Skip to content

feat(office): optionally extract authored picture alternative text - #4830

Draft
Afloat16 wants to merge 1 commit into
Unstructured-IO:mainfrom
Afloat16:Afloat16/office-picture-alt-text-20261121
Draft

Afloat16 wants to merge 1 commit into
Unstructured-IO:mainfrom
Afloat16:Afloat16/office-picture-alt-text-20261121

Conversation

@Afloat16

@Afloat16 Afloat16 commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Add include_image_alt_text=True to the DOCX and PPTX partitioners. Their default picture subpartitioners then emit an Image containing the picture's authored title and description. This makes accessible descriptions of diagrams and charts available to text retrieval without loading a model or running OCR.

Blank descriptions are skipped and repeated title/description text is included once. Existing image filenames are not treated as authored descriptions. Defaults remain unchanged, and registered picture subpartitioners keep complete control of their output. This covers pictures reached by the existing paragraph/slide picture traversal.

Validation

Eight extraction cases fail on main. All new tests and the existing DOCX/PPTX suites pass: 354 passed, one long DOCX vertical-merge stress case deselected. Controls cover opt-in behavior, blank descriptions, duplicate labels and custom plugins. Whole-repository Ruff 0.15.10 lint and format checks pass. Full repository tests remain pending, so this is a draft.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant