[RaphaelFakhri]

Summary

• _resolve_url_source_file_name now keeps the request file_name only when its extension matches the extension detected for the URL. Otherwise it appends the detected extension, the same way the URL basename path already does. For example, annual-report.docx on a URL that serves a PDF becomes annual-report.docx.pdf.
• Before this change, any name with an extension was kept as-is. The page-memory track routes on the source_file_name extension, so a v2 URL job for a PDF with file_name="annual-report.docx" was accepted and then failed in normalize_to_pdf with page_memory parse track only supports PDF and PPTX.
• Names that already match, such as hipo-upload-optimization_a884bd1b.plan.md on an .md URL, are unchanged. The comparison ignores case, so Report.PDF is kept.
• No API schema, worker, or migration change.
• Fixes [Bug]: URL job file_name with a different extension breaks the page-memory parse #498

Verification

• uv run pytest apps/api/tests/unit/test_document_ingestion_filename.py -q: 6 passed. With the source change reverted, the new mismatch test fails (1 failed, 5 passed).
• ruff check on the changed files: passed.
• Did not run the PostgreSQL-backed contract tests or a live URL ingest.

Deployment Notes

• No new environment variables.
• No database migrations, queue changes, or storage changes.
• Jobs that were already created keep their stored source_file_name.

Checklist

• Tests were added or updated when behavior changed
• Public docs, examples, or OpenAPI contracts were updated when needed
• Database migrations are idempotent and safe to deploy
• Logs, errors, and validation paths avoid leaking secrets or user data
• The pull request description explains any breaking or user-visible change