github docling-project/docling v2.146.0

4 hours ago

Feature

  • vlm: Add Unlimited-OCR 2 preset and layout parser (#5107) (bdb1d20)
  • Add LightOnOCR-3 presets and grounded output parsing (#4789) (4b382ee)
  • vlm: Add guarded Nemotron 3.5 multi-pass parsing (#4933) (84599f9)

Fix

  • service-client: Keep the presigned rejection when the InBody fallback fails (#5003) (421b3ba)
  • iwork: Read decimal128 cells without context rounding (#5101) (2fea2c9)
  • md: Stop the parser hanging on pipes in lazy continuation lines (#5110) (2092759)
  • asciidoc: Turn links in paragraphs into hyperlinks (#5115) (07154cd)
  • pptx: Render lettered, Roman and parenthesized list markers (#5122) (26a5bcf)
  • docx: Apply the w:lvlOverride of a list numbering instance (#5123) (bb7340d)
  • utils: Drop leading UTF-8 BOM for explicit encoding="utf-8" (#4849) (7346d1e)
  • xlsx: Read images placed in cells (#5015) (764b5d4)
  • Allow chunks artifact type in ArtifactRef (#5111) (7b12671)
  • iwork: Keep converting when a date cell holds a span no datetime can hold (#5087) (eb0bb24)
  • vlm: Keep the vLLM stop reason so pages are not all read as incomplete (#4901) (af2dbad)

Don't miss a new docling release

NewReleases is sending notifications on new releases.