FabledCurator/tests/test_cleanup_service_audit.py

"""Tests for cleanup_service's library-audit additions.

Covers:
  - project_min_dimension_violations: SQL-only query against width/height
  - delete_min_dimension_violations: routes through existing delete_images
  - audit lifecycle (start_audit_run / apply_audit_run / cancel_audit_run)

Tests assert via column selects per reference-async-coredml-test-assertions
(post-DML ORM entity access via session.get() raises MissingGreenlet on
async sessions; we use db_sync here but the convention is preserved for
consistency).
"""

from pathlib import Path

import pytest
from PIL import Image
from sqlalchemy import func, select

from backend.app.models import ImageRecord, LibraryAuditRun
from backend.app.services import cleanup_service

pytestmark = pytest.mark.integration


def _make_image_record(db_sync, tmp_path, *, width, height, color):
    """Helper: write a real PIL file to a UNIQUE path (reference-image-record-path-unique)
    and insert an ImageRecord row with all required NOT-NULL columns set."""
    path = tmp_path / f"img_{width}x{height}_{color[0]}.png"
    Image.new("RGB", (width, height), color).save(path)
    # sha256 column is varchar(64) — use a deterministic 64-char pseudo
    # hash built from the unique inputs.
    sha = f"{width:04d}{height:04d}{color[0]:03d}".ljust(64, "0")[:64]
    rec = ImageRecord(
        path=str(path),
        sha256=sha,
        size_bytes=path.stat().st_size,
        mime="image/png",  # reference-image-record-required-columns
        width=width,
        height=height,
        origin="imported_filesystem",  # feedback-check-existing-enums
        integrity_status="ok",
    )
    db_sync.add(rec)
    db_sync.flush()
    return rec


def test_project_min_dimension_violations_returns_count_and_samples(db_sync, tmp_path):
    _make_image_record(db_sync, tmp_path, width=100, height=100, color=(10, 0, 0))
    _make_image_record(db_sync, tmp_path, width=50, height=50, color=(20, 0, 0))
    _make_image_record(db_sync, tmp_path, width=400, height=400, color=(30, 0, 0))
    db_sync.commit()

    result = cleanup_service.project_min_dimension_violations(
        db_sync, min_width=200, min_height=200,
    )
    assert result["count"] == 2  # 100x100 and 50x50 violate
    assert len(result["sample_ids"]) == 2


def test_delete_min_dimension_violations_unlinks_and_cascades(db_sync, tmp_path):
    rec_small = _make_image_record(db_sync, tmp_path, width=50, height=50, color=(40, 0, 0))
    rec_big = _make_image_record(db_sync, tmp_path, width=500, height=500, color=(50, 0, 0))
    small_path = Path(rec_small.path)
    db_sync.commit()

    # images_root is tmp_path here because the test fixtures stored files there;
    # delete_images() uses it to unlink originals + thumbs from the right tree.
    deleted = cleanup_service.delete_min_dimension_violations(
        db_sync, min_width=200, min_height=200, images_root=tmp_path,
    )
    assert deleted == 1
    # Verify via column selects per banked rule.
    remaining_ids = db_sync.execute(
        select(ImageRecord.id).order_by(ImageRecord.id)
    ).scalars().all()
    assert remaining_ids == [rec_big.id]
    assert not small_path.exists()  # file unlinked


# --- Audit lifecycle tests (Task 5) ---

import backend.app.tasks.library_audit  # noqa: F401, E402 — celery registration


def test_start_audit_run_creates_row_and_dispatches(db_sync, monkeypatch):
    dispatched = []
    from backend.app.tasks import library_audit as la_mod
    monkeypatch.setattr(
        la_mod.scan_library_for_rule, "delay",
        lambda audit_id: dispatched.append(audit_id),
    )
    audit_id = cleanup_service.start_audit_run(
        db_sync, rule="transparency", params={"threshold": 0.9},
    )
    db_sync.commit()
    row_status = db_sync.execute(
        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit_id)
    ).scalar_one()
    assert row_status == "running"
    assert dispatched == [audit_id]


def test_start_audit_run_rejects_when_another_is_running(db_sync, monkeypatch):
    from backend.app.tasks import library_audit as la_mod
    monkeypatch.setattr(
        la_mod.scan_library_for_rule, "delay", lambda audit_id: None,
    )
    cleanup_service.start_audit_run(
        db_sync, rule="transparency", params={"threshold": 0.9},
    )
    db_sync.commit()
    with pytest.raises(cleanup_service.AuditAlreadyRunning):
        cleanup_service.start_audit_run(
            db_sync, rule="single_color",
            params={"threshold": 0.95, "tolerance": 30},
        )


def test_apply_audit_run_with_correct_token_deletes_matched(db_sync, tmp_path):
    rec = _make_image_record(
        db_sync, tmp_path, width=100, height=100, color=(60, 0, 0),
    )
    db_sync.flush()
    audit = LibraryAuditRun(
        rule="transparency", params={"threshold": 0.9},
        status="ready", scanned_count=1, matched_count=1,
        matched_ids=[rec.id],
    )
    db_sync.add(audit)
    db_sync.commit()

    deleted = cleanup_service.apply_audit_run(
        db_sync, audit_id=audit.id,
        confirm_token=f"delete-audit-{audit.id}",
        images_root=tmp_path,
    )
    assert deleted == 1
    remaining_count = db_sync.execute(
        select(func.count()).select_from(ImageRecord)
    ).scalar_one()
    assert remaining_count == 0
    new_status = db_sync.execute(
        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit.id)
    ).scalar_one()
    assert new_status == "applied"


def test_apply_audit_run_with_wrong_token_raises(db_sync, tmp_path):
    audit = LibraryAuditRun(
        rule="transparency", params={"threshold": 0.9},
        status="ready", matched_ids=[],
    )
    db_sync.add(audit)
    db_sync.commit()
    with pytest.raises(cleanup_service.ConfirmTokenMismatch):
        cleanup_service.apply_audit_run(
            db_sync, audit_id=audit.id,
            confirm_token="wrong-token", images_root=tmp_path,
        )


def test_cancel_audit_run_flips_status(db_sync):
    audit = LibraryAuditRun(
        rule="transparency", params={"threshold": 0.9},
        status="running", matched_ids=[],
    )
    db_sync.add(audit)
    db_sync.commit()
    cleanup_service.cancel_audit_run(db_sync, audit_id=audit.id)
    new_status = db_sync.execute(
        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit.id)
    ).scalar_one()
    assert new_status == "cancelled"