Fix follow ups in thread + fix user name (#3686)

* Fix follow ups in thread + fix user name

* Add back single history str

* Remove newline

backend/alembic/versions/c5eae4a75a1b_add_chat_message__standard_answer_table.py

@@ -0,0 +1,36 @@
+"""Add chat_message__standard_answer table
+
+Revision ID: c5eae4a75a1b
+Revises: 0f7ff6d75b57
+Create Date: 2025-01-15 14:08:49.688998
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision = "c5eae4a75a1b"
+down_revision = "0f7ff6d75b57"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "chat_message__standard_answer",
+        sa.Column("chat_message_id", sa.Integer(), nullable=False),
+        sa.Column("standard_answer_id", sa.Integer(), nullable=False),
+        sa.ForeignKeyConstraint(
+            ["chat_message_id"],
+            ["chat_message.id"],
+        ),
+        sa.ForeignKeyConstraint(
+            ["standard_answer_id"],
+            ["standard_answer.id"],
+        ),
+        sa.PrimaryKeyConstraint("chat_message_id", "standard_answer_id"),
+    )
+
+
+def downgrade() -> None:
+    op.drop_table("chat_message__standard_answer")

backend/ee/onyx/onyxbot/slack/handlers/handle_standard_answers.py

@@ -150,9 +150,9 @@ def _handle_standard_answers(
             db_session=db_session,
             description="",
             user_id=None,
-            persona_id=slack_channel_config.persona.id
-            if slack_channel_config.persona
-            else 0,
+            persona_id=(
+                slack_channel_config.persona.id if slack_channel_config.persona else 0
+            ),
             onyxbot_flow=True,
             slack_thread_id=slack_thread_id,
         )
@@ -182,7 +182,7 @@ def _handle_standard_answers(
             formatted_answers.append(formatted_answer)
         answer_message = "\n\n".join(formatted_answers)
 
-        _ = create_new_chat_message(
+        chat_message = create_new_chat_message(
             chat_session_id=chat_session.id,
             parent_message=new_user_message,
             prompt_id=prompt.id if prompt else None,
@@ -191,8 +191,13 @@ def _handle_standard_answers(
             message_type=MessageType.ASSISTANT,
             error=None,
             db_session=db_session,
-            commit=True,
+            commit=False,
         )
+        # attach the standard answers to the chat message
+        chat_message.standard_answers = [
+            standard_answer for standard_answer, _ in matching_standard_answers
+        ]
+        db_session.commit()
 
         update_emote_react(
             emoji=DANSWER_REACT_EMOJI,

backend/onyx/chat/answer.py

@@ -12,10 +12,10 @@
 from onyx.chat.models import CitationInfo
 from onyx.chat.models import OnyxAnswerPiece
 from onyx.chat.models import PromptConfig
-from onyx.chat.prompt_builder.build import AnswerPromptBuilder
-from onyx.chat.prompt_builder.build import default_build_system_message
-from onyx.chat.prompt_builder.build import default_build_user_message
-from onyx.chat.prompt_builder.build import LLMCall
+from onyx.chat.prompt_builder.answer_prompt_builder import AnswerPromptBuilder
+from onyx.chat.prompt_builder.answer_prompt_builder import default_build_system_message
+from onyx.chat.prompt_builder.answer_prompt_builder import default_build_user_message
+from onyx.chat.prompt_builder.answer_prompt_builder import LLMCall
 from onyx.chat.stream_processing.answer_response_handler import (
     CitationResponseHandler,
 )
@@ -212,19 +212,6 @@ def _get_response(self, llm_calls: list[LLMCall]) -> AnswerStream:
             current_llm_call
         ) or ([], [])
 
-        # Quotes are no longer supported
-        # answer_handler: AnswerResponseHandler
-        # if self.answer_style_config.citation_config:
-        #     answer_handler = CitationResponseHandler(
-        #         context_docs=search_result,
-        #         doc_id_to_rank_map=map_document_id_order(search_result),
-        #     )
-        # elif self.answer_style_config.quotes_config:
-        #     answer_handler = QuotesResponseHandler(
-        #         context_docs=search_result,
-        #     )
-        # else:
-        #     raise ValueError("No answer style config provided")
         answer_handler = CitationResponseHandler(
             context_docs=final_search_results,
             final_doc_id_to_rank_map=map_document_id_order(final_search_results),
@@ -265,11 +252,13 @@ def processed_streamed_output(self) -> AnswerStream:
                 user_query=self.question,
                 prompt_config=self.prompt_config,
                 files=self.latest_query_files,
+                single_message_history=self.single_message_history,
             ),
             message_history=self.message_history,
             llm_config=self.llm.config,
+            raw_user_query=self.question,
+            raw_user_uploaded_files=self.latest_query_files or [],
             single_message_history=self.single_message_history,
-            raw_user_text=self.question,
         )
         prompt_builder.update_system_prompt(
             default_build_system_message(self.prompt_config)

