[codex] add PDF extraction preview

[Parsenal]

Summary

• record the PDF extraction engine and page count on Source, expose them in source detail, and add the migration
• add a side-by-side extraction preview from the completed Extract pipeline step, using the existing PDF viewer plus stored extracted text
• fix the source-content API client path and cover docling/fallback metadata in focused tests

Notes

Draft sample PR for #829.

The code/test branch is published through the connector. The regenerated docs/openapi.yaml is present in my local commit and make check-openapi passed locally with it, but the generated file is not included on this draft branch yet because local GitHub CLI auth is not configured and the connector cannot reliably apply a 261 KB generated-file update. Before this leaves draft, push the exact local branch or add the regenerated OpenAPI file.

Validation

• .venv/bin/python -m pytest tests/unit/test_pdf_adapter.py tests/unit/test_source_content.py
• npm --prefix apps/web run build
• .venv/bin/ruff check alembic/versions/0024_add_source_extraction_metadata.py src/wikimind/api/routes/ingest.py src/wikimind/ingest/adapters/pdf.py src/wikimind/models/dto/ingest.py src/wikimind/models/tables/ingest.py tests/unit/test_pdf_adapter.py tests/unit/test_source_content.py
• npm --prefix apps/web run lint (existing warnings only)
• .venv/bin/alembic heads
• make check-openapi locally with regenerated docs

Refs #829

[github-actions]

Marking stale: no activity for 26 days. Will auto-close at 30 days.

[github-actions]

Closing: no activity for 35 days. Reopen if still needed.