feat: expand RF4 media inventory

rf4_research/community_cli.py

@@ -37,9 +37,9 @@ DEFAULT_STATE_FILE = Path(".cache/community-fetch-state.json")
 ALLOWED_HOSTS = frozenset({
     "download.rf4db.com", "rf4db.com",
     "rf4-stat.ru",
-    "rf4map.ru",
+    "rf4map.ru", "gw.rf4map.ru", "hb.ru-msk.vkcloud-storage.ru",
     "rf4-posts.com",
-    "rf4game.de",
+    "rf4game.de", "rf4game.ru",
 })
 MAX_RESPONSE_BYTES = 5 * 1024 * 1024  # 5 MB
 
@@ -105,6 +105,8 @@ def fetch_site_key(url: str) -> str:
         hostname = hostname[4:]
     if hostname.startswith("cdn."):
         hostname = hostname[4:]
+    if hostname == "gw.rf4map.ru":
+        hostname = "rf4map.ru"
     if not hostname:
         raise ValueError("source URL must include a hostname")
     return hostname

rf4_research/media_assets.py

@@ -41,7 +41,7 @@ def _kind(url: str, label: str | None, context: str) -> str:
     value = f"{label or ''} {context}".casefold()
     if "/flags/" in path or "/themes/" in path or path.endswith(("/logo.png", "/logo.svg")):
         return "reference"
-    if re.search(r"/(?:fish|fishes|fische|species)/", path) or "/media/fische/" in value or any(word in value for word in ("рыба ", "fish: ", "fisch: ")):
+    if re.search(r"/(?:fish|fishes|fische|species)/", path) or re.search(r"/(?:fish|species)[_-]", path) or "/media/fische/" in value or any(word in value for word in ("рыба ", "fish: ", "fisch: ")):
         return "fish"
     if re.search(r"/(?:maps?|levels?|lakes?|waterbodies)/", path) or "/media/levels/" in value or any(word in value for word in ("карта ", "map of ", "gewässerkarte")):
         return "waterbody"

rf4_research/media_cli.py

@@ -15,12 +15,12 @@ DEFAULT_ROOT = Path("data/media")
 MAX_ASSET_BYTES = 15 * 1024 * 1024
 
 
-def _download_one(root: Path, state_file: Path) -> str:
+def _download_one(root: Path, state_file: Path, asset_url: str | None = None) -> str:
     manifest_path = root / "manifest.json"
     manifest = json.loads(manifest_path.read_text(encoding="utf-8"))
-    queued = next((item for item in manifest["assets"] if item.get("status") == "queued"), None)
+    queued = next((item for item in manifest["assets"] if item.get("status") == "queued" and (asset_url is None or item["asset_url"] == asset_url)), None)
     if queued is None:
-        return "queue is empty"
+        return "queue is empty" if asset_url is None else "asset is not queued"
     url = queued["asset_url"]
     _validate_url_before_io(url)
     check_and_reserve(fetch_site_key(url), state_file=state_file)
@@ -51,6 +51,7 @@ def main(argv: list[str] | None = None) -> int:
     parser = argparse.ArgumentParser(description="Index authorized RF4 media without hotlinking")
     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("--asset-url", help="Download this exact queued asset instead of the first queue item")
     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"))
@@ -77,7 +78,7 @@ def main(argv: list[str] | None = None) -> int:
             print(f"reclassified {len(manifest['assets'])} candidates")
             return 0
         if args.download_one:
-            print(_download_one(args.root, args.state_file))
+            print(_download_one(args.root, args.state_file, args.asset_url))
             return 0
         if not args.url:
             parser.error("url is required unless --download-one is used")