[books] Add an openlibrary source

vrobbler/apps/charts/management/commands/cleanup_duplicate_chartrecords.py

@@ -0,0 +1,87 @@
+import logging
+
+from django.core.management.base import BaseCommand
+from django.db.models import Count, Max
+
+logger = logging.getLogger(__name__)
+
+MEDIA_FIELDS = [
+    "artist",
+    "album",
+    "track",
+    "tv_series",
+    "video",
+    "podcast",
+    "podcast_episode",
+    "board_game",
+    "trail",
+    "geo_location",
+    "food",
+    "book",
+]
+
+
+class Command(BaseCommand):
+    help = "Clean up duplicate ChartRecords"
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--dry-run",
+            action="store_true",
+            help="Show what would be deleted without actually deleting",
+        )
+
+    def handle(self, *args, **options):
+        from charts.models import ChartRecord
+
+        dry_run = options["dry_run"]
+
+        total_deleted = 0
+
+        for media_field in MEDIA_FIELDS:
+            field_id = f"{media_field}_id"
+
+            duplicates = (
+                ChartRecord.objects.filter(**{f"{field_id}__isnull": False})
+                .values("user_id", "year", "month", "week", "day", field_id)
+                .annotate(
+                    count=Count("id"),
+                    max_id=Max("id"),
+                )
+                .filter(count__gt=1)
+            )
+
+            for dup in duplicates:
+                max_id = dup["max_id"]
+                to_delete = list(
+                    ChartRecord.objects.filter(
+                        user_id=dup["user_id"],
+                        year=dup["year"],
+                        month=dup["month"],
+                        week=dup["week"],
+                        day=dup["day"],
+                        **{field_id: dup[field_id]},
+                    )
+                    .exclude(id=max_id)
+                    .values_list("id", flat=True)
+                )
+
+                if dry_run:
+                    self.stdout.write(
+                        f"Would delete {len(to_delete)} duplicate ChartRecords "
+                        f"({media_field}, user={dup['user_id']}, year={dup['year']}, "
+                        f"month={dup['month']}, week={dup['week']}, day={dup['day']})"
+                    )
+                else:
+                    deleted, _ = ChartRecord.objects.filter(id__in=to_delete).delete()
+                    total_deleted += deleted
+                    self.stdout.write(
+                        f"Deleted {deleted} duplicate ChartRecords "
+                        f"({media_field}, user={dup['user_id']}, year={dup['year']}, "
+                        f"month={dup['month']}, week={dup['week']}, day={dup['day']})"
+                    )
+
+        if dry_run:
+            self.stdout.write(self.style.WARNING("Dry run complete - no changes made"))
+        else:
+            self.stdout.write(self.style.SUCCESS(f"Total deleted: {total_deleted}"))