[Bug]: Enhance PDF Processing with OCR for Unreadable Documents

[MridulTi]

Describe the issue

Currently, our PDF processing pipeline extracts text from documents. However, if text extraction fails, we need a fallback mechanism to perform OCR and retrieve the content. This ensures users can still query the document even if it's an image-based PDF or has unreadable text.

Steps to Reproduce

1. Upload a scanned PDF (or an image-based PDF).
2. Attempt to extract text from the document.
3. Observe that the system fails to retrieve any content and does not try OCR.

Expected Behavior

• If text extraction fails, the system should automatically attempt OCR.
• OCR-extracted text should be processed for search queries.
• Errors should be logged if OCR also fails.
• Users should receive a meaningful message if the document is entirely unreadable.

Relevant Logs/Error Messages

NA

Priority

Medium