upd: better onboarding doc

README.md

@@ -23,24 +23,35 @@ pip install -r requirements.txt
 
 You will need to set some env vars, either in a .env file at the project root, or just by exporting them like so:
 ```shell
+export PYTHONPATH=.
 export OPENAI_API_KEY="xxx"  # API key used to query the LLM
 export EMBEDDING_API_KEY="xxx"  # API key used to query the embedding model
 export DATABASE_URL="sqlite:///$(pwd)/database/db.sqlite3"  # For local developement only. You will need a real, cloud-based SQL database URL for prod.
 ```
 
 Start the backend server locally
 ```shell
-python backend/main.py
+python -m uvicorn backend.main:app
 ```
 
 Start the frontend demo
 ```shell
-streamlit run frontend/app.py
+python -m streamlit run frontend/app.py
 ```
 
 You should than be able to login and chat to the bot:
+
 ![](docs/login_and_chat.gif)
 
+Right now the RAG does not have any document loaded, let's add a sample:
+```shell
+python data_sample/add_data_sample_to_rag.py
+```
+
+The RAG now has access to thn information from your loaded documents:
+
+![](docs/query_with_knowledge.gif)
+
 ## Documentation
 
 To deep dive into under the hood, take a look at the documentation

backend/config.yaml

@@ -1,7 +1,12 @@
 LLMConfig: &LLMConfig
-  source: ChatOllama
+  source: AzureChatOpenAI
   source_config:
-    model: llama2
+    openai_api_type: azure
+    openai_api_key: {{ OPENAI_API_KEY }}
+    openai_api_base: https://genai-ds.openai.azure.com/
+    openai_api_version: 2023-07-01-preview
+    deployment_name: gpt4
+  temperature: 0.1
 
 VectorStoreConfig: &VectorStoreConfig
   source: Chroma
@@ -10,12 +15,12 @@ VectorStoreConfig: &VectorStoreConfig
     collection_metadata:
       hnsw:space: cosine
 
-  retriever_search_type: similarity
+  retriever_search_type: similarity_score_threshold
   retriever_config:
-    top_k: 20
+    k: 20
     score_threshold: 0.5
 
-  insertion_mode: full
+  insertion_mode: null
 
 EmbeddingModelConfig: &EmbeddingModelConfig
   source: OpenAIEmbeddings

data_sample/add_data_sample_to_rag.py

@@ -0,0 +1,10 @@
+from pathlib import Path
+
+from backend.rag_components.rag import RAG
+
+
+config_directory = Path("backend/config.yaml")
+rag = RAG(config_directory)
+
+data_sample_path = Path("data_sample/billionaires_csv.csv")
+print(rag.load_file(data_sample_path))

docs/recipe_vector_stores_configs.md

@@ -8,12 +8,12 @@ VectorStoreConfig: &VectorStoreConfig
     collection_metadata:
       hnsw:space: cosine
 
-  retreiver_search_type: similarity
-  retreiver_config:
-    top_k: 20
+  retriever_search_type: similarity_score_threshold
+  retriever_config:
+    k: 20
     score_threshold: 0.5
 
-  insertion_mode: full
+  insertion_mode: null
 ```
 
 `persist_directory`: where, locally the Chroma database will be persisted.
@@ -24,4 +24,4 @@ VectorStoreConfig: &VectorStoreConfig
 
 `score_threshold`: score below which a document is deemed irrelevant and not fetched.
 
-`ìnsertion_mode`: `null` | `full` | `incremental`. [How document insertion in the vector store is handled.](https://python.langchain.com/docs/modules/data_connection/indexing#deletion-modes)
+`ìnsertion_mode`: `null` | `full` | `incremental`. [How document indexing and insertion in the vector store is handled.](https://python.langchain.com/docs/modules/data_connection/indexing#deletion-modes)

requirements.in

@@ -12,7 +12,7 @@ streamlit
 extra_streamlit_components
 google-cloud-storage
 openai==0.28.1
-langchain==0.0.316
+langchain==0.0.354
 chromadb==0.4.14
 tiktoken
 gcsfs