feat(Unplug): chatting without brain streaming

[StepanLebedevTheodo]

Description

Adds the possibility to chat with Quivr without selecting some brain

[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 18, 2023 7:53am
quivrapp	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Aug 18, 2023 7:53am

[github-actions]

Risk Level 2 - /home/runner/work/quivr/quivr/backend/core/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. Instead, consider logging the exceptions with more context or re-throwing them after logging.

2. The check_user_limit function has an else: pass statement which is unnecessary and can be removed for cleaner code.

3. The create_question_handler function has a large amount of code within a try block. Consider reducing the amount of code within try blocks to only include code that might throw an exception.

4. The create_question_handler function is using the print function. Consider using a logger for any debugging or informational messages, which can be configured to write to different outputs.

Example changes:

import logging
logger = logging.getLogger(__name__)

# In delete_chat_from_db
except Exception as e:
    logger.error(f'Error deleting chat history: {e}')
    raise

# In check_user_limit
if int(user.requests_count) >= int(max_requests_number):
    raise HTTPException(
        status_code=429,
        detail=\"You have reached the maximum number of requests for today.\"
    )

# In create_question_handler
try:
    check_user_limit(current_user)
    LLMSettings()
except HTTPException as e:
    raise e

# Replace print with logger
logger.info('streaming')

---

Risk Level 5 - /home/runner/work/quivr/quivr/backend/core/llm/qa_headless.py

The openai_api_key and user_openai_api_key are being stored as plain text which is a high security risk. It's recommended to use environment variables or a secure vault service to store sensitive information. For example:

import os
self.openai_api_key = os.getenv('OPENAI_API_KEY')

Also, the generate_answer and generate_stream methods are quite long and complex. Consider breaking them down into smaller, more manageable functions.

---

🔒🐛🧹

---

Powered by Code Review GPT

[gozineb]

Remove the comments

[StepanLebedevTheodo]

Remove the comments
Done

[StanGirard]

Replace with only -> "Your name is Quivr. You're a helpful assistant."