[scrobblers] Use MediaSourceId to avoid multiple Jellyfin scrobbles

vrobbler/apps/music/constants.py

@@ -90,6 +90,7 @@ JELLYFIN_POST_KEYS = {
     "TAGLINE": "TAGLINE",
     "PLAYBACK_POSITION_TICKS": "PlaybackPositionTicks",
     "PLAYBACK_POSITION": "PlaybackPosition",
+    "MEDIA_SOURCE_ID": "MediaSourceId",
     "ARTIST_MB_ID": "Provider_musicbrainzartist",
     "ALBUM_MB_ID": "Provider_musicbrainzalbum",
     "RELEASEGROUP_MB_ID": "Provider_musicbrainzreleasegroup",

vrobbler/apps/scrobbles/migrations/0071_add_scrobble_source_id.py

@@ -0,0 +1,18 @@
+# Generated manually for deduplication fix
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("scrobbles", "0070_rename_brickset_scrobble_brick_set"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="scrobble",
+            name="source_id",
+            field=models.CharField(max_length=255, null=True, blank=True),
+        ),
+    ]

vrobbler/apps/scrobbles/mixins.py

@@ -79,6 +79,7 @@ class ScrobblableMixin(TimeStampedModel):
         self,
         user_id,
         source: str = "Vrobbler",
+        source_id: Optional[str] = None,
         playback_position_seconds: int = 0,
         status: str = "started",
         log: Optional[dict] = None,
@@ -88,6 +89,7 @@ class ScrobblableMixin(TimeStampedModel):
             "user_id": user_id,
             "timestamp": timezone.now(),
             "source": source,
+            "source_id": source_id,
             "status": status,
             "playback_position_seconds": playback_position_seconds,
         }

vrobbler/apps/scrobbles/models.py

@@ -582,6 +582,7 @@ class Scrobble(TimeStampedModel):
 
     # Metadata
     source = models.CharField(max_length=255, **BNULL)
+    source_id = models.CharField(max_length=255, **BNULL)
     log = models.JSONField(
         **BNULL,
         default=dict,
@@ -1136,6 +1137,23 @@ class Scrobble(TimeStampedModel):
         skip_in_progress_check = kwargs.get("skip_in_progress_check", False)
         read_log_page = kwargs.get("read_log_page", None)
 
+        source_id = scrobble_data.get("source_id")
+
+        # If source_id is provided (e.g., MediaSourceId from Jellyfin), check for existing scrobble first
+        # This prevents duplicates when webhooks arrive faster than scrobble creation
+        if source_id:
+            existing_by_source_id = cls.objects.filter(
+                media_query,
+                user_id=user_id,
+                source_id=source_id,
+            ).first()
+            if existing_by_source_id:
+                logger.info(
+                    "[create_or_update] found existing scrobble by source_id, updating",
+                    extra={"scrobble_id": existing_by_source_id.id, "source_id": source_id},
+                )
+                scrobble_data["playback_status"] = scrobble_data.pop("status", None)
+                return existing_by_source_id.update(scrobble_data)
 
         # Find our last scrobble of this media item (track, video, etc)
         scrobble = (

vrobbler/apps/scrobbles/scrobblers.py

@@ -153,6 +153,7 @@ def jellyfin_scrobble_media(
     return media_obj.scrobble_for_user(
         user_id,
         source=post_data.get(JELLYFIN_POST_KEYS.get("SOURCE")),
+        source_id=post_data.get(JELLYFIN_POST_KEYS.get("MEDIA_SOURCE_ID")),
         playback_position_seconds=playback_position_seconds,
         status=playback_status,
     )