Skip to content

fix(docx): retain base text from East Asian ruby annotations - #4827

Draft
Afloat16 wants to merge 1 commit into
Unstructured-IO:mainfrom
Afloat16:Afloat16/docx-ruby-base-text-20261121
Draft

Afloat16 wants to merge 1 commit into
Unstructured-IO:mainfrom
Afloat16:Afloat16/docx-ruby-base-text-20261121

Conversation

@Afloat16

@Afloat16 Afloat16 commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Summary

DOCX ruby annotations hold the document text in w:rubyBase and pronunciation in w:rt. Python-docx skips the ruby node, causing the base text to disappear and ruby-only paragraphs to be omitted entirely.

Lift the visible base-run content into the enclosing run before partitioning. Text stays in its original position, pronunciation is not duplicated, and tabs and breaks remain available to existing readers. The enclosing run's formatting applies; separate formatting inside the base runs is not resolved.

Validation

All seven regression cases fail on main and pass after the change, covering body, table, header, footer, hyperlink, ruby-only paragraphs and embedded tabs. Whole-repository Ruff 0.15.10 lint and format checks pass. The new tests and existing DOCX suite pass: 229 passed, one long vertical-merge stress case deselected. Full repository tests remain pending, so this is a draft.

Reference: https://learn.microsoft.com/en-us/dotnet/api/documentformat.openxml.wordprocessing.rubybase

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