✨(ai): add ai summarization for threads

[MathisRouget]

Purpose

Some threads and messages can be very long. Summarizing them helps users quickly retrieve key information.
We also want to assist users in avoiding common mistakes when sending messages.

Proposal

Introduce an AI-powered summarization feature for threads.
Also add a simple rule-based system to detect forgotten attachments.

AI Summarization

• Uses the AlbertAPI with the albert-large model.
• If a thread is too long, a summary is automatically generated upon reception.
• If no API key is configured, the feature is hidden from the frontend.
• Users can manually update the generated summary.

Attachment Reminder (non-AI)

• Detects when users mention an attachment (e.g. “see attached”, “voir PJ”) but forget to include a file.
• Displays a warning before sending the message.

[Copilot]

Pull Request Overview

This PR introduces AI-powered thread summarization and a rule-based attachment reminder system. The changes enhance user experience by automatically summarizing long threads and preventing users from forgetting to attach files when mentioned in their messages.

Key changes:

• Adds AI thread summarization using the AlbertAPI with automatic generation for threads with sufficient content and manual refresh capabilities
• Implements attachment reminder detection that warns users when they mention attachments but haven't included any files
• Adds comprehensive configuration system to conditionally enable AI features based on required settings

Reviewed Changes

Copilot reviewed 29 out of 31 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
src/frontend/src/features/layouts/components/thread-view/components/thread-summary/index.tsx	New React component for displaying and refreshing AI-generated thread summaries
src/frontend/src/features/forms/components/message-form/index.tsx	Enhanced message form with attachment detection logic using keywords from multiple languages
src/backend/core/ai/thread_summarizer.py	Core AI service integration for thread summarization using OpenAI-compatible API
src/backend/core/models.py	Added summary field to Thread model for storing AI-generated summaries
src/backend/core/api/viewsets/thread.py	New API endpoints for retrieving and refreshing thread summaries with proper permission checks

Files not reviewed (1)

• src/frontend/package-lock.json: Language not supported

Comments suppressed due to low confidence (1)

[sylvinus]

Great work!

I focused mainly on the backend

We should add `AI_FEATURE_THREAD_SUMMARY=1" env var to enable specific features individually ?

[jbpenrath]

Last few nitpicks from my side

[sylvinus]

Good for me once JB's feedbacks are done