Honor user's chat settings when running the extract questions phase

- Add marginally better error handling when GPT gives a messed up respones to the extract questions method
- Remove debug log lines
This commit is contained in:
sabaimran 2023-11-18 13:31:51 -08:00
parent 67156e6aec
commit a8a25ceac2
3 changed files with 46 additions and 21 deletions

View file

@ -240,10 +240,18 @@ class ConversationAdapters:
def get_openai_conversation_config():
return OpenAIProcessorConversationConfig.objects.filter().first()
@staticmethod
async def aget_openai_conversation_config():
return await OpenAIProcessorConversationConfig.objects.filter().afirst()
@staticmethod
def get_offline_chat_conversation_config():
return OfflineChatProcessorConversationConfig.objects.filter().first()
@staticmethod
async def aget_offline_chat_conversation_config():
return await OfflineChatProcessorConversationConfig.objects.filter().afirst()
@staticmethod
def has_valid_offline_conversation_config():
return OfflineChatProcessorConversationConfig.objects.filter(enabled=True).exists()
@ -267,10 +275,21 @@ class ConversationAdapters:
return None
return config.setting
@staticmethod
async def aget_conversation_config(user: KhojUser):
config = await UserConversationConfig.objects.filter(user=user).prefetch_related("setting").afirst()
if not config:
return None
return config.setting
@staticmethod
def get_default_conversation_config():
return ChatModelOptions.objects.filter().first()
@staticmethod
async def aget_default_conversation_config():
return await ChatModelOptions.objects.filter().afirst()
@staticmethod
def save_conversation(user: KhojUser, conversation_log: dict):
conversation = Conversation.objects.filter(user=user)

View file

@ -1,5 +1,6 @@
# Standard Packages
import logging
import json
from datetime import datetime, timedelta
from typing import Optional
@ -31,6 +32,10 @@ def extract_questions(
"""
Infer search queries to retrieve relevant notes to answer user query
"""
def _valid_question(question: str):
return not is_none_or_empty(question) and question != "[]"
# Extract Past User Message and Inferred Questions from Conversation Log
chat_history = "".join(
[
@ -70,7 +75,7 @@ def extract_questions(
# Extract, Clean Message from GPT's Response
try:
questions = (
split_questions = (
response.content.strip(empty_escape_sequences)
.replace("['", '["')
.replace("']", '"]')
@ -79,9 +84,18 @@ def extract_questions(
.replace('"]', "")
.split('", "')
)
questions = []
for question in split_questions:
if question not in questions and _valid_question(question):
questions.append(question)
if is_none_or_empty(questions):
raise ValueError("GPT returned empty JSON")
except:
logger.warning(f"GPT returned invalid JSON. Falling back to using user message as search query.\n{response}")
questions = [text]
logger.debug(f"Extracted Questions by GPT: {questions}")
return questions

View file

@ -55,6 +55,7 @@ from database.models import (
Entry as DbEntry,
GithubConfig,
NotionConfig,
ChatModelOptions,
)
@ -669,7 +670,16 @@ async def extract_references_and_questions(
# Infer search queries from user message
with timer("Extracting search queries took", logger):
# If we've reached here, either the user has enabled offline chat or the openai model is enabled.
if await ConversationAdapters.ahas_offline_chat():
offline_chat_config = await ConversationAdapters.aget_offline_chat_conversation_config()
conversation_config = await ConversationAdapters.aget_conversation_config(user)
if conversation_config is None:
conversation_config = await ConversationAdapters.aget_default_conversation_config()
openai_chat_config = await ConversationAdapters.aget_openai_conversation_config()
if (
offline_chat_config
and offline_chat_config.enabled
and conversation_config.model_type == ChatModelOptions.ModelType.OFFLINE.value
):
using_offline_chat = True
offline_chat = await ConversationAdapters.get_offline_chat()
chat_model = offline_chat.chat_model
@ -681,7 +691,7 @@ async def extract_references_and_questions(
inferred_queries = extract_questions_offline(
defiltered_query, loaded_model=loaded_model, conversation_log=meta_log, should_extract_questions=False
)
elif await ConversationAdapters.has_openai_chat():
elif openai_chat_config and conversation_config.model_type == ChatModelOptions.ModelType.OPENAI.value:
openai_chat_config = await ConversationAdapters.get_openai_chat_config()
openai_chat = await ConversationAdapters.get_openai_chat()
api_key = openai_chat_config.api_key
@ -690,11 +700,6 @@ async def extract_references_and_questions(
defiltered_query, model=chat_model, api_key=api_key, conversation_log=meta_log
)
logger.info(f"🔍 Inferred queries: {inferred_queries}")
logger.info(f"🔍 Defiltered query: {defiltered_query}")
logger.info(f"using max distance: {d}")
logger.info(f"using filters: {filters_in_query}")
logger.info(f"Max results: {n}")
# Collate search results as context for GPT
with timer("Searching knowledge base took", logger):
result_list = []
@ -711,20 +716,7 @@ async def extract_references_and_questions(
common=common,
)
)
logger.info(f"🔍 Found {len(result_list)} results")
logger.info(f"Confidence scores: {[item.score for item in result_list]}")
# Dedupe the results again, as duplicates may be returned across queries.
with open("compiled_references_pre_deduped.txt", "w") as f:
for item in compiled_references:
f.write(f"{item}\n")
result_list = text_search.deduplicated_search_responses(result_list)
compiled_references = [item.additional["compiled"] for item in result_list]
with open("compiled_references_deduped.txt", "w") as f:
for item in compiled_references:
f.write(f"{item}\n")
logger.info(f"🔍 Deduped results: {len(result_list)}")
return compiled_references, inferred_queries, defiltered_query