Refactor tests, scripts and the rules tool, then select every ruff rule (ALL)

[anth-volk]

Fixes #3931

Summary

Phase 1 step 1.8, group 8b, part 3 of 3: tests, scripts and the rules tool, refactored within ruff's complexity limits, then select = ["ALL"]. Stacked on #3933. This completes step 1.8.

• All 64 findings in scope are fixed, with no complexity suppressions.
• Rules tool (scripts/rules/, still standard-library only and runnable on Python 3.11):
  • The command line, registry loading, enforcer descriptions, the Mermaid graph, validation and the workflow parser are split into helpers and lookup tables.
  • 35 new unittests pin its command line, schema, validation (all 41 ordered violations from a fixture registry) and rendered output.
  • Old and new versions gave byte-identical output across every real workflow, about 3,000 generated workflows and 500 generated registries.
• Scripts: the collector verifier, baseline capture, Cloud Run load generator, household mirror command and local boot smoke test now take grouped values, and status codes are HTTPStatus members. Each command-line interface is unchanged: the same --help text, output and exit codes. New tests cover each script.
• Tests:
  • Long tests split into assertion helpers.
  • Fixtures grouped into composite fixtures or case values.
  • Route tables replace long if chains in the v1 contract test.
  • No assertion was dropped, and collected test IDs are unchanged apart from the 71 added.
• ALL:
  • pyproject.toml selects every rule, and ignores only the two that conflict with ruff format (COM812, ISC001) and one of each mutually exclusive docstring pair (D203, D213).
  • The registered per-file exemptions are unchanged: tests, generated Alembic revisions, and the rules tool's unittest tests for PT.
  • ALL found one new finding, a tuple-valued test enum without __slots__, which gets its own commit.
• Registry and docs:
  • api.quality.ruff-lint now lists ALL, and its statement gives the four ignores and why.
  • The rules-registry skill doc says every rule applies and that a ruff upgrade PR fixes any new rule's violations.

Testing

• The make test set at this branch's tip: 2,919 passed, which is the base's 2,433 plus 166, 249 and this PR's 71.
• The rules tool's unittests (109) pass on Python 3.11 and on the venv's Python.
• uvx ruff@0.15.12 check . (with ALL) and ruff format --check . are clean. Ruff reports no formatter conflicts.
• make typecheck-v2 is clean.
• mypy outside the gate:
  • The changed files that already existed show no new errors, and 9 fewer than the base.
  • The 23 new characterization test files from parts 1 and 2 have 152 mypy errors, mostly no-untyped-def. Tests are outside the mypy gate and exempt from annotation rules.
• Every model schema and pre-existing adapter, the FastAPI OpenAPI, the /v2/openapi.json preview and the legacy spec are byte-identical to ruff-idioms-arguments, with and without OLD_SIMULATION_GATEWAY_URL set.
• CI/CD: no workflow changed. The changed .github scripts read the same environment variables in the same order. The collector's COLLECTOR_ID_TOKEN and GOOGLE_OAUTH_ACCESS_TOKEN are exported by verify_observability_collector.sh from gcloud. OBSERVABILITY_PROJECT_ID and OBSERVABILITY_METRIC_LOCATION come from vars.* in deploy-observability-collector.yml. I haven't checked that those repository variables exist in GitHub's settings.
• CI: above the bottom of its stack, so only CI complete runs until the PRs below merge.

🤖 Generated with Claude Code