yfedoseev/pdf_oxide v0.2.6

v0.2.6 | CJK Support & Structure Tree Enhancements

on GitHub

v0.2.6 Release

This release adds comprehensive CJK (Chinese, Japanese, Korean) language support and enhances structure tree handling for Tagged PDFs.

🌍 CJK Language Support

• Predefined CMap support for CJK fonts (PDF Spec Section 9.7.5.2)
  • Adobe-GB1 (Simplified Chinese) - ~500 common character mappings
  • Adobe-Japan1 (Japanese) - Hiragana, Katakana, Kanji mappings
  • Adobe-CNS1 (Traditional Chinese) - Bopomofo and CJK mappings
  • Adobe-Korea1 (Korean) - Hangul and Hanja mappings
  • Fallback identity mapping for common Unicode ranges
• Type 0 /W array parsing for CIDFont glyph widths
  • Proper spacing for CJK text using CIDFont width specifications

🏷️ Structure Tree Enhancements

• TagSuspect/MarkInfo support (ISO 32000-1 Section 14.7.1)
  • Parse MarkInfo dictionary from document catalog
  • marked, suspects, user_properties fields
  • Automatic fallback to geometric ordering when structure tree is marked suspect
• Word Break /WB structure element (Section 14.8.4.4)
  • Support for explicit word boundaries in CJK text
  • Word break markers emitted during structure tree traversal
• Abbreviation expansion /E support (Section 14.9.5)
  • Parse /E entry from marked content properties
  • expansion field on StructElem for abbreviations

✨ New Features

• Object reference resolution utility
  • PdfDocument::resolve_references() for recursive reference handling
  • Simplifies working with complex PDF structures
• ActualText verification tests - Comprehensive test coverage for Section 14.9.4
• Enhanced artifact filtering with subtype support
  • ArtifactType::Pagination with Header, Footer, Watermark, PageNumber
  • ArtifactType::Layout and ArtifactType::Background classification

🐛 Bug Fixes

• Soft hyphen handling (U+00AD) - Correctly treated as valid continuation hyphen for word reconstruction

✅ Verification

• CJK text extraction tested on Chinese, Japanese, Korean documents
• Structure tree handling validated
• Word break detection verified
• All tests pass

📥 Installation

Rust (crates.io)

cargo add pdf_oxide

Python (PyPI)

pip install pdf_oxide

Release Date: January 9, 2026
Branch: main
Tag: v0.2.6