NZ Parliament Votes

Use the read-only Python CLI in scripts/cli.py to search weekly Journals and extract supported recorded party divisions. Every data command supports --json, bounded requests, stable exit codes, source URLs and retrieval timestamps.

Commands

# List one bounded page of weekly Journals; optionally filter titles on that page
python3 scripts/cli.py search [QUERY] --page 1 --limit 20 --json

# Extract printed divisions from a Journal UUID returned by search
python3 scripts/cli.py votes JOURNAL_ID --json

The common lookup path is positional (QUERY or JOURNAL_ID). Human-readable output is the default. --limit is restricted to 1–50 and each request has a 10-second timeout and 4 MiB response ceiling.

Interpretation rules

• Weekly Journals are drafts, not the authoritative Sessional Journal record.
• Return source-printed party labels/counts; preserve ambiguous singleton labels as entity_kind: unknown.
• Never infer an individual MP's vote from party membership, classify a singleton as a person, resolve a printed label to an identity, or add roles/electorates.
• Participant rows are emitted only when their printed counts reconcile with the side total.
• Missing abstentions remain null; unsupported layouts produce diagnostics, not inferred results or empty-source claims.
• Search filtering applies only to the one requested API page; it is not a full-archive search.

Resources

• scripts/cli.py — canonical command entrypoint
• scripts/parliament_votes.py — strict stdlib Journal and search parser
• scripts/test_contract.py — deterministic contract, adversarial parser, and normal/empty/blocked tests
• scripts/smoke_test.py — synthetic fixture assertions and bounded live search probe
• tests/fixtures/ — minimal scrubbed synthetic Journal, empty response, and contract sentinel
• references/source-profile.json — machine-readable commands and source constraints
• references/source-notes.md — provenance, parser semantics, and coverage limits