>^WeX*gd(E|gB|K4n>-Zj_jxmJG3rDVDP?Jme+1 z?o}gRl$x(4COk|M-U(z+jg6s$mI#YpAF@Vu3$U+S(irdSuVS{y0}{k$(oVl9{XA-1 zJ?OK@zoe}Ex@!AV3JZT8VP%X$6~^?xYa0;odvi_hRjxIADEII^o6h{2;3!Y0GQeD$ z`@pEIeXn%uy{Z+~UWX z#q}7immlAntCI3g#b4TbrlqJqa!-L9JpJ)j~;?BJyju{gS@JW@2qier{q}-BuTY#XinX?$y@` zQ|;M%pSx1r<;PXK1=;6wUi?&2bi91EkEHLeUcU3T8x@nCjeC$l2| zn^I4(IODywH`-Lj$xtYB(y8u7r_AQp8|2FPEpg`7_li+wFgLlP`o~nbm8%EZRNc6T z$Td{}fZ#h&lS;p+({)?*k%?v{kY1- zhNIALt8{(Sp%x~pQ0TU29M~KXxC7#c>cM-Plr+r+DfB(|H#^Ax858*fYVwut)KFt) zlG*QVq1w*ZUp!(Bp{tplJ@%-VVVvKy-~tIT`h?|=1tOz8n_82~)4qvoD262bU#`ag zM-^cqSa658qat(;rPWstTDEx+6|9b>x9wfX1neZzMjLN-69O9Mz0R~0ma(t6?C#!} z%ZsN}xg{bBi1nc$S7ni*OP4Vu+Ras^W zT$;DZX%ee6un$;~K$W9Wl&B#T&H1z?ljyL^teoU|$SanKebKjy@U|8YoSe{hOL<8w(1y!~3J>ui}usu`g zfO}pYg+8wMpFpDDS9H4xp}fZZ^Gbq7xEM`Wx6fDdv*2Fp`-4&`;76n z2gbgrFS#1d@|Pk8o><58w~HtK`l^VCY==3u#THXWp^j#4VpXvlU3lKrr0~pWv#Mc0(+KZSXVBPa@;~66P&I*tlkl4yY(m2gDeb?kCX3iI>GeY z1v7y%5=pSOL`V)hw?RIU9=5twcQD|j2>*5`$_Ue5QVogG> z0RTQN!{ctIB1=pB;33s<7Ek~_F1Hjrz6zxDY5Rt2GY2(1o!eV@SVB@lz?I-~hVJON zW2w$T)6C3CuX<={mUYE>d*{Ezc@N34U1g&Mb$=PjI|D7l<b%cRhl=RvL(WW@eqQ(**w(9yp2 zaJ64w)L;Q6dgudbFXSTFo%Q@EkU_tm3B3;iou`;?XymTGyyRzfWCO!^ZWa2FD9d?% zD6t~x|6}hf|Ki$WwOfi)T#FQ!BE{X^-Q9~@ahKv0FYfN{6nA$i?k>e;uy<(BdGC>) z+kfElVSY2r?CiagtgIyKNtUN+UoAeiV&>O4xR%P;56Rln)0g(u7oAT6oDfPfo<{7U zsay$cNelV<*N9>QgeXTxN?5KUpsx{41NHbo(R(_)ouX@5=%)$N@}lP84UHhlyXHlN zxVzEUm5oXK*q3c)-BCiihrlgc_Byv&rG7bCn~+Zk^T)J~%v9L8kK)xBLHwCv zuds>e@8`eXo+fcC9086<4}WG~{NaHX$$zw&Px?m^e-oG&>JDMS-{ARdv|~bI>9Is2 zVtF-CkV=!A+Ll*Cj6PqMRWT7fpE_iaqWXt4C87ap(A6;gb*2eDN!hDn}$l0P< z0z`jR90yjupbxWKhDEd-&EdQMzS5aNw4E;&?1^hY6)^mPp~wbKup z^ACKX&MoDGI88J3$s@HY-@xE}i&Tik9Y#-4@O2A;k-IsuL-R;si6QtyiY;_W!*}?p=IeCwviv|YU4JLk zTOo~XOv=jDC&kKUL50-lh>r4Tu>#Bj`hr-G9Bk+N)3^ZKtbn)9Q`L`iCEPTlHgqf? z4z=q?Y20@vqXZ#OEn7-eufD6@!uhmyba^BHkw5&Ox;;uA5F{Tsd&&0<6|hl{rA6T6Mf2k0eDZE*G)> z28F1`Ad$jTC9NtZKgtO3w@~K}1B8MBVsFCixls-kOd9Y9X5kO1zh1eP)tNAQ%-4 z*R`m7(65psuUWm_V2D1xi72Zf(bw&+Nswrvi)mmTQs@n-Cy;lqW3qqL**C>y@oBTh zqtU%rF*P?&KR$lzvcI=ekaLOPzWeI?TR2T&TwWa&B(F9-m1QQoAA3*0vwbcjINg)! z-hD(y-&D(08$%wHklu~ftdJgB1Vk^N-)qg?;VlQa(2br!9P{&7Z17TMMo{r0q4L5R z$%@uw36J0Ar3jomvqwN^X-PBc@0)B-I9I!l@`_#f8@IzM18o*{H(#>?`y(nej^$21 zdefeWoL{tvz&H9kW;;X!ZNE0j%G^LU^_#n# zg)`ch)za_j0gj(!Np69gcHY?XkrV>M5ju(j+D6PJRu);?GQPD|%hJ&wLh}+=rJx5x_X)`C=W)?@`>m1-l8&qAK2ft$u4lq3jxq5qstG zad2rlKxu`A`n@-mn9kZ_ZXh+<2>VC_Lc`Cn=x?tw{-B2k(k=e^vtQ{H2DZ878}wB_ zzU-->oP$@z8j7Wx9#ul$!ULY_%fK`>*~<;^KvP|meuiZ^YPz#3f`4ML1<;z1v9X07 zaek0U66?ibXD=B#S|3|u+K0U?SVvpFQq5Be`+|R@Hd$Q|T@Z8h^f*;DbNF({D7h9J zB;=(+1nqMeL#uI&0)ToB7X(T0kmEB@VB(Lkq=zSktYUMvFwqkRx(L)8&N4}zuJTTW zkl%O)d*(5Gp-cfpKXKFq1>h^myd+TBjL;DUZS~m@l6RV6Fty6huo>r?zv_djPRYdH z^~q9~ND#;%6wYt7K~bDhlK0J05iy2F{D46wXx;7sav$8y8U}_7M*E#muPbA$=~9Vc zk;tg~7S({eJrT1ozc?G?ns8i!$lk?gma%~TdZl=bep$q3T zbQqYuuhk{d)|kbfM=-z2g##->EDyw3FJ-9{3<{De5Cl z+R!1Li1B=a;XO~5niNH-VF>sx#yX|56=N9pfv`|TEh0=T8B4}gMl-gQa)XS4#s{O; zbfx9}xzxv-Ho_}Pz_rzVS~bU9qX2;T-PEi!Fw1JPRS!(8i##XRoo2_3et31@z|3{D zqyirN7`-5T-zMQK9d90~${ZjLjZ;5jbk#oTQ-i(g9XrO(r(z0qCy^)%jCD8pS8i5< zAMsh+6UEq?JkV>bmPFl!=g9{60s#HUpXHcX8*38C%~W(`mHG8QyyZ_BzEg?ASNH9^ z>pUSqM}0Yhu;p7{Qkp*C4-V~kapH&3Q&3ouN~kN5?k{VvFufr3kb!YI8@`6a`?*9M z$sGqJ43uob&6g2--C#EwjYb6wbVf_GB@kBN+o~W2z8*vo3-28gcOxwI<08NiP4OyLL;3b)K2RqnNm-?4y$o`@#;>}`>n&M^ zA}Pq0H|1{zwbwgW6%@84t`|FSHJVWeq+{_*o*R9%4#Bvrag?1McL=r%C`cq+ z8ipH9zpABS>~)+NnwzdJBkeQ9zan<8A~|l3OK7eM=Bk@M-`L?rgf&AYjDvIa@adPSK{p!Z~+`m~}Ose!MK@bv_1_S1$a^<)jDFxF&@ zvPNzX@y4Wy4pw6q%ltybbsi^mGkED=SAsv@>tG;>qA_$w(>MSs({|=IEZtRAng|T!S*7k5O`AjE+U(~;~@I9 zoajLESv4@tUBZnJY?5xO80x&~Z{rI#LpW2-bKUqllG*sR610O_QfEj=dR#r?VJFH4fo zoaMtek0R+C6z^N`H^PsB5XU386C1VrHOI-0Yi&_=TsI%KTJxOS|9mEA*O`UNPtc6zD0s#)Iq=T@THp zDhcm%m{Saz(V;*GQh?FjmPY_B*B?fYBZTNP>Gdn2FOs*Vq)Kz3C|XD0c06=+4@>dA zP`UaG#E~Y8*c+rF6e?tq`8wlZ0pT04{3;@d*`cj25g)W%-b!-ag0i6$E2+CE%+8)V z;BTT&D{t}EJAcOrO- zcoMG*RYWTa;8}!3v|&o}!9WYV&<$P0^8s&&6+vU0>&8@jo#7ss!c`~SP!5AyY+>@` zxa<(Z#rPnieRPy4^oa#A%~G@{9SrB=m$NuPKXh85B4Y!pOM(h|uH!sJ8enMwg zPTHK0GjpyyH8*pb0|Y+f>Wn~k(p*U-F}TA#KJHM*v|DgD$35QK;wnc{W>#Q4!{p=F zgU?5}_=70AtoNt-u~XJLFV0;(fX8PrPlkKCz`LL29u%a`nlH&oTa~FQXj=(1h;&8a zf@t`7?vT7q{zB`00Xb*LA@lkt$Oq36WR65Qfy1>V=?I>BQhML2rusw9VznrRsHRf| zqM1e9u=ap$EPVpT4q#xt(JnajWnl@%&D(wpJ$FAuE#L5uo)jQ23$fS}KP0JIv)gnmrpEf6lI^k5EdTN7Nx-|nwgH(pk%pXC zzDJx?Z)zHO!@)i;d!<{_0nuAiL-!~Y60%qi`rRbKD+|)dcf}@Lvo^?Yztv%k^MMsu zeUICmTZrXI5inSr39Ef`I8M$M35hvFw2naZuYe4Qj1mRup@S;pxBhhF zLh<`ep#K1yR9kqT6M{Yw6Q?PXf^AbDH>*o5njIoe&8D3a(zlNFbjp3PK3Ha(t@{X` z!as4Do!`=;rP}43WlV&C0c`8lu1BZ6QCl4m#oEb`T0@MqFV}SSHDc*+PBHRjIo-CrN&4 z=h86OIT72{&8>p6B7Z>G)!16^%-F+JfX5OO!6P4MC37QkcOzlw-);7e%0-3hSuZ}#D>`4;M4 zhj%q!iwesVi&|TmWK&77R8nnDCbRis;s81?ACspxz>OT>cV&b<*_#Q9vS^uWhkEV4 zK~T` zL#W-F_uQ{e)kw26p@JyxKjYNm$&`Y|QYH&)13E~Jfekx$x0JQ>w4WY%ZS(1kB(mUv zndHVDO*4Zoh%jz# zHXNGKu7bLJiPUu5TvWo#Fon;c|6B*@L9_(W=&~jFAUID44G6iK=7UP48pB#*^=lo9 zYqwn9!&ecz_z~R}ytO=+KZaHs%0J>W=)eh+%|9XwO<+ti6$;M{^o8`P7yc!&nyrV4zBc*vJr>zRv z_@YoG;~lG#U~;Tvss!o=sl2o%b!e$YHva?f{L;b*F=b&oMkst2*$i&pa_12Wmz~!g zMzm#|T@)EQur}YEwlbm)szSI5V+aozmZmDXx8FOP%u1mnwd>(8sNQWG>=1u%Z_R;` zuAGi!s*n_65+nNqr}7i6^5d0^8AQrV-lJ!?#zTlbyaReKuK4;nF9`$|6w5O+P072u z8YR}RAJR+bn3cB7;q(#x*$d!vVNoGAp?qKKHO2ep2q_j1+NHZ9?y>e*B{sKBGRAUh z*ZF3nw1Gqho2G^q%TNczO|l*lJelKXM8`Gz=hNrpCr>J5AFk>rM5LPnYUDDBNJ;~d zT<(Oi28k(&B+A!aQgWpmlb0!p{nL5TWpX1$cn<+pHvHr&8i*30WaN<)H~Gh;==QkA z8HN6)Egvcb4-knh^YdTJ)RO{3tRn?>m)FA_adil+-nhrMInbrr(_?MwUJA11YAEAS zXrjP0ABJ>Ru5@0)=waRiDq^N>?xc?sSm^Qsuf-Zm%?Oro7Ky@mq$ZW!ZV8owy8%37 zPb1mHig18BAip{3z&f6NT;5E#Pv)e?(mPW8E{RH(>SIA#fmBA_Q+b^Qx>Iq}I^9>O z1}I5z41!m`$np8pdP|r|Hoy&rkAAKNLsB*{pX^MBP+#Fkrg`ZdMOtog_a3nJv@B@5 zGbE?(eSaBAy4O5gnmt#QM`7bP`9c62|H(s3z1d&O;|P{5sLlZnG&aiZv?MGGQ9ZiY z-fFIv-*A}Ip;i?IxGCF)w(9aP=0hPya8ETs!wH4FBB)-yUv@7+$uCUzf44zqcLV9y zi4T2V`!--%MSAHDequ^KJkJG=eU`;J>RoaqBnG%Je-l2Kci{^#lCZ3Y?+koX-JPC@JZ4gR6 zA$K@IvL~2Mx=*HJ0iIh-prVqcQx7G5h)54CXrjk-`e#>-=M?8XQlFs(nKtLif`if> z5~}iM2>QKOn>8sO2$bl8<8f{h_Z2UjWA|P)VA%dbd*y+aYWs+v>nid(2ln)l`zch* zWeaxs;|;<5LQ~r8P3`2!((nba@Qhp$9K!O_ru;HO)pMkHZ`+w8RV^JJ&&RJHjCAW6 z4Sv>Khw%d>zM0}BQ3-x`vJwtoMgTKS`oNq#SoM*U_hR1Dq72`qB<~2qwz}@9Tlu|x z)6H;q&u$35x{e7LF^GRMuiIr!&G{9!EGE4x=4CB*j!J&`dL+iQWGYV#BjePxTVQ4e znQ%#d7LOtEbeKi96HCX!=W=c(?f}=0;o%z)81|#@!8E++GEWvj2U`_PcTz=z`1Th?I{D|DWAx08JdT7W87U4PTwu0 zJBdm?d~rWN!PNkiCS)3 z_V&#VOC(_i6370EAMf>{F zcXSYwucL$t?DhHM*bOBNnp_MOpse1OO%|f(ZSWIy^gLbGyQh{0QDqyyRMo|#uQ2PfH^C@b!m3nN+VvVo29Y8xg+AKhaFKtM>uKPH zw_kAp_rqaGp$lvXbFNg1d1LSyTZz<-V~Oah$_QvVq7>_j=I0+y@_?^bAY~KLc{p>z z4Kb;wB@p@1x@b|5aN6i6@Jp;gx*Y_@rx^WO@>cT_1x7Tp)6Xhxfqw6s8T#TNsSeZ-(7OleKHlE z9-~!GmEIW9~i=PSl zFfmSBEqyfjR*`QX5|#_vu~5Rqrhfl%%hk6nYXqD6&1Xi zEsX#www>b1>DyQf_WYd7awd7Gzwya^^GNTBosb!d@} z>3-tOor&;}3*7{)P7Z`gj47jr6EA$FFJ%{u5~%mt);VzXk-3%9KBm0eDjJrsKxrJh zu~p;2stgYHU8GsVqdI58?{}4Lk0jJ_O!ll>M-Y>e}~-PP+)9 zWya|bu6M>(!t^ytxY;ADS@}j!`1^Nv=U;ZiW%1C~QpK!7<1Q*yM15Z?$$AB%&(o_w)%38&xh7W*K(p1{P5f1kT=l7(;!w^%K7+Re*dAN&LwuNOI`Q6EEY+sMM>%qNs zMS-h|uJgi}$&a*v8VH?De*jrq87^>D$;cvAE=+UL%YS=i^davcNRg{2zJW1oYCad; zl{Cri#`^4lQ5u0X2X&uJ-CsmZCc-oLzQc)eC{av&*%MZfpHpFpCDjj^5k+qch)+eG zmM+ZlZ*wvvzh)b`d$K?N#;g0-`PJdj7?bk=!xV_K(*J}yrK~A_hs5jR?}$_Y4Q?Vw zDRm@WnsuouTtKWKK1r@W)p|}GW^EI=AATX1PcJ0Qo<&()?pqWM7ku;SX)pozdJ<53 zwr+$716kT~M(1);)>Df?Ky%Ya$~}l2JGi?~@m{%kzPdZm6dwfQU0rR3_2Yf@qsJCZ zS`Pe5@;m|mz6jWnIJVkD&K=wP;kX)YIqAD7+aqZ=(Z^gW^$7qmQxpr6M546G=zMNU z3pb`h>@`I$@J>bh+1?5BityEan%c?QMRR`~J`7;j`!Pk!aQRZSm|V zhW|<<9sG6lJ;3oig__&)F%ll6)31%9ti{CPa&wC`{~Y28P@mw~zc5Na3Tz!QY|E-z z>Ew;1V&`~0w%mL-J%eY}RFj&zX2DkqjA06ki?h+S)FiQ(-spwq^ipU0`@5hHXI9dQ z957kp2RY%JtP5x@)DWLiP!l1Lkv{M6OfGe)JR6SmKu`maX!w^t+T1Q3xr77;Se2>9 zIZ#yDQ2{||r9}&Uy-qb{AK7!aR|!fG>XWl9W6j@mtOo~!PD=uJ&&oX5(|V2Xv{Q5Z zzBl-G;4FO!>?IrU{B|9nPgxi!)A@yD-^wfpT1jm-3|De`1Cwmz(lxwV{<;ONP|F5s z`UFeIg-n2+(8hFbQnbhsC#NS!(ym9ecg$CX_{d{}=AlFvyqjb+m%rrT-AqLqc|A^+ zV`H6=^tm&2F(1TN?IJI}F6g8>b7cZB3sN5#s>U(rtE`*nXw;R-{T3!I>-+s`z=xLU zI7m8y*SgX(kGphKndMjfX$M)g!@Z~8t(?i*LV!(Ts8>fu=j6N3dEbOc4VfZs%;P-*2=%ZSj#>i156>YjRpH24fHJ|Lj z9u8Ai#`>EW10k_3*F=zNYHI8vvNS~!X*`nI-m<%re9zV3f4DcSn_X7n2X_9*N%?v; z(T$!uIQaD?n_YdVhwam+0VR(n+pVo1yL)%&t<2{aK&{|bPfP5bl+wxteG_GKbAqV$ zRX1AI9~l9o6c`j?lsdva{HW_f!ox#Oa0`U{cuI)O0-V&mhADPAK~x-&R1Bs?@F{WO zr)Junx5nMZUB0Vt*kCApRU`O2C0Vag5^av(pqL({A9=d_=cc-V*dMaB%^pqJdet4X z*!k`Q`zD#zmYC;f=`5iH3umHiavvY`+GfKj@he|m%PJ~*^%U>)LEZPdGh3mb4J|7b z29p@`U6SoD+r=+0BIJhkdF+4S6%x!3nSZ;)KmO=2l$kO9fhwpS|V28cfWA54S7&*^`S-?up`bXW!KjFk8m9@U`~JIA$_kYQ0X2>glISv z&lGcY%+530`aOkm2t#}tJM>#3W6}R|mD<^m5)cI6ZjESHU!N%|K zG{YAr2-;qWSszJ~^GE==`Gn;MS2uZ46)T@!n(CsyunQSNOeQsDJ!D{-%<3nEL+=fu zIIWWy865FYdlH#DY=Jr!KwQEAic>? z)n=FK)#qEoDBsQUoG8V{jX(R`gCoB4lbYb)Adkr#40f-EhEDE6M9Sgy1-0yeHuLO` zYoZHcY61*SQJ>Al9=24GDMe|vX|zJH)BIe`1|lW3U5yr&Utt~bamDt^06 z?zQG4?{w2+7N4+xAOuCe?yX*8^*s^W&Aw3L$9;lCmjS)`>j^bz91cGH?UD2#@ek)K zAgCyNP!;u={%lu=nVn{Ja z9UUl?k$Xb>8TV!Mxv&OQ7Q91^N?c<2?kr8JC>JgKx_?JE3vEx8jX9oLc0W2eN@~Unmet z`f%;AALKsp+DH^+29|}$!;euLr%Q0r+E=$8!4{kMO3&VBXN9cdw`&q{+ zB|TJ-0nX}?@oex!L|T4rampGNq69kxry||OGG8OCQ!|ia(}E1ByM@gE3(>y`WuV&8 zm-nK*u_qQL&h1KYezdMtqsN!+4;{h-u0?HkKquqb&oF>`tyG;s>hlRv&6ve9nnhkA z%Fw6i;WbNaC=KC2A(%dm51-b{x#CY%0y>eq-<1~J>{`x9aI|W@a<%Q)aOSEb0iU2S zyCca9F%MYO1bxGb7o<*8-=;EDy(Vb)3Cb=PG(dq`o*WecKSW_;rB|yNAeL($k0p2z zQ4KD^FQvIBPTBa=3ICgpzX0(grs>Optg|n4vr{X}gY8kqzhjQ~$awz;G;;G3aP#ao z9TNBFWx;JaV2Dx}m%c+NbS!TS`>Uj0pX1k-Gv(bG4coFg9zEfaPJK|nJ=15qaFlCb zho;h}A$m=qiNMS6<)Y3`HZd$qs<+sWF+Sq@C@_BAZL$vxr6^$D&(B)uzZCMM1AxA< zAwe?7h~o`<-629w;6EU$MS5;u487-FtLXDu(Ncu$qNa9?4<^z({}GP1byLRrZ$9Rq z8Ci^W(56!QEv!iUBXgP-nCU0JTbv;kS#;1!M3EBi4^;xr8Bn{78+>)Io(P|@pM*PF z`q|GTltBTZTeCq$ew_(%dsamH`4&yh080%hIlrqk3%44^nyHmh0_aCGMX0wb<#-c* zEFt7!EftZO8E>I|>yvGfn`!6IdxcHzgD3M{zw>53V(cgtQVD*(4omeQ0z-47=^>rf z_;5c8+iy=^aCIXmPX0NIPWhCotH>rt>WgNww3%5U;)e+ypR|8S>E}nnKrb!Q!QKd% z8YV@u)>{mEH!mI%khA5Onh_SyAW8%pB@+6%EXpTP-)17Wu+6k1k!!M1f6fX7l+XW4H5}59g!0z4wG0lztc2AEY5JcXTFO8d?y1 z&meZhB-FCG53pgg9M@U(lY`Ly~NKx4~`^b9Xm|o9W4`?hJ z0*7sXMN!2nJhGKP8cK(#X07@~5QP>^`6MJ~vDi7V!kS=Y0%G~CZ#0tpiL`!;RUpBp z#>i)m&YlP{<8jbCSu`Ql0lwnTPv7tovf;+3XG&Xx(b1RmoFt&Sy4L zJy&`_Dky=+ELaoCL0+Hid3yni0Q>^TX9(8}Q zz7~coJ5NsdP|iZVdG{^G)`XSwzdAYn+=vGzXk|V#mhhGIs?MRiNRH2=jqs++fRa-b zeK}}(44fy)mN{xGqklQ~R^XfX#C69UQUl&<8j#(O2^ru&Oo10XNOwE}s*zr=EHaC^ z&*UIPSS8Z~z|-B2L^7f6lys9Dkq$!-;?@M6Be1ExZ;}tPVjB|e9?9FEG1w2kou&!0 z?E(YUKe+=z2jeo-wy25ER0K)-O#TpQaW?4U@(EBl&BY;d46M|oaU839~)LL0pi%u*4w^j9uRLgpqELI#;^wS!~V8K z@>Gxrbtmr5K@^Zy%!OUKY&kt-ef8uKkE1fY^+7b^>oM-bOcpXT$Bsgh>BDUV!