Add chat_model data for logging selected models to telemetry

This commit is contained in:
sabaimran
2023-11-10 16:46:34 -08:00
parent e695b9ab8c
commit 54d4fd0e08
2 changed files with 19 additions and 12 deletions

View File

@@ -606,7 +606,7 @@ async def chat(
return StreamingResponse(iter([formatted_help]), media_type="text/event-stream", status_code=200)
# Get the (streamed) chat response from the LLM of choice.
llm_response = await agenerate_chat_response(
llm_response, chat_metadata = await agenerate_chat_response(
defiltered_query,
meta_log,
compiled_references,
@@ -615,6 +615,19 @@ async def chat(
user,
)
chat_metadata.update({"conversation_command": conversation_command.value})
update_telemetry_state(
request=request,
telemetry_type="api",
api="chat",
client=client,
user_agent=user_agent,
referer=referer,
host=host,
metadata=chat_metadata,
)
if llm_response is None:
return Response(content=llm_response, media_type="text/plain", status_code=500)
@@ -634,16 +647,6 @@ async def chat(
response_obj = {"response": actual_response, "context": compiled_references}
update_telemetry_state(
request=request,
telemetry_type="api",
api="chat",
client=client,
user_agent=user_agent,
referer=referer,
host=host,
)
return Response(content=json.dumps(response_obj), media_type="application/json", status_code=200)

View File

@@ -132,6 +132,8 @@ def generate_chat_response(
chat_response = None
logger.debug(f"Conversation Type: {conversation_command.name}")
metadata = {}
try:
partial_completion = partial(
_save_to_conversation_log,
@@ -179,8 +181,10 @@ def generate_chat_response(
tokenizer_name=conversation_config.tokenizer,
)
metadata.update({"chat_model": conversation_config.chat_model})
except Exception as e:
logger.error(e, exc_info=True)
raise HTTPException(status_code=500, detail=str(e))
return chat_response
return chat_response, metadata