[MaxGhenis]

What

Re-pins microcosm's NZ Chronicle feed (nz/chronicle_feed.json) to Chronicle main acbedb5f, which includes chronicle#379 (G8-d4): Stats NZ estimated resident population at 30 June 2023 for New Zealand and its 16 regions. (External planning context, not in this repo: the NZ hub's G8-r work uses these rows to scale Census 2023 structure to June 2025 by the ERP 2025 / ERP 2023 ratio.)

Field	Was (5c4912aa)	Now (acbedb5f)
source_commit	5c4912aa…	acbedb5fef490137219207cf40945d70c2db34ea
fact_row_count	12,318	12,335
facts_sha256	2dc87f91…	5a016eaebdeb289d09470bdc99fa637db383eeeb5b762d2c80db0e8952ecf1a8
manifest_sha256	d79a4a0d…	678638048ba8a0abef14af27ab96977a8c796df27e0ef43436f409663ceb4178
consumer_fact_schema_sha256	cc9efb90…	d953306d6f01fbc64db4372d0262d3012ee590269f8a25b276b18dca0626c6a0
artifact_schema_version / fact schema versions	v2 / consumer_fact.v4	unchanged

The schema digest moves because chronicle 06315995 and cbd3423d added gas_year to the v4 consumer-fact period vocabulary. The UK pin already carries d953306d. No NZ row uses the new period type.

tests/golden/nz_country_spec.json moves only the chronicle_feed.json resource hash and the fingerprint. It was regenerated with canonical_json_bytes(_loaded_spec_summary("nz")).

How the artifact was built

• The hub ran nz_suite_build.sh from a clean detached Chronicle worktree at acbedb5f (dirty=0). That's one build-bundle --suite nz (82 s) and one build-consumer-artifact (8 s).
• The run read 67 source packages, skipped none, and produced 12,335 facts. The artifact's facts file and manifest hash to the pinned values.
• Row diff against the 5c4912aa artifact that Re-pin the NZ Chronicle feed to chronicle main 5c4912aa (G8-d2 TA population) #1272 pinned, keyed by aggregate_fact_key:
  • 17 rows added, 0 removed, 0 changed;
  • the added rows are Stats NZ ERP at 30 June 2023: 16 regions (regc_2025 ids 01–09 and 12–18; no row for 99, Area Outside Region) and New Zealand (5,200,000).
• Build record: ~/reviews/nz-hub/jobs/nz-suite-build-acbedb5f/summary.txt, with source_commit and dirty=0.

Tests

One file per run, worktree code, NZ_CHRONICLE_ARTIFACT pointing at the new artifact:

File	Result
test_ledger_artifact.py (including the real-artifact pin test)	40 passed
test_country_spec.py (golden byte equality)	136 passed
test_nz_spec_package.py	1189 passed
test_spec_engine_country_bundles.py	21 passed
test_transport_target_kernels.py	15 passed

Control: with NZ_CHRONICLE_ARTIFACT pointing at the old 5c4912aa artifact, test_real_nz_artifact_matches_committed_feed_pin fails (1 failed); at the new artifact it passes. ruff check and git diff --check pass.

Invariants

• The pin's six identity fields equal the artifact built from the declared source commit.
• The new pin admits only the new artifact.
• Only the feed resource hash and the fingerprint move in the golden.
• The existing 12,318 rows are byte-identical between the two artifacts.

Inactive NZ package; no published number. Needs Max: no.

🤖 Generated with Claude Code

[MaxGhenis]

Merge audit (NZ hub): head 0078734. Independent review: GPT-6.1 Sol (Subfleet 20261022-214144), VERDICT APPROVE at this head. It recomputed all six pin identity fields, diffed the old and new artifacts (17 added, 0 removed, 0 changed), matched the added rows to the Stats NZ workbook, and ran 1,401 tests plus both negative controls. CI: gh pr checks exit 0. Its "Needs Max" flag is the changelog sentence, which describes an input-data change verified by that review; it is not a published estimate. Merging with a merge commit because G8-r1 (nzhub/g8r1-references) is stacked on 0078734.