feat(accuracy): build hardened phase 4 evaluation harness

tests/test_build_regional_yolo_dataset.py

@@ -35,11 +35,12 @@ def test_regional_dataset_balances_contexts_and_keeps_validation_closed() -> Non
     assert evidence["protected_samples_in_training"] == []
 
 
-def test_regional_dataset_rejects_protected_tiles() -> None:
+@pytest.mark.parametrize("protected_split", ["calibration", "test", "background-test", "challenge"])
+def test_regional_dataset_rejects_protected_tiles(protected_split: str) -> None:
     manifest = {"samples": [
-        {"sample_slug": "f-cal", "region": "flanders", "split": "calibration", "context": "ribbon"},
+        {"sample_slug": "f-cal", "region": "flanders", "split": protected_split, "context": "ribbon"},
     ]}
-    summary = {"tiles": [{"sample_slug": "f-cal", "split": "calibration", "image_path": "/cal.png"}]}
+    summary = {"tiles": [{"sample_slug": "f-cal", "split": protected_split, "image_path": "/cal.png"}]}
     with pytest.raises(ValueError, match="protected"):
         module.build(
             summary=summary, manifest=manifest, region="flanders",

tests/test_building_proposal_classifier.py

@@ -114,8 +114,9 @@ def test_classify_proposals_consumes_reference_once() -> None:
     assert [item[0] for item in miner.classify_proposals(proposals, reference, 0.25)] == ["positive", "negative"]
 
 
-def test_eligible_tiles_rejects_protected_manifest_split() -> None:
-    manifest = {"samples": [{"sample_slug": "x", "region": "flanders", "split": "test"}]}
+@pytest.mark.parametrize("protected_split", ["calibration", "test", "background-test", "challenge"])
+def test_eligible_tiles_rejects_protected_manifest_split(protected_split: str) -> None:
+    manifest = {"samples": [{"sample_slug": "x", "region": "flanders", "split": protected_split}]}
     summary = {"tiles": [{"sample_slug": "x", "split": "train", "image_path": "x.png"}]}
     with pytest.raises(ValueError, match="protected"):
         miner.eligible_tiles(summary, manifest, "flanders")