Repository Issue Activity (beta)

#720 [Bug] Hybrid auto mode: TableFormer (hardcoded ACCURATE) collapses table rows on dense datasheets — orphan-cell fallback corrupts structure (cf. #627)

#650 Your project is on StackMap — a curated map of the AI stack

#719 markdown_with_html escapes generated HTML table opening tags

#711 Python/Node wrappers silently drop empty-string hybrid_chunk_size

#709 test: assert hybrid CLI flags appear in the built JAR --help

#708 test: add sync check for hybrid-chunk-size default across bindings

#707 test: cover hybridChunkSize forwarding in Node bindings

#706 test: drive the production hybrid chunking loop with a configured chunk size

#694 StringIndexOutOfBoundsException in TableStructureNormalizer aborts whole document (empty TextChunk substring)

#695 Chart gridlines are detected as a table, producing a /Table with only empty cells

#618 image_output = "external" The generated image is not clear enough

#568 [M1] Cap page-render DPI to bound single-page peak memory (the real lever for #458)

#563 Make the word spacing threshold (TEXT_LINE_SPACE_RATIO) configurable

#584 unneccessary extra "-" and dot added in md files generated from pdf for points in bullet

#662 Comparison method violates its general contract! in DocumentProcessor.sortPageContents

#635 pdf having content like below AWS_TeraVMm_v2.55.0-<build-number>: not handled properly in converted md

#658 opendataloader hangs indefinitely

#686 opendataloarder gets stuck

#652 After upgrading to latest version 2.50 seeing issues in image extraction,extra grey scale image seen

#676 hyperlinks and links are not handled properly in converted md

#674 Add `py.typed` to Python package to denote that the package supports typing

#673 Hybrid backend fails to start on Korean-locale Windows - 'cp949' codec can't decode byte 0xe2 while loading docling model config

#663 IEEE roman-numeral section titles emitted as paragraphs while an equation is tagged as a heading

#619 Hybrid triage routes full-page scanned pages to the Java path -> empty output

#515 I've noticed that whenever a PDF contains underlined text (similar to fill-in-the-blank formatting), it triggers a parsing error that causes the fillable content to be split across lines. I'm not sure what causes this issue.