FabledCurator/tests/test_importer_attach_in_place.py

"""attach_in_place — Importer seam used by FC-3c (downloader)."""

from pathlib import Path

import pytest
from sqlalchemy import select

from backend.app.models import Artist, ImageRecord, ImportSettings
from backend.app.services.importer import Importer
from backend.app.services.thumbnailer import Thumbnailer

pytestmark = pytest.mark.integration


def _make_jpg(path: Path, color=(255, 0, 0), size=(64, 64)) -> bytes:
    from PIL import Image
    path.parent.mkdir(parents=True, exist_ok=True)
    Image.new("RGB", size, color).save(path, "JPEG")
    return path.read_bytes()


@pytest.fixture
def importer(db_sync, tmp_path):
    settings = db_sync.execute(
        select(ImportSettings).where(ImportSettings.id == 1)
    ).scalar_one()
    images_root = tmp_path / "images"
    images_root.mkdir()
    return Importer(
        session=db_sync,
        images_root=images_root,
        import_root=tmp_path / "ignored",
        thumbnailer=Thumbnailer(images_root=images_root),
        settings=settings,
    )


def test_attach_in_place_imports_happy(importer, db_sync):
    images_root = importer.images_root
    artist = Artist(name="Alice", slug="alice")
    db_sync.add(artist)
    db_sync.flush()

    target = images_root / "alice" / "patreon" / "post1" / "img.jpg"
    _make_jpg(target)

    result = importer.attach_in_place(target, artist=artist)
    assert result.status == "imported"
    assert result.image_id is not None

    rec_path = db_sync.execute(
        select(ImageRecord.path, ImageRecord.origin, ImageRecord.artist_id)
        .where(ImageRecord.id == result.image_id)
    ).one()
    assert rec_path.path == str(target)
    assert rec_path.origin == "imported_download"
    assert rec_path.artist_id == artist.id


def test_attach_in_place_with_sidecar(importer, db_sync):
    images_root = importer.images_root
    artist = Artist(name="Bob", slug="bob")
    db_sync.add(artist)
    db_sync.flush()

    target = images_root / "bob" / "patreon" / "post1" / "img.jpg"
    _make_jpg(target)
    sidecar = target.with_suffix(target.suffix + ".json")
    sidecar.write_text(
        '{"category": "patreon", "id": "12345", "url": "https://patreon.com/post/12345", '
        '"title": "Test Post", "published_at": "2026-05-01T00:00:00+00:00"}'
    )

    result = importer.attach_in_place(target, artist=artist)
    assert result.status == "imported"

    primary_post_id = db_sync.execute(
        select(ImageRecord.primary_post_id).where(ImageRecord.id == result.image_id)
    ).scalar_one()
    assert primary_post_id is not None  # post created from sidecar


def test_attach_in_place_sha256_dedup_does_not_modify_disk(importer, db_sync):
    images_root = importer.images_root
    artist = Artist(name="Cara", slug="cara")
    db_sync.add(artist)
    db_sync.flush()

    first = images_root / "cara" / "p" / "a.jpg"
    bytes_ = _make_jpg(first)
    importer.attach_in_place(first, artist=artist)

    second = images_root / "cara" / "p" / "b.jpg"
    second.parent.mkdir(parents=True, exist_ok=True)
    second.write_bytes(bytes_)
    assert second.exists()

    result = importer.attach_in_place(second, artist=artist)
    assert result.status == "skipped"
    assert result.skip_reason.value == "duplicate_hash"
    # Caller is responsible for deleting the on-disk file; we don't touch it.
    assert second.exists()


def test_attach_in_place_supersede_keeps_file_at_new_path(importer, db_sync):
    """phash near-dup where the new file is LARGER → existing record
    swaps to the new path; old file deleted."""
    images_root = importer.images_root
    artist = Artist(name="Dee", slug="dee")
    db_sync.add(artist)
    db_sync.flush()

    smaller = images_root / "dee" / "p" / "small.jpg"
    _make_jpg(smaller, color=(50, 50, 200), size=(64, 64))
    first = importer.attach_in_place(smaller, artist=artist)
    assert first.status == "imported"

    bigger = images_root / "dee" / "p" / "big.jpg"
    _make_jpg(bigger, color=(50, 50, 200), size=(256, 256))

    result = importer.attach_in_place(bigger, artist=artist)
    assert result.status == "superseded"

    new_path = db_sync.execute(
        select(ImageRecord.path).where(ImageRecord.id == first.image_id)
    ).scalar_one()
    assert new_path == str(bigger)
    assert not smaller.exists()
    assert bigger.exists()


def test_attach_in_place_phash_smaller_skips_without_modifying_disk(
    importer, db_sync
):
    images_root = importer.images_root
    artist = Artist(name="Eve", slug="eve")
    db_sync.add(artist)
    db_sync.flush()

    bigger = images_root / "eve" / "p" / "big.jpg"
    _make_jpg(bigger, color=(0, 200, 0), size=(256, 256))
    importer.attach_in_place(bigger, artist=artist)

    smaller = images_root / "eve" / "p" / "small.jpg"
    _make_jpg(smaller, color=(0, 200, 0), size=(64, 64))

    result = importer.attach_in_place(smaller, artist=artist)
    assert result.status == "skipped"
    assert result.skip_reason.value == "duplicate_phash"
    assert smaller.exists()


def test_attach_in_place_invalid_image_returns_skipped(importer):
    images_root = importer.images_root
    bad = images_root / "fred" / "bad.jpg"
    bad.parent.mkdir(parents=True, exist_ok=True)
    bad.write_bytes(b"not a jpeg, just some bytes")

    result = importer.attach_in_place(bad)
    assert result.status == "skipped"
    assert result.skip_reason.value == "invalid_image"