~l>5 zyAXt;ZqTltvb)rJPZEkFI4~~lQq8q0zho(iX#dx;fC6ctt~{~TeQZWMyAWLCuVU_R zY_*A$Ur-=R5r-t2QP%3Y=|gtb=T}1$r>NfmA%YiS1>$c!w0xp*s*fEGpjxZX4ZFB9{p-97x!tL^JP32gD*M|q*Yds zXKY`R$iX6t7(20H`dti=#nBg_$~epbWw2yqJVLY!Qgf4 zx4}#yj7-(9K={%d8^zrFDW}kt;2*QbF{WpyWnU~PZC=76g_DZZf<*#jlYP@7)Vn4o z5F@+@V1XRiGW1XCf}}DMk$!KJ?JL16{BnJM1Ca3xJGOpuI)B&wpaqO74IaTgkE3#WhrVyMtN>_{q=*SdaPbg_M?1m?UT z-h9B$2AMJjYv?~j2yEob>5U0p@ySx%(V7g9zR-jChiL!lExiB@7Hn&IuT<<%a`c4{ zxA>u*6#1-~4F?aKH^ z5XIvKMDLsCgTYxwH)b58FXR0ehw&m{Ue;ZfYPfAxNflgCr0M%oBUlsS!9N}xn5u%} zMWPej2g_2TDIaV=)rs{v65&OXZ=MrjytYP(z-v`$^RZCIR>h}(0-CWtZh5Dl>T@E> z^N2t$Nc(Bhw+!WO51D!BXMw@LhdL1GHHoS84$i9nq6%d$LwZY)$M*&AGE;VLV^fdc zxL2D9KCq%#q+t?+^x5KkVeytjRsB|Bmkr(cJWTz znCk@oN1+;x+cOf??HJYNM;8l2={B};%rHGLvnLpvDu2RgK5H-OUy}ruzWQmQlh>TO z?7_$XPr>$@&@*__p?6IC$MS;@oTjfceWqY%CFEEOsw52GQ6&Gm66JY^Ru@VqrLf#> z4d;g@=omjY&#As2XZvAWf$gVepx$($sp@24{HLb)dwJQzT0tfXNB<#S83tf33k8-D z^XKATVPgdVYS~m)hLA|v_Yl*OcfBNR-6`>WX$h9P6 z2vGlQCGJ9gxxo zKD;|POOgR4k{@K_r9nU_wdC_zLwHFN6DuHS{p=w7~ioc5XzbWL(rUUM(O(@l5ffS6jy>FJ9^?9Kn z%T;7L3+SJT%18sr1j10Q5150L9w3eM z5>PkC)-iu|uWmjp*@r*I5^v_|9V%O%rQcsxp}N=dofjKYyP2sf7dyn zh!S`fYqKZG(cT0||8t0c2+&#*blnk{@F_Yq1F3jHfvQvS0zxtFnQpo}*u8g&d#~2$ zzcuEEAS@woNd@OQRD}+K4_&=%&iRb z{)KaYQ>&p=ukYJNqX4yOgR3pI{-U@a&pDLNv$#@HH25?;idT{O>mhpasN#8$;EoSH%<68WwmkzQH>pukt%!-20_UXwgD`Vi` zhCNep%?j{V9(4Y64Bt>{yogE)X?dbtk zlSnI!`UQdKiHr}^)fRk1O-J&?0%WMy|7R%A9dsA>i`Yjzm3gW|F^zG(Vt3AqIm{@z=94?07`upAX8P;D51URMvkeo}2 z++<1c|N6!MC@Ie!B=B(cw2B$|-~KjT@a0Q1s*Zl4|GD}<*iju0oK;tE%s7dEDf++8 z3pEr3opFL>^FQX-U%&dh!2a(+|Bs{pr=$NpKc1(e8cMwFOEerD94ZGvrr*wIo%fKR z!@@@VK1i5hZEJXSfBW{W+CfJ5w==Qr5c)Md{HUd0mCQmiWO-ki8g1>1ER_$s*oyzT z`u(}iY=3TBUDauzx{dcb8+LZ0BL$&6vy4O6K@o599Qmb5{EG!?<3q1^I-T%2SWkv2 z$Eas1UyZO2{0|fK=M5pNqj{cqo_r3DUxTT9p60Q)n_47qvH?u(b=+D0hiC8?CZF@K z-uuBwGFG$1rZa7)7j%4J83z(#Fuo*$dw38<86+~n{Bn%=mn(kFK5JMoU{}G8u~nH3 zw4u&=uQ~W~V~KdK=CPGo0dmp5Te@G2$h(2WdVeUPVJ6$eWHkA8Ig8!NYPFpd{;zNL zSC$<@C$rgnbf8Gincp_>>RuhwVR1OTp+wC0|HWT^kLgg6k55Lzh2e49y(kD}W*bkL;^k_OZjG@PnIJ&!MGAuep5kagQ9K<6P>!44weu9c)S3G7+bOkV!jR9#$B{*5CY{9Qjm`B@AVQG6DH8D0_55%3Ip@OzSMF-K&s8cgCScMjbp*^qs4(Dv zVeR#<&=f(|iTXFs|NbBEHzx+h-lCoOa*%3kY zO!&qWtWOZ2Lx6W2sawL&YPU9>Wl;QHKYBplYwW_fULV1_KfLm@?`n7y4acC5z&@v< z(Fa#9$3TDAAaU@WMJCPTN|nWWb+tOSh~~Fc(^c?B#(Nt8F72q@4Y@9k+j_9M^HKX3 z^o>4(i$;c&nw6Q2^99=IG4>~88P;c01=fRrOWg;6y*#x7-0!Jpw2f!=c0^_gcLMF0 zWFt!)HxQCk%miC*jA`TcJz^g(*k!>tdn>mc3;#`9M!C`V7Hgup1;dS=g^shWL4M(# z3s{{=_lSt!%|1#5?|Rg}y~(`JplH(0lM|836k!b={5_3NUcviEyaVm%uxCT0rEfIb z_1N&mmZvH7KF5I56zs|&8J}L3YMR6h#%ZF@6ssbas@GE5IXVW6jEtPNr+$rmKWUOf z`MYtBeQt}PQ0~|o?A+T&L8sPy?T!xUyoEvU7#oujyqY*hHSra^^>_{Bbp`Sk7b9B& z^Gh_plt99m&)R^VWudg793`9Qa=l?bganq-4BV!e`JVZ8BX5n z<0=xjllLq9 zkrlbs?Zr^<$Fj>#6{ko7JuccONs$-!b*Yf)n6-GIP@ z0kBOMp>oGncYyB)6iU>Qq4am5tnAp1f{-~B0w?+-u zCyRAEoe5Pte9Fqo-vCM!-5kpqQh8S|vl1 z()uiyyRgiTv*}**+XD>%B@2P0**Zg#QMhBz&g6#1y@Bx3i%>V`Xw)dKv<2IZVFA(X zB^J*|J}2PW6BFm2&TBimmxVEs_MU}7>pC7+-6JRCq@I#hTz41nojtGzWRyc{4GS#^` zIifxLSUfl<;LM&s@Pzw%bqG_|boJrpuuN37|Dig$y7*pQsgz`~#@%d>K-BeQb#21g zW+F&p!>9r_#~}w+8&j^4=2_0lhPy^6a@@_qQa3us(nKwwHA=$9*{ILfmswq! zoWyzSvf6U;7=RnFK`P~soHycIW2S`8gijz{j7ghH#4H2vcAaW{ym)JEVol?AKew~C zShwO&n&w3LTJtS*Ui8Ne*^4ku$E67x8tZq5_Y0}z^3=e;wn557!st>*-5w+5~uCVNvS%a z*)r0MzU6vKgm#(YdJMNvx1`n5dX<~hLH%^vaI&4Y0^1vg3xC(jisJ9@-p#0P zFDn*TlaXw;zB!dZzb0+_;qH0Il1-tVgMc=buqE9o!ke8splEEsE9Mo89UCb&-W-8Vv2a}Ya7mSPM@t-sr zky2Z1u-zFjkxJo;E`DLt;FiMOUq)iHQHVvZYh<7nK zW|l)rL-9_joD`lU)PmC20;|He9`vNJTKP@7^n#-N|C_x7nnGJ~9X-^5H}~jAMR`@x zg~5-SPsLQQuh0s&%h`L&%nC*JjTNMnW~YR+8gmQcXiUNixug51F4H zKTaRbP@lKIq`tqOE5F)i6xGeK;225eL?0npuaD&60|Yo6t|-fm(`Z~z?nR*SFUUUyRkiM3et_uEc5p*U z_1HfxUyHrmIug;Vq?Fuo@)gPs|3t&YM$4CyLBE8Vt#H%zJ%0TN2REBJR;!N4jT9=g z4_8@8sME^hIsid=f+ZR>h~q9A_lwrZy?8P=ZW*1nzwK{y7H_PY#{-Tnt-zG45ql~^ zn%-ks!HbWsOiNQpc3Rb(ubFPp&;Y7vfu7c0`@vvYLm5=%KK$vb$a-ml*e{gZ^?2dr zV5Q{?*IgWXDtwWpifa7v#H!2<_tW*P-{}4lhsV9aO(*V(1a9l&Hv3xZ(?J=DQYFw3 zdSkqK3-`U{S{8HjL@es+wD+X}mYuIVtsnWc zk#0_SL>ym>Ms<5jqBmoBHnc2PQDbwta8Ru}Qnz5HIs*dM{9`D{p`W(f)=rioaC5GY zTg=qISsyfBaB!;-Dwq1IN8c4Gmxw8qvtV&O=^52K4<}k5_{+@)gB)n0T)ZQ}5r5qH z-+dflsSkBTT0p9tRnc~mI(eQBUF&j_r@6YDQc6z#vXCLeWUt(#YCLn^f+k(hji zkKlCGn^q2YfVtBtyeE0daYPfsDCeC*HyXj7ai|qC5?0kZU~3J1^ogMfkY?pCZ#3cy zJS0#Z$$T8$Rm>VfXdHF?*fgAb(t^aGJ+k@b_BKz~8AkJ?0_XAI{e> zABuo5iWT(oIXXHzeeaa+$Ptf=N^8F`!eb3DGE%^0Uw*)AvYxcfsLR_a^dSn?YTIx8 zuWgk-Z2^xfjK=ihM_y}QdmRDH&d5hok5B&lbfyG!mH1fUX=KJiv3go=QUb(7dbavQ zU2*J=ex7#^xE3U z`@xB{y3ZkwTbVg7H#bb?*CN+WCMe1RtVYwMQ|*0$F!Rm*14v+yCHU+C={_+-OqTGB zXxGaiYE+otR+c+7g~owHbV58`nj~_!^v!vG1Y`ScVwVaA=m?!wZ4o93;8Np_#5j_0 zS9IM=D|I_j3P=_U*khUcP^ek1-uJT*Qm6>@%^&fIbuygJ`*4vQ?R(3U4O>Jw*O|_)s8aH`W=GFJUNozg2g{qa=~QVKar zQ4g263x8tqsL|Brv4sU>G*b01*}@aC@Z10fx76a)3etkRHHcQMSQ3+}EI`h{@!hXC zHU1PT(9%zmBmBR@&ODk8Y>(rrqm)70X_U6GPShQ)ly4nwHS@Hh^5j} zORcR8seMW*jjaiSHd0%WRuHwslGGBRgb-U0UV7d+Z{B0#XkcHJr9Q9=T-6h7CiYEp$9-bX#fiS2@Vc-m(%$f)@{@|UY5 zrlvU37Fad2xg{lJqFp4|taNURN*dXS+~jhNnMD$%roa5t75jfl>%RG)j>n8k>#V{- zzI0cYrP$p2*a6!^PA|MZS7aWz#W1QT)WTZb@N33$_V4C%soJ@WgfI=yL5K!0MYo}Ayq?rP$!z&!&q^=;ZrLi3{{F0IMtjbr3Omea(=GxaB`ESpGLeaP6TRSuhoEqx_S<^alFWQlU@2z&N z9F=;sbLU9?9OJ-+PmMZO#WqKdpchAkC1VSasJLUF&w9f8Ea@ChO|ZcogEtuB@g}<2 zsF@d?q-vevsMj7{x!f=Cusxe&`F#z*w$!83eI-}AUz0CNd4D%Z5lI&s7ptNteEGZujj`-I?guEefgvwUe#6z2TadBHvFf7^zCiIgY7qCLEh zC0YX0OS25frbs0XMSx{aSn1dr!mc=6si;&jE6OaMcqV6QwPiw~8{j0$<>jm8M>||D z5vF1(bm4=TSdl1jyIvsDqKXfdPt1ZUFa+Q7gni6_RZ2U4nO~h|BO=x72C}lf7}dvz zs^(J{b?GQWfryC2x|zwvo8c`^G1;@(+y?b*3#*Dk`{yScTAr9UfN6_4q09H6p^G+9 z-$fJDL<_uNC2ChRt%EVk(3HpvSf#b(Bj)c>NG>(Xm2VREN!ArE`9}Y>uz!ovf9-9K zVyByB&F<)Q!{~5b|2IUQ?5Hj(nV9x5>i%~yByVIqH&O{dDx@{CB7ao-(SnBa#@O}9m142y>l_c=fe2S zX{`?CfHDFhu`?~CD=y^R3o;jsC3Ax!GWb1O2NikSVetDlPUbtryE{`K^aj&%zbk$r zLhK5UrS|xG!+!B&E@3J(2ld_iTs7)*LmxLSBQqNPEwFOEffF-CmPniq(Mwub(!rMl z-Mz~X@44nx6Af0$t$nOyfqj}|urkISbipNC7#oi)5jU?};-PfO@fz10Nj;LDk8IL{ z;68krV{p^5*4fxyEFLx15Gge!U@&GEcr&l5I#@kR1D<;U<85qidTlG2d*$Bq?bb3; zVi;x_OtaK9&CK!0{@okqOj+j9&b_h2dg_XEBuj2 zR2$Fsu?Ywxfk;EO{(d$?q{0ehtIefa2h;eL8fTWY?E%U0SGim!_Sd?6yCm>w$vbX$Mzpyhc+3&N@;HQYzyx zHhACu%FZ#Np?W8%4Q}tB3vF-#+YaTT{z%G!0vyp}ob;GfgSus_)!@Etp>R)cGX6SI zV(mEdxefEi7{N3gC_xgv)QCl8W$HgG+od;+6=>n6t4?1zy9vTw@Kjy+p$3LSR@g`T zGFwu$*Kc6x(%b0wmwlq`)B}Cga5#-q{B2j;0&nuy@)J^}F}&4gKYnE->$8eM`zh*Y zh_CV(ZqHYX>u1UWio)*?VsJ$@-NDlCYJsq~Fo1O*_m0!I>!_{9B+TPTFG`6`O7Y`& zXLf(>C(LCi?zyA*p>j;Xc;Z0js4x zt@-i&XMHc{_jamc;)AYL_uq8kxW!!c3#f>x7gz2ZLTEw`0Ix}^3E0vRN3Dm1m9M0 z+8bh9F52{v=*Fd8?MHvR+K+iq<}=>k7CTDxtWnin;0m0D&06!zJSv^lOm8wKHOwh5vqVrpmU zp}v@5_4a$#r>GMo!uyh+)c2>@L4H4hb4xzYy`F|qq)^gJJAz*vQB6fCICb4Q-tH0? zAr}Fx+xXpjAqa}iUz{kmC0*DM+?CD0xR2iGTYTxxP#0Zcmc*uNJ|87?kR{QBmCD1~ zik`p)7IoO23;Q`B26tN*wMBA&4g<5N9oA{1o;Ecnif>nH&MH-8@lc))7Ia8 zNdcw4VfUz8h0f!#US?i_=@WVI&4WK&Kw$%F%yeS;oo3WA)3t|^qds`~Ln!1zL^H_W>^nZfX8 zFPFG?HYGq9SG)Ui#k77L;B}WLjYmR~hH`#7e^}~$vPi5zTvjPary>7UeW%V5iM>+q zH7*=n6W>Miof`@ZSyPD6R+E+~U}r$QerB$}7pL*s!jR@wd!OMvo9^~y3s~bsK_686sIEjHYVw|v`wKMbN`7`-<4ET^giRPHQ~b#DFY&L zxrI>$aUPl6#F7l5-O^&@ia*%QA$mzp_cHz{zWjnUWg4JdB&G^c)~A%Lb<*ho&q0*O z+BqS#npvN4f(i+Nci%G7_Q@+y5gUPCk{y|xA9L=Zq)Z7)sfpY#-S8;5zCc5Ftu^8m z!RhJNW>@qNmu1p{b|Z{m=u&gMtthpHA3CA;0!Ve0VJe zdjy8`|LyAFe`(5!1jD==1T+3CR0bh0Duw|YQ#*K+!@ h3*@_&N(iHYy(NxrwrA!9 Union[LLMResult, Generator]: + """ + Invoke large language model + + :param model: model name + :param credentials: model credentials + :param prompt_messages: prompt messages + :param model_parameters: model parameters + :param tools: tools for tool calling + :param stop: stop words + :param stream: is stream response + :param user: unique user id + :return: full response or stream response chunk generator result + """ +``` + +Ensure to use two functions for returning data, one for synchronous returns and the other for streaming returns, because Python identifies functions containing the `yield` keyword as generator functions, fixing the return type to `Generator`. Thus, synchronous and streaming returns need to be implemented separately, as shown below (note that the example uses simplified parameters, for actual implementation follow the above parameter list): + +```python + def _invoke(self, stream: bool, **kwargs) \ + -> Union[LLMResult, Generator]: + if stream: + return self._handle_stream_response(**kwargs) + return self._handle_sync_response(**kwargs) + + def _handle_stream_response(self, **kwargs) -> Generator: + for chunk in response: + yield chunk + def _handle_sync_response(self, **kwargs) -> LLMResult: + return LLMResult(**response) +``` + +- Pre-compute Input Tokens + +If the model does not provide an interface to precompute tokens, return 0 directly. + +```python + def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage], + tools: Optional[list[PromptMessageTool]] = None) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param prompt_messages: prompt messages + :param tools: tools for tool calling + :return: + """ +``` + +- Validate Model Credentials + +Similar to vendor credential validation, but specific to a single model. + +```python + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ +``` + +- Map Invoke Errors + +When a model call fails, map it to a specific `InvokeError` type as required by Runtime, allowing Dify to handle different errors accordingly. + +Runtime Errors: + +- `InvokeConnectionError` Connection error + +- `InvokeServerUnavailableError` Service provider unavailable +- `InvokeRateLimitError` Rate limit reached +- `InvokeAuthorizationError` Authorization failed +- `InvokeBadRequestError` Parameter error + +```python + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + """ + Map model invoke error to unified error + The key is the error type thrown to the caller + The value is the error type thrown by the model, + which needs to be converted into a unified error type for the caller. + + :return: Invoke error mapping + """ +``` + +For interface method explanations, see: [Interfaces](./interfaces.md). For detailed implementation, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py). \ No newline at end of file diff --git a/api/core/model_runtime/docs/en_US/provider_scale_out.md b/api/core/model_runtime/docs/en_US/provider_scale_out.md index ba356c5cab63d0..07be5811d30137 100644 --- a/api/core/model_runtime/docs/en_US/provider_scale_out.md +++ b/api/core/model_runtime/docs/en_US/provider_scale_out.md @@ -58,7 +58,7 @@ provider_credential_schema: # Provider credential rules, as Anthropic only supp en_US: Enter your API URL ``` -You can also refer to the YAML configuration information under other provider directories in `model_providers`. The complete YAML rules are available at: [Schema](schema.md#Provider). +You can also refer to the YAML configuration information under other provider directories in `model_providers`. The complete YAML rules are available at: [Schema](schema.md#provider). ### Implementing Provider Code diff --git a/api/core/model_runtime/docs/zh_Hans/provider_scale_out.md b/api/core/model_runtime/docs/zh_Hans/provider_scale_out.md index b34544c789fa76..78aad8876f4b84 100644 --- a/api/core/model_runtime/docs/zh_Hans/provider_scale_out.md +++ b/api/core/model_runtime/docs/zh_Hans/provider_scale_out.md @@ -117,7 +117,7 @@ model_credential_schema: en_US: Enter your API Base ``` -也可以参考 `model_providers` 目录下其他供应商目录下的 YAML 配置信息，完整的 YAML 规则见：[Schema](schema.md#Provider)。 +也可以参考 `model_providers` 目录下其他供应商目录下的 YAML 配置信息，完整的 YAML 规则见：[Schema](schema.md#provider)。 #### 实现供应商代码 From fb49413a41fd22828fe306aadd2e2fea2648f6ec Mon Sep 17 00:00:00 2001 From: zhuhao <37029601+hwzhuhao@users.noreply.github.com> Date: Sun, 29 Sep 2024 16:55:59 +0800 Subject: [PATCH 77/79] feat: add voyage ai as a new model provider (#8747) --- .../model_providers/_position.yaml | 1 + .../model_providers/voyage/__init__.py | 0 .../voyage/_assets/icon_l_en.svg | 21 +++ .../voyage/_assets/icon_s_en.svg | 8 + .../model_providers/voyage/rerank/__init__.py | 0 .../voyage/rerank/rerank-1.yaml | 4 + .../voyage/rerank/rerank-lite-1.yaml | 4 + .../model_providers/voyage/rerank/rerank.py | 123 +++++++++++++ .../voyage/text_embedding/__init__.py | 0 .../voyage/text_embedding/text_embedding.py | 172 ++++++++++++++++++ .../voyage/text_embedding/voyage-3-lite.yaml | 8 + .../voyage/text_embedding/voyage-3.yaml | 8 + .../model_providers/voyage/voyage.py | 28 +++ .../model_providers/voyage/voyage.yaml | 31 ++++ api/pyproject.toml | 1 + .../model_runtime/voyage/__init__.py | 0 .../model_runtime/voyage/test_provider.py | 25 +++ .../model_runtime/voyage/test_rerank.py | 92 ++++++++++ .../voyage/test_text_embedding.py | 70 +++++++ dev/pytest/pytest_model_runtime.sh | 3 +- 20 files changed, 598 insertions(+), 1 deletion(-) create mode 100644 api/core/model_runtime/model_providers/voyage/__init__.py create mode 100644 api/core/model_runtime/model_providers/voyage/_assets/icon_l_en.svg create mode 100644 api/core/model_runtime/model_providers/voyage/_assets/icon_s_en.svg create mode 100644 api/core/model_runtime/model_providers/voyage/rerank/__init__.py create mode 100644 api/core/model_runtime/model_providers/voyage/rerank/rerank-1.yaml create mode 100644 api/core/model_runtime/model_providers/voyage/rerank/rerank-lite-1.yaml create mode 100644 api/core/model_runtime/model_providers/voyage/rerank/rerank.py create mode 100644 api/core/model_runtime/model_providers/voyage/text_embedding/__init__.py create mode 100644 api/core/model_runtime/model_providers/voyage/text_embedding/text_embedding.py create mode 100644 api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3-lite.yaml create mode 100644 api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3.yaml create mode 100644 api/core/model_runtime/model_providers/voyage/voyage.py create mode 100644 api/core/model_runtime/model_providers/voyage/voyage.yaml create mode 100644 api/tests/integration_tests/model_runtime/voyage/__init__.py create mode 100644 api/tests/integration_tests/model_runtime/voyage/test_provider.py create mode 100644 api/tests/integration_tests/model_runtime/voyage/test_rerank.py create mode 100644 api/tests/integration_tests/model_runtime/voyage/test_text_embedding.py diff --git a/api/core/model_runtime/model_providers/_position.yaml b/api/core/model_runtime/model_providers/_position.yaml index 80db22ea84fe63..89fccef6598fdd 100644 --- a/api/core/model_runtime/model_providers/_position.yaml +++ b/api/core/model_runtime/model_providers/_position.yaml @@ -40,3 +40,4 @@ - fireworks - mixedbread - nomic +- voyage diff --git a/api/core/model_runtime/model_providers/voyage/__init__.py b/api/core/model_runtime/model_providers/voyage/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/voyage/_assets/icon_l_en.svg b/api/core/model_runtime/model_providers/voyage/_assets/icon_l_en.svg new file mode 100644 index 00000000000000..a961f5e4355eea --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/_assets/icon_l_en.svg @@ -0,0 +1,21 @@ + \ No newline at end of file diff --git a/api/core/model_runtime/model_providers/voyage/_assets/icon_s_en.svg b/api/core/model_runtime/model_providers/voyage/_assets/icon_s_en.svg new file mode 100644 index 00000000000000..2c4e121dd71f0b --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/_assets/icon_s_en.svg @@ -0,0 +1,8 @@ + + + voyage + + + + + \ No newline at end of file diff --git a/api/core/model_runtime/model_providers/voyage/rerank/__init__.py b/api/core/model_runtime/model_providers/voyage/rerank/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/voyage/rerank/rerank-1.yaml b/api/core/model_runtime/model_providers/voyage/rerank/rerank-1.yaml new file mode 100644 index 00000000000000..9c894eda85203b --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/rerank/rerank-1.yaml @@ -0,0 +1,4 @@ +model: rerank-1 +model_type: rerank +model_properties: + context_size: 8000 diff --git a/api/core/model_runtime/model_providers/voyage/rerank/rerank-lite-1.yaml b/api/core/model_runtime/model_providers/voyage/rerank/rerank-lite-1.yaml new file mode 100644 index 00000000000000..b052d6f00028cb --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/rerank/rerank-lite-1.yaml @@ -0,0 +1,4 @@ +model: rerank-lite-1 +model_type: rerank +model_properties: + context_size: 4000 diff --git a/api/core/model_runtime/model_providers/voyage/rerank/rerank.py b/api/core/model_runtime/model_providers/voyage/rerank/rerank.py new file mode 100644 index 00000000000000..33fdebbb45ef36 --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/rerank/rerank.py @@ -0,0 +1,123 @@ +from typing import Optional + +import httpx + +from core.model_runtime.entities.common_entities import I18nObject +from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType +from core.model_runtime.entities.rerank_entities import RerankDocument, RerankResult +from core.model_runtime.errors.invoke import ( + InvokeAuthorizationError, + InvokeBadRequestError, + InvokeConnectionError, + InvokeError, + InvokeRateLimitError, + InvokeServerUnavailableError, +) +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.rerank_model import RerankModel + + +class VoyageRerankModel(RerankModel): + """ + Model class for Voyage rerank model. + """ + + def _invoke( + self, + model: str, + credentials: dict, + query: str, + docs: list[str], + score_threshold: Optional[float] = None, + top_n: Optional[int] = None, + user: Optional[str] = None, + ) -> RerankResult: + """ + Invoke rerank model + :param model: model name + :param credentials: model credentials + :param query: search query + :param docs: docs for reranking + :param score_threshold: score threshold + :param top_n: top n documents to return + :param user: unique user id + :return: rerank result + """ + if len(docs) == 0: + return RerankResult(model=model, docs=[]) + + base_url = credentials.get("base_url", "https://api.voyageai.com/v1") + base_url = base_url.removesuffix("/") + + try: + response = httpx.post( + base_url + "/rerank", + json={"model": model, "query": query, "documents": docs, "top_k": top_n, "return_documents": True}, + headers={"Authorization": f"Bearer {credentials.get('api_key')}", "Content-Type": "application/json"}, + ) + response.raise_for_status() + results = response.json() + + rerank_documents = [] + for result in results["data"]: + rerank_document = RerankDocument( + index=result["index"], + text=result["document"], + score=result["relevance_score"], + ) + if score_threshold is None or result["relevance_score"] >= score_threshold: + rerank_documents.append(rerank_document) + + return RerankResult(model=model, docs=rerank_documents) + except httpx.HTTPStatusError as e: + raise InvokeServerUnavailableError(str(e)) + + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + :param model: model name + :param credentials: model credentials + :return: + """ + try: + self._invoke( + model=model, + credentials=credentials, + query="What is the capital of the United States?", + docs=[ + "Carson City is the capital city of the American state of Nevada. At the 2010 United States " + "Census, Carson City had a population of 55,274.", + "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific Ocean that " + "are a political division controlled by the United States. Its capital is Saipan.", + ], + score_threshold=0.8, + ) + except Exception as ex: + raise CredentialsValidateFailedError(str(ex)) + + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + """ + Map model invoke error to unified error + """ + return { + InvokeConnectionError: [httpx.ConnectError], + InvokeServerUnavailableError: [httpx.RemoteProtocolError], + InvokeRateLimitError: [], + InvokeAuthorizationError: [httpx.HTTPStatusError], + InvokeBadRequestError: [httpx.RequestError], + } + + def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity: + """ + generate custom model entities from credentials + """ + entity = AIModelEntity( + model=model, + label=I18nObject(en_US=model), + model_type=ModelType.RERANK, + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_properties={ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size", "8000"))}, + ) + + return entity diff --git a/api/core/model_runtime/model_providers/voyage/text_embedding/__init__.py b/api/core/model_runtime/model_providers/voyage/text_embedding/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/core/model_runtime/model_providers/voyage/text_embedding/text_embedding.py b/api/core/model_runtime/model_providers/voyage/text_embedding/text_embedding.py new file mode 100644 index 00000000000000..a8a4d3c15bbb13 --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/text_embedding/text_embedding.py @@ -0,0 +1,172 @@ +import time +from json import JSONDecodeError, dumps +from typing import Optional + +import requests + +from core.embedding.embedding_constant import EmbeddingInputType +from core.model_runtime.entities.common_entities import I18nObject +from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType +from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult +from core.model_runtime.errors.invoke