scripts/test__eval_threshold_gate.py

• _task · function · L15-L24 — def _task(task_name="citation_extraction", agg_metric="accuracy", agg_passed=True, per_class=None, status="measured")
• _report · function · L27-L28 — def _report(*tasks)
• test_aggregate_pass_no_failure · function · L34-L35 — def test_aggregate_pass_no_failure()
• test_aggregate_fail_reported · function · L38-L41 — def test_aggregate_fail_reported()
• test_no_threshold_not_gated · function · L44-L47 — def test_no_threshold_not_gated(): # aggregate without a passed key (task declares no threshold) is not gated
• test_pending_task_never_gated · function · L50-L51 — def test_pending_task_never_gated()
• test_unknown_nonmeasured_status_never_gated · function · L54-L56 — def test_unknown_nonmeasured_status_never_gated(): # symmetric with the lift gate: a future non-measured status is skipped, not gated
• test_missing_status_treated_as_measured · function · L59-L62 — def test_missing_status_treated_as_measured(): # back-compat: a task with no status key is gated (pre-status reports stay valid)
• test_per_class_fail_with_aggregate_pass_is_reported · function · L68-L82 — def test_per_class_fail_with_aggregate_pass_is_reported()
• test_per_class_key_shape_matches_lift_gate · function · L85-L96 — def test_per_class_key_shape_matches_lift_gate()
• test_both_aggregate_and_per_class_fail_reported · function · L99-L107 — def test_both_aggregate_and_per_class_fail_reported()
• test_per_class_without_threshold_not_gated · function · L110-L117 — def test_per_class_without_threshold_not_gated(): # a per_class row that carries no passed key (no per-class threshold) is # not gated — only passed is False blocks
• test_per_class_on_pending_task_never_gated · function · L120-L125 — def test_per_class_on_pending_task_never_gated()