Add GET /movies/{id}/recommendations

Ingestion already stored movie_recommendations edges from TMDb's
recommendations list (linking only to movies already ingested into
Aurora, per the never-fabricate-a-stub rule), but nothing exposed them.
Adds the repository lookup, service assembly, and route to read them
back in ranked order.

Co-Authored-By: Claude Sonnet 5 <[email protected]>

app/repositories/movie_repository.py

@@ -11,7 +11,7 @@ from sqlalchemy.orm import selectinload
 from app.models.company import Company, MovieCompany
 from app.models.genre import Genre, MovieGenre
 from app.models.keyword import Keyword, MovieKeyword
-from app.models.movie import Movie
+from app.models.movie import Movie, MovieRecommendation
 
 
 async def get_movie_by_id(session: AsyncSession, movie_id: uuid.UUID) -> Movie | None:
@@ -59,3 +59,20 @@ async def get_movie_companies(
     )
     result = await session.execute(stmt)
     return [(company, role) for company, role in result.all()]
+
+
+async def get_movie_recommendation_ids(
+    session: AsyncSession, movie_id: uuid.UUID
+) -> list[uuid.UUID]:
+    """Aurora ids of movies TMDb recommends alongside this one, ordered by
+    rank (section 15) -- only ever movies already ingested into Aurora, since
+    `ingest_movie` (movie_ingestion.py) links these edges purely by
+    external-id lookup and never creates a stub for an unseen movie."""
+
+    stmt = (
+        select(MovieRecommendation.recommended_movie_id)
+        .where(MovieRecommendation.movie_id == movie_id)
+        .order_by(MovieRecommendation.rank.asc().nulls_last())
+    )
+    result = await session.execute(stmt)
+    return list(result.scalars().all())