[EPAC-1880]: Backfill unittest tests for backend/cpp-oas-statistics

[riddim-developer-bot]

Why

backend/cpp-oas-statistics/cpp_oas_statistics.py was the only Python ingest pipeline without a unittest suite. This PR closes that gap so CI can gate CPP/OAS parser regressions the same way it gates every other pipeline.

What changed

• backend/cpp-oas-statistics/test_cpp_oas_statistics.py — four required named test cases: test_parses_happy_path_fixture, test_handles_missing_required_field, test_handles_malformed_upstream, test_emits_required_log_events.
• backend/cpp-oas-statistics/fixtures/cpp_fixture.csv and oas_fixture.csv — minimal fixtures (Ontario + Alberta, 2022–2024) so tests run without network access.
• backend/cpp-oas-statistics/fixtures/expected_output.json — golden snapshot (13 provinces, national totals CPP=1,530,000 OAS=1,330,000).
• .github/workflows/backend-python.yml — new workflow (did not previously exist); matrix over 10 Python pipelines; cpp-oas-statistics included without skip guard.

Trade-offs not taken

• Fixture coverage limited to Ontario + Alberta (2 of 13 provinces). Broader coverage adds noise without exercising new code paths.
• Did not migrate to pytest (out of scope per issue).

Test plan

cd backend/cpp-oas-statistics && python3 -m unittest discover -v

Verification evidence

$ cd backend/cpp-oas-statistics && python3 -m unittest discover -v
test_emits_required_log_events ... ok
test_handles_malformed_upstream ... ok
test_handles_missing_required_field ... ok
test_parses_happy_path_fixture ... ok
test_returns_none_for_empty ... ok
test_returns_none_for_unparseable ... ok
test_standard_bilingual_format ... ok

Ran 7 tests in 0.004s
OK

actionlint passed on backend-python.yml (pre-commit hook). Skipped: iOS build — no Swift sources changed.

Resolves EPAC-1880

Reviewer-Boundary: review-only

[riddim-reviewer-bot]

⚠️ Autonomous review liveness watchdog

Symphony expected a reviewer-bot review for the current autonomous PR head within the configured SLA window, but none was found.

• PR: [EPAC-1880]: Backfill unittest tests for backend/cpp-oas-statistics #451
• Repo: RiddimSoftware/epac
• Head SHA: f0e82a54ffa725fe6add86965cbfb54e0d2556ad
• Suspected missing reviewer owner: RiddimSoftware/epac
• Review SLA window: 300000 ms
• PR last activity: 2026-05-14T21:45:38Z
• Last heartbeat: fresh at 2026-05-14T22:13:13Z from riddim1.local @ 6360becfc8d5dc6e2d4a2c265ca6884c5d23e8c6

[riddim-reviewer-bot]

ReviewAutonomousPR

• Verdict: approve
• Reviewer boundary: review_only
• Acceptance criteria coverage: covered=6, missing=0, unclear=1

Summary

Reviewed against the provided diff and acceptance criteria; no blocking correctness issues identified in this review-only pass.

Actionable findings

• None.

Acceptance criteria coverage

• covered — test_parses_happy_path_fixture exists and asserts fixture-driven payload equals committed golden output.
  • Implemented in test_cpp_oas_statistics.py; compares against fixtures/expected_output.json (excluding generated timestamp).
  • Actionability: none
  • Evidence: backend/cpp-oas-statistics/test_cpp_oas_statistics.py, backend/cpp-oas-statistics/fixtures/expected_output.json
• covered — test_handles_missing_required_field asserts a clear non-KeyError failure on required-field omission.
  • Added bad_cpp inline fixture and asserts ValueError from build_payload.
  • Actionability: none
  • Evidence: backend/cpp-oas-statistics/test_cpp_oas_statistics.py
• covered — test_handles_malformed_upstream asserts parse failure and that pipeline finished is not logged.
  • Patches fetch_csv to empty payload and verifies main raises and log output lacks pipeline finished.
  • Actionability: none
  • Evidence: backend/cpp-oas-statistics/test_cpp_oas_statistics.py
• covered — test_emits_required_log_events validates JSON stderr events include pipeline started and pipeline finished.
  • Captures sys.stderr, patches fetch_csv, runs main(["--output", ...]), and parses each log line as JSON with required event/fields.
  • Actionability: none
  • Evidence: backend/cpp-oas-statistics/test_cpp_oas_statistics.py
• covered — At least one fixture file committed under backend/cpp-oas-statistics/fixtures/ so tests run offline.
  • Added cpp_fixture.csv, oas_fixture.csv, and golden snapshot under fixtures directory.
  • Actionability: none
  • Evidence: backend/cpp-oas-statistics/fixtures/cpp_fixture.csv, backend/cpp-oas-statistics/fixtures/oas_fixture.csv
• covered — backend-python.yml removes the cpp-oas-statistics skip guard and includes it in the matrix job.
  • New workflow matrix lists cpp-oas-statistics directly without conditional skip logic.
  • Actionability: none
  • Evidence: .github/workflows/backend-python.yml
• unclear — A regression in assertions should fail the cpp-oas-statistics matrix job.
  • This is an end-to-end CI behavior that should be validated in the parent workflow/run, but is not directly asserted in the added unit tests.
  • Actionability: follow_up
  • Suggested follow-up: Add/confirm a CI-level regression mutation test scenario or comment in PR that parent workflow behavior was observed.