feat(backend): add custom prompt

[mamadoudicko]

Description

Activate custom prompt feature (backend)

[vercel]

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

Name	Status	Preview	Updated (UTC)
docs	🔄 Building (Inspect)	Visit Preview	Aug 7, 2023 2:30pm
quivrapp	🔄 Building (Inspect)	Visit Preview	Aug 7, 2023 2:30pm

[github-actions]

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

1. The get_prompt method could be improved by handling the case where brain.prompt_id is not None but get_prompt_by_id(brain.prompt_id) returns None. This could lead to a NoneType error when trying to access brain_prompt_object.content. Consider adding an additional check to ensure brain_prompt_object is not None before trying to access its content attribute.

if brain and brain.prompt_id:
    brain_prompt_object = get_prompt_by_id(brain.prompt_id)
    if brain_prompt_object:
        brain_prompt = brain_prompt_object.content

2. The generate_stream method is quite long and complex. Consider breaking it down into smaller, more manageable functions. This would improve readability and maintainability of the code.

3. The generate_stream method uses asyncio.create_task to create a background task. However, it does not appear to handle cancellation of this task if an error occurs or if the parent task is cancelled. Consider adding error handling and cancellation logic to ensure the task is properly cleaned up.

4. The generate_stream method uses json.dumps to serialize streamed_chat_history.to_dict() to a JSON formatted string. However, it does not handle the case where streamed_chat_history.to_dict() contains non-serializable types. Consider adding error handling or ensuring that streamed_chat_history.to_dict() only contains serializable types.

---

🐛💡🔧

---

Powered by Code Review GPT