Skip to content

v0.9.7

Latest
Latest

Choose a tag to compare

View all tags
@AdemBoukhris457 AdemBoukhris457 released this 16 Nov 13:18
· 17 commits to main since this release
v0.9.7
0d2ebf7
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

🚀 What's new in v0.9.7

  • PaddleOCR-VL PDF Parser (with Restoration + Split Tables): New PaddleOCRVL-powered PDF parser that combines layout-aware OCR, visual-language understanding, page restoration, and split table merging in a single high-level pipeline.
  • Split Table Merging Everywhere: Split table detection & merging is now available across ChartTablePDFParser and EnhancedPDFParser, so multi-page tables are reconstructed consistently whether you’re extracting text, tables, or charts.
  • Restoration-Friendly Flow: The new parser plays nicely with restoration steps (denoising, deblurring, cleanups), improving OCR and structure extraction on noisy reports and scanned PDFs.
  • Docs Upgrade: Documentation updated to explain when to use the new PaddleOCR-VL parser, how split table merging works across parsers, and how to configure these features in real-world workflows.

✅ Motivation

Doctra is increasingly used on messy, real-world PDFs where tables are split across pages and visual context matters (charts, complex layouts, degraded scans). This release focuses on:

  • Making split-table merging a first-class feature across multiple parsers.
  • Introducing a PaddleOCR-VL–based parser that can better understand visual + textual context.
  • Tightening the integration with restoration so that users get more reliable structured outputs from imperfect documents.

🛠 What’s Changed

  • feat: Add PaddleOCRVL PDF parser with restoration and split table merging by @AdemBoukhris457 in #82
  • feat: Add split table merging to ChartTablePDFParser by @AdemBoukhris457 in #81
  • feat: Add split table merging support to EnhancedPDFParser by @AdemBoukhris457 in #80
  • docs: Document new PaddleOCR-VL parser & split-table merging behavior (usage, configuration, and examples)

📦 Version

v0.8.0 → v0.9.7
Minor feature-focused release that extends split-table merging across parsers and introduces a PaddleOCR-VL–powered PDF parser with restoration support. No breaking changes to the public API — existing workflows keep working, but gain access to smarter parsing options.

Contributors

AdemBoukhris457
Assets 2
Loading