[SSoC'26]Feat: document ingestion queue for large files

[diksha78dev]

Closes #291.

Built a lightweight document ingestion queue for large files using FastAPI BackgroundTasks and SQLite.

• Background Processing: Document embedding now runs asynchronously, preventing server timeouts on massive files.
• Batched Indexing: Yields the GIL between batches to keep the main event loop responsive.
• Real-time UI Tracking: Added status tracking (queued, processing, failed) and UI polling to display live chunk progress.
• Crash Safety: Wrapped AI imports in error boundaries to cleanly mark documents as failed if the process crashes.

[vercel]

@diksha78dev is attempting to deploy a commit to the Darshan's projects Team on Vercel.

A member of the Team first needs to authorize it.