From bf0bfe59766ccf7501f0551422d8888f83c8f16f Mon Sep 17 00:00:00 2001
From: Antti Kettunen <antti-kettunen@hotmail.com>
Date: Tue, 5 May 2026 08:04:50 +0300
Subject: [PATCH] Fix Deezer and iTunes normalization edge cases

- derive Deezer track album types from album metadata instead of the raw track type
- inject Deezer artist data into artist-album payloads
- keep the iTunes explicit-duplicate guard from swapping in broken releases
---
 core/metadata/normalize.py        |  6 ++++--
 core/metadata/providers/deezer.py | 20 +++++++++++++++++++-
 core/metadata/providers/itunes.py | 17 +++++++++++++++++
 3 files changed, 40 insertions(+), 3 deletions(-)

diff --git a/core/metadata/normalize.py b/core/metadata/normalize.py
index 4cb670c5..222a1678 100644
--- a/core/metadata/normalize.py
+++ b/core/metadata/normalize.py
@@ -230,10 +230,12 @@ def normalize_deezer_track(raw: dict[str, Any]) -> MetadataTrack:
     contributors = raw.get("contributors") or []
     artists = _normalize_artists(contributors if len(_as_list(contributors)) > 1 else [artist] if artist else [])
     image_url = _first_non_empty(album.get("cover_xl"), album.get("cover_big"), album.get("cover_medium"))
-    album_type = raw.get("type") or album.get("type")
-    if not album_type:
+    album_type = str(album.get("type") or "").lower()
+    if not album_type or album_type == "track":
         nb_tracks = int(album.get("nb_tracks", 0) or 0)
         album_type = _infer_album_type_from_track_count(nb_tracks)
+    elif album_type == "compile":
+        album_type = "compilation"
     return MetadataTrack(
         id=str(raw.get("id", "")),
         name=str(raw.get("title", "")),
diff --git a/core/metadata/providers/deezer.py b/core/metadata/providers/deezer.py
index 4b953e29..3abb0e71 100644
--- a/core/metadata/providers/deezer.py
+++ b/core/metadata/providers/deezer.py
@@ -86,12 +86,30 @@ class DeezerMetadataAdapter(BaseMetadataAdapter):
         albums: list[dict[str, Any]] = []
         offset = 0
         page_size = 100
+        artist_data = self.get_artist_raw(artist_id) or {}
+        artist_name = str(artist_data.get("name", "") or "").strip()
+        artist_stub: dict[str, Any] = {"id": artist_data.get("id") or artist_id}
+        if artist_name:
+            artist_stub["name"] = artist_name
         while offset < limit:
             fetch_limit = min(page_size, limit - offset)
             data = self._request_api(f"artist/{artist_id}/albums", {"limit": fetch_limit, "index": offset})
             if not data or not data.get("data"):
                 break
-            albums.extend(list(data.get("data") or []))
+            for item in list(data.get("data") or []):
+                if not isinstance(item, dict):
+                    continue
+                enriched = dict(item)
+                album_artist = enriched.get("artist")
+                if isinstance(album_artist, dict):
+                    merged_artist = dict(artist_stub)
+                    merged_artist.update(album_artist)
+                    enriched["artist"] = merged_artist
+                else:
+                    enriched["artist"] = dict(artist_stub)
+                if artist_name and not enriched.get("artist_name"):
+                    enriched["artist_name"] = artist_name
+                albums.append(enriched)
             if len(data.get("data") or []) < fetch_limit:
                 break
             offset += len(data.get("data") or [])
diff --git a/core/metadata/providers/itunes.py b/core/metadata/providers/itunes.py
index cbb1501d..3a683334 100644
--- a/core/metadata/providers/itunes.py
+++ b/core/metadata/providers/itunes.py
@@ -157,6 +157,7 @@ class ITunesMetadataAdapter(BaseMetadataAdapter):
             return []
 
         seen: dict[str, dict[str, Any]] = {}
+        track_count_cache: dict[str, int] = {}
 
         def _normalize_album_name(name: str) -> str:
             normalized = (name or "").lower().strip()
@@ -170,12 +171,28 @@ class ITunesMetadataAdapter(BaseMetadataAdapter):
             normalized = re.sub(r"\s+", " ", normalized).strip()
             return normalized
 
+        def _album_has_tracks(collection_id: str) -> bool:
+            if not collection_id:
+                return False
+            if collection_id not in track_count_cache:
+                items = self._lookup_raw(id=collection_id, entity="song", limit=200)
+                track_count_cache[collection_id] = sum(
+                    1
+                    for item in items
+                    if item.get("wrapperType") == "track" and item.get("kind") == "song"
+                )
+            return track_count_cache[collection_id] > 0
+
         for album_data in results:
             if album_data.get("wrapperType") != "collection":
                 continue
             normalized_name = _normalize_album_name(str(album_data.get("collectionName", "") or ""))
+            collection_id = str(album_data.get("collectionId", "") or "")
             current = seen.get(normalized_name)
             is_explicit = album_data.get("collectionExplicitness") == "explicit"
+            has_tracks = not is_explicit or _album_has_tracks(collection_id)
+            if not has_tracks:
+                continue
             if current is None:
                 seen[normalized_name] = {"data": album_data, "is_explicit": is_explicit}
                 continue