import ( + InvokeAuthorizationError, + InvokeBadRequestError, + InvokeConnectionError, + InvokeError, + InvokeRateLimitError, + InvokeServerUnavailableError, +) +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel + + +class VoyageTextEmbeddingModel(TextEmbeddingModel): + """ + Model class for Voyage text embedding model. + """ + + api_base: str = "https://api.voyageai.com/v1" + + def _invoke( + self, + model: str, + credentials: dict, + texts: list[str], + user: Optional[str] = None, + input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT, + ) -> TextEmbeddingResult: + """ + Invoke text embedding model + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :param user: unique user id + :param input_type: input type + :return: embeddings result + """ + api_key = credentials["api_key"] + if not api_key: + raise CredentialsValidateFailedError("api_key is required") + + base_url = credentials.get("base_url", self.api_base) + base_url = base_url.removesuffix("/") + + url = base_url + "/embeddings" + headers = {"Authorization": "Bearer " + api_key, "Content-Type": "application/json"} + voyage_input_type = "null" + if input_type is not None: + voyage_input_type = input_type.value + data = {"model": model, "input": texts, "input_type": voyage_input_type} + + try: + response = requests.post(url, headers=headers, data=dumps(data)) + except Exception as e: + raise InvokeConnectionError(str(e)) + + if response.status_code != 200: + try: + resp = response.json() + msg = resp["detail"] + if response.status_code == 401: + raise InvokeAuthorizationError(msg) + elif response.status_code == 429: + raise InvokeRateLimitError(msg) + elif response.status_code == 500: + raise InvokeServerUnavailableError(msg) + else: + raise InvokeBadRequestError(msg) + except JSONDecodeError as e: + raise InvokeServerUnavailableError( + f"Failed to convert response to json: {e} with text: {response.text}" + ) + + try: + resp = response.json() + embeddings = resp["data"] + usage = resp["usage"] + except Exception as e: + raise InvokeServerUnavailableError(f"Failed to convert response to json: {e} with text: {response.text}") + + usage = self._calc_response_usage(model=model, credentials=credentials, tokens=usage["total_tokens"]) + + result = TextEmbeddingResult( + model=model, embeddings=[[float(data) for data in x["embedding"]] for x in embeddings], usage=usage + ) + + return result + + def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int: + """ + Get number of tokens for given prompt messages + + :param model: model name + :param credentials: model credentials + :param texts: texts to embed + :return: + """ + return sum(self._get_num_tokens_by_gpt2(text) for text in texts) + + def validate_credentials(self, model: str, credentials: dict) -> None: + """ + Validate model credentials + + :param model: model name + :param credentials: model credentials + :return: + """ + try: + self._invoke(model=model, credentials=credentials, texts=["ping"]) + except Exception as e: + raise CredentialsValidateFailedError(f"Credentials validation failed: {e}") + + @property + def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]: + return { + InvokeConnectionError: [InvokeConnectionError], + InvokeServerUnavailableError: [InvokeServerUnavailableError], + InvokeRateLimitError: [InvokeRateLimitError], + InvokeAuthorizationError: [InvokeAuthorizationError], + InvokeBadRequestError: [KeyError, InvokeBadRequestError], + } + + def _calc_response_usage(self, model: str, credentials: dict, tokens: int) -> EmbeddingUsage: + """ + Calculate response usage + + :param model: model name + :param credentials: model credentials + :param tokens: input tokens + :return: usage + """ + # get input price info + input_price_info = self.get_price( + model=model, credentials=credentials, price_type=PriceType.INPUT, tokens=tokens + ) + + # transform usage + usage = EmbeddingUsage( + tokens=tokens, + total_tokens=tokens, + unit_price=input_price_info.unit_price, + price_unit=input_price_info.unit, + total_price=input_price_info.total_amount, + currency=input_price_info.currency, + latency=time.perf_counter() - self.started_at, + ) + + return usage + + def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity: + """ + generate custom model entities from credentials + """ + entity = AIModelEntity( + model=model, + label=I18nObject(en_US=model), + model_type=ModelType.TEXT_EMBEDDING, + fetch_from=FetchFrom.CUSTOMIZABLE_MODEL, + model_properties={ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size"))}, + ) + + return entity diff --git a/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3-lite.yaml b/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3-lite.yaml new file mode 100644 index 00000000000000..a06bb7639feacd --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3-lite.yaml @@ -0,0 +1,8 @@ +model: voyage-3-lite +model_type: text-embedding +model_properties: + context_size: 32000 +pricing: + input: '0.00002' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3.yaml b/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3.yaml new file mode 100644 index 00000000000000..117afbcaf3c808 --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/text_embedding/voyage-3.yaml @@ -0,0 +1,8 @@ +model: voyage-3 +model_type: text-embedding +model_properties: + context_size: 32000 +pricing: + input: '0.00006' + unit: '0.001' + currency: USD diff --git a/api/core/model_runtime/model_providers/voyage/voyage.py b/api/core/model_runtime/model_providers/voyage/voyage.py new file mode 100644 index 00000000000000..3e33b45e110d56 --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/voyage.py @@ -0,0 +1,28 @@ +import logging + +from core.model_runtime.entities.model_entities import ModelType +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.__base.model_provider import ModelProvider + +logger = logging.getLogger(__name__) + + +class VoyageProvider(ModelProvider): + def validate_provider_credentials(self, credentials: dict) -> None: + """ + Validate provider credentials + if validate failed, raise exception + + :param credentials: provider credentials, credentials form defined in `provider_credential_schema`. + """ + try: + model_instance = self.get_model_instance(ModelType.TEXT_EMBEDDING) + + # Use `voyage-3` model for validate, + # no matter what model you pass in, text completion model or chat model + model_instance.validate_credentials(model="voyage-3", credentials=credentials) + except CredentialsValidateFailedError as ex: + raise ex + except Exception as ex: + logger.exception(f"{self.get_provider_schema().provider} credentials validate failed") + raise ex diff --git a/api/core/model_runtime/model_providers/voyage/voyage.yaml b/api/core/model_runtime/model_providers/voyage/voyage.yaml new file mode 100644 index 00000000000000..c64707800eebe0 --- /dev/null +++ b/api/core/model_runtime/model_providers/voyage/voyage.yaml @@ -0,0 +1,31 @@ +provider: voyage +label: + en_US: Voyage +description: + en_US: Embedding and Rerank Model Supported +icon_small: + en_US: icon_s_en.svg +icon_large: + en_US: icon_l_en.svg +background: "#EFFDFD" +help: + title: + en_US: Get your API key from Voyage AI + zh_Hans: 从 Voyage 获取 API Key + url: + en_US: https://dash.voyageai.com/ +supported_model_types: + - text-embedding + - rerank +configurate_methods: + - predefined-model +provider_credential_schema: + credential_form_schemas: + - variable: api_key + label: + en_US: API Key + type: secret-input + required: true + placeholder: + zh_Hans: 在此输入您的 API Key + en_US: Enter your API Key diff --git a/api/pyproject.toml b/api/pyproject.toml index 64b35621b2ec60..e737761f3b2c0b 100644 --- a/api/pyproject.toml +++ b/api/pyproject.toml @@ -123,6 +123,7 @@ FIRECRAWL_API_KEY = "fc-" TEI_EMBEDDING_SERVER_URL = "http://a.abc.com:11451" TEI_RERANK_SERVER_URL = "http://a.abc.com:11451" MIXEDBREAD_API_KEY = "mk-aaaaaaaaaaaaaaaaaaaa" +VOYAGE_API_KEY = "va-aaaaaaaaaaaaaaaaaaaa" [tool.poetry] name = "dify-api" diff --git a/api/tests/integration_tests/model_runtime/voyage/__init__.py b/api/tests/integration_tests/model_runtime/voyage/__init__.py new file mode 100644 index 00000000000000..e69de29bb2d1d6 diff --git a/api/tests/integration_tests/model_runtime/voyage/test_provider.py b/api/tests/integration_tests/model_runtime/voyage/test_provider.py new file mode 100644 index 00000000000000..08978c88a961e7 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/voyage/test_provider.py @@ -0,0 +1,25 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.voyage.voyage import VoyageProvider + + +def test_validate_provider_credentials(): + provider = VoyageProvider() + + with pytest.raises(CredentialsValidateFailedError): + provider.validate_provider_credentials(credentials={"api_key": "hahahaha"}) + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "object": "list", + "data": [{"object": "embedding", "embedding": [0.23333 for _ in range(1024)], "index": 0}], + "model": "voyage-3", + "usage": {"total_tokens": 1}, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + provider.validate_provider_credentials(credentials={"api_key": os.environ.get("VOYAGE_API_KEY")}) diff --git a/api/tests/integration_tests/model_runtime/voyage/test_rerank.py b/api/tests/integration_tests/model_runtime/voyage/test_rerank.py new file mode 100644 index 00000000000000..e97a9e4c811c82 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/voyage/test_rerank.py @@ -0,0 +1,92 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.entities.rerank_entities import RerankResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.voyage.rerank.rerank import VoyageRerankModel + + +def test_validate_credentials(): + model = VoyageRerankModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials( + model="rerank-lite-1", + credentials={"api_key": "invalid_key"}, + ) + with patch("httpx.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "object": "list", + "data": [ + { + "relevance_score": 0.546875, + "index": 0, + "document": "Carson City is the capital city of the American state of Nevada. At the 2010 United " + "States Census, Carson City had a population of 55,274.", + }, + { + "relevance_score": 0.4765625, + "index": 1, + "document": "The Commonwealth of the Northern Mariana Islands is a group of islands in the " + "Pacific Ocean that are a political division controlled by the United States. Its " + "capital is Saipan.", + }, + ], + "model": "rerank-lite-1", + "usage": {"total_tokens": 96}, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + model.validate_credentials( + model="rerank-lite-1", + credentials={ + "api_key": os.environ.get("VOYAGE_API_KEY"), + }, + ) + + +def test_invoke_model(): + model = VoyageRerankModel() + with patch("httpx.