All pull requests

Pull requests

• feat(iwork): parse Keynote charts as classified pictures with data

  #4376·DanielNg0729 opened Sep 25, 2026Member·

  ·
  3

• fix(odf): stop splicing footnote/endnote body text into the citing paragraph`\

  bugodf
  #4375·avanish-garg opened Sep 25, 2026·

  ·
  1

  4

• fix(docx): read footnote/endnote body text from the separate OPC parts

  bugdocx
  #4374·avanish-garg opened Sep 25, 2026·

  ·
  1

  4

• fix(docx): read moved text, bidirectional spans and block-level custom XML

  #4369·L4XB opened Sep 24, 2026·

  ·
  3

• fix: preserve mixed and numeric chart headers

  #4367·amarjaleelbanbhan opened Sep 24, 2026·

  ·
  1

  3

• fix(docx): exclude '<w:bCs>' from bold detection

  bugdocx
  #4366·ceberam opened Sep 24, 2026Member·

  ·
  1

  3

• feat(acroform): key form fields with a geometric solver in the form_field stage

  #4364·DanieldeAlcala opened Sep 24, 2026·

  ·
  1

• feat: Introduce cap in image DPI to avoid over-downsampling

  #4363·nikos-livathinos opened Sep 24, 2026Member·

  ·
  4

• fix(csv): unescape doubled quotes in quoted fields

  bugcsv
  #4362·ceberam opened Sep 24, 2026Member·

  ·
  1

  7

• fix(jats): preserve styling in abstracts and footnotes

  xml
  #4361·Kylinny opened Sep 24, 2026·

  ·
  7

• fix(chart extraction): keep numeric series labels as column headers

  #4358·amarjaleelbanbhan opened Sep 23, 2026·

  ·
  1

  3

• fix(xlsx): keep every message of a threaded comment

  bugxlsx
  #4353·devYRPauli opened Sep 22, 2026·

  ·
  5

• feat(ocr): configure EasyOCR detector canvas size for large pages

  #4342·amarjaleelbanbhan opened Sep 22, 2026·

  ·
  4

• fix(docx): render East Asian w:numFmt values in list markers

  bugdocx
  #4336·tao-hpu opened Sep 21, 2026·

  ·
  1

  11

• fix(picture-description): report failed remote descriptions as errors, not empty text

  #4334·Ulli2060 opened Sep 21, 2026·

  ·
  3

• fix: leave API parameters set to None out of the provider request

  #4333·Ulli2060 opened Sep 21, 2026·

  ·
  3

• fix(vlm): report failed remote API calls as INFERENCE_ERROR instead of empty success

  #4332·Ulli2060 opened Sep 21, 2026·

  ·
  1

  3

• fix(md): keep inline code spans inside GFM table cells

  bugmarkdown
  #4327·kevin9327 opened Sep 20, 2026·

  ·
  7

• fix(image): recognize .gif files as images

  #4326·kevin9327 opened Sep 20, 2026·

  ·
  7

• fix(latex): parse tabularx and longtable as tables

  #4325·kevin9327 opened Sep 20, 2026·

  ·
  4

• fix(cli): skip Excel and PowerPoint lock files in directories

  bugCLIpptxxlsx
  #4324·kevin9327 opened Sep 20, 2026·

  ·
  4

• fix(html): keep rows stored in tfoot

  bughtml
  #4323·kevin9327 opened Sep 20, 2026·

  ·
  3

• fix(md): preserve whitespace around inline emphasis in table cells

  bugmarkdown
  #4318·Anurag-M1 opened Sep 19, 2026·

  ·
  1

  10

• fix: prefer explicit text extensions over weak sniffing

  #4309·nurikk opened Sep 18, 2026Contributor·

  ·
  3

• fix: reject undeclared NUL-bearing text

  #4308·nurikk opened Sep 18, 2026Contributor·

  ·
  4