[videos] More dedup scripts

vrobbler/apps/videos/management/commands/merge_duplicate_series.py

@@ -0,0 +1,84 @@
+from django.core.management.base import BaseCommand
+from django.db import models
+
+
+class Command(BaseCommand):
+    help = "Find TV Series with duplicate imdb_id values and merge videos to the newest series"
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--dry-run",
+            action="store_true",
+            help="Show what would be done without making changes",
+        )
+        parser.add_argument(
+            "--auto",
+            action="store_true",
+            help="Automatically fix duplicates without prompting",
+        )
+
+    def handle(self, *args, **options):
+        from videos.models import Series
+
+        dry_run = options["dry_run"]
+        auto = options["auto"]
+
+        duplicate_imdb_ids = (
+            Series.objects.filter(imdb_id__isnull=False)
+            .exclude(imdb_id="")
+            .values("imdb_id")
+            .annotate(count=models.Count("id"))
+            .filter(count__gt=1)
+        )
+
+        if not duplicate_imdb_ids.exists():
+            self.stdout.write("No duplicate imdb_id values found in Series")
+            return
+
+        self.stdout.write(
+            f"Found {duplicate_imdb_ids.count()} imdb_ids with duplicates in Series"
+        )
+
+        for dup in duplicate_imdb_ids:
+            imdb_id = dup["imdb_id"]
+            series_list = list(
+                Series.objects.filter(imdb_id=imdb_id).order_by("created")
+            )
+
+            self.stdout.write(f"\nimdb_id: {imdb_id}")
+            self.stdout.write(f"  Found {len(series_list)} series:")
+
+            for s in series_list:
+                video_count = s.video_set.count()
+                self.stdout.write(
+                    f"    id={s.id}, name='{s.name}', created={s.created}, videos={video_count}"
+                )
+
+            keeper = series_list[-1]
+            duplicates = series_list[:-1]
+
+            self.stdout.write(f"  Keeping: id={keeper.id} ('{keeper.name}')")
+            self.stdout.write(f"  Merging from: {[d.id for d in duplicates]}")
+
+            if dry_run:
+                self.stdout.write(self.style.WARNING("  [DRY RUN] Would move videos"))
+                continue
+
+            if not auto:
+                confirm = input("  Move videos from duplicates to keeper? [y/N] ")
+                if confirm.lower() != "y":
+                    self.stdout.write("  Skipped")
+                    continue
+
+            for dup_series in duplicates:
+                video_count = dup_series.video_set.count()
+                if video_count > 0:
+                    dup_series.video_set.update(tv_series=keeper)
+                    self.stdout.write(
+                        f"    Moved {video_count} videos from id={dup_series.id} to id={keeper.id}"
+                    )
+
+                dup_series.delete()
+                self.stdout.write(f"    Deleted series id={dup_series.id}")
+
+        self.stdout.write(self.style.SUCCESS("\nDone!"))

vrobbler/apps/videos/management/commands/merge_duplicate_series_by_name.py

@@ -0,0 +1,98 @@
+from django.core.management.base import BaseCommand
+from django.db import models
+
+
+class Command(BaseCommand):
+    help = "Find TV Series with duplicate names, merge Videos into the one with 'tt' IMDB id"
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--dry-run",
+            action="store_true",
+            help="Show what would be done without making changes",
+        )
+        parser.add_argument(
+            "--auto",
+            action="store_true",
+            help="Automatically fix duplicates without prompting",
+        )
+
+    def handle(self, *args, **options):
+        from videos.models import Series
+
+        dry_run = options["dry_run"]
+        auto = options["auto"]
+
+        duplicate_names = (
+            Series.objects.values("name")
+            .annotate(count=models.Count("id"))
+            .filter(count__gt=1)
+        )
+
+        if not duplicate_names.exists():
+            self.stdout.write("No duplicate Series names found")
+            return
+
+        self.stdout.write(f"Found {duplicate_names.count()} names with duplicates")
+
+        for dup in duplicate_names:
+            name = dup["name"]
+            series_list = list(Series.objects.filter(name=name).order_by("created"))
+
+            self.stdout.write(f"\nSeries name: {name}")
+            self.stdout.write(f"  Found {len(series_list)} series:")
+
+            for s in series_list:
+                video_count = s.video_set.count()
+                self.stdout.write(
+                    f"    id={s.id}, imdb_id='{s.imdb_id}', created={s.created}, videos={video_count}"
+                )
+
+            tt_series = None
+            for s in series_list:
+                if s.imdb_id and s.imdb_id.startswith("tt"):
+                    tt_series = s
+                    break
+
+            if not tt_series:
+                self.stdout.write(
+                    self.style.ERROR(
+                        f"  ERROR: No series with 'tt' IMDB id found, skipping"
+                    )
+                )
+                continue
+
+            duplicates = [s for s in series_list if s.id != tt_series.id]
+            self.stdout.write(
+                f"  Keeper (tt): id={tt_series.id} ('{tt_series.imdb_id}')"
+            )
+            self.stdout.write(f"  Merging from: {[d.id for d in duplicates]}")
+
+            if dry_run:
+                self.stdout.write(self.style.WARNING("  [DRY RUN] Would move videos"))
+                continue
+
+            if not auto:
+                confirm = input("  Move videos from duplicates to keeper? [y/N] ")
+                if confirm.lower() != "y":
+                    self.stdout.write("  Skipped")
+                    continue
+
+            total_moved = 0
+            for dup_series in duplicates:
+                video_count = dup_series.video_set.count()
+                if video_count > 0:
+                    dup_series.video_set.update(tv_series=tt_series)
+                    self.stdout.write(
+                        f"    Moved {video_count} videos from id={dup_series.id} to id={tt_series.id}"
+                    )
+                    total_moved += video_count
+
+                dup_series.delete()
+                self.stdout.write(f"    Deleted series id={dup_series.id}")
+
+            self.stdout.write(
+                self.style.SUCCESS(f"  Merged {total_moved} videos total")
+            )
+
+        self.stdout.write(self.style.SUCCESS("\nDone!"))