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "object": "list", + "data": [ + { + "relevance_score": 0.84375, + "index": 0, + "document": "Kasumi is a girl name of Japanese origin meaning mist.", + }, + { + "relevance_score": 0.4765625, + "index": 1, + "document": "Her music is a kawaii bass, a mix of future bass, pop, and kawaii music and she " + "leads a team named PopiParty.", + }, + ], + "model": "rerank-lite-1", + "usage": {"total_tokens": 59}, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + result = model.invoke( + model="rerank-lite-1", + credentials={ + "api_key": os.environ.get("VOYAGE_API_KEY"), + }, + query="Who is Kasumi?", + docs=[ + "Kasumi is a girl name of Japanese origin meaning mist.", + "Her music is a kawaii bass, a mix of future bass, pop, and kawaii music and she leads a team named " + "PopiParty.", + ], + score_threshold=0.5, + ) + + assert isinstance(result, RerankResult) + assert len(result.docs) == 1 + assert result.docs[0].index == 0 + assert result.docs[0].score >= 0.5 diff --git a/api/tests/integration_tests/model_runtime/voyage/test_text_embedding.py b/api/tests/integration_tests/model_runtime/voyage/test_text_embedding.py new file mode 100644 index 00000000000000..75719672a9ecc9 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/voyage/test_text_embedding.py @@ -0,0 +1,70 @@ +import os +from unittest.mock import Mock, patch + +import pytest + +from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.voyage.text_embedding.text_embedding import VoyageTextEmbeddingModel + + +def test_validate_credentials(): + model = VoyageTextEmbeddingModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials(model="voyage-3", credentials={"api_key": "invalid_key"}) + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "object": "list", + "data": [{"object": "embedding", "embedding": [0.23333 for _ in range(1024)], "index": 0}], + "model": "voyage-3", + "usage": {"total_tokens": 1}, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + model.validate_credentials(model="voyage-3", credentials={"api_key": os.environ.get("VOYAGE_API_KEY")}) + + +def test_invoke_model(): + model = VoyageTextEmbeddingModel() + + with patch("requests.post") as mock_post: + mock_response = Mock() + mock_response.json.return_value = { + "object": "list", + "data": [ + {"object": "embedding", "embedding": [0.23333 for _ in range(1024)], "index": 0}, + {"object": "embedding", "embedding": [0.23333 for _ in range(1024)], "index": 1}, + ], + "model": "voyage-3", + "usage": {"total_tokens": 2}, + } + mock_response.status_code = 200 + mock_post.return_value = mock_response + result = model.invoke( + model="voyage-3", + credentials={ + "api_key": os.environ.get("VOYAGE_API_KEY"), + }, + texts=["hello", "world"], + user="abc-123", + ) + + assert isinstance(result, TextEmbeddingResult) + assert len(result.embeddings) == 2 + assert result.usage.total_tokens == 2 + + +def test_get_num_tokens(): + model = VoyageTextEmbeddingModel() + + num_tokens = model.get_num_tokens( + model="voyage-3", + credentials={ + "api_key": os.environ.get("VOYAGE_API_KEY"), + }, + texts=["ping"], + ) + + assert num_tokens == 1 diff --git a/dev/pytest/pytest_model_runtime.sh b/dev/pytest/pytest_model_runtime.sh index b60ff64fdcd901..63891eb9f8d13f 100755 --- a/dev/pytest/pytest_model_runtime.sh +++ b/dev/pytest/pytest_model_runtime.sh @@ -9,4 +9,5 @@ pytest api/tests/integration_tests/model_runtime/anthropic \ api/tests/integration_tests/model_runtime/upstage \ api/tests/integration_tests/model_runtime/fireworks \ api/tests/integration_tests/model_runtime/nomic \ - api/tests/integration_tests/model_runtime/mixedbread + api/tests/integration_tests/model_runtime/mixedbread \ + api/tests/integration_tests/model_runtime/voyage \ No newline at end of file From 067955c2b1cb496c92b1b1b1ea2b404c527bca1e Mon Sep 17 00:00:00 2001 From: Joe <1264204425@qq.com> Date: Sun, 29 Sep 2024 20:59:28 +0800 Subject: [PATCH 78/79] feat: remove self-host --- api/services/account_service.py | 28 +++++++--------------------- 1 file changed, 7 insertions(+), 21 deletions(-) diff --git a/api/services/account_service.py b/api/services/account_service.py index ad2a27f74f4bce..47455615c09ec3 100644 --- a/api/services/account_service.py +++ b/api/services/account_service.py @@ -188,15 +188,7 @@ def create_account_and_tenant( email=email, name=name, interface_language=interface_language, password=password ) - # SELF_HOSTED Just create account, not create tenant - if dify_config.EDITION != "SELF_HOSTED": - TenantService.create_owner_tenant_if_not_exist(account=account) - else: - # SElF_HOST just have one tenant - tenant = Tenant.query.first() - TenantService.create_tenant_member(tenant, account, role="user") - account.current_tenant = tenant - db.session.commit() + TenantService.create_owner_tenant_if_not_exist(account=account) return account @@ -684,18 +676,12 @@ def register( if open_id is not None or provider is not None: AccountService.link_account_integrate(provider, open_id, account) - if dify_config.EDITION != "SELF_HOSTED": - if not dify_config.ALLOW_CREATE_WORKSPACE: - raise WorkSpaceNotAllowedCreateError() - tenant = TenantService.create_tenant(f"{account.name}'s Workspace") - TenantService.create_tenant_member(tenant, account, role="owner") - account.current_tenant = tenant - tenant_was_created.send(tenant) - else: - # SELF_HOSTED just have one tenant - tenant = Tenant.query.first() - TenantService.create_tenant_member(tenant, account, role="user") - account.current_tenant = tenant + if not dify_config.ALLOW_CREATE_WORKSPACE: + raise WorkSpaceNotAllowedCreateError() + tenant = TenantService.create_tenant(f"{account.name}'s Workspace") + TenantService.create_tenant_member(tenant, account, role="owner") + account.current_tenant = tenant + tenant_was_created.send(tenant) db.session.commit() except WorkSpaceNotAllowedCreateError: From 7e4d105e8b77a72a91ea7c9edbe0df40ef1e0617 Mon Sep 17 00:00:00 2001 From: Joe <1264204425@qq.com> Date: Sun, 29 Sep 2024 21:04:54 +0800 Subject: [PATCH 79/79] feat: add setup only --- api/services/account_service.py | 26 +++++++++++++++++--------- 1 file changed, 17 insertions(+), 9 deletions(-) diff --git a/api/services/account_service.py b/api/services/account_service.py index 47455615c09ec3..caebeb9401ee5c 100644 --- a/api/services/account_service.py +++ b/api/services/account_service.py @@ -146,10 +146,15 @@ def update_account_password(account, password, new_password): @staticmethod def create_account( - email: str, name: str, interface_language: str, password: Optional[str] = None, interface_theme: str = "light" + email: str, + name: str, + interface_language: str, + password: Optional[str] = None, + interface_theme: str = "light", + is_setup: Optional[bool] = False, ) -> Account: """create account""" - if not dify_config.ALLOW_REGISTER: + if not dify_config.ALLOW_REGISTER and not is_setup: from controllers.console.error import NotAllowedRegister raise NotAllowedRegister() @@ -371,9 +376,9 @@ def _get_login_cache_key(*, account_id: str, token: str): class TenantService: @staticmethod - def create_tenant(name: str) -> Tenant: + def create_tenant(name: str, is_setup: Optional[bool] = False) -> Tenant: """Create tenant""" - if not dify_config.ALLOW_CREATE_WORKSPACE: + if not dify_config.ALLOW_CREATE_WORKSPACE and not is_setup: from controllers.console.error import NotAllowedCreateWorkspace raise NotAllowedCreateWorkspace() @@ -387,9 +392,11 @@ def create_tenant(name: str) -> Tenant: return tenant @staticmethod - def create_owner_tenant_if_not_exist(account: Account, name: Optional[str] = None): + def create_owner_tenant_if_not_exist( + account: Account, name: Optional[str] = None, is_setup: Optional[bool] = False + ): """Create owner tenant if not exist""" - if not dify_config.ALLOW_CREATE_WORKSPACE: + if not dify_config.ALLOW_CREATE_WORKSPACE and not is_setup: raise WorkSpaceNotAllowedCreateError() available_ta = ( TenantAccountJoin.query.filter_by(account_id=account.id).order_by(TenantAccountJoin.id.asc()).first() @@ -399,9 +406,9 @@ def create_owner_tenant_if_not_exist(account: Account, name: Optional[str] = Non return if name: - tenant = TenantService.create_tenant(name) + tenant = TenantService.create_tenant(name=name, is_setup=is_setup) else: - tenant = TenantService.create_tenant(f"{account.name}'s Workspace") + tenant = TenantService.create_tenant(name=f"{account.name}'s Workspace", is_setup=is_setup) TenantService.create_tenant_member(tenant, account, role="owner") account.current_tenant = tenant db.session.commit() @@ -633,12 +640,13 @@ def setup(cls, email: str, name: str, password: str, ip_address: str) -> None: name=name, interface_language=languages[0], password=password, + is_setup=True, ) account.last_login_ip = ip_address account.initialized_at = datetime.now(timezone.utc).replace(tzinfo=None) - TenantService.create_owner_tenant_if_not_exist(account) + TenantService.create_owner_tenant_if_not_exist(account=account, is_setup=True) dify_setup = DifySetup(version=dify_config.CURRENT_VERSION) db.session.add(dify_setup)

