Aurora-Core/tests/unit/test_normalizer.py

"""Unit tests for the TMDb normalizer (app/services/metadata/normalizer.py)
— section 13. Uses hand-built payload fragments shaped like real TMDb
responses rather than calling the network."""

from __future__ import annotations

from datetime import date

from app.services.metadata.normalizer import normalize_movie, normalize_person

FAKE_MOVIE = {
    "id": 603,
    "title": "The Matrix",
    "original_title": "The Matrix",
    "overview": "A computer hacker learns...",
    "tagline": "Welcome to the Real World",
    "status": "Released",
    "original_language": "en",
    "release_date": "1999-03-30",
    "runtime": 136,
    "budget": 63000000,
    "revenue": 463517383,
    "popularity": 45.5,
    "vote_average": 8.2,
    "vote_count": 24000,
    "adult": False,
    "imdb_id": "tt0133093",
    "genres": [{"id": 28, "name": "Action"}, {"id": 878, "name": "Science Fiction"}],
    "production_companies": [
        {"id": 79, "name": "Village Roadshow Pictures", "origin_country": "US"}
    ],
    "poster_path": "/poster.jpg",
    "backdrop_path": "/backdrop.jpg",
    "belongs_to_collection": {"id": 2344, "name": "The Matrix Collection"},
    "credits": {
        "cast": [
            {
                "id": 6384,
                "name": "Keanu Reeves",
                "character": "Neo",
                "order": 0,
                "profile_path": "/keanu.jpg",
                "gender": 2,
            }
        ],
        "crew": [
            {
                "id": 9339,
                "name": "Lana Wachowski",
                "department": "Directing",
                "job": "Director",
                "gender": 1,
            }
        ],
    },
    "images": {
        "posters": [
            {"file_path": "/poster.jpg", "width": 2000, "height": 3000, "aspect_ratio": 0.667, "iso_639_1": "en", "vote_average": 8.1}
        ],
        "backdrops": [],
        "logos": [],
    },
    "videos": {
        "results": [
            {
                "id": "abc123",
                "type": "Trailer",
                "site": "YouTube",
                "key": "vKQi3bBA1y8",
                "iso_639_1": "en",
                "iso_3166_1": "US",
                "official": True,
                "published_at": "1999-03-01T00:00:00.000Z",
            }
        ]
    },
    "keywords": {"keywords": [{"id": 1, "name": "dystopia"}]},
    "alternative_titles": {"titles": [{"title": "Matrix", "iso_3166_1": "US"}]},
    "external_ids": {"imdb_id": "tt0133093"},
    "recommendations": {"results": [{"id": 604}]},
}

FAKE_PERSON = {
    "id": 6384,
    "name": "Keanu Reeves",
    "biography": "Keanu Charles Reeves is a Canadian actor.",
    "birthday": "1964-09-02",
    "deathday": None,
    "place_of_birth": "Beirut, Lebanon",
    "gender": 2,
    "known_for_department": "Acting",
    "popularity": 30.2,
    "also_known_as": ["Keanu Charles Reeves"],
    "profile_path": "/keanu.jpg",
    "images": {"profiles": [{"file_path": "/keanu.jpg", "width": 1000, "height": 1500, "aspect_ratio": 0.667, "iso_639_1": None, "vote_average": 5.0}]},
    "external_ids": {"imdb_id": "nm0000206"},
}


def test_normalize_movie_maps_core_fields():
    movie = normalize_movie(FAKE_MOVIE)
    assert movie.external_id == "603"
    assert movie.title == "The Matrix"
    assert movie.status == "released"
    assert movie.release_date == date(1999, 3, 30)
    assert movie.runtime_minutes == 136
    assert movie.budget == 63000000
    assert movie.adult is False


def test_normalize_movie_maps_genres_and_keywords():
    movie = normalize_movie(FAKE_MOVIE)
    assert movie.genres == ["Action", "Science Fiction"]
    assert movie.keywords == ["dystopia"]
    assert movie.alternate_titles == ["Matrix"]


def test_normalize_movie_maps_cast_and_crew():
    movie = normalize_movie(FAKE_MOVIE)
    assert len(movie.cast) == 1
    assert movie.cast[0].external_id == "6384"
    assert movie.cast[0].department == "Acting"
    assert movie.cast[0].character == "Neo"
    assert movie.cast[0].gender == "male"

    assert len(movie.crew) == 1
    assert movie.crew[0].department == "Directing"
    assert movie.crew[0].job == "Director"
    assert movie.crew[0].gender == "female"


def test_normalize_movie_maps_artwork_and_videos():
    movie = normalize_movie(FAKE_MOVIE)
    poster_types = [a.artwork_type for a in movie.artwork]
    assert poster_types.count("poster") >= 1
    assert all(a.url.startswith("https://image.tmdb.org/t/p/original/") for a in movie.artwork)

    assert len(movie.videos) == 1
    assert movie.videos[0].url == "https://www.youtube.com/watch?v=vKQi3bBA1y8"
    assert movie.videos[0].is_official is True


def test_normalize_movie_external_ids_never_include_tmdb_itself():
    # TMDb's own id is handled separately by the ingestion pipeline via
    # `normalized.external_id`, not folded into `external_ids` (section 3:
    # Aurora doesn't conflate its own identity with provider identity).
    movie = normalize_movie(FAKE_MOVIE)
    assert "tmdb" not in movie.external_ids
    assert movie.external_ids["imdb"] == "tt0133093"


def test_normalize_movie_maps_collection_and_recommendations():
    movie = normalize_movie(FAKE_MOVIE)
    assert movie.collection == ("2344", "The Matrix Collection")
    assert movie.recommendation_external_ids == ["604"]


def test_normalize_movie_handles_missing_optional_sections():
    minimal = {"id": 1, "title": "Minimal Movie"}
    movie = normalize_movie(minimal)
    assert movie.title == "Minimal Movie"
    assert movie.genres == []
    assert movie.cast == []
    assert movie.artwork == []
    assert movie.collection is None


def test_normalize_person_maps_core_fields():
    person = normalize_person(FAKE_PERSON)
    assert person.external_id == "6384"
    assert person.name == "Keanu Reeves"
    assert person.birthday == date(1964, 9, 2)
    assert person.deathday is None
    assert person.gender == "male"
    assert person.external_ids["imdb"] == "nm0000206"
    assert "Keanu Charles Reeves" in person.aliases


def test_normalize_person_profile_artwork_deduplicated():
    person = normalize_person(FAKE_PERSON)
    # profile_path duplicates an entry already in images.profiles; it must
    # not be inserted twice.
    paths = [a.provider_asset_id for a in person.artwork]
    assert paths.count("/keanu.jpg") == 1