kreuzberg-dev/kreuzberg v4.1.1

on GitHub

Bug Fixes

• pptx: Resolve extraction failures for shapes without txBody and add PPSX/PPTM support (#321)
• ruby: Handle unknown config keys in from_file and discover methods
• ci: Add Windows tessdata setup for Elixir CI

Details

PPTX/PPSX/PPTM Fix

Fixed PPTX extraction failures caused by shapes without text bodies (e.g., images, decorative elements). Also added support for PPSX (PowerPoint Show) and PPTM (macro-enabled PowerPoint) file formats.

Ruby Config Fix

Fixed ArgumentError: unknown keyword: :images when using Config.from_file or Config.discover with config files containing keys not directly mapped to Ruby's Extraction options. Added key filtering and aliasing for backward compatibility.

CI Fix

Added Tesseract tessdata download for Windows in Elixir CI to resolve test failures.