backend/onyx/chat/llm_response_handler.py

@@ -7,7 +7,7 @@
 from onyx.chat.models import ResponsePart
 from onyx.chat.models import StreamStopInfo
 from onyx.chat.models import StreamStopReason
-from onyx.chat.prompt_builder.build import LLMCall
+from onyx.chat.prompt_builder.answer_prompt_builder import LLMCall
 from onyx.chat.stream_processing.answer_response_handler import AnswerResponseHandler
 from onyx.chat.tool_handling.tool_response_handler import ToolResponseHandler
 

backend/onyx/chat/models.py

@@ -8,7 +8,6 @@
 from pydantic import BaseModel
 from pydantic import ConfigDict
 from pydantic import Field
-from pydantic import model_validator
 
 from onyx.configs.constants import DocumentSource
 from onyx.configs.constants import MessageType
@@ -261,13 +260,8 @@ class CitationConfig(BaseModel):
     all_docs_useful: bool = False
 
 
-class QuotesConfig(BaseModel):
-    pass
-
-
 class AnswerStyleConfig(BaseModel):
-    citation_config: CitationConfig | None = None
-    quotes_config: QuotesConfig | None = None
+    citation_config: CitationConfig
     document_pruning_config: DocumentPruningConfig = Field(
         default_factory=DocumentPruningConfig
     )
@@ -276,20 +270,6 @@ class AnswerStyleConfig(BaseModel):
     # right now, only used by the simple chat API
     structured_response_format: dict | None = None
 
-    @model_validator(mode="after")
-    def check_quotes_and_citation(self) -> "AnswerStyleConfig":
-        if self.citation_config is None and self.quotes_config is None:
-            raise ValueError(
-                "One of `citation_config` or `quotes_config` must be provided"
-            )
-
-        if self.citation_config is not None and self.quotes_config is not None:
-            raise ValueError(
-                "Only one of `citation_config` or `quotes_config` must be provided"
-            )
-
-        return self
-
 
 class PromptConfig(BaseModel):
     """Final representation of the Prompt configuration passed

backend/onyx/chat/process_message.py

@@ -302,6 +302,11 @@ def stream_chat_message_objects(
     enforce_chat_session_id_for_search_docs: bool = True,
     bypass_acl: bool = False,
     include_contexts: bool = False,
+    # a string which represents the history of a conversation. Used in cases like
+    # Slack threads where the conversation cannot be represented by a chain of User/Assistant
+    # messages.
+    # NOTE: is not stored in the database at all.
+    single_message_history: str | None = None,
 ) -> ChatPacketStream:
     """Streams in order:
     1. [conditional] Retrieved documents if a search needs to be run
@@ -707,6 +712,7 @@ def stream_chat_message_objects(
             ],
             tools=tools,
             force_use_tool=_get_force_search_settings(new_msg_req, tools),
+            single_message_history=single_message_history,
         )
 
         reference_db_search_docs = None

backend/onyx/chat/prompt_builder/build.py → backend/onyx/chat/prompt_builder/answer_prompt_builder.py

@@ -17,6 +17,7 @@
 from onyx.llm.utils import message_to_prompt_and_imgs
 from onyx.natural_language_processing.utils import get_tokenizer
 from onyx.prompts.chat_prompts import CHAT_USER_CONTEXT_FREE_PROMPT
+from onyx.prompts.direct_qa_prompts import HISTORY_BLOCK
 from onyx.prompts.prompt_utils import add_date_time_to_prompt
 from onyx.prompts.prompt_utils import drop_messages_history_overflow
 from onyx.tools.force import ForceUseTool
@@ -42,11 +43,22 @@ def default_build_system_message(
 
 
 def default_build_user_message(
-    user_query: str, prompt_config: PromptConfig, files: list[InMemoryChatFile] = []
+    user_query: str,
+    prompt_config: PromptConfig,
+    files: list[InMemoryChatFile] = [],
+    single_message_history: str | None = None,
 ) -> HumanMessage:
+    history_block = (
+        HISTORY_BLOCK.format(history_str=single_message_history)
+        if single_message_history
+        else ""
+    )
+
     user_prompt = (
         CHAT_USER_CONTEXT_FREE_PROMPT.format(
-            task_prompt=prompt_config.task_prompt, user_query=user_query
+            history_block=history_block,
+            task_prompt=prompt_config.task_prompt,
+            user_query=user_query,
         )
         if prompt_config.task_prompt
         else user_query
@@ -64,7 +76,8 @@ def __init__(
         user_message: HumanMessage,
         message_history: list[PreviousMessage],
         llm_config: LLMConfig,
-        raw_user_text: str,
+        raw_user_query: str,
+        raw_user_uploaded_files: list[InMemoryChatFile],
         single_message_history: str | None = None,
     ) -> None:
         self.max_tokens = compute_max_llm_input_tokens(llm_config)
@@ -83,10 +96,6 @@ def __init__(
             self.history_token_cnts,
         ) = translate_history_to_basemessages(message_history)
 
