github docling-project/docling v2.134.0

4 hours ago

Feature

  • iwork: Add Apple Numbers (.numbers) support (#4113) (8d7cc9f)
  • vlm: Expose logprobs on the streaming API path (#4528) (3073a27)

Fix

  • docx: Preserve text from dangling fragment hyperlinks (#4563) (d15cf10)
  • html: Keep text placed directly inside a figure (#4574) (caba660)
  • docx: Keep hyperlink runs nested in a wrapper (#4576) (7833b32)
  • docx: Add use_outline_level_for_headings option to backend (#4480) (b38ed49)
  • html: Separate inline text from a following block tag in get_text (#4484) (d451d71)
  • md: Drop whitespace-only lines before code block content (#4543) (eda5b5f)
  • datamodel: Let the extension pick between OLE2 legacy Office formats (#4476) (58f1f8d)
  • xbrl: Tolerate open taxonomy packages during temp dir cleanup (#4482) (d65cbd6)
  • cli: Write plain text for --to text (#4508) (e34b74d)
  • cli: Detect LaTeX documents by content when the name gives no format (#4551) (93ce581)
  • code-formula: Stop repetition loops and make the crop margin configurable (#4548) (012b611)
  • cli: Keep local LaTeX sources in place so included files resolve (#4532) (734e8f0)
  • latex: Keep the first-line indentation of code environments (#4502) (51fe9ff)
  • email: Keep Cc recipients in the converted document (#4505) (53eec17)
  • image: Flatten transparent images onto white instead of black (#4511) (0cd61e0)
  • odf: Respect page_range in ODS and ODP (#4512) (7bfbdb9)
  • mets-gbs: Rewind the stream so archives passed as DocumentStream load (#4521) (31aca0f)
  • cli: Apply the video options to videos found in a directory (#4523) (1cd56d8)
  • md: Reset pending line breaks at block boundaries (#4525) (dd7e1ef)
  • html,jats,boxnote: Count the columns a rowspan shifts cells into (#4527) (c87c9ce)
  • opendocument: Emit the embedded chart title as a caption (#4529) (dfa8728)
  • md: Keep text after a line break out of a code span (#4509) (5c5a8c1)

Documentation

  • agents: Require simplified technical English (#4552) (b91e6b2)

Don't miss a new docling release

NewReleases is sending notifications on new releases.