Feature
- Expose max_concurrent_results on ThreadedDoclingParseBackendOptions (#4804) (
ce5621b) - Expose per-page confidence breakdown in ConfidenceScores (#4233) (
5bdabb9) - progress: Add conversion progress bars and an event callback (#4560) (
fdb6321) - iwork: Read pictures in Numbers table cells (#4736) (
dd477a9) - Allow docling-tools to download whisper artifacts and for docling to run ASR from an artifactory directory (#4760) (
765dc86) - pptx: Read every plot of a combination chart (#4773) (
070a6d3)
Fix
- md: Keep pictures through HTML conversion (#4777) (
774387e) - vlm: Report the stop reason of streaming requests (#4778) (
fd346c7) - docx: Read the active branch of a body-level mc:AlternateContent (#4615) (
b35ad57) - docx: Keep table rows after a list resumes in a cell (#4674) (
0f23355) - odf: Keep ODS pictures anchored to the sheet (#4678) (
f3eda5f) - pptx: Move footer placeholders to the furniture layer (#4746) (
3acd5ca) - html: Decode a rendered stream with the encoding the document declares (#4468) (
85dcebc) - pptx: Keep a hyperlink attached to a whole shape (#4787) (
8508141) - md: Keep the hyperlink of autolinks outside table cells (#4791) (
4757220) - pypdfium2: Keep letters of adjacent lines out of text cells (#4802) (
006c098) - email: Route an Outlook .msg that filetype reports as XLS to the email backend (#4798) (
385ddd4) - markdown: Resolve relative images for local sources (#4816) (
8119cdc) - asciidoc: Nest lists with different bullet markers (#4822) (
f1d1725) - asciidoc: Stop an unclosed block delimiter from hanging the parser (#4662) (
2104df5) - odf: Keep image references out of list item text (#4693) (
4b8231f) - vlm: Report streaming request failures as per-page inference errors (#4561) (
ab4419b) - asciidoc: Strengthen the parsing of nested lists (#4721) (
c0be875) - html: Resolve the line break sentinel outside paragraphs (#4675) (
7be1913) - docx: Resolve linked style numbering (#4727) (
ffce191) - pdf: Run threaded pipeline models on long-lived threads (#4707) (
cb1df98) - cli: Keep convert-remote working on docling-client installs (#4803) (
711fc47) - csv: Rewind content before repeated dialect detection (#4754) (
f0d9098) - email: Decode the HTML body once when a meta tag declares a charset (#4756) (
ff2342c) - asciidoc: Keep a table whole across blank and wrapped lines (#4761) (
78d1251) - html: Don't spread one run's formatting or link to a whole heading (#4670) (
930bf20) - html: Parse the start of an ordered list by the HTML rules for integers (#4713) (
a057e8d) - ebcdic: Preserve scaled numbers without context rounding (#4759) (
3807f7f)