feat: add brain prompt overwriting from chat

[mamadoudicko]

Description

• Update backend add prompt overwriting from chat for doc based and headless

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Aug 22, 2023 10:51am
quivrapp	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Aug 22, 2023 10:51am

[github-actions]

Risk Level 2 - /home/runner/work/quivr/quivr/backend/llm/openai.py

The code changes seem to be safe and follow the SOLID principles. However, there is a potential issue with the Optional[UUID] type for prompt_id. If prompt_id is not provided, it might cause issues in the code where it's used without checking for None. Always check if prompt_id is not None before using it.

Example:

if self.prompt_id is not None:
    # use prompt_id

---

Risk Level 2 - /home/runner/work/quivr/quivr/backend/routes/chat_routes.py

1. The delete_chat_from_db function is catching all exceptions and simply printing them. This is not a good practice as it can make debugging difficult. It would be better to log the exceptions with more context or re-raise them after logging. For example:

try:
    supabase_db.delete_chat_history(chat_id)
except Exception as e:
    logger.error(f'Failed to delete chat history for chat_id {chat_id}: {e}')
    raise

2. The check_user_requests_limit function is using os.getenv to get the MAX_REQUESTS_NUMBER environment variable. It would be better to use a configuration file or a configuration management system to manage these values.
3. The create_question_handler and create_stream_question_handler functions have a lot of duplicated code. Consider refactoring this into a separate function to improve maintainability and readability.
4. The create_question_handler and create_stream_question_handler functions are using the print function. It would be better to use a logger to log these messages.

---

Risk Level 3 - /home/runner/work/quivr/quivr/backend/llm/qa_base.py

1. The code changes seem to be safe and follow the SOLID principles. However, there is a potential issue with the Optional[UUID] type for prompt_id. If prompt_id is not provided, it might cause issues in the code where it's used without checking for None. Always check if prompt_id is not None before using it.

Example:

if self.prompt_id is not None:
    # use prompt_id

2. The prompt_to_use and prompt_to_use_id properties are using the UUID constructor directly on self.brain_id. This might cause a ValueError if self.brain_id is not a valid UUID. Consider using a try-except block to handle this potential error.

Example:

try:
    uuid_obj = UUID(self.brain_id)
except ValueError:
    # handle invalid UUID

---

🔍🐛🔄

---

Powered by Code Review GPT