Documents: Migrate transformation job to celery

[kartpop]

The documents/upload/ endpoint allows an original pdf doc to be converted to markdown for easy vector store ingestion.

Currently this transformation job is being run in fastapi background task. This should be migrated to run in celery worker instead.

[avirajsingh7]

We can introduce a separate endpoint for handling document transformation jobs.
Additionally, to support document transformation through the existing /documents/upload/ endpoint, we can make an internal call to the transformation logic.

This way, both approaches are supported :

• a dedicated endpoint for transformation jobs, and
• optional transformation support directly through the upload flow.

Just a opinion we can discuss on this