feat(FR): polish department + region metadata vs data.gouv.fr (#1352 PR-B)

[dr5hn]

Summary

PR-B of the four-PR Option-C plan for France data review (#1352). Polishes only department + region metadata in contributions/states/states.json, leaving cities (PR-A) and overseas city files (PR-D) untouched.

Authoritative source: data.gouv.fr open feeds (INSEE-derived):

• departements.min.json (101 entries) — https://github.com/user-attachments/files/25721909/departements.min.json
• regions.min.json (18 entries) — https://github.com/user-attachments/files/25721911/regions.min.json

A new read-only diff tool, bin/scripts/fixes/france_states_diff.py, compares the feeds against repo country_code='FR' records and reports real deltas, advisory typography, coverage gaps, and out-of-scope records.

Changes (14 fields, 13 records — all in contributions/states/states.json)

11 metropolitan department native repairs — the field contained unrelated French text fragments (looked like translation-pipeline garbage) on 11 records. Repaired against data.gouv.fr nom:

ISO2	name	native before → after
01	Ain	Se faire → Ain
11	Aude	entendus → Aude
27	Eure	Ton → Eure
30	Gard	Gardien → Gard
32	Gers	Génie → Gers
36	Indre	Interne → Indre
46	Lot	Parcelle → Lot
50	Manche	Quelques → Manche
55	Meuse	Miuse → Meuse
68	Haut-Rhin	Peau → Haut-Rhin
83	Var	Notre → Var

1 overseas region native repair:

ISO2	name	native before → after
973	French Guiana	French Guiana → Guyane

Corse (20R) reclassification + native typo — Wikipedia/INSEE confirm Corse is one of France's 18 regions; data.gouv.fr lists it in the regions feed (regional code 94). Reclassifying brings the metropolitan region count to 13 and the special-status collectivity count to 2 (Lyon Métropole + Paris):

ISO2	field	before → after
20R	type	metropolitan collectivity with special status → metropolitan region
20R	native	Corsée → Corse

Out of scope (flagged for follow-up)

• Region native typography drift vs data.gouv.fr (Grand-Est vs Grand Est, Pays-de-la-Loire vs Pays de la Loire, PAC apostrophe — both forms in use officially).
• Overseas collectivity native/name cleanups (PM, NC, TF, WF) — PR-D scope.
• fips_code coverage gap (27/124 FR records have FIPS) — not in data.gouv.fr feeds.
• Likely-stale timezone on 973 (set to Europe/Paris, should be America/Cayenne).

Full audit in .github/fixes-docs/FIX_1352_PR_B_SUMMARY.md.

Test plan

• python3 bin/scripts/fixes/france_states_diff.py reports Department deltas: 0, Region deltas: 0.
• python3 -m json.tool contributions/states/states.json — JSON valid.
• FR record count unchanged: 124 before, 124 after.
• All id values still globally unique across states.json (5,296 ids).
• Every FR country_id/country_code resolves correctly against contributions/countries/countries.json.
• .github/scripts/utils.js validateRecord against the 13 touched records: 0 errors, identical warning footprint to untouched FR records.
• FR type distribution after change: 95 dept / 13 metro region / 5 overseas region / 5 overseas collectivity / 2 metro collectivity w/ special status / 1 European collectivity / 1 overseas collectivity w/ special status / 1 overseas territory / 1 dependency = 124.

Refs: #1352 (do not close — only PR-B of 4)

🤖 Generated with Claude Code

[dr5hn]

Weekly data-quality review (2026-04-27)

Verdict: clean

Checks

• Schema: ✅ Only existing records modified; no new records; no auto-managed fields (flag, created_at, updated_at, id) touched.
• FK integrity: ✅ Only native and type fields changed; no FK fields modified.
• Coordinates: ✅ No coordinate changes.
• Wikidata: N/A (no Wikidata field changes)
• Naming convention: ✅ Correctly fixes:
  • 11 FR department native values that contained garbage text (e.g. "Se faire" → "Ain", "Miuse" → "Meuse", "Peau" → "Haut-Rhin") against authoritative data.gouv.fr nom.
  • French Guiana native from English "French Guiana" → French "Guyane" ✅ (canonical fix: English belongs in name, French in native).
  • Corse native typo "Corsée" → "Corse" ✅; type reclassified to "metropolitan region" in line with data.gouv.fr regions feed and ISO 3166-2.

Advisory (non-blocking)

• Cross-PR FIX_1039 postal conflict — contributions/countries/countries.json postal-code changes (+25/−25) and .github/fixes-docs/FIX_1039_SUMMARY.md are identical to those in PRs docs: multi-level territories policy (FR overseas, dual representation) (#1352 PR-C) #1392, feat(FR): diff cities against data.gouv.fr, add missing metropolitan communes (#1352 PR-A) #1394, and feat(FR-overseas): populate missing city files for GF, BL, MF, PM, TF (#1352 PR-D) #1400. The first to merge wins cleanly; the others will conflict on those two files. Coordinate with the sibling PRs to land FIX_1039 exactly once.
• French Guiana timezone (pre-existing) — The existing states.json record for French Guiana as an FR subdivision (id ≈ 4822) still carries timezone: "Europe/Paris"; the correct timezone is America/Cayenne. PR feat(FR-overseas): populate missing city files for GF, BL, MF, PM, TF (#1352 PR-D) #1400 adds a new GF-country state with the correct timezone, but this FR-subdivision record remains stale. Out of scope here, but worth a follow-up.

🤖 Automated weekly review — Claude (sonnet-4-6).

---

Generated by Claude Code

[Copilot]

Pull request overview

Polishes France subdivision metadata by aligning selected native names and the Corse region classification in the contributions dataset, and adds a read-only diagnostic script plus documentation to audit FR department/region deltas against data.gouv.fr feeds.

Changes:

• Fixes corrupted/incorrect native values for 11 metropolitan departments and French Guiana; corrects Corse native typo.
• Reclassifies Corse (FR-20R) from special-status collectivity to metropolitan region.
• Adds bin/scripts/fixes/france_states_diff.py (read-only comparison tool) and a companion audit summary doc.

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
contributions/states/states.json	Updates FR department/region native values and changes Corse type to metropolitan region.
bin/scripts/fixes/france_states_diff.py	New script to diff FR repo records vs data.gouv.fr department/region feeds and report deltas/coverage.
.github/fixes-docs/FIX_1352_PR_B_SUMMARY.md	Documents scope, rationale, and validation results for PR-B of issue #1352.

Comments suppressed due to low confidence (1)

contributions/states/states.json:181883

• The French Guiana (iso2=973) state record still has timezone set to Europe/Paris, but French Guiana’s IANA timezone is America/Cayenne (also reflected in contributions/countries/countries.json for country_code=GF). Since this PR already touches this record, consider correcting the timezone here (and potentially auditing the other overseas regions’ timezones in a follow-up).