-        # for cases where like the QA flow where we want to condense the chat history
-        # into a single message rather than a sequence of User / Assistant messages
-        self.single_message_history = single_message_history
-
         self.system_message_and_token_cnt: tuple[SystemMessage, int] | None = None
         self.user_message_and_token_cnt = (
             user_message,
@@ -95,7 +104,10 @@ def __init__(
 
         self.new_messages_and_token_cnts: list[tuple[BaseMessage, int]] = []
 
-        self.raw_user_message = raw_user_text
+        # used for building a new prompt after a tool-call
+        self.raw_user_query = raw_user_query
+        self.raw_user_uploaded_files = raw_user_uploaded_files
+        self.single_message_history = single_message_history
 
     def update_system_prompt(self, system_message: SystemMessage | None) -> None:
         if not system_message:

backend/onyx/chat/prompt_builder/citations_prompt.py

@@ -144,9 +144,7 @@ def build_citations_user_message(
     )
 
     history_block = (
-        HISTORY_BLOCK.format(history_str=history_message) + "\n"
-        if history_message
-        else ""
+        HISTORY_BLOCK.format(history_str=history_message) if history_message else ""
     )
     query, img_urls = message_to_prompt_and_imgs(message)
 

backend/onyx/chat/tool_handling/tool_response_handler.py

@@ -5,7 +5,7 @@
 from langchain_core.messages import ToolCall
 
 from onyx.chat.models import ResponsePart
-from onyx.chat.prompt_builder.build import LLMCall
+from onyx.chat.prompt_builder.answer_prompt_builder import LLMCall
 from onyx.llm.interfaces import LLM
 from onyx.tools.force import ForceUseTool
 from onyx.tools.message import build_tool_message
@@ -62,7 +62,7 @@ def get_tool_call_for_non_tool_calling_llm(
                 llm_call.force_use_tool.args
                 if llm_call.force_use_tool.args is not None
                 else tool.get_args_for_non_tool_calling_llm(
-                    query=llm_call.prompt_builder.raw_user_message,
+                    query=llm_call.prompt_builder.raw_user_query,
                     history=llm_call.prompt_builder.raw_message_history,
                     llm=llm,
                     force_run=True,
@@ -76,7 +76,7 @@ def get_tool_call_for_non_tool_calling_llm(
         else:
             tool_options = check_which_tools_should_run_for_non_tool_calling_llm(
                 tools=llm_call.tools,
-                query=llm_call.prompt_builder.raw_user_message,
+                query=llm_call.prompt_builder.raw_user_query,
                 history=llm_call.prompt_builder.raw_message_history,
                 llm=llm,
             )
@@ -95,7 +95,7 @@ def get_tool_call_for_non_tool_calling_llm(
                 select_single_tool_for_non_tool_calling_llm(
                     tools_and_args=available_tools_and_args,
                     history=llm_call.prompt_builder.raw_message_history,
-                    query=llm_call.prompt_builder.raw_user_message,
+                    query=llm_call.prompt_builder.raw_user_query,
                     llm=llm,
                 )
                 if available_tools_and_args

backend/onyx/onyxbot/slack/handlers/handle_buttons.py

@@ -127,7 +127,7 @@ def handle_generate_answer_button(
                 channel_to_respond=channel_id,
                 msg_to_respond=cast(str, message_ts or thread_ts),
                 thread_to_respond=cast(str, thread_ts or message_ts),
-                sender=user_id or None,
+                sender_id=user_id or None,
                 email=email or None,
                 bypass_filters=True,
                 is_bot_msg=False,

backend/onyx/onyxbot/slack/handlers/handle_message.py

@@ -28,12 +28,12 @@
 
 
 def send_msg_ack_to_user(details: SlackMessageInfo, client: WebClient) -> None:
-    if details.is_bot_msg and details.sender:
+    if details.is_bot_msg and details.sender_id:
         respond_in_thread(
             client=client,
             channel=details.channel_to_respond,
             thread_ts=details.msg_to_respond,
-            receiver_ids=[details.sender],
+            receiver_ids=[details.sender_id],
             text="Hi, we're evaluating your query :face_with_monocle:",
         )
         return
@@ -70,7 +70,7 @@ def schedule_feedback_reminder(
 
     try:
         response = client.chat_scheduleMessage(
-            channel=details.sender,  # type:ignore
+            channel=details.sender_id,  # type:ignore
             post_at=int(future.timestamp()),
             blocks=[
                 get_feedback_reminder_blocks(
@@ -123,7 +123,7 @@ def handle_message(
     logger = setup_logger(extra={SLACK_CHANNEL_ID: channel})
 
     messages = message_info.thread_messages
-    sender_id = message_info.sender
+    sender_id = message_info.sender_id
     bypass_filters = message_info.bypass_filters
     is_bot_msg = message_info.is_bot_msg
     is_bot_dm = message_info.is_bot_dm