feat: track media catalog coverage

rf4_research/media_assets.py

@@ -162,6 +162,25 @@ def audit_media_catalog(root: Path) -> dict:
     return {"total": sum(statuses.values()), "statuses": statuses, "issues": issues, "orphaned_files": orphaned}
 
 
+def media_coverage(root: Path) -> dict:
+    """Compare quarantined and approved media with the versioned catalog baseline."""
+    manifest = json.loads((root / "manifest.json").read_text(encoding="utf-8"))
+    baseline = json.loads((root / "catalog-baseline.json").read_text(encoding="utf-8"))
+    result: dict[str, dict] = {}
+    for entity_type, target in baseline["entities"].items():
+        candidates = [item for item in manifest.get("assets", []) if item.get("entity_type") == entity_type]
+        approved_keys = {item.get("entity_key") for item in candidates if item.get("status") == "approved" and item.get("entity_key")}
+        expected = target.get("count")
+        result[entity_type] = {
+            "expected": expected,
+            "candidates": len(candidates),
+            "approved": len(approved_keys),
+            "candidate_gap": max(expected - len(candidates), 0) if isinstance(expected, int) else None,
+            "approved_gap": max(expected - len(approved_keys), 0) if isinstance(expected, int) else None,
+        }
+    return {"baseline_date": baseline["verified_at"], "entities": result}
+
+
 def inspect_image(body: bytes) -> tuple[int, int, str]:
     try:
         with Image.open(io.BytesIO(body)) as image: