feat(api): extract PDF hyperlinks and URLs into document metadata

[nancysangani]

🔗 Related Issue

Closes #441

---

📝 What does this PR do?

Adds automated PDF link and URL extraction during document ingestion and
exposes the results in the document metadata API.

backend/app/rag/url_extractor.py (new):

• extract_urls_from_pdf() — two-pass extraction per page using PyMuPDF
  (already in requirements.txt): link annotations (page.get_links())
  first, then plain-text regex fallback for inline URLs not marked as
  hyperlinks. Deduplicates while preserving first-seen order. Returns []
  safely for non-PDF files or any extraction error.

backend/migrate_add_extracted_urls.py (new):

• Standalone migration script consistent with the project's existing
  migrate_add_role.py / migrate_add_google_refresh_token.py pattern.
• Adds extracted_urls TEXT DEFAULT NULL to the documents table.
• Safe to re-run — skips if the column already exists.

backend/app/models.py:

• Adds extracted_urls = Column(Text, nullable=True) to Document.

backend/app/schemas.py:

• Adds extracted_urls: Optional[List[str]] = None to DocumentResponse.

backend/app/services/document_ingestion.py:

• Adds a URL extraction pass after the summary step. Calls
  extract_urls_from_pdf(), JSON-serialises the result, and stores it in
  doc.extracted_urls. Wrapped in try/except — a failure never blocks
  ingestion.

backend/app/routes/documents.py:

• Adds _deserialize_doc() helper that parses the JSON string from the DB
  back into List[str] before returning DocumentResponse. Applied to
  get_document, list_documents, and rename_document.

---

🗂️ Type of Change

• ✨ New feature

---

🧪 How was this tested?

• Ran python migrate_add_extracted_urls.py — column added successfully
• Ran the backend locally (uvicorn app.main:app --reload)
• Uploaded a PDF containing hyperlinks; confirmed extracted_urls field
  in GET /api/v1/documents/{id} response contains correct URLs
• Uploaded a PDF with no links; confirmed extracted_urls is null
• Uploaded a non-PDF (DOCX/TXT); confirmed extraction is skipped and
  extracted_urls is null
• Confirmed existing text/table/image ingestion is unaffected

---

✅ Self-Review Checklist

• My branch is based on dev, not main
• I have not added any secrets / API keys
• I have not modified main branch or any HuggingFace deployment config
• My code follows the existing style (no unnecessary formatting changes)
• I have updated relevant docs / comments if needed

[github-actions]

🎉 Congratulations on getting your Pull Request merged! 🎉

Thank you for contributing to PDF-Assistant-RAG as part of GSSoC '26! 🚀

Keep up the great work! ✨