All pull requests

Pull requests

• fix(pdf): warn when host fonts leave page text unpainted

  ai:ci-safeai:review-lgtmbugdocumentationpdf parsingpdftests
  #5009·Yun-0000 opened Oct 25, 2026·2 of 3 tasks·

  ·
  3

• fix(asciidoc): don't read a leading named image attribute as the alt text

  ai:ci-safeai:review-lgtmasciidocbug
  #5008·kevin9327 opened Oct 25, 2026Contributor·1 of 3 tasks·

  ·
  3

• fix(docx): keep tables whose w:tblGrid is missing or under-declared

  bugdocxtable structure
  #5007·inchang-ing opened Oct 25, 2026Contributor·

  ·
  1

  8

• feat(pptx): keep shape groups in the document tree

  ai:ci-safeai:review-lgtmenhancementpptx
  #5005·uommou opened Oct 25, 2026·3 tasks done·

  ·
  1

  6

• fix(service-client): keep the presigned rejection when the InBody fallback fails

  ai:ci-safeai:review-lgtmbugerror-handling
  #5003·dolfim-ibm opened Oct 25, 2026Member·3 tasks done·

  ·
  7

• fix(ocr): restore OCR of scanned pages only with page_bitmap_area_threshold

  ai:ci-safebugocr
  #4999·yashikam19 opened Oct 24, 2026Contributor·2 of 3 tasks·

  ·
  1

  7

• fix(html): preserve preformatted code whitespace

  #4998·teachershuang opened Oct 24, 2026Contributor·

  ·
  1

  2

• fix(deepseekocr): keep table cells that a rowspan shifts to the right

  ai:ci-needs-carebugocrtable structuretests
  #4994·simpleqt opened Oct 23, 2026Contributor·

  ·
  10

• fix(md): keep backslash escapes inside the surrounding text

  ai:ci-safebugmarkdown
  #4991·Krina-Paghadar opened Oct 23, 2026·3 tasks done·

  ·
  1

  8

• fix(pptx): extract the picture of a picture-filled shape

  ai:ci-safebugpptx
  #4989·sarankumar1325 opened Oct 23, 2026Contributor·3 tasks done·

  ·
  1

  9

• ci: summarize reference data changes and build on fixes in deps update

  ai:ci-needs-careai:review-changes
  #4987·dolfim-ibm opened Oct 23, 2026Member·

  ·
  6

• chore: update locked docling dependencies

  #4985·docling-ops[bot] opened Oct 23, 2026Bot·

  ·
  4

• fix(xlsx): do not join tables across a blank merged separator row

  ai:ci-safeai:review-lgtmbugxlsx
  #4981·mr-shakib opened Oct 22, 2026Contributor·3 tasks done·

  ·
  1

  6

• fix(vlm): defer optional local provider imports

  bugvlm-pipeline
  #4967·YamonBot opened Oct 22, 2026·3 tasks done·

  ·
  1

  5

• fix(odt): preserve text box block content

  #4963·Vamsi-2581 opened Oct 22, 2026Contributor·

  ·
  1

  2

• fix(odt): emit paragraphs styled with style:list-style-name as list items

  bugodf
  #4959·Kylinny opened Oct 21, 2026Contributor·

  ·
  1

  4

• fix(odf): keep headings numbered through a list as section headers

  bugodf
  #4956·bercedev opened Oct 21, 2026·1 of 3 tasks·

  ·
  1

  3

• fix(vlm): fall back from bfloat16 on devices without native support

  bugtests:fullvlm-pipeline
  #4955·Abelo9996 opened Oct 21, 2026·2 of 3 tasks·

  ·
  1

  4

• fix(docx): preserve nested list hierarchy across empty spacer paragraphs

  ai:ci-safebugdocx
  #4951·taran-dev4u opened Oct 21, 2026Contributor·3 tasks done·

  ·
  1

  9

• fix(utils): retain BOM decoding error diagnostics

  bug
  #4949·JackyHsu-X opened Oct 21, 2026·2 tasks done·

  ·
  1

  5

• feat(ocr): improve Tesseract OCR reliability

  enhancementocr
  #4942·BBC-Esq opened Oct 21, 2026Contributor·3 tasks done·

  ·
  4

• fix(ocr): stop concurrent conversions from crashing the tesserocr engine

  ai:ci-safebugocr
  #4941·RachelWanggg opened Oct 21, 2026Contributor·1 of 3 tasks·

  ·
  1

  5

• ci: add a Python 3.15 lane for the extras that install

  ai:ci-needs-careai:review-lgtmCLI
  #4937·dolfim-ibm opened Oct 21, 2026Member·

  ·
  4

• fix(markdown): keep GFM task-list checkboxes and paragraph [x] markers

  ai:ci-safeai:review-lgtmbugmarkdown
  #4935·Siddharth-DWT opened Oct 21, 2026·3 tasks done·

  ·
  1

  11

• feat(vlm): add native MLX support for Nemotron 3.5 Parse

  ai:ci-safedependency mgmtenhancementvlm-pipeline
  #4934·geoHeil opened Oct 21, 2026Contributor·

  ·
  1

  17