dfd

helper/metadata_search.py

@@ -7,6 +7,11 @@ import sys
 
 from helper.logger import log, debug
 
+try:  # Optional dependency
+    import musicbrainzngs  # type: ignore
+except ImportError:  # pragma: no cover - optional
+    musicbrainzngs = None
+
 
 class MetadataProvider(ABC):
     """Base class for metadata providers (music, movies, books, etc.)."""
@@ -266,6 +271,86 @@ class GoogleBooksMetadataProvider(MetadataProvider):
         return tags
 
 
+class MusicBrainzMetadataProvider(MetadataProvider):
+    """Metadata provider for MusicBrainz recordings."""
+
+    @property
+    def name(self) -> str:  # type: ignore[override]
+        return "musicbrainz"
+
+    def search(self, query: str, limit: int = 10) -> List[Dict[str, Any]]:
+        if not musicbrainzngs:
+            log("musicbrainzngs is not installed; skipping MusicBrainz scrape", file=sys.stderr)
+            return []
+
+        q = (query or "").strip()
+        if not q:
+            return []
+
+        try:
+            # Ensure user agent is set (required by MusicBrainz)
+            musicbrainzngs.set_useragent("Medeia-Macina", "0.1")
+        except Exception:
+            pass
+
+        try:
+            resp = musicbrainzngs.search_recordings(query=q, limit=limit)
+            recordings = resp.get("recording-list") or resp.get("recordings") or []
+        except Exception as exc:
+            log(f"MusicBrainz search failed: {exc}", file=sys.stderr)
+            return []
+
+        items: List[Dict[str, Any]] = []
+        for rec in recordings[:limit]:
+            if not isinstance(rec, dict):
+                continue
+            title = rec.get("title") or ""
+
+            artist = ""
+            artist_credit = rec.get("artist-credit") or rec.get("artist_credit")
+            if isinstance(artist_credit, list) and artist_credit:
+                first = artist_credit[0]
+                if isinstance(first, dict):
+                    artist = first.get("name") or first.get("artist", {}).get("name", "")
+                elif isinstance(first, str):
+                    artist = first
+
+            album = ""
+            release_list = rec.get("release-list") or rec.get("releases") or rec.get("release")
+            if isinstance(release_list, list) and release_list:
+                first_rel = release_list[0]
+                if isinstance(first_rel, dict):
+                    album = first_rel.get("title", "") or ""
+                    release_date = first_rel.get("date") or ""
+                else:
+                    album = str(first_rel)
+                    release_date = ""
+            else:
+                release_date = rec.get("first-release-date") or ""
+
+            year = str(release_date)[:4] if release_date else ""
+            mbid = rec.get("id") or ""
+
+            items.append({
+                "title": title,
+                "artist": artist,
+                "album": album,
+                "year": year,
+                "provider": self.name,
+                "mbid": mbid,
+                "raw": rec,
+            })
+
+        return items
+
+    def to_tags(self, item: Dict[str, Any]) -> List[str]:
+        tags = super().to_tags(item)
+        mbid = item.get("mbid")
+        if mbid:
+            tags.append(f"musicbrainz:{mbid}")
+        return tags
+
+
 # Registry ---------------------------------------------------------------
 
 _METADATA_PROVIDERS: Dict[str, Type[MetadataProvider]] = {
@@ -273,6 +358,7 @@ _METADATA_PROVIDERS: Dict[str, Type[MetadataProvider]] = {
     "openlibrary": OpenLibraryMetadataProvider,
     "googlebooks": GoogleBooksMetadataProvider,
     "google": GoogleBooksMetadataProvider,
+    "musicbrainz": MusicBrainzMetadataProvider,
 }