pypi docling 2.143.0

4 hours ago

Feature

  • service: Add optional filename to ZipArchiveResult (#4996) (87de795)
  • Support mixed-format ZIP inputs (#4865) (aff98aa)
  • asr: Run the WhisperS2T presets offline from --artifacts-path (#4898) (0dbf500)
  • ASR: Add transformers and MLX presets for the ASR pipelines (#4771) (53c11a2)
  • vlm: Let a GenerationStopper declare a valid end of the output (#4927) (e38fceb)

Fix

  • vlm: Defer optional local provider imports (#4967) (2eaca3a)
  • utils: Retain BOM decoding error diagnostics (#4949) (6ddd22e)
  • asciidoc: Keep a block title separate from the paragraph it titles (#4391) (821526e)
  • pptx: Extract the picture of a picture-filled shape (#4989) (4ad544a)
  • pptx: Read the series of every chart plot (#4811) (90234a7)
  • asciidoc: Drop the closing marks of symmetric section titles (#4938) (7b729c6)
  • ods: Do not join tables across a blank merged separator row (#4992) (90661c4)
  • utils: Split a tuple or range into chunks only once in chunkify (#5001) (15c67d8)
  • Accept file URI bases for image resources (#4912) (59acbb6)
  • asciidoc: Keep /// as text and support longer comment delimiters (#4394) (9ccf607)
  • docx: Leave hidden text out of the default export (#4958) (36be1d2)
  • csv: Use following rows when the header holds two delimiters (#4896) (db79955)

Documentation

  • Fix typos and grammar in FAQ, installation, and advanced options (#4986) (5de229e)
  • Add RAG example with TopK (#4983) (4d97a18)

Don't miss a new docling release

NewReleases is sending notifications on new releases.