from pathlib import Path

from image_prediction.config import CONFIG
from image_prediction.pipeline import load_pipeline


def test_all_duplicate_images_are_filtered():
    """See RED-10765 (RM-241): Removed redactions reappear."""
    pdf_path = (
        Path(__file__).parents[1]
        / "data"
        / "RED-10765"
        / "RM-241-461c90d6d6dc0416ad5f0b05feef4dfc.UNTOUCHED_shortened.pdf"
    )
    pdf_bytes = pdf_path.read_bytes()

    pipeline = load_pipeline(verbose=True, batch_size=CONFIG.service.batch_size)
    predictions = list(pipeline(pdf_bytes))

    seen = set()
    for prediction in predictions:
        key = (
            prediction["position"]["x1"],
            prediction["position"]["x2"],
            prediction["position"]["y1"],
            prediction["position"]["y2"],
            prediction["position"]["pageNumber"],
        )
        assert key not in seen, f"Duplicate found: {key}"
        seen.add(key)

    all_passed = sum(1 for prediction in predictions if prediction["filters"]["allPassed"])
    assert all_passed == 1, f"Expected 1 image with allPassed flag set to True, but got {all_passed}"

    assert len(predictions) == 177, f"Expected 177 images, but got {len(predictions)}"