feat: add explicit media review workflow

rf4_research/media_assets.py

@@ -95,6 +95,31 @@ def reclassify_manifest(path: Path) -> dict:
     return manifest
 
 
+def review_asset(
+    path: Path, *, asset_url: str, decision: str, entity_type: str | None = None,
+    entity_key: str | None = None, note: str | None = None,
+) -> dict:
+    if decision not in {"approved", "rejected"}:
+        raise ValueError("decision must be approved or rejected")
+    manifest = json.loads(path.read_text(encoding="utf-8"))
+    item = next((asset for asset in manifest.get("assets", []) if asset["asset_url"] == asset_url), None)
+    if item is None:
+        raise ValueError("asset URL is not present in manifest")
+    if decision == "approved":
+        if item.get("status") != "stored":
+            raise ValueError("only a stored asset can be approved")
+        if entity_type not in {"fish", "waterbody", "tackle", "reference"} or not entity_key:
+            raise ValueError("approved asset requires entity type and canonical key")
+        item.update({"entity_type": entity_type, "entity_key": entity_key})
+    item.update({
+        "status": decision,
+        "reviewed_at": datetime.now(timezone.utc).isoformat(),
+        "review_note": note or None,
+    })
+    path.write_text(json.dumps(manifest, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
+    return item
+
+
 def inspect_image(body: bytes) -> tuple[int, int, str]:
     try:
         with Image.open(io.BytesIO(body)) as image:

rf4_research/media_cli.py

@@ -8,7 +8,7 @@ import urllib.error
 import urllib.request
 
 from .community_cli import USER_AGENT, _StrictRedirectHandler, _validate_url_before_io, check_and_reserve, fetch_html, fetch_site_key
-from .media_assets import extract_media_candidates, merge_manifest, reclassify_manifest, store_asset
+from .media_assets import extract_media_candidates, merge_manifest, reclassify_manifest, review_asset, store_asset
 
 
 DEFAULT_ROOT = Path("data/media")
@@ -52,10 +52,21 @@ def main(argv: list[str] | None = None) -> int:
     parser.add_argument("url", nargs="?")
     parser.add_argument("--download-one", action="store_true", help="Store one queued asset while respecting site cooldown")
     parser.add_argument("--reclassify", action="store_true", help="Reapply current conservative classifier without network access")
+    parser.add_argument("--review-url", help="Review an asset already present in the manifest")
+    parser.add_argument("--decision", choices=("approved", "rejected"))
+    parser.add_argument("--entity-type", choices=("fish", "waterbody", "tackle", "reference"))
+    parser.add_argument("--entity-key")
+    parser.add_argument("--note")
     parser.add_argument("--root", type=Path, default=DEFAULT_ROOT)
     parser.add_argument("--state-file", type=Path, default=Path(".cache/community-fetch-state.json"))
     args = parser.parse_args(argv)
     try:
+        if args.review_url:
+            if not args.decision:
+                parser.error("--decision is required with --review-url")
+            item = review_asset(args.root / "manifest.json", asset_url=args.review_url, decision=args.decision, entity_type=args.entity_type, entity_key=args.entity_key, note=args.note)
+            print(f"{item['status']} {item['asset_url']}")
+            return 0
         if args.reclassify:
             manifest = reclassify_manifest(args.root / "manifest.json")
             print(f"reclassified {len(manifest['assets'])} candidates")