!$Z~){Tb9H*cb81j#cx0&M(IA7V1rSRQbbBJRl&roGPRJTSWXd>Gswp+P9Q*&0&U0_kcvS3|Oyv zR5aoHIe-=2;nd;|%UNpx6j%OO>~I1N=TD7m6F^{xmTR$b$GL3 z%MD&u$%6wos}Y-M@9_gS$(<7N1ZC^Dt&$STwFZFWRPUVIt=zbOFDcDrv58qq|s56InjHSwTe#0H9O_)D?I}_NGS@Y`+P6^6cCSMS8Xt8Lk`JiV`br~q?09_Z*uUHX zgF<~=@o*drF48XO(t({|V zJHwV0e;&k^;Dl*uLetvVF8TkX>ph_1?Aq^PglGvON<@u_UJ`<6GeYzrh+Y#x^xn%L zL=q95=tPa)JEM0ddhbT>ooU~lJkRr%@Av-Ky4NhrEMv~O&wb9m_O-8lefF$wQpPFn z^_$W(3H31!0cudJrqTgRK@^L!fw?e$C>5W(l@;lEb9h&?NR*M^5r?#+u`A$`ti`Da zbh{m2X{F)-%8BM5QF7_M<^nq&SAMOSWua7z?Yl(3^1QYl{!ykE1$!k4=>5n&Dy`k# zMjGQiIxeWJzZqIvGGTg*D99!QO3(@J;Lt$Nf_nnXUXmRu;fy#L54`+>9709H<9AxJ z{uRziG65`iDAL7UTz!pN3c8S(R&=a8wZRpNwz}IwK&XG|tgjj@D>b+N!G1dTs#n9J z<9SUN&Q8V$x#YyQ&(XP`e9vcz%6c>q~ zx=-EZ>#+eP`Oo*(a*gF_$?R~dY}BL)i%)Q4#pO!f9pm03x63ul4|a&4=Qsoa-OHa_5kP+ zfy%q9{Wtrq(*E-&r5C8jR2OE<9ups9|BCJY+{*~y4D&_HTccczulwL6h*@n%{r+`j zERwm9Tp+R$d_PeF>DiCx)tA9|Up{ZoK33s2NI<&WpB#>39wNG zzuad!k#Y@Pe5#c=Uz_=Ajf$=Vlq4x97YVeA^2_#%(@_ilpm+Lb(@ijHW97Kc-w}(Q zq3Be;;&oV<)E0ifli9h_!FHkz$=ghS!w1tu8$&5InlV3-gZecZfZP0F4Gjt1v#c}y zrh3cKhmhu58Ag!J3N1D2=@M)E{@c;TqCc3Ho?`1WgWX%SP@JVz(ENhqKbnUC$p-)h z^J<@Bon{UMI29I#pC?HnD{X9edz z3Z=f&CL1I$SSea#znB3S{)~tmDX1u9#|>^9)ll#kBcLhr_Hh(1mfX*6rq= z$4;lSK-4UlZ(LH{7d0uPr1cA*5%MO^g6rrYuycf30&!Y%Yu0@?r@}%x&;en(Aw2a+>)9i zhxTX*@2@gE8cs!r2PJ{!2HIPNMz&@7>_f^^T*$eTS(?dW!(8!|WUT&DvmKTGLhju7 zPhrfN?e@_f8C^bTe ztMuStV!U~89Ak}k(rdIyDO8uI`+C`TBF!!;+GljPqFl8Ll1qpz3HJ#kT5})taz3e+ zcxrtP)k~|gdi}%4vj~s42#5KBoIgZpozFmKd*g>j0~P3v!;ThnN(0$OWZB8ukW4U< zxlZFYew{PhXVA6H{cHl_8NlyL4Yq%ws5ArA113H_qpk-$ZM5uPm5AQnoitpea18L@4!hp|9T?^2Isk-&*A}MhR|g+hNX#ZyX7fRN>hSe)yAF>uP!LebP30+8X`h)b zK)dXCX^Xv}bLqHlKUQpOEq&wxTR=UvJa)1DZa#S5+A$R|wgnzZ5ZKs^Y`u|JKQuq| zVWr6bq652$`v>Hioj9QG_)*z6c|Wx0n9}L(FKl>=2!2SRfzxA-$w|7}ujvL0@1rAXtm8;-`2EtEhM z67id*U26!b!@rHa^-zH4B>|HA2x#O$5X2d96m33|@sI`w?Cfz}hE7c1;^0M)(2aNa zGCWjiJjR^WKu0vf&jju|@{9RNNiqA1T(LC_PSXd1zTTz1>FK1e|9N}I0r~R{%0^F9 zZ$d`Yfp8NdGzKFeMyrAs7I#9g-G@2#nysScx@=!VStZ753f?S+cHE@X?{zYI@*pA{kM&4>1S2b%bO9 zLkF6DjNWL(qo8(A5Oudw%2tfc+LIbfK_QG2&p`SXqzvy+&bqQ2pb%j?Kz{mL1fQWN zm~S={rtIPIOoIPCRv6_9f5~0|KK^KWfN!S=cGxGLEm#VaeM5M^;udzQFm~c%?1Mv9 zvbJ*aw#mYI$%0^*&wjjg_o?p(foS*RsjIGtWZ#q^N005kBs#CoLL*%u1+gHM}5}nRlc)6m+cMSekcU?_z(=W znCgrs-|+mdVC?r!Xa*3V4A-9}eG+D__1j=TE>f-L;IvX7A8An z!EW;Z$r!>DmsV`T6SukZ?vdZ#44@RhwkhD%d4bZ^8N2XKkJg-?TtrJ*3>n2-!N6Cg zhgHel`QJbv4EK{o_Y(O`QmG+yc>A;uY6%Zv{^Vz8TySyTlj=v%xtK}S4q`_)NR>mc zo^2MoNvvNOQ++U_S}C?LI}+yrcN=;%;5Xv~D+T6KW1)-y>{akRQLVpkuynjSty=!NXLx zYEf-1v^ga=xqTBTu&eP3RE7tFaNp1$g-9&ly+WDmPDp*a>YGXJACqn0)Hcpq4!>Pp zxbJdJI4q>@1A!i$hm;SV9za;p_wcANixk$s@KLQL$d5P;p&G>=m#yX3&y9;=Fq>o<6ri-1 zxAs}M_W-RE^m!HhBk}F1EaT-tGUc9$2+L>W2H-t^WL?x z$j9xeaszmQ>t?HnWSbi>~jtq-E5$#$P%Nk#&wr7Ir<*Fd%w+WutQ<=lkcAX`Cy4N`Ihn8Mi=C@Mx0n%!dV`B4Z$E40bZ_tIp<~ty~v8leeu#Xt0;Iu zMPb)hFNM_ZCkJW4$Awsej~i-f9u7TxuLJ#7jcww|t5W+6Ybt@Pc-97x(94eIe9cwR z$QhMeZGLJ!lq7?^3G62mve*WZ7~90t@n;;nC+1Ob2|!Ap>gU8>Ro{@H3;jcV|IsHr zlWQ5aM-H6@>H+b-E;EyHS~SF%O9{>G8qx*)23tz(2CY}x1&)y5h!JFe^r476D$&-6PjQn!EK8DZN74Bm> z%~zb9d(mDpGu3cBwNFsfk;F*NYY=02^q>%rXpK5az7tTOVSp$?u(hgs-^9q5o)S>g zlb6K`UihwUKKViro*tF`MPuKf?loo7WO-#pFES0@X5PGE#I22b(H?5TkG&(uqv_$RL8e~>w%tT%M!%IUV2P_B$aTRya*^N zizNL!QP`>i2m}83moO2?D~Xx4rtKH9F|mGA0D{eTfV)%TDl~QLH9`@tB&G?ARkS&4 ze0yeIL>RikUFn<$m;jz_O-K4AoRqw*cy)!;xbf3wYlVNWj~XO?0fWZt2`b3qxIX6 zMoJGOGuI8yT7KRdGk0a=za1VONCB#IfBBPUs*Sy+U`;;qb*=9Z`Ij&nBy#keXt0cO6O3T$2Yf9nj-ibsfeUrqxXK zj8cwj>}Pw^Q%LLEt-u_881`6^5`MmZ11tQF zT_%vlc;I_bQ6+}IfNN62`;YeHPyRYU23V6O^LztB@`MDFxl(h~#+MN18xu^1PvY=! zGL%{<9icx;rQB*fdhwlx#&yTE*|opayEbvYcAifIjyqaGEJL3&S)@LoCgcX6<9zniu9hc zm>ELdP(s@iKPll9T7yF$@j|t%HIsykEX)&HlWvHwcP;{dV}bW6SQciRZ^ zp2Jo?DuMv8`~g2OJ1m`YN|7T})Kr_PRG_MP8l6JS*#%~xzL2Hk7iEEP0(|T4ZgB{n zA4$#KgI7_9p$Ce2+PV2;Y|3cFG$pL;ZmyxFU4>mQc^lq`XS);+Y6;kDG(#>CP^qp9 zF!4F}b%B$TV;V;u|2e#5k?D{kqx$0MIIbUqX>a-2+ls;}jdt7l=J3mn=Xei=rI(lA zyw22m<)!|EqH@E&QGcq#Vk)Sgsl}qAJ z`oRi;{`vq}=}s)`f`g;`FoIS2Qijo{7ih@GL>7ZPRXzoPN9z za@H-(cHt+@RCtwESYO-5#S^dPt$-NqTGt+%sZvii9)ohBiowB|Dr;wI%7dvsZqm;! ztS#V&oTc(GOADT(Gut{+l^tP%*UXyUN`7_P|%h-e~|67V)5-Z_9YLtSq99WY^=yIR>}8CD?aC)llc4`AeqScr4<>REp(!G5y)5;E)gv00Y`+OA`LcTmC^v{*o<#T%FOO z=GI*jMVz)kW4)Qre21-!G4A(5aj-2v7aDw;YJaol)nVkT-9KfFO3F;l0=|NkvQpfq z>|&~kLxbjn7B$_FMSiE6+Qqu|`L^&{-v_UgOuXI=kGPWY@?_vD=zT_P|C&2md1YbN zdC)^{4E5}?Z_cYmo}6EWJQP>_M4DnQEV|1Of;yN5kxGH!a05#^ zLKpN^(s}!V3u1^qAy!PAq`txPM&ilFTTPj7DS+T$yqCfBI5RcELw|xpJ16UO74{h~ zbmG1J(V_W^;i*7CVe@r@n)zS1@UI{Kz2oWtH8tqR4~~h!o;H1do&pM79gnpe($j+k z`s5WJ_Nr?am9_T7R-fONbu^!!1>h+?I`m#y%!UL7myx@kt?S$J2p| zPo;Sg5}ur;^PpIa2#gV09B5KO)sGJ%51bM1iyjjL_!ure@we+A(iU+6r2LW23KL)o zloMBY$#5E1XJvAv)BU4?Z3)2@KWeBYjsK7V$gjjD0bMeF`VS?5=#oq%D;miPXLxWV zz2r{Gxc$N3nEY46?PbF9&G)Kn7LFsM`!A5UDR8|Z4G$O6!r`#q3d}|GI7>_cLjg`? z#v;o~Eoks;elOo5;8}%~t&i@I#oqOvxo$4Z;MYiKTyZwf?D>03@5ubveg$K+L5SgA z;q3e4qmr3>s7T2AF5cc_o8oSKQ#O|Gg%_G)2LF;0;zauoSAZXD2kHW+zsDCK^`sgp z1Bil%um=Ov=8r6NuIfhcwgGT=E1X9vQeUMZKWFMCqw z*Nj4B5E!Zw#~ec6!>t$f9%GIz}~HU&p4{h|S?QK(}0G61MrRk$Y^a zDXu?3oCq*>W*Q|Xt;*f}o_ih>}vsD!UU(<5G#=M z_aplq*tNUeaR@H410IuWA#7@4_rbR=QA^)Uh?tf{v0Y1Zo8q*XA;IYLeD5l55c<9{ zg%(t#yj%W>+Lhpj!V3b&;rR#qBb0-rwG9YaNPx8D0pH@dZiNg6s`At+cc%a(Ax1PQ zJY?&4^q>}>gg#net=)z;8`0wa6u!G+5pg5*Sd%M0`*%&jnHA+G;MpQ^Zw?k*_bT$hjRSK+y`EYj%S9cf^swb5AmD?50O-o6z*u(m8Q89DD&5&bwZcNGD~g=@Ds z{KM^jgU>+aMRH#p$D0}jMMa_%ao2BOsCixvPyeu;2;m-(WIZ%bAOcEwGv2Ov&x{{BK=++SZ8!FBm@BKMiKuc- z+6nfHIh|(388z8)!F!{Xl$^8XC6s@f1o)VH2BemH<~XMF-8%U@y9YF&zT+{q>gKHo1z&sRG$&ezAvlb&zOcfKVauCf+>zu21Tv@K&wtU>ccs3IA? z%E-?r&YYa;R|x)HV8#4Uv%+iUKM%l)R1fsy1q^|F`02$3sG%vT@PzFfNh<>?M& zs{+GXFct2sSVEc*It-2Xx#2_^Rr!JSYbkjabq=L3>hB6{)Lv(;l*DHJP$tNZdKW6C z{_b=6i@0}u4M*U`ll8f}gF)Zf0q@np#)FcDhK*E|&pz$Ry1U=TMLUOH*QS!SYV6K~ zfJ3tMK(M&MG#4SR%u@OBVD!F|=w1d}-p9h1I4lmGgp#_9c`Y;dzMx*|QQ(WZF`Zq5 z$20*iCO&FDqp>5&atm0_Tu?{PKsY$QP%+X7h?O|h3QH-t)a4O4osn?iZ+_JbKkk{>^&gr zW71emc<3986YtnPN0wbri+uU=J|<{0`mt3sKGjt#vQV=3V48}|$M#m1sC;s%HqNCm zr%LCOausQ>qFx+wChXsryaaCgRQJI%BEkqPLR=|-BD5rX83|!fl0W|(G?~QSiphY3 zMH0vIo1_x_EsHuO;-(P#Q|)#>ZVfXlWY%7+YQ?uR|Lpkmuz93u$2NyOedlA?EH@ma zf@h#n9P>=~-r~gi&7wwOUTr^3B21B$4KA%OW;IwHz@|Sno1(IpZ{IwT~_$@{&hquuFx;mmtJy1$r2XIL|`@^J1ce?0al-J zAfs5InL0sEGX^!J9?N+Uj4SW|`x*hPKkYJsF-1tG<6iQ~rf2DdJjjSY++oZ4eKQSr#JbL+Uz1qtZ0VhGzteN8uj&$i0j5?scsmPLn{~k;efq zX#}>@?jj3)8hnb9oEZDjj_u|Nm)p0RXg-R1WGL%VB4;Cq+f%gl47ny{aDb_1i8v=* za=&kQl7-qBpp24aCuI447F4c%4!lc75#aKJ{_rH;1Nf>5xXd3UAM<)Kn|2lJPGR0p zo$iqAje@+-3+8e>&UZa^8~mbU@gF^yTShvseMjaZw=9NY)c6Y|YeUeWl^l~G=&OgJ zu}~x+^GMLQoJ-G9ndJ#xcCwi6bDdFV4CaZ2A6*esXG@~Ry^wrVMU^w%obDi7eZg3I z$ZpxS@6Nmk|8{SP$Jgl}F`{=5vw+TOzOVT2lwAov-HxR$j^dz;86g@qjZUrygOUTY%*oekIChq;6!ew;$ypp8N32r z08#-fyC$Oy_8_0j-uRgx3qRjh7nIlI!3|OEHw*NdW;D4U`E99p7W%M&P&S`un`&q) zrI+Rr`*v*6MoB*l3^jNs^j+6;fi`NTHUkyHbs5mk8V7PIGu$8let z2c#K31WH*W?6x;>9G{XEdI45G^_$^_-s8RTw#Q+jdpKnc$473=a@sG0?PYr*3LIB7 z9>SC$NtJ}(YxKu=tXuCkoxxwgSO%hmEYVn$$X?u@brv!roEbmky@^~6Uav||8+K7M zMquz9QdnOUi~~(FA~(Ik)i2ccs#gK$vF4luEjQ`W`fCOvS9_v~`{SIoRWzct@dFs7 zzAj?5WTg1)wieA?@SKOQp>DZPSnLL$T9yO$PxY<)hD9zmPX?=MiLQwY%oMj-v1G)B z+K?!5yo$cc{5kv^Hcr#Ah_fB7=plN&_u|hqPfWs9%AL+s`2n*Y^-@VARKuZiOKnx1 zmSUohfxYeg%S1zCIU!`Za^2by13p%@KQkxcSJ9JE?K04P^D7x zc-^rxR(^X_F%9(L_{qWb3<(}ch(v6`t-`0`)Y~@^kJc!4r1tr3&5~N8$|N-|pkqWn zH-X1(ytXR1lERHDj)*wCoq-4=lyq(k%JJ3d&NDgA?K@D5Ykx4_5NYnEUIvY1l52hm5^RVj&Z>?+K2Q2=uhmM+ooV zV5d68P!&`VgCj9U_utGXy;#?xf7HEHR;*tGdKvc12!n!U%hkGUMJ0h=z!;prO?rPX zpqW2eSUXX~8GhkUg&aPuteg1s2rIZ8Q8M)@g_BM8&u#!b{{QzPdw~0v`5i;@O~Mz# z*zq(;Tzz(pc=|3kmbd>}dVd`T|G!t{opd}XSc;o1fPyKV=aO&r7ah-)29|KhUq|Bq zcnx^^95@-D9^I?^G^WDT`j8<83C!uA1N48s9;*8DdLaeg-XHyppuZHm@`=khJW z0sWk%#pf+06iAnm)g(%#zvh9QU=0>8w)~Of|9tC!yoZw15^5JcbxZP((p`{x1A^)s zy`I09>HoQ2@CF1ugG7qS4SmIxXL7-P-^wT&xBD44>Um4{e25d6UimV*HB$NiIrU6l z1g}~dC18nMVefGi3re0Mw#dxm@1sBYsV5tZ5WW|FtzbVC_djm)pMf&qZvN=LL13@2 zq)6sZ>Ziu%+(hoY^qqyw4RE(25_?PX&vg9P2ap-!<}HO6Zw!vhvIn@l@@WeMqwi;o zu2sr2>Eh-ss#1Z3SpVmZ40r+jp=wYN4NJr^ZY92-&Y|ghof8dtrrYG76uON?EDy_^ z6#na9{qz5|Be1G?w8=mrtk|Lc!&^xOnk~45S@bwho_u9WaZ6>@minJdnGpBhJ7Mxv zoDH%w+{y!AnSd^Y8b(e21S@?lJakn>D-rU`fa$;QN(<$KbVH0$WTXf?VMI$Tvm_oS zmKo-#&?1Wk^$&wMIag5%DBxIrl)V0aFMA0ZV?NOy{MT)LfgeE6cG2q|BGk8T+5kVC zUUk}XGAO4?HL5G(ig;A>y6-Mm7?EbAlSt^J0`PxL(I+Ed_m(@T*ywxs#s3%zq1jg& z5d6W~-3#C;0uupZMYNJ*ug2FP8>!t-f1@!Z+4tsEV%>5K8@MWvtLtWKtAN}YvG1Zz z)<+jUcO{3EQlbAE7(32sD=ZwF@C6RGZ7V4rnEPrgSDc)IaDxnlvo+d%=*$3W-2HQg z$}i@c4lQ9=K+AB3`K>tWA=o$Jy*GeIHe21R7lQ9)`8i7oyIkFC9LrL zy!Yos>av=1ahYiZ%wxXNj83FYS_~z}ZoAr#^&gD?1JR!+dFe({tZ|iq($vOC=8 zH;R4Cv~G3WWcYz*R-bV=jx>1lZhLB^8J8BD-nAgSzt1!r#7-|(sr-7KZ?!o!yhnKP zX0^g-LAS*5-1lX4{Yy>w)Z?p-w1-|p-QVa{HJxiz-S)z1*UwTL@^<$6am=hu9apAG8e3gtukw05 zp=|Xc<@j`p8s0CXBp!N-ik7}sz{oKWymxLZyJ7y<`I)5)JVd$Ea{a;CoH(UR;o3(H z#Xqgiler05lCW?Nx;nB(716cNQr{{Fy>k0&Ax3cHgurg*4D6Eal+T28q79O&KqCSr zjQi@R4PUR%e8hJ|l=**^rOe1$Er=cXvP2uF<>U(c_OSrk=C?gL^+4mut%V)6qIa&_ z6mh?r5F~Lt`ZfKfY}feNbP^6r5gPH|JSQv-x<^_B+Iu+XDx2)37Ucaed@I^}1!_WLyEJv>|5aT=-#N zP#+RKvA2JGEh3w;(WM$CXg%2eV)uA}*Wux^-4zUy~l|#$yh~f6kbxK8{0rpYM;_(hBFZ@gxdYbU7-%E`Ht0A-37* zP)6j%{Ijuj!55Kz(f;AxV0{{nQSSw~FnXW#v_^#ocjU0nF`66hhelqlbvkBdDP&$d zZ7)L1wks3(uc~TXgh?;FrHVC*-8eNHh16=gRg+yb_9qMUl|D)=sP@HkJ!Df!jnXWb z?D;gjT6xS!OIVTz3IU5SseuRnLLC3G3x8@c%N}?t)bG;D_?+~kM%`OnlR?HC%D_|g zduwA_A=o_N_G>Qn(Cew6G2tk-7tbLbb9SY~81|?_BeB}qw|O9lqd3I55BqsWM-B1u zVW1LY(HeN)$gZT{#RM0Wf(h1ioSV8cQh}JC@clVOjw8bnlO@5pTk^Zuj-j-GK?l-) zMugjkT4>rIwQQCFVb^NbLoZAxPCg{%B-vY8am=P@9gbuPuT4ANEn==6Yqe-eYM^^9 z?y^~`pLSva<+?{41(cuCS70?G23bchsf}aZbL)hXU9vB0IP9A&2CMA$?w6Ykd2he7 z{>nQT&sP$|tEp~(f2?rPU_&`sT;>|&0xr7TBTM*wmItu2EnOGgfs@txkk^jjR`P53 zeioSIl3@I%`X4)b#CNR1lZ^C`d%+STzlR>?WXy%qq*T0UJY>{ds=n9FjSs$k*Fb+&qy%F&>o)n|kueO7$fBE4sypqkiui92mPupR^ zAmKH@$@@CbKbAzB`JNV`>wH@L#@wwPVrt%c$_F0YeP{&V@>9Ij&W~xrTQGc;JD21> zMnb6VYi`@cqmUM5<@CTh`}u0=@EePlrIn;rOCm{w35f9=n6U9sT;Jps=*# z_NsTiEBxr`NcEp>g4cN3693@e#{denmp@d*(T?!%->B{XF{o2**t+HL{y&i>K8{d*NW$|=gejn*3s)nAr*%@+ds&QR!i z+H6ev9e(R`p8{k`4XaX;KDm(WOKgx5Uo$eX5d^1?&MsLJdvUgyIcsQlbxvF9YjhO> zvA#x}+oXIPM??hpW*g4NeeR1&41fs-ZdF@3nzk64FpMrFvrihWmbYC@{@QnkDs7N_ z8$Y_*nMUz?u209iTx~P@jrDnHc+OwnX5}C~-A&xI=;3@#czD)1WfX7uXVL>U0U!u* zk+3fok|J!msHNJx-1-2PN+Zwxi(mqd6S6^ElWfv5{885x>`rjZwHp}G#Rr{cdaUS| zd!8}y!Auaxrx>{(@?yTNt!~l(O*sV9`ofI0ec#G*zW(;{n82!XaiK1RTOauo zu*?0$*|6EIx3B;mXK%r!$h}w??X}+<)UVbBjhEOP7xz3`m|{g3V5nR=ZMI@ng+wV16z8Fu`QAgG5=U-uF1)} zo?p3-OuKH4ob7kOaM`5yx6UJwrR#FAoNI=F2t&IV){ifGA=0C=$5WZ4dT}S;K6t(R z;n#g`d&(Li7rDwa@owV7V9;Bc61T#T0V~atiasq&saCy6M$&6v^e6 zz8iMFeK>4bUknV{mj*cs?y~*`5`TOI&MLK&yt-fVtp0q{oQ=iDiUm)RoWFSg;T`|% zxa9<|do)L^O$-Jo`n&<7#x}V1qRyVF1{PoDrkw9xAL=cK-c{+evKT~Z`Tj&z#Zx;S z)gey!@^NedyeXlA8?