Enable ruff argument, idiom and pytest-style rules (ARG, B, C4, FBT, PERF, PLW, PT, PTH, RUF, SIM and others)

[anth-volk]

Fixes #3923

Summary

Phase 1 step 1.8, group 8a: arguments, idioms and pytest style. Stacked on #3920. Group 8b follows with the complexity refactors and the switch to select = ["ALL"].

• Test stand-ins exempt from unused-argument rules (added 2026-10-16): test fakes and stand-ins mirror the signatures of what they replace, because callers pass those arguments by keyword.
  • ARG001, ARG002 and ARG005 join the **/tests/** file-level exceptions (api.quality.tests-ruff-exemptions). That removes 32 inline codes from 16 test files here and 37 by Register weighted-microdata rules, with AST guards and #767 proofs #3938.
  • Application code keeps its unused-argument suppressions and reasons.
• Unittest exemption (decided for this group): the shared rules tool's tests use unittest on purpose, because the tool needs only the standard library. scripts/rules/tests/** is exempt from ruff's pytest-style rules (PT), registered as api.quality.unittest-ruff-exemptions.
• Rules: select gains ARG, ASYNC, B, C4, FBT, PD, PERF, PGH, PLW, PT, PTH, RUF and SIM. All 912 findings are fixed or kept with an inline reason, one commit per family.
• Unused arguments (ARG, 468):
  • 118 tests now request side-effect fixtures with @pytest.mark.usefixtures; the collected test IDs are identical.
  • Unused *args/**kwargs became *_args/**_kwargs. A named lambda or callback parameter is underscored only where every caller passes it positionally.
  • The rest keep their names: methods that implement a protocol or override a base method (typing.override needs Python 3.12), Flask views that receive URL variables by keyword, and SQLAlchemy and Alembic hooks keep an inline reason. Test fakes called by keyword are covered by the test exceptions above.
• Keyword-only booleans (FBT): PolicyEngineCountry.calculate and prepare_calculation (spm_requested, prepared), EconomyService._setup_sim_options, and the cache backend protocol's pipeline(*, transaction=True). Every caller already passed these by keyword. The FastAPI current query parameter and Alembic's positional include_object hook are kept, with reasons.
• Bugbear (B):
  • B904: the simulation and report-output routes' create and update handlers now raise BadRequest(...) from e. The 400 responses are unchanged; only the traceback's chaining text changes.
  • B905: zip(..., strict=True) where the lengths are fixed by literals.
  • B023: loop variables bound in closures that only run within their own iteration.
• Best-effort swallows (SIM105, 21): every try: … except Exception: pass around a telemetry or logging call is now with contextlib.suppress(Exception):, with the reason as a plain comment above it. Behaviour is identical.
  • Count: 19 blocks that groups 5 and 6 had kept as try/except with suppressions, plus two converted earlier.
  • Sites: the ASGI factory, request context, migration logging, the observability runtime, the household and policy routes, the economy service's SPM error handling, the three mirroring services, and the household-event command script.
  • Suppressions removed: 19 SIM105, 20 BLE001 and 17 S110. One BLE001 was on migration logging's outer handler, which ruff now sees log the traceback.
  • Trade-off: ruff's BLE001 doesn't flag contextlib.suppress(Exception). These catch-alls are documented by their comments rather than by a reasoned suppression.
• Idioms: 154 dict() calls become literals, plus comprehensions (C4, PERF), simplified conditions and combined with statements in the same order (SIM), ClassVar annotations, explicit Optional types, raw regex patterns and sorted __all__ (RUF), Path.open (PTH) and check=False on subprocess.run (PLW1510).
• Pytest style (PT): 37 pytest.raises gain match= with the real message; where a test only asserted that text separately, the match replaces that assertion. Parametrize names and values use tuples and lists, and composite assertions are split.
• Kept with a reason:
  • Module-level state that tests patch: the Cloud SQL, v2 engine and readiness globals (PLW0603).
  • Non-ASCII digits under test: RUF001.
  • .values: MicroSeries and a pandas extension array, where .to_numpy() would differ (PD011).
  • A single startup read in an async function: ASYNC240.
• Registry: api.quality.ruff-lint now lists the 13 families. The rules tool changed (comprehensions, check=False, a bound loop variable and an unused parser parameter removed); sim takes the same changes in Enable ruff argument, idiom and pytest-style rules (ARG, B, C4, FBT, PERF, PLW, PT, PTH, RUF, SIM and others) policyengine-sim-api#802.

Testing

• The make test set (tests/unit, tests/contract, to_refactor and the dedupe integration test): 2,416 passed, 4 failed, the same as the base today.
  • The 4 failures are a date-dependent test on master. tests/unit/v2/test_stage12_comparison_models.py fixes retention_expires_at at 2026-10-14 but leaves created_at on the real clock, so its inserts fail the retention check since that date. It's unrelated to this PR and needs its own fix against master.
  • With the model clock frozen at 2026-09-20, every test passes on both base and branch.
• make typecheck-v2 is clean. mypy on the 128 changed files outside that gate shows no new errors.
• ruff check and ruff format --check (0.15.12) are clean repo-wide.
• Every Pydantic model schema and adapter, the FastAPI OpenAPI with every native router enabled, the /v2/openapi.json preview and the legacy spec are byte-identical to the base.
• The quality guards, the rules guard and the rules tool unittest (also on Python 3.11) pass.
• For the 2026-10-16 changes, at each tip from Enable ruff argument, idiom and pytest-style rules (ARG, B, C4, FBT, PERF, PLW, PT, PTH, RUF, SIM and others) #3924 to Register weighted-microdata rules, with AST guards and #767 proofs #3938:
  • ruff check (including RUF100) and ruff format --check are clean, and so are the guards and the rules tool unittests.
  • Rendered rules are fresh at all 104 commits.
  • Project Python differs from the previous version only in noqa comments, apart from:
    • the twelve files with contextlib.suppress rewrites;
    • one test lambda that ruff format re-wrapped once its comment was gone.
  • mypy is clean at Enable ruff argument, idiom and pytest-style rules (ARG, B, C4, FBT, PERF, PLW, PT, PTH, RUF, SIM and others) #3924.
  • The test modules that cover the rewritten code pass: 809 tests in 31 modules at Enable ruff argument, idiom and pytest-style rules (ARG, B, C4, FBT, PERF, PLW, PT, PTH, RUF, SIM and others) #3924, 980 at Register weighted-microdata rules, with AST guards and #767 proofs #3938, and 159 in the v1 route contract and deploy-script tests. The full suites weren't re-run for this change.
• CI/CD: no workflow changed. .github/find-api-model-versions.py changed only in idioms; it reads only GITHUB_ENV, which the Actions runner provides, and only when run without --shell, which its CI caller always passes.
• CI: this PR is above the bottom of its stack, so it runs only CI complete, which fails and says why. Its full run starts once the PRs below it merge.

🤖 Generated with Claude Code