j

ProviderCore/registry.py

@@ -68,6 +68,13 @@ def _supports_upload(provider: Provider) -> bool:
     return provider.__class__.upload is not Provider.upload
 
 
+def _provider_url_patterns(provider_class: Type[Provider]) -> Sequence[str]:
+    try:
+        return list(provider_class.url_patterns())
+    except Exception:
+        return []
+
+
 def get_provider(name: str,
                  config: Optional[Dict[str,
                                        Any]] = None) -> Optional[Provider]:
@@ -166,47 +173,53 @@ def list_file_providers(config: Optional[Dict[str, Any]] = None) -> Dict[str, bo
 def match_provider_name_for_url(url: str) -> Optional[str]:
     """Return a registered provider name that claims the URL's domain.
 
-    Providers can declare domains via a class attribute `URL_DOMAINS` (sequence of strings).
+    Providers can declare domains via class attribute `URL` (preferred) or `URL_DOMAINS`.
     This matcher is intentionally cheap (no provider instantiation, no network).
     """
 
+    raw_url = str(url or "").strip()
+    raw_url_lower = raw_url.lower()
     try:
-        parsed = urlparse(str(url))
+        parsed = urlparse(raw_url)
         host = (parsed.hostname or "").strip().lower()
         path = (parsed.path or "").strip()
     except Exception:
         host = ""
         path = ""
 
-    if not host:
-        return None
-
     # Prefer Internet Archive for archive.org links unless the URL clearly refers
     # to a borrow/loan flow (handled by OpenLibrary provider).
     #
     # This keeps direct downloads and item pages routed to `internetarchive`, while
     # preserving OpenLibrary's scripted borrow pipeline for loan/reader URLs.
-    if host == "openlibrary.org" or host.endswith(".openlibrary.org"):
-        return "openlibrary" if "openlibrary" in _PROVIDERS else None
-
-    if host == "archive.org" or host.endswith(".archive.org"):
-        low_path = str(path or "").lower()
-        is_borrowish = (
-            low_path.startswith("/borrow/") or low_path.startswith("/stream/")
-            or low_path.startswith("/services/loans/") or "/services/loans/" in low_path
-        )
-        if is_borrowish:
+    if host:
+        if host == "openlibrary.org" or host.endswith(".openlibrary.org"):
             return "openlibrary" if "openlibrary" in _PROVIDERS else None
-        return "internetarchive" if "internetarchive" in _PROVIDERS else None
+
+        if host == "archive.org" or host.endswith(".archive.org"):
+            low_path = str(path or "").lower()
+            is_borrowish = (
+                low_path.startswith("/borrow/") or low_path.startswith("/stream/")
+                or low_path.startswith("/services/loans/") or "/services/loans/" in low_path
+            )
+            if is_borrowish:
+                return "openlibrary" if "openlibrary" in _PROVIDERS else None
+            return "internetarchive" if "internetarchive" in _PROVIDERS else None
 
     for name, provider_class in _PROVIDERS.items():
-        domains = getattr(provider_class, "URL_DOMAINS", None)
-        if not isinstance(domains, (list, tuple)):
+        domains = _provider_url_patterns(provider_class)
+        if not domains:
             continue
         for d in domains:
             dom = str(d or "").strip().lower()
             if not dom:
                 continue
+            if raw_url_lower.startswith(dom):
+                return name
+        for d in domains:
+            dom = str(d or "").strip().lower()
+            if not dom or not host:
+                continue
             if host == dom or host.endswith("." + dom):
                 return name