fix(jats): preserve structured abstract sections and fix nested list parent

[ceberam]

Summary

This PR addresses two gaps in the JATS backend. In particular, it addressed a gap not fully covered by #3584 and documented on #3584 (comment)

Structured abstracts: JATS articles commonly use a structured abstract — an <abstract> element whose content is split into named <sec> children (e.g. Background, Results, Conclusion). The previous implementation collected all section text into a single TextItem, prepending each section title inline as "Title: text…". This lost the document structure that JATS explicitly encodes. Each <sec> inside an <abstract> is now emitted as a dedicated section heading (DocItemLabel.SECTION_HEADER) with its paragraph(s) as separate TextItem children beneath it. Plain (un-sectioned) abstracts are unaffected.

Nested list parent: Nested <list> elements inside a <list-item> were walked with the ListItem itself as the docling parent, causing add_list_item() to receive a non-list-group parent. This triggered a DeprecationWarning from docling-core and a silent on-the-fly group creation. The list group is now created explicitly before recursing into the nested list.

Changes

docling/backend/xml/jats_backend.py

• Adds a new AbstractSection TypedDict (title, paragraphs) to represent a parsed abstract sub-section.
• Extends the Abstract TypedDict with a sections: list[AbstractSection] field alongside the existing content field (which is retained for plain paragraph-only abstracts).
• Rewrites _parse_abstract_section() to return an AbstractSection instead of a flat string.
• Rewrites _parse_abstract() to separate <sec> children (→ sections) from bare <p> children (→ content).
• Rewrites _add_abstract() to emit each section as doc.add_heading(level=hlevel+2) with individual doc.add_text() calls underneath, falling back to the original single-TextItem path for plain abstracts.
• Fixes nested list handling: explicitly creates a GroupLabel.LIST group under the ListItem before walking a nested <list>, so that child list-item elements always receive a proper list-group parent.

tests/test_backend_jats.py

• Updates test_jats_structured_abstract_sections_are_preserved to assert the new structure (sub-headings present, flat "Title: text" format absent) and to verify the DoclingDocument item labels directly.

tests/data/jats/sources/pmc2231364.nxml

• Adds a minimal JATS fixture derived from the BMC Microbiology article PMC2231364 (CC BY 2.0), which has the three-section Background/Results/Conclusion abstract structure that originally exposed the gap.

tests/data/jats/groundtruth/pmc2231364.nxml.*

• Groundtruth files generated from the new fixture. The markdown output now matches the structure visible on the PubMed page:

## Abstract

### Background

Environmental modulation of gene expression…

### Results

To provide us with a comprehensive view…

### Conclusion

The comparative transcriptomics analysis…

Behaviour before and after

Structured abstract

Before — all sections collapsed into one TextItem:

## Abstract

Background: Environmental modulation… Results: To provide… Conclusion: The comparative…

After — each section is a separate heading with its own paragraph:

## Abstract

### Background

Environmental modulation…

### Results

To provide…

### Conclusion

The comparative…

Nested lists

Before — ListItem passed directly as parent to nested list walk, triggering:

DeprecationWarning: ListItem parent must be a list group, creating one on the fly.

After — a GroupLabel.LIST group is created explicitly under the ListItem before recursing, matching the required parent contract.

Testing

All 40 JATS tests pass with -W error::DeprecationWarning. The new fixture is picked up automatically by test_e2e_jats_conversions.

Checklist:

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

[github-actions]

✅ DCO Check Passed

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

[mergify]

Merge Protections

🔴 1 of 2 protections blocking · waiting on 👀 reviews

	Protection	Waiting on
🔴	Require two reviewer for test updates	👀 reviews
🟢	Enforce conventional commit	—

🔴 Require two reviewer for test updates

Waiting for

• #approved-reviews-by >= 2

This rule is failing.

When test data is updated, we require two reviewers

• #approved-reviews-by >= 2

---

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)(?:\(.+\))?(!)?:

[codecov]

Codecov Report

❌ Patch coverage is 86.48649% with 5 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
docling/backend/xml/jats_backend.py	86.48%	1 Missing and 4 partials ⚠️

📢 Thoughts on this report? Let us know!

[PeterStaar-IBM]

lgtm!