fix(doctags): accept the paragraph tag when loading DocTags

[manticore999]

Problem

export_to_doctags() emits <paragraph> for items labelled PARAGRAPH, but the
tag is missing from the pattern load_from_doctags() matches against, so those
elements are skipped and their text is silently dropped. Export and import
disagree about the format.

PARAGRAPH is produced by the PPTX, JATS, USPTO, AsciiDoc and LaTeX backends, so
this affects ordinary documents. For a PowerPoint deck, every slide body is lost
and only the titles survive:

from docling.document_converter import DocumentConverter
from docling_core.types.doc import DoclingDocument
from docling_core.types.doc.document import DocTagsDocument

doc = DocumentConverter().convert("deck.pptx").document
print(doc.export_to_text())
# Quarterly Report
# Revenue grew by twelve percent this quarter.
# Costs remained flat across all regions.

dt = doc.export_to_doctags()          # contains <paragraph>Revenue grew ...</paragraph>
back = DoclingDocument.load_from_doctags(
    DocTagsDocument.from_doctags_and_image_pairs([dt], [None]), document_name="d")
print(back.export_to_text())
# Quarterly Report          <- slide text is gone

Minimal version, no files needed:

doc = DoclingDocument(name="x")
doc.add_text(label=DocItemLabel.PARAGRAPH, text="PARA")
doc.add_text(label=DocItemLabel.TEXT, text="TEXT")
# round-trip through DocTags -> "TEXT"   ("PARA" dropped)

Fix

Add the tag to the parser's pattern and map it to DocItemLabel.PARAGRAPH, so
both the content and the label survive.

Impact

Measured over the 192 groundtruth DoclingDocuments in the docling repo,
comparing text length before and after a DocTags round-trip:

	documents losing >10% of text
before	29
after	17

The six worst cases (patents ipa20200022300, pftaps057006474, pg06442728,
pa20010031492, ipa20180000016, and powerpoint_malformed_pictures) returned
0–3% of their text and now round-trip correctly. The remaining 17 lose text for
unrelated reasons (drawingml, some tables, checkboxes) and are untouched here.

Tests

Two tests in test/test_doctags_load.py, covering the plain and <loc_...>
forms. Both fail without the change and pass with it; the existing doctags tests
are unaffected (12 passed).

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 @manticore999, 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)(?:\(.+\))?(!)?:

[manticore999]

Updated: widened from <paragraph> to also cover <handwritten_text> and <reference>.
All three are emitted by export_to_doctags but were missing from the parser's tag
pattern, so they hit the same silent-drop path. REFERENCE is produced by the LaTeX
backend; PARAGRAPH by PPTX, JATS, USPTO, AsciiDoc and LaTeX.

Tests are now parametrized over the three labels and fail without the change.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!