%I7Hv!)K^wHIQnm^kGYjoNfJX#D%U)!5Oy}3e;CPaxio6pam z{*1=um0$2Yc$jdN(Vs#x_jN4KNG2Yk{^dGJl?e{)GV!wqyOx@JxyN8G7=PHQ`|pwoASNf66PpI%z2zaCpU`*_*gJADB+Lf#!knH>)ode-ljwLZ*le2W76a(PNCbrwOV^0R7U2b<-E#7iajsM0<^XqJ>6I4HrFs$GN4a`#|+KKu*+$PD(>CrsG=~ z6tMH=m)P7ez|ovLWPDvV4Zl8Nm&H&cD%Lmf6&*~|F2Ui9ZHh(j*@173#(jqZjUOIQ z`on!n7H+&-y*AU*{uKQfBM?vfK<5YFcQNT>_@HU5X+<93j=+C-l6kKF70N%Tj+_+KUf^>NUOeyAlzGH*>sy1+5lGjjiGc^Pq9#CSl?9sd;+e=%(TGmL>9; z=C&>gxe)Y=+ifw9$737`>Bx3(B8th9Mr7ewiS}$o&jxC_osicHfOk-YQJ2cb7qSAw_6fxz7IhK!Z(5tN8kPH=h4a9wiqveqC5lu=9?mgx^Y#V9847 z7ahyo7ZZiLryPWZd%qS#=JTg0(Xw^nX;tTYc6V&j_vR7tAX`%*g;WN_52xUmC)2cw zcF{|>|AeuU`E0-joDT^ykW6I4x(z+XzYkHktF*%Wcc}RrmGRv1cWpi@AiD zaGyHb$6Tda%YwGItNW}VyGJy&UeUCPLL~C7kP7u$X8EG% z78yc{tyeznL?SxV4far?xg;{1v9hM?;<uvcA#ZEb<3n?;So$uInRUONWxz4_wAZfwzpw9{MCC zGyI!9$kM~n`&>eC^=fR^cgOU3m(*hOGu)YRKg24456$F5`kfZPno{hcas1m3+?mvu zXPiI01VC3XRfBd9{ulMry!K4*WG#mch)5YAc>NS0L zLY2K_KPL$xd6~)w2W_`l&3~Hjlb}euJ!;Qh0@U`F*CU({KwNt<{s?xR18E{k7Yy=k zJ0Vzr3|Tnjvmw|W_SA;ocDB0xi`f%MYy7c88bdw(;@_*V1tUz$&y4+>qV!norZ#uJ zjqs=YYybRv7k!V@rTBXTJ&<0~$E3ndEAhaESFTKkn&yS54-70XZj;@c_{4XE+}JEf zqPX+RwVSj$Akr)lEi5jneAg3Uv^|#3TM={Mn`vP5bFRk+HCXPSGm-7=?P)jQ;Q972 z^IMTc`&wmwb@i*Qi$~hjyzeL(!G0r( za#qaRLv#~k=N%=JeTVWbwvmxY23q5~@taY1F1Ih-6mPmFE>h5^3ATUG{tFGY5WXX5 z=mI!MyfsRGPt=pXr`MBN)&1ODx%+G}XO-o@v;CuAI3^UOJv% zNbVr4E~onmad#EnWh~+C3GgJYzZMxtPe-SrHgnuOv0@3Y6tWd3^$Z_O72mP(DAY$) z)t!0q804i-makFZ);r$aya&aU^;?YlRx78dXDe>kyK!9@dXs|~Rp|E+TyHdb_GLMe z({AT_;wEa>Iwo;FzWi`+gw&1zcsG z%6u+94-3P&I?;t6DlzUR7ItB?E)n%6(LM9%(N~K)z9TNe%R97V`O$XPT?b{PJzTXF z$);+@1@i$F-U{!iUi7PR+O_*0bv#`vN@R*dEBDDZK;yyGi`gcGYDrq_85s3;Fzd|+ zQLX!gHZ}=qqhD!!3M`gL`sAOJYJ*wR{$4(tEM#|6`q%Rnn92#*8+_xB{E&j>=H1hO z&(a@;&L4|=zsla%%(vtqHjshH*S52y@2eA|YkG(K@gS)%Xt{c!doEwINd7N`e7guZC@KM16nz~*q1AIl0JURJvr#>RI;EG z%cHO8e!KJdbM(-@%><_MNnJae6?9F1f zf`|&O4_Zdb8<#=lwi5uQC!SjPJ)2dWy$fVjf3<#BHUb*vDXSaX*SEsAiG@SQV5@b` z8*_bo!b)SZDUs#brpi%&%^Ws48RzBI1EzpaOzA17c?sXGU%dPh{h#@g(mVcUAS1U< ziTyX9?F-K{ri7`k^{yE0v+ad$zJjIN?|`y;fnq%(9OwIWs38Xa&ChCjjuJq`<+(ZB zo!#Hmq+GVRlPcj21Y*l8Ti1Aso8I21RWo9BGre(poCR}9QIq+AZfW%%^x&Wg++T3) zaS~^HyiGe=WJD{NXPy1z;)t)A8NdhBV--sAL-eOS6r-@PG!}qKEvMc&9f23Pcx#Y- zF{ukE&G#{s5y)!9EYoq>k~{Q1-HPQ{7kW2oM0$H{KUG}J$7>gN`tXD`#ZZ6^%{*ZA6ELpzsAyl75hN3IWL=)Pd5Y__t-QN9J3O*AQd z^?p-M`QqaI-1ap$@x^V{4hBgw3xdGCg+*_FllBe`bDzV{jVB9oT5gD0yRwm&e=k7) z0|ApYX+So>j2kojg&bwxkkj*vF6eLO{ZBkXfKxgVg2NbhDNox8ib20xk*<`14x+vt zbsi)zc9bgQi&}M%O#_i`k$CWr`$SR=QDiA35GVh}CufFAgj?3>p2W=*muNfZ{=vZX zpR|6{EEVhArG*z6-rf{$)9B()aK8w#H8wXe7)7 zt4Zu9kBqj4_joOV@|qYe>^GzgNd3VBZDBm0(DlJAwRT)Q3|EPA1;sL#c1fgb0hKQA z`snJ)2*D5HLBdzESQVGG(Kvc zNEP=*KKGxuS#d*TTda~RpBu}{lF+`lwM!8H2jV3ChBzo>-8I332$B=|+=Yy-?%02L z(%&fEp_3p>mHJ^kZ@0)d(%@#o^a7qMLs8!X-o9O&Z16T-{#^G|26?T#=gBzX@o&Nn zQljFzLlxj4M?%kDDjG5}pZ3+)EAJNVRbh3EdwE-kAZpfL1;_~8#e60-hO-OX`3%Um z=PiH8K9H6`-QKHe1RQ3fo?r1WKH2s(#P56EAT_=~)`uy}44J8TcaaJ%@=kvB%zeM{ z-C&~3#sl`Fwkt;YjRtQ^Gm)4=50PGg`tEH(`zC)093lp)yTWMYfvBfd z*D%$;ccq{#{WDxS8bJ2MLb67YmY?erudhSjf(`wNw#F8P?skR6NrzyDk%u&gNbbaM zphS_*!K#?v@UyAn9=Pg=x|zR%4eTsSqB{?L69^DZ+}ww7_FBDkwf?j+=JEk4p)f z$~7yDVcLNj+u(&ucuzEhdkWwbe%AIwBux2!OezO51#xv1IZDkwp2~5vl!_-;&I@;S zI88wruz zu=SSwYu@h8@px>C{#ShIQ4;I+u8)otrpJSnFnsHF*e5Czl_wZFd{o+KZIQAryf?Y)=LZ7qSB7Pz?1N6Wlm397Aqwr48DyI=mD?_#>bER1C! znfPxgXK)9Hw;gv8BK3W-$ULxEsMfhkz&AbBQS<%rH#)ILfT}@Q8NQY$2cJ8}rDn)s z>hAd*m96W6e5Ja06wE!1iKgLPDsb#H^*OcfY#t0#Wul+|lcS7Bx~6C?c9Wodfwry! z79_&Z&2!HdrzUgMx@dv%BK|wT{9EhD1ule(?e)Nz`|=e4eTzpo5#cEfxQa0H&M%5U z-0SMcWDf`ffl*;{)4X{FOI|_3QPvpj=?aUq8JnKWn8oavV5q&B(7N-{3SI!tY&_HD zR!LG#?>W8LnJl)?&&gMgByC$Ntgp1>VnS0aZvJ4HtGTE#b#(F?ftL1 zubOZ)6EpV~LrD9dz*F&QZ_9k4h%yjy_9BVoJUF7Owz*PHd)*tuks0%ahhtJRf9XPC zvu+BIf%R3}&c=EdHC1r9A3hIZnT)xiA zL{Goc6Fq-CDtcRgv)BaHuvRdwUirz}UzJ@y?cSY|Mm51C!-M11eS2gixMaXbnr59IUkHkoyRk)C`BQ~p#yHsJA| z*j<@b(&@47!zSKTY?ZGp+D;pN9B11TBPDC+o#%g%CBwf3A)6jU24+Eb9c*9Fj)PvK z{-MNwL!1CEnRJ7pw>V1ISvp5-7DM;$nubUz#&dEM=?lgO8}m;1HR)IF{g{WejFWI* z-|?YtzIfA0+R7@qr#CM3GSq)%{W26rRDXF<%tbVtgx=Ud7Cq@o^(*{Y~aJimkM*9|&@stc0%Tm6v zI+sSg+!xUT2wJ`K5Yw;7`)UEsM+sT=#N;H={p!D8N9`0WzLO%f(A-ry2)>)J2)g4b zS;_AkYXfhb;>BQN^bV;*43G%6%ezpFUx*SeIb1W%fW%FybE)QUKYZrsopd`^(Cmu$ z;k?lqH4RNR?BAx#Zg#lhO84!SMZH=vuf;fW=YR=~-{)S=r z{V^gS&nSUgeqjdFFbjfWCZSc!@FNMs!^eFZ^N;@>6#Cuy2F_VN69RCVAK}5#?{47q z1kP413nFrKAJQoNdw%@hnV07gY=Eh75Yp2}Lf#N&xQ%bMYJ}fi&i;?K{0-I1ng_rh_)auXl#A%0E7)gqs`LYqdA}R_f1Jaq*Z>cesAXA=Z7vq;J#qwAh*kS_ z4%zb-?{Am={o{3DRQRZr05A-z^8?ZbWI0{q;34bqf5hki>skDO!I>3hj*EhXkjv39 z>mW-j7F8o}J|OsQa{R5=Alw{HEVa34Y}YfJpr)l>{*_0jg*Ilo26+F+Z)gd>2ZG-_ zj|&rm6q+oWi~M#wQ9sGPGH(1M+UnP!Jr=;$`x9;R2Eb%{;f)C+2tLwh|Gq2yg%9}G zkv!oN^3=D|qWgpnz|T0DZhW+h5u9`VpSY|5P~as9ta;DWfr5Y-UR709eAuhU#DVu4^I=L9U0EQad;8Vgzgmh@+?6U%Q$y6=cc<(2fG7@fb_ z#~&XvC@w*L>|`NBN|PXi>*zH$#_65xE&TqG|2g0gg3CEC^i&s~-+CQ8v{p&{t36@tsa%04N_1 zFdF`4L4P|TX9KQJNI{zy4bdTJdpMg-c2WG@BKv=SKjQYKp({&h@pG471E6|usFiWB(H<^UtFVc?1~X=~DnN%wmGG*+IWA+w}fE z-f!>2eek-mCE{$2OW_;v6_R$PLT`jIlofi21nJ)Z&9wY(d zy3_Znnl0ipun#VS7K~<}`iHo#nEdJ$?blc_XJDr(Zkg)T7uP4={xIY-%j^r?WNAnxhaWebWK@fb#B3kQ+ zf@^BECOpu}Gud(2ZUg|mmUOh6G%z*w3$E5hZ+W?{55uB>cdm_ZZ)h`V=NVfBH{pr) zX03o-t4#do*+s1)sV9xko9%k8{R#ccjv#$Q5*TC^woY@_P$EaNX0#m_5!U?e_fGo! zuNeOuj=BBjaF*J8Kf_Og+X-(6HNk^mz3<6ZytkgKHKLZ#r^G;VLq-T=gaZmG6K;yV8C%b1-0C4f|aAUaRJ92A3xu~^Cojzi}j z6T1F{U%L-Id%rndIcvAX-mBHSV(g#CJLeQWl0xS>SL#FS8o3S(0!Z%vKd> zw4YcXC8Jtj>S?Q4I<0-+y>6U(Ac+dH)wuP5f`*1eT2?s%z*H#&`3Psow~BSIg*nKx zXi_+9W&7cf zOO!4F`{+%0G3A@eIqj4an`UOuF$NYXg2lcBX;<@s+r{g_6bv}T&phiw)(PmqSeR)2F}%Lmk7;`Xrm=gac{PZ zvCSnmogS<#t(Gj+r$Ki+AI=9SG7%9z#)TjE zs6q>FXqqp6h+i5jp;XUb9Z@e*5l%Q?i=R-3?N#QfQ%U=J*v`GqBFPuTe!YUBxHwATHavXAuTZ4W08=$P znpJ&az*g4#Y-mcOnk?_!g!&*SJicCl^AxI!$-V(evTmdJ|AXgzmXz`*p3}D75y(`Z z*Yq8%wrm{(H7H%z)J?Esg)IAW7S^jgB6^&54YPa@~x5b$59#VLfjsa86pU^+3A- zGJ|6JS-JUqyy6m2ICR3N5z{(dp2q|lg)W66-NjF4cSZ`zR3>LPB^~CM@fDojJYWns z1OTAs=w6KoD$W)&#p@$$84c&3{I;Dm7a%{Zid7bw=wwf&N#tMV z=ysGjg>l_eI4X;o-ME4^F>!u?$@e=mo~|od`>0Jp!1V`kbZWNvi($DbY-3OSvoXi9 zwMF9^mN&VZ@aarp^ZTuWaO+;odAQ+zTd|e-c*w}xIDh5S(o5`!LMN#MW$bBhRnZFT zet#w0VeoL>CH#@7WmO};hUNnirHj**Hz@~#79^>9<^Ue3+?DwN>hg1AAHY&Em_5vl zJ919>^K~~D4g~cs2uhMb>MybUc5AN$3SZ(~Up0k+J@pj}oE>zx^f9BOqJ+FDic~ze zfSOl`IIoRS4juK;JPpGh&ZRnmQwNla?ChCyt(hs7dEW;>8uEg9)RDYS;kp3#nove% zmU#>SJw)k}GJBpsw*5Y(Sj?Yj&%{;{|&Cs*WF| ziQuqgHMlG+6gVwML(~aa<{VfCZ;o1&lf65y7O{h7SQ&w*GFL0ld-n5J?#ZN?JuD9$ zt^?HYmhqt zo6l%9XfPQ^-s5PMbUKm^j<$!kBR@OdJl)PNw@Mrxjh2BiRPkP-_ih4)8S}`RUuck! zlh|MQNzSoNBlFro$N@1se0KPn4{N5`yA3v+AC}!wy45?m_938VuTWmy8ihE1%*>P? zBEFNP@*Ou5~NbcKK48w@>!BQ_D@Omh%;nKgY<*%JcV;2*M zL02DW`8U*g=O0FKzp3uoa4mQa%px@Y+~oOj7Vr*B*(pM90P$)&@RI!VaZw^4MC z?-5~MW!z#J8h!sL)atl$g7(2766{l#H$wP!6MGke={BV4m?=>4VZm~eyvE6J56$Cx|~N5 zRF_+^S2rdj0Ke8i3-*|M<yi*8kF*qgP|TDjZBTb1C$!;tG-`P4grH0%!~b*wOW zQ0=z#+pAAX3?IcFAK%SDNxUeE751HpF9oWbL&+3TlIOm<7T|MqUaX2j3+KtfqKo#f zc9Gswo|CP((^wuuRx$M=flMrr>~PG%);!ouD|uf9(tZ`JgnfED9d0wUe0EAxBBxTiwj!)%t6i5O$SnEf+`oL1DOYe+jXyRL zy(v{v$f|(-s-Ly2@VK%fg1)e*=mF$;qY^Z18G2l<>4*h4gnE=_xOIw*-o_WK-p8b_|`)8kD#ND_9Fd?-r+UoH8CKi&gs0h$dxF9vBZnxclEgG?W)F{7s4TJ zya7qiC@#H1eCv74uJzBgj-E1l&TZeRp&6mcy_p6&EFK_FoN*63J*N`A)v3QddGKj0 zm|~W{tCg4^`ZWqu@D>tz1rH{-sRg?pi@H70X?+zX`Ux>l> zqL#3EP6UkV%pDq}s%L8vP_0>|#{hWPvgKdk@E;JsF|dBoV_?=s`TZERTTe#m&wavr z;q4Om=_VGj6Q|@&fW>F~Y8RoXRpUXL=a+u+7@I2B6NBp2E-0rOa3oo9O3Qgg%*g6{ ze8H}$?N@|8tBt9+4bIgz<4RTKI`P_?vGZ^C@xnbIz1vIlH@@0)HN3*f9jW=I5SCGRbigz>HSTlA)yRkScp)3M)IhOqhwl9F%IS)sS6 zUz<#Es{9r_$P7>%Iyo=4{87rx^9?EYz12HEHxrir>>)9-@1|IPtmPxlD>5?g#JIgt zYA%2OCWsxGAXM1z)z;L!^Sn#ag0azUh3Tz4Gb_6yRdgOP&ERQ@)0^_3A(bq$m<9Lf z2lFwS$vSu4cI`KIm@voR&?XgyabTE4bcmaklZX1g8Cq83UGqko=(cw zN->pqGJN59$V01@#ji3vuIlcs(&F49)LteW7;U^VVElc6a)3I}svbywu0Q@MpR1P{ zwWvAkP)8^s0Nb2m$1ps|QrV-9(2WR6ud)4z(nF`q0mVr zQnY(ud)3T@d|^unYLvX$=H-aq{AlHnnkasbxWJvrxst8QW1z3Hztn3dP%+q?-)Y+h z`osH%=n%4@o#^Y9EWX2Y z9-LL($$ap@ZTU`R*tSN#RBXVA&_zy3@6_qLuf$Y3l^n8t%wemnMXoMp@w`34q*OoD ztflX)xLTNch{ZUG3fr-aDV?~q-X71zxsaOU&bBc_xVTq^3%c%pf6k*OxIcF{+kg5- zl;ODxN#@xAedX#GV?Luo!{bZif>e{)q0f3^)wc~EEOn-ebYHw+ZgDS;@bnGBdKmfIkG<6~u9Gi|6F;dO&)5|m z1!9OcJo1@TUJ>reLW_RnIxl;Ro!cgROz;f`L_OOsCDk5Bia z7Yo*QCd)i-#2o`@vRI%d@ojEJg)2Y#G>hDV97C{rTT@>2t#>pWHzFN=@F&eAKjLLUj6joB3OBBdym8FHjG5(FyvMjzM9@&jT63 z-^=nIq5AIV?r$knZ%7~p2EU4O_*jUVbjh7sk3JH=mm(~3>ygEtXsUE2oT|-|04tgFTMtZnCz{`1qneaO-pd_?pZ!SLxe-VF^q#8 z?b2Qbn|=k{)WJ`N#HarKSGHss?F^32!{5f8wX|`)-LJD~$_$C_LtCd@NaOWjD>{eV zCV-E8jI8r^O~~-z(-zT6%Xv+oxhsT^F(@aJq5K2F)a<4Dx)L6IB zV(zC!ml>H3?5k;+*~QAcC3@;vth)eDi0kANE#2 zMa>e3K#E2=^m?etL_}HiE6)q5(Qlw8DM5zKWZDuc5-4TZrz#(P zSx#L%p@Ehjz&qc&+OfgYRj2_Tqa^^ieYt!GaxAns#e>1Lc79F{p&kv>QL{QHi6l7b zi^x&goc#fwcv}KU?S3sYg~!_}rt>RTqtBTfcQ>iGvLn;1{niuqj!k(&FPwI!q~rK5 zo>eb$TQ3F%@86g z=kteGQee51cG)AjS9DTPrLJD9gH~GVI;Xa{#7?X!Nn1!(*>E=iWh;_`5V9Ire}SK( zU@{!E?+xJdkCtB`YAM0dWnZ?$rQ;se$jgg zn1M)GYiFEzG^7b+kb)e_Ewb5dVor(@oRidv1l@{&m0@D|E%baKqF>*Vo%g}{~NDAYWA}zaj&M0-B zfzlY5?D)Z)+gh)~HLBAxu!pYH_ESOD!i*C=*&H3b+J+#1^rCMZ{^#P(ENs4$ASgwC zssRqQYwY0-S{kUdw-eW{YL(v>eIC@FXwHJ_JSV&q@Bw~fCQSI{b6CZ`ivwzl_4@X> zj)OiOiPEG`IgR|3ahNu&P~s{$Nfbv4jaS*c)7tmkx+wO$SW{wLWIR_+6_dwEz{?N% z$PBmLJHQXsZJ!l=4wgp8Irm(r&R#x#=&L7vxIXW{-oouH$55iXYo*rs?lt%*=@N8Q z`LXy>@T(Vh;J(vh#nT2{TE#h?Hl+4(!}&&<; zBOd%=!iILmINj?%!%3L9mZ+m^%I&Id&THV;lL_cwuL+k#y#@n;+xX7++ z7F>-_VsCcw0FW+iV#)8Mn-rYZS zQGO;EaP>W2??cuO#_(!U?b?T|IuqK@F0{onzv_i&k4@jE=2Gt})2T(v$E-zQi^@1- z7PyL2I%#h&W8(-coW623q~?h^Vi?_H@_*-rIgR(U*h}4)ybAv6{Ls9uTkKm*X4}Z5 zm{d1+gvS`}{8`@Y7jLsu*S(%Js^gSIoQn>V2ZR|QYyhxp;{-7Jz6T3q`Zj_${L&XY z4FH58Nsu$q)auyue&6|vnrHcHGdK1-EJHIb4(Jz0CvmlXq<`GsYZugFR;2$qg-u!mOr>2=&Jp94ac1%dV)4*sFU+t z^*l>Ww+*DNisdI$vdkB+qlPq0yk}~6Ogwrv-)K<)v(POyDu44Ju*JX7wfAKnxcT;v z7Vn+f3Y z5ijdmbhmE!+F!WX?-X)qZFg&wl@s8gW0Qa9hMvqJ-eTP-27JStm+!B@*%Oya0lsXa zXRm3f(Bz!0Rs(Lj3Jd~Ikk1!v!*t#vl|xJmKKRR?C#sKJ@hYdr&y z;ti6vK?$9Ae!i)N{Vz?d+5xEm9~}MIkp>_QH4}*smty`biXHQl`Ex9q8>`5Q-jFu1 z(pcoe_UP{E(#li}Dl1L$d}9AlBb<2Z>j|Zf5g6?G;yce{B7^f3lBU+|$BmWRyFLTb zI4a*1-ta%x%s1TS#e7@x#Z1`sSc`&>uc$%DV0A^F*KlDN834f3&L-{Sq4fgA+>&~0 z8b(GaJlI}^sWq*vQq5u4gNJG0-;czik>=tXYudXV@32Rq6wnrrO(7-Usv90!YxJ&GZbW!?1bKs<|1Q8IalKn zL#OC18aU_{(+eH*@{c>2Qs-R}lskj?$oYsPdJ(yO4dl(AAuxVly@~QNjThFljWoI zR2}V!g{uo&JFSGxExt(g!k;k$p0$DyAXZ|BSykmU-Xi~#grkO653Kg;GF)Xbnex<*pT7x(s&vzQMV4Wqe+(-UJhv$=`ynFMw8Dyk%tiTc;H z1gJ!vv#I$@FiAgGdi*N)8OXMymPMVbm-qgH4M7$!_3>O&@T@4>K|klh^$M6JPPo{} z`9S^{v+y&sTlMPOJH>HH!dUFSK()exGt{%Rhqd~2RAYP}=EFMW-w&RDpMMFSAnMTn z_+UhqDR9jHk=3VtRo~=&dVP&j9xpJwtCh^%BVPmv!F_#4hn5ruQc7VSQn7@hPh{@Q ze2BzlfX*d~X|IQ9;Lez*@)4N7{rSYb9IyhKcX4_{Eh|M!%F5527j$^Sn#t@$yHLyn zmFsLAE*EIL&SeZYg0edjFpX4JsvQ|R<~PRBA)uREf|glqK={e#c&U7$!2^q~riL(wZEXershP6FiSwC!o zQAEa5=s9nmR&K^Y6w1GI7Det`K{0`_$q4}?glpr zxRBI^Bb|&un}HhSKUFNBHE{A+f6A;`tr&$0qQ4kNfv2W@gzeP9*LXN*it==fs^htM zo)oEUlFiJ_%}FAdBWMwcnzgxP@eiC?L85n3nthd%<*IHuf6j(f<0`rl>*iycL=A?CALHSb<6QM`J{zwX z7b}8@do@b@??6)o!#SSGevKU-@0sIXA*ZLC3#Jexe*3KxiW z5h=^OX`s@>V!{fCKF6&qM;q+yiflR;`qZr-nrz;44S}O%QBzjBgtvBLMJcDW1H^I0 z=?IZEk5YxvM3X^63+9~w1y5vX&-ck4n}N`5Q*P2%*X*0}&-Z9@^=m(NQzQX89q>hk zDqZhyWY^K2JQu04t{1k{Y6y8XveA>9J^q zA5N9a!I~0Os4A&7ClTzVj@Pfi6R#EL6hK-XyEYfuL?Xt6*H=d+mXBvK#j`%jJ5Lq# zo%r>>TgRUvzss>;>nj@?OjbDF^oV&?1r`L*s3T?Zty>I%7sr59%C`%P1xuUUsqU8w z+?opaooyKLr}~Mrq91X>f3zUtDzZr}iI@i+BLExZI89Z?U_m ziJ|y6QnX7D82qA!QTVr_pXhW_)B&jJD1i9BiDdc2~{ z?_g=O7!)%a1Um8TK0Y6r$p|3&U~i60^~DWqHJJdk8!lhq@7z9 zgzaqUG{cX(vTJ}(6lYkdlz-1oU3GtDfTy@GI7waM4cK$5#H7jat)aIAk%?sE$>dI7 z!D4rGZ+{IxdHcPX92xGhu-L>$0~@c12@5sqPBBOJ(dtM~q?8&tICl#3tiV8$Ta%MR zv(Sx;$+58@cJsq-jX~9mr%v|oeLSO{X?Tz?o7c=MJ^ z_h@wXn04yZX_^C>ub<4s?9c2gH(J}#I3;2oiz!7zstB5FnFSmQn0FLstES+M(CA|k z2@!0TRaOJ_f*ud8wC_&wg)!Sm4;Kd89|%Rp1972iCPVQM7g9KTC6skkn_j~ckZ5X8 ztfkWEJUELE5j4r)+o@X+d3*HJ5pRo#Mn&!OBz1QR^rLST()$$Vv^u%J)FW40!KeK6 za~{;Jwea36<1-UPOpY6v&iU1%GAYBO#R5+G(_*__NRjZ0V(4!FxM}YYZTcUlu9noxC6hvkNAH~7;ME+2m zNB(dQzc!pwr@ArkF*>L58pGOWrSU?_f0#`~4MAH*;hJj&Tf}QZ&gq6T@C6@9QM4i} z`~}WfZ;jr0vqKCD#?!lMm6FKY8OF{)$0T$?y(EClIWtq&)S`%Gu>0DY_oC$1 z;L*3qpS~oUfnYUlzsk`sR~yeRK-afrcJU3_wO<<)yQMUIFUcdcvp*F4(IA>k`d=YIm<=!!1w#>)YKXsNo>-LJmued zlsk9?39~I%n2zZ6_j4i@Q7-!5a~gcHT7lInYKaHYRSlHJ_TQ=RJqH2gAHVV!)yn}2 zh!`G+5oOz3>^k1x`gN8Q{dDFU#Sb(JRzp5e> z!4s!Cn+P$Er33i8kheRFiOi0EEe`eJ)tY%9cWWJXqth+URSD@QIfHpGEE6sXfg2qI zsFtFJAKBPRU(SJmK~Lzkq~WtN8R~3qt(0luH?avX+W`2B1yI6!4KcYVv0Lv^Cc|YE zO2y||F7H#{Qa`X+o!Si~U^8c%o`-EOC@pAm`v+Vzt9IOu_C4zOVn424W0&;pKA>*1 zaAyN|EWp!#mDR1gl7~UDOD*_n)iPOjl)tjoU|m-$6|F0h(vxBRb2HqGwbzUm-X;Tjh-lJ}NU z?X@g?ty8s`ffQt{C(4wWs*ip+?`Gv!>#TzA>7K!A5v_=LqmhFc+R1L#Y}bLanQg^5 zUhdEv!-Lp%E?=4mmPe~39oD{odqD8KYSf~mNISd6L<3(kE|eYILynjWBduH|;jCJR zzql)F^e*v+;%I(?#dwIYD^Lo_yt=J}#Q(f@dg$q8;&(}mD7M%$pMJEs0|DOOq*z!v z9ox7a8(SZ4GSqQ*wlGc9MrYnvV_3R`Oxqx_9}VT6~UQ*ExjYF3x&b^uz!QQxV7n z-6o-DIP@{FjEmh3n^`>@FK_8MvgC6cYd>9v9mg^qoRh6|0ff8oRQLu?ywWZ{ zv+df^cZhC@IXtGgUB8J|H9VPRSCWiNLre?*(WvP9R@kJ|NUHD@a`XJdrHsv)M7;B7 zoYSH@3FYg2Mj;tu;|T+M57t|F(64u}mbhyqz#mrj+Yc!4pcR;uBHd|d8Bc^2{VGLo zET4lW6g5p|Zi108-C<6=ePS=wy~FlK5pIQ7DfPlrHIymaOl54tl<$Ql!R@OF_vh(+ zOD11@I;-}i4-oz3K4V`p1Wcag%i+wkVp*_1HvTyTSY$NBo&3aP6i0EMRB;L4$34syYKCzz{)jAGUrmy1SR+40=&Ozu{PeSv}q?@*b zriYPrEmmN^Z2Vg@>xMN{G;(f+*>8wA-(wY@1d$eoI+vryL>Xhb4Aq&Kchiz2pfX&E z;k{4F;n?wybfTkOZ)}|%)=b{>s*}PFixlO&x*V>FT=B}u^3b^bQaWA7V0QQ6} zH6#*C8^3{#VWTuWScCP&^M&PIuq?Z5m^}_sT*g+b$Uf?A;fa!ngMn&e@pE_iYX7be zg3_Ob=h+C2#&!cu>0vpn5-(4ce9cyOPG6{Sa^*il5ggCdM$LNkfD|rT)-C*YeTlmi zm;8$tb-@gvw4`7PfY<8zL_Fp~=4}z*nZvf%djbCev~%CU%u3ER>EWFLaQQw}NW&#^ zsd3ZeHTVI<35S+FrM^!{+^IQ!eYS~pa5*j!Pf|k^T`F(cuWO>aBi-x|L;5~Sl+TSK z#MUdB4+z)=5=+tCu{;Xu(_?e9GMr1Y;R$K6AByj81DsZeg!j>!_?iL7Bg5^Tf%9X+FK?@b6 ziu2w!H#*!o**v8e(5;HyXwN$uSiooQk94X_s)8U+z9B!ELZ#o4)|?MlTCKx)ichtq z+BceV<5gE+&$*EEuXF4Z)E&!+mVBnfmjBd9`3`J1$!^fmM?h8yGo<6!6|X~ggH!l? zl^8)EMGEuy2Ihn7w2=?XXU|{v6cAoOQupQ`ohs^%(&f23@dN^zbJHkFw4{bK*HZ)l zeXBiz2hPgtRoZ`vo!foon4#8;9_&~yY+rn6DD=f_O-69z8IxSfMdi(ikP0y!W~y~y zU$GXsY}+NVlCMy*<6hcg)XBxT?eTmlqiRd1ZivBei4LLOux_$7wn|X>GBI`q{b@0{ zy&)R#3(jU@&Q|g1vBsJXEVp(ZT;GvrsRj+^MGM9c6AE&`m+AppuHkNsd(o@AwUD)N zA2>Hs&@%_buRi1F5m9gr zI>VUOwZ8g>ZyZwJb&<9SZauqhZ^V*c z2&ju(COnK%_rtdR0WNxAyy523F*+mb({N`Np^CqW*&=m9TuvUaymQ_+G@*gEs2SDePh(ci-)$W z%ww}paoG+-@5HrwZLD%7=zQAtrxki)-(p(CWRqGU&!npi?5B?pl93PrC5hk_CkO$& z*W)ZP#`Pn%8v%dv#;_@OiwFs;c(mAn+6CO$Iuo|3tgB5Q4miq%Dp=tJw*vwSD=Ql0 z_2Z>)I)SdYCp$T%Q=PR>#|-WM056u{yO&32U&OM?h7s~EJjdSRDwgu10IfNUW;D2# zyGe#n*X!x?PN(}Fw=T&8x#|(iz_PoKU=4K>uIEuV$h5Y{9_?B@Aw@!VT}UMj-#=V? z)dsuzpg+?4GUSz(R>`v_KE3n8HO+K3nqWsG1On%+UYbitmk1Vudxio5<;B(-*v4D& z8i!W~?B$w`p1o#etH=?-4$@3>?Zg$gk+?s)ii%I~@qPzBq#dco0sEQa z7N8g&JL{jtyKF3%0kWSwv!pw2Gh1)dG#aoxWBP+X9KF_~EitwB!j1mt*qH3?m`*iw zZ6IS%>n$0fG0^boO{`}N%k>+on`MRoO{UQhXwuxSpUnN!ADBq|%<&@< zcZX%P!ln_EWeO|vKgmd(iRRwnG507^8mk_oIhtTcX;2xF=;!gLg_Wi*hO8l{i0~4n)Hmq1FXCN+y77LHI5~ zj&ZC%2}R$ImXMh?{j{1@f`r<@5rV)U#r;%NHY_>n)+@?D&eT7I*thpZ3+X!>muMe6 zExlLUo*hTtzAKWj* zkJ}L55Jgy=pGmN=z9ubRm%JH{RSyuKHaIf{K~PMITYMWQ(uO^RTGO$&bOQ~U;s2t1 z_6J+JRsiV?KU{jY0$^BUJdPx`W(t(n)D@aH3e%44gH8>e$zZ9OprYJ*3cmkESmwf` zB^mw~VR<*U!MLCn0!Lrg$UYZ+SyawedG%nTaz|jvXwO*OQc}1OD%7Nnu|^^1Te-gt zx{Tber{YqVt0q)(_FlE5Nz{ZF zKg@h@)0eAzc&$%<00Pc~_XT-G$^2q2Gf{IWtz9vgz)mD>?vb8V<6XrR!2iFP%U85K zp>K??Ei)7$?`R9vtkid}#DYG*1>BSLH5<2rNJT8z)R)=PLN7$oLRT`c)gL`^FzU=J_CoA_(5WUL|2b0{}(l zB`|oetstM4U+lf$fXl())?R07M){K)By?q>C7II5MzYvji)%$#Qh@!t6}@%m>cEh! zf0~1`xc#u9ud=Qi{jl)x2BEF9)-R02pL#8FENNb?v8Kieg)Ybk6ifswmjGOK&<)`n z_7_ht={hV?1Tnq3N+$BDR6G}p`M74Ln`#78UQP8KEdn%lH1`T7JR$WWb>C!_Ae$Liv_?oNxhr&fQayf%~`h^{ucb9aJuCrzfbBn zIu{Uu4@gXEwF>X;NPpNWas5h8RHeI|!NC{@Pyp6sJkKDPg_hIrRbpEHv6!6n8rbYC zZ$e?Z>^<_=-ul-uQkms;X?Ev$B(SFr7-=}CX3SgW{x{2!rIhe$u3vn`A{mnRd3AV+ z6klcxee9Fefk8>nOI=)?wo5(I*H<_5a6PUAf|o}5Lzi_Y_)yQMQ^yt6&K`$;_?`L+ z@Vu;>Q$JTdpe4Ld>5TU|Xb1W-KAp_pN}=yhy2>TjCWZq`(f_0!v;AFW+|RzUV!hG* z0lzM#d@8>Tk$E>??0(yNgY;5dr;R$G+#Lbv^<^5nxVZF0VIneAnOW6VQEnvT0GOcp zUx3*^R=4WQP~MbPYr>Oj*aw}|I9uxV%umQt6#sQ&IciW^09iJLL?%mHTbo?TXE=VE za;~=v!|#2bap{q)ZH`5Oq-0oBPUl3UBJMu;tNsHPAv(1lh4uck4(|O6_D>4YGkMP< zzu>5E-t+A&%z5PdMC}??i3#-BNUem2+jB8@OwFD^;aUuTJu+^wU(~B?ymrcvzk5mU zbintymzAZJ7mQ}M)oG;xD9s8Yhy44%Qj3Lsv3Yx#CBhwBGzh##vm8jk5`cyCUk5dF zJc3=C@$fKu0TrCZIRTZ_+XI3B@Vx+tD@Z-sjjj;*=L2H%33g9h;;rSWM5h0`;vaWW zE5-ZJY+MS4r8Gnkn4eMXpgf7Yr(Nj($G2%O8_}Wq%9r~o&W$M!eAwHwE^nRxc%r|a zjp7z00DSZP)j-G$o_mKs3qI^nP+k3V<&S@CH5LxLMgRkv{lIM{mpC#YxmkbiuS&=5zR4{lE9yA9qRnPIxD%@jH}lN==^fI~MHqDQmmEaKnIpWz;c!)4{EO>c_ay<(iz} z!LbAIvu3Bj^H`o0hB*QLfp2dAIA{UDvlHyDOos_w>#BtKy#6qbfhxbmAZ{B~=&DH|l9krw*lk-`tkuCA^ax3dC5wpqlasNls}=wKfy{QNW0#^%NN4acDzOD^2=Q?4t* z2orBjq$}B!IE7@$pdCrd&WJZ#O-%lAutU-yZFfws`H?!(!9R0E(8)FJxbAFtrfS)! z*ITkg?6@9S!S3Lkk1OZ=^FS^VlFtcv1GV4`2tb}&B0|T?Nf(<9&1jV6^~N|{-x$(O z5JsuwsXxxyO-_b?eue;QmM3J51O#llEYnI;s>*~ihss#Stw^63ThAJ1rkbWg*!4yj zxVy4%>Gz@icb}(`69J8Q$xKxFtPbz)-MjX8FuV_aXW3V;JwTths(*@ljha?%x|!YN zZxRRF6IodM=ga(SMyr|NeEcYDwQlEZ+{tFnL$Gz>QNLnJ0bb z{7*&-{H@&>7{hjq&@v(MTL?=Pvr@mAh^Ii6_*>Y0)IMbD%73pG zmlxf9ify^O^VFZ;NF9W3J}*NR3QM9PO%g-)A1hU&(^Zv;M=k2zh}{XvHiMbtsU%Y< z|8r^jt7lpO(+Ga&{2mgBhtmu#D=qA9b<;`}hEt(Dm;JMphuk{SN-LBF5~+vDK>I;E z*nvkALH}cp6M#Sjy9bE@5NyJX#19SnY6v`}FrTWz9X{0%!X4um4i9YVR6{a8zD0@U zfqBvfl+JO^bZD>tQ@h}%tiS{r^NV>T_ng9+%iCRYTF7J6KkMzf-Jwws>PD^TBsNWK ztYh#hJZtI4T*Ed#&bepjx1CC1xb)=m)&!%MiZ!3?KTaPOH431a@h?>qtdu@BZM(bW z_Feta)Noz7+sw;hcV=_#5UL?G_xY}Cn_#ZUa#0ylqn{9d;?MG(`TfzPVCivC(1hN_ z|CrWM{6MQSFjscbr`HKGe*W+~Tb9K=xqS;}xsz{z5@~T-ID;plXpHprY7U7&hZ^cQ zWk})Gq9?fLh30t2h9vs`IQ__h^49b0r5>+eys7 zmDwn%Z{5n+_PORH9V|qWKxf%L7(v<~J=E?=JQF^nHtI6EWo0;hL?)pn2P2tScf%VH zdI;~e4dp6qW?FSy&3Blt@t@!rLl}G?oorYp^TLQV*b#Vo7)n*{eORXxOW6~gv}a%R z$1T`CEF?%WLO@TDjxsjZ}$YDVvS@ z%hmmW_WIX0KmJM9U91O3#9xDF-ilA7T2~5^w^nQmnQ0}8%c1?#bj-;auom6SOicHU zIs~2vF4*TI>!E_{Ey~|4N|CraFM7?eE;;0ZK6PiN!j@>ZS7b1DCb@~ zRjBR92KW#zd?O;EonK|$AqcvuiHkkZbOI;6S@~RoLEKZ)9i@A^tvXP~_0*jn&Ehd& zsE}8>jw@R@k|&xR6i!}8GV~cEd<0+F4P8z)?-;bzm@2DTL7-f2T1|@IN%B%n$Vwl@ zml5_#ry-ijED#MoP6#UB7TP_-pHy~Bs~Xyp&e5DpNlI||R%5c;k~R8JHTVZG6SJrt~+6jBUhPoL}fh}cU>1txYWyE8-a!I>l;5@gRWlnpRO5x1cxFYE-@>iP87A6VSK3nF-N073O0{5 zf)XP`uH40q?3XLh%YG^gy*5WLiJNwD4XRuLRLQQt(t5sfHZU|F#Ga=*=Tj)LF@vTN zEN>PRCdfll_-l=N!y^q~Ua;fpKr`#aE81J5?g-n5DUJ^HxmUMAg=RN!Fy%pshO;8h zU`Id1^0sWaL-Hx>_v@R?`B#S%L@Fs69y*I_D_hK>lPcj;-#lsi}_2g2>Ftw2)Cyn6gFM zv?t+c4HPq)w1&zGI-fe4ntN3*=zh<4?lqCPXW_ORnZ2?xRQLHst?I_TaZK8?HP>Rf zOm6I2h!_*e9KOGS57YSwVc$*w_;ZC$4dMDt0)LZ9?B0FQEaoZD#gAOJfcQo`;f27P!Uio(5IC~_D z3-Y-#86-PXtV=^+)OSQ0#{B52(YL8ex`8WUM@;u1jJtgkR~&^mN&>a-(w|QqPBn?_ zTHokX5d>bH6dweSz>;bC4s*x#R^2tdCK!`iuat)sG|u>0MfD|f;?=gZA`vYeUnhyp zkC5QTnUt|-kjM$#|3}w*M>Vx|U!a131re+$ND~zi0TmFbK@pLzpdd9W(o2-yBXR{5 zX`<4LRFM+tC4^9<_ZoWWEhK?J5=ebVz1Mqx-+S+$j4&8U&e>=0Rpy#=tq_N}`chji z!Ht8em3bpkccEa*ezDEEn@_%#pKqGHn{E}Al9pOvmoWbro--wGE~RnFSnE?_xnY)T z=7XOeNfwggV2i;tRZUWCkIPk*61(@?qS?vV6R(I3oTL6j?&aY0mFfn|>;Epb|Lw=P zmh*(EcdOT|r&vwP0})vhTB5$lPqJIy^@)q5&lL}o?)Hvjvo#hue1+c29Qorgw9C}; zVwl5|ScOk&_1E%Ec8V*0drz^D%Dj+4 zzy2g1(e}Yt-9a(ZpBbO~o)Fm^2rf?8sK~(Eh@NZTo9`b$aD<5b32;6h3N}Z#poz|tvx*O=bF_wrd?`9?C(AK~2EAy$K$NYk$l`4&t6$Mk&7Jgq*OWMENt^If(yQjO+`1;f%eaC?BE5@C0PUHb+ zUbYI7a1UsE{BjMynDL{^?QM@;yJ^-08Sm7gJWCyt%b+1?JU`$OY=_%>eN<$1b+f=! z%&^A8VawQhd*^lngMeME$4YF@NlHq( zX@aq6W>*|?i6#m3-i<-z8~=nzbn=ZQv)=C(WPU2`fj^4w5HtzQA2EI{Z5ejZ0f3WRix9eO{3SE|c7UPKGy^wDLD<;Q4w>Ba-dFb2~C$QZ*5|dO2qr%QyQ|4LW)?KS?dQA@Zi12|4 ztWGwXk>@%gZJ7qD5D2781`vRIQ7^0;BPkUB(>GQwjn(?<{J84LK{{`8C?=tFsSW5% ze(dT+{>|G))y`w@?DfXhWMpJYhZnOmgbN97AdTHKDdo(*OoPrTvdU4ysbbaq)1{0L znNc8_HGJ%0S;};nO@{%pL&kt%LMk#;c;}orE`8R}&O+ZGcccXYl~{INT2*ky4HyAXdq$ z`t_s?w%keS7wRoP#{9vqG-oD5^`C8QCY=QoetM(L9y!E3JByVFZ;FFh+h*%_>{549 z`@0hoixc?VHWf0+aYD$KW|~wYCNHzgD`oYx7du>A2D+;ycXSe@%FL`{RP(y?RC}&cq<1J znY^8%O#0z|jh2I%6HJTzeI$>6+#|EYsr~syc($6cFj~cGW>L*P(8sdiagcoy|HDc) zGM#Vz_W=M-!E5z;_BBKRwIPAZ@Z(x{;z5JU^r5oZQCtKVw+oqU15ynpw?OASYnOYD zxXsp=*3JWM2;Xwp%iZvwZAOh!-hu-JyuS&i(6MzV(i`YH94X5Y-TJd9`^vO=E!_&q zyAFvts+xkAq7q*Q6c7N$f&a3H%Jxf1J+5cUU#$>{vQFcV4&!{I*P-_yN?qzv zs-h+qUPnrItiKF|{xq!7$61ce7<716?Pt}vD@+0%A7dsQ;9kLG-YMc#iJJ*JLdl*n z{$na%lx|smpN`qV&`Y;~`~@Tx7{y7WZjs)KaG5q!* z5wSi}IMMaAWYsHX6qqjHTbXea2Uwr6`GY4&gjXq@-RnEPTPL+0&A!RH4ymn;n;_b8Zgjt7B9GcL;k#|sJfLeR={67f^T@S)?Gi!htf~q+wjOG z`O2mQHF@*$_nMI#)rN2{y-M@P6O)gs>$0^!T?;i$Gui3RD?drsrE~7V%rz)^<4V$U zgNn?yOBwUM2^poMN@Wm<;cqw>j41sJQkvP7B%O~ch@}wM?jGB^c78f}%MZ`lRjIJ; zeHT`URH9Jh+cMHh{T19}d{^Bu8zT3&R^ga6nJH?$HSQh`K{dvcr}AZ$BKYM$`r~Uf ztoZmJshJoN7q+p7nY%Sen@;0?G+44`zlSn1szL_(8~fnVsAt=;-Y1iptSQHO{;n_q zBs#|Ie|z#fH!SKnt#@uAt^G3RA0rancy7j^J@cj?bTfDDMd@z&6c+BqMR3v@xg5kb zI_=d;zeXCaT=$$3Iw4^kt)$5nE2FP;pM)J9cZ{D0I^o-@8k19)|E=eISM0X2su- zS2}$nvGC!OkOYy69JZq@Qu9bVlyry?N{2tsDs3A$OW zxfq{WV)ed?gKp_J(;&9cZfE83&?~=Ay?ci@8EShM!Bp!-Oe_>ZUAyubj&PXFdf4H* zTalm&a79I}@e00DDrW+TKk|ZWa|>pK2x3-mBs(KJtun{@m){c$77rX#F>hbV660oc>T!sAZ;fggtlFoziH(tEZi?TNuJ4&ao)75{i_wb`S6 zrWCVDXPUkDvh6a}jsQQ+z_AB&5;XZm?-P~0u_c^+LDZ?JvCyH@Jtko9(#w`L%H~P= z8`5_uTfIh$BNKQ}JrQ#O<^l_ETh?2=R?2vJZN=1*IZN?7DD4sbY`Mb>W3tpBeCW8L zy!DbZ|0JqI-A~}SH(%GV(}>UfP=RR^E`J%xl!`4}GUrmkJ`?+bm7PB~ULRS^ic#6MocL;sRDMSp3!GX^=3>~2izeFj$`uWw1m3gyy1 zU@Z#!UPteB_$a7-H=4Pj?%z>vVtEo$Ojwf&(B5vGAD#{_d2qVwhxXa|D|^I#G5n8y{k{G z+40up>@DZ3(qg5t|Ey_!>M?tx>=#3?Ul5CAeemoX@@#HcxG9xls0Ou22UFqA-k!+x z4z4CwUf^e5&h0KAindazY4N9H0IoN{e+ek>O_tNt zSEb*AyuZo&R}iMr>CN}j7Wok51GB_KNY655;<}EAisw=_#|QV)=a9BNbs;J?*)nUM z7pm(nk0FiriJMov_gMd)hVNm8®^aQ5#KOZ7pfex#L4ou3P%--2~N$WQyBD=d{3 zPC7W~%0dwml#H8p1@M`=M6Me|pVjHYaJ@m9)g(L?A4Wn6LnGFtb)s;VldHbtElN&k zP!*%7Dg!w|>5hFOij(EUFc!XHjnkkTa&a9Ad-V_I^{&oLkwKqzU}qfRtKqEOcvxP< z#oudv?iK(}zP3Jzy*PC7a~BW%Mn$UNdN*T=VN zhHIBxj|>wr^1%E@-EwKcNgqfZrQpnW=sxfGd9UVlDSVBJ~NO z1D;1)pr`JNKHEEikNmVY{LK$bDfwzSasAp(sJY9O=`#2dlzw%r)_#BOIfCNJjYw$s zYjF_s0_99h$F0(!`iH2)rP5_ZH39dsq{h~!`I*JE<&uU9tCUvcR>|@0&O>o9sWK zUA5wq!Wss}3?JXBg@X$hc3{%`ODJMw)oQ*!z^_^cyKRCgRjLZ#GNLHS*N1uvW{s~$ zp5$MU=iCTQv*tgx(OR|&^~+SI%&bJu)H$sL^#xyB8{xk_Shn!MV~IHBeYkg|&@lTQ zVPGpttG6k#*kyKiw>0IR+6jJnmk(6Epq-5jmf=>xTAjmdUXOi@y5=9|ml$r_KL>A3 zabK$*ON!r=ZBtrJYM*W_6umPoxu_S!-0D?PS&DyMy!rV_$oxSZ1_5FYg@^!6fRq(8 z>HM)dT6U|_mK7F5MYi2*6}v85T-+vBevE$L`tm`&0i_EQcxV}h)qd?OU~X^v3Hg3? z((f2SF1_N!VequKtwQH1XPq|Zpx8>Q$*Kqb*B!2 zlrOMWsd%BFG&5Krw%T@tk%DRE7$CA4QC42u{0Pp~Cx;GeJkR{p#Em^+@)8+!pla=G zd5kuojB|;8v=^;NRj9Pt{E%%sJ6)imFPaKr3