feat: add opt-in OCR PDF indexing

[raoabinav]

Evaluator Summary

• Abi added opt-in OCR indexing for scanned and image-heavy PDFs, so LEANN can retrieve content from documents that normal embedded-text extraction would miss.
• Design choice: OCR is only enabled through leann build --enable-ocr and only runs on pages without embedded text, preserving the fast default path for ordinary PDFs.
• The implementation covers directory-discovered PDFs and explicit PDF file inputs, adds an OCR example app, documents the optional dependency and Tesseract binary requirement, and keeps OCR behavior isolated from normal text ingestion.
• Quality bar: 7 targeted tests cover blank-page OCR, text-page skipping, explicit-file handling, and dependency stubs, plus lint, format, type, lockfile, and whitespace checks.

[raoabinav]

Parking this PR for now to keep the review queue focused on #347. I’ll revisit/reopen this work later as a smaller clean PR once the content-hash PR lands.