fix(scenes): use ON CONFLICT for tag slug upsert in enrich_tags_from_tube

Replace SAVEPOINT + IntegrityError fallback in resolve_tag with
postgres INSERT ... ON CONFLICT (slug) DO NOTHING + re-SELECT.
Postgres serializes on the unique index, so concurrent inserts of
the same slug no longer race on lookup→insert and the second caller
no longer raises uq_tags_slug. Mirrors the on_conflict pattern
already used for SceneTag/MovieTag inserts.

app/resolve/tag_resolver.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 import uuid
 
 from sqlalchemy import select
-from sqlalchemy.exc import IntegrityError
+from sqlalchemy.dialects.postgresql import insert as pg_insert
 from sqlalchemy.orm import Session
 
 from app.models.tag import Tag
@@ -27,21 +27,17 @@ def resolve_tag(session: Session, *, norm: NormalizedTag) -> Tag | None:
     name = (norm.name or "").strip() or slug.replace("-", " ").title()
     if len(name) > 120:
         name = name[:120]
-    tag = Tag(name=name, slug=slug)
-    # SAVEPOINT — chroni outer transaction przed concurrent insert race:
-    # gdy worker scrapuje sceny + API endpoint enrich_tags_from_tube robią
-    # `resolve_tag('hardcore-sex')` jednocześnie, jeden INSERT się uda,
-    # drugi → UniqueViolation slug. Bez savepoint cała transakcja API
-    # rzucała 500 (Sentry GOON-H, 5x od 2026-05-12). Z savepoint rollback
-    # na savepoint + re-SELECT zwraca już-istniejący tag.
-    sp = session.begin_nested()
-    try:
-        session.add(tag)
-        session.flush()
-        sp.commit()
-        return tag
-    except IntegrityError:
-        sp.rollback()
+    # Concurrent insert race: worker scraper + API enrich_tags_from_tube oba
+    # robią lookup→insert na tym samym slug ('hardcore-sex'), oba widzą NULL,
+    # oba próbują INSERT, drugi pada UniqueViolation uq_tags_slug (Sentry GOON-H,
+    # 5 events od 2026-05-12). ON CONFLICT DO NOTHING + re-SELECT po slug
+    # załatwia atomowo — Postgres serializuje na unique index.
+    stmt = (
+        pg_insert(Tag.__table__)
+        .values(name=name, slug=slug)
+        .on_conflict_do_nothing(index_elements=["slug"])
+    )
+    session.execute(stmt)
     return session.execute(select(Tag).where(Tag.slug == slug)).scalar_one_or_none()