Recognize the Japanese maiden marker 旧姓 (#309)

[derek73]

Reopens the useful half of #309, which was closed in favour of #329.

parse("山田花子 旧姓 佐藤")    # family 山田花子, maiden 佐藤
parse("山田 花子 旧姓 佐藤")   # given 花子, family 山田, maiden 佐藤

1.4.0 left the marker in the name — first 山田花子 / middle 旧姓 / last 佐藤.

Why the default vocabulary and not locales.JA

旧姓 sits beside the Cyrillic урожд. and German geb. entries, on the rule that admitted those:

Native-script entries cannot collide with Latin-script names, which is what makes them safe to enable by default.

Verified: matching is whole-token, _normalize("旧姓") is 旧姓, and neither character appears in any shipped surname, title, suffix, conjunction, particle or bound-given vocabulary.

The pack exists for what needs the my-data-is-Japanese declaration — segmentation, where a pure-Han string cannot say which language wrote it. A Han-script marker needs no such declaration, because it can only ever match Han text. Patching maiden_delimiters in the pack would also have been wrong for an independent reason: （） is already a nickname delimiter and _extract resolves the collision by exclusion, so it would trade fullwidth-paren nicknames for maiden names wholesale (measured — 山田 太郎（マイケル・ジャクソン） moves from nickname to maiden).

Scope, stated in three places rather than left to discovery

This reaches the spaced form only. Japanese more often brackets the marker, and 山田（旧姓：佐藤） under Policy(maiden_delimiters=...) still returns maiden "旧姓：佐藤" with the marker and its colon attached — _extract claims delimited content whole, before classify has tagged anything inside it, so _group's marker-consuming rule never sees it.

That is not a Japanese limitation. Jane Smith (née Jones) keeps its marker the same way, where the bare Jane Smith née Jones drops it. Closing that asymmetry is #329.

The limit is recorded in the vocabulary file's docstring, both case-row notes, and the release-log entry.

The differential rule is its own, deliberately

fix(#274) maiden markers consumed declares fields = ["maiden", "middle", "last"]. The Japanese diff touches four:

first:  '山田花子' -> ''
middle: '旧姓'    -> ''
last:   '佐藤'    -> '山田花子'
maiden: ''        -> '佐藤'

first moves because the name left after the marker is consumed is wholly Han, so it reads family-first (#271) — the maiden fix compounding with the order flip, the same pattern already named by cjk-comma-compound and cjk-honorific-suffix.

Adding first to #274's field list would have made that rule broader than its prose and let it absorb diffs unrelated to markers, which is exactly the hazard filed as #328. The new rule's name_regex is the marker itself, so it can claim nothing else.

Verification

• Suite 2942, mypy / ruff / Sphinx -W clean
• Differential 750 names, unexplained: 0; both new rows claimed by the new narrow rule
• Two case rows, classified fix(#309) against measured 1.4.0 values
• corpus_cjk.jsonl regenerated from the case table

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.47%. Comparing base (7f4074d) to head (0fa17c8).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #330   +/-   ##
=======================================
  Coverage   98.47%   98.47%           
=======================================
  Files          41       41           
  Lines        2823     2823           
=======================================
  Hits         2780     2780           
  Misses         43       43           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.