[Yuilona]

The Word backend dropped hyperlink text that sits in a wrapper inside the w:hyperlink. A tracked insertion made inside link text (<w:hyperlink><w:ins><w:r>…) is the common case, and smart tags, custom XML and simple fields behave the same way.

_iter_paragraph_content already reads runs inside w:ins, w:smartTag, w:customXml and w:fldSimple when they wrap runs directly in the paragraph. A hyperlink's text, however, came from python-docx's Hyperlink.text, and its formatting from Hyperlink.runs. Both only read the direct w:r children (CT_Hyperlink.text is "".join(r.text for r in self.xpath("w:r"))). This PR collects the hyperlink's runs with the same _get_children_recursive helper, so text and formatting come from all of its runs, in document order. The hyperlink target is unchanged.

<w:hyperlink> content	Before	After
[the ] <w:ins>[docs]</w:ins>	the	the docs
<w:smartTag>[docs]</w:smartTag>	(empty)	docs
<w:customXml>[docs]</w:customXml>	(empty)	docs
<w:fldSimple>[docs]</w:fldSimple>	(empty)	docs

This touches a different part of _iter_paragraph_content from #4369, which adds moveTo, dir and bdo to the paragraph-level wrappers. Once both are in, those wrappers will be read inside hyperlinks too. #4455 covers wrapped runs in plain table cells, which take a separate path.

Issue resolved by this Pull Request:
Resolves #4575

Testing (Linux, Python 3.11):

• I added test_hyperlink_keeps_runs_nested_in_a_wrapper, parametrized over the four wrappers. It checks that the link item reads the docs and keeps its https://example.com/ target. All four cases fail on main and pass with the change.
• tests/test_backend_msword.py passes, and the DOCX ground-truth files are unchanged.
• I ran the other backend suites as well: html, mhtml, epub, email, markdown, xlsx, pptx, odf, jats, uspto, mets-gbs, vtt, asciidoc, csv, latex and docling-json. 523 passed. One test, test_email_backend_collapses_line_breaks_in_headers, fails the same way on main in my environment.
• ruff format and ruff check are clean. ty check on the file gives the same diagnostics as on main.

Checklist:

• Documentation has been updated, if necessary.
• Examples have been added, if necessary.
• Tests have been added, if necessary.

Disclosure: this PR is AI-assisted. Claude Code found the bug while converting randomly generated DOCX files and checking that every word survived, then wrote the fix and the tests and ran the checks above.

🤖 Generated with Claude Code

[github-actions]

✅ DCO Check Passed

Thanks @Yuilona, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🟢 Merge protection satisfied — ready to merge.

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[PeterStaar-IBM]

lgtm!