test/test_annotator_box_collapse_guard.py

"""Regression: collapsed/empty annotator boxes must not wipe page redactions."""

from __future__ import annotations

import os

os.environ.setdefault("PYTHONUTF8", "1")

import pytest

pytest.importorskip("gradio_image_annotation_redaction")

from tools.redaction_review import (
    _annotator_box_has_area,
    _annotator_boxes_are_collapsed,
    update_all_page_annotation_object_based_on_previous_page,
)


def test_annotator_box_has_area():
    assert _annotator_box_has_area({"xmin": 0.1, "ymin": 0.2, "xmax": 0.3, "ymax": 0.4})
    assert not _annotator_box_has_area(
        {"xmin": 0.0, "ymin": 0.0, "xmax": 0.0, "ymax": 0.0}
    )
    assert not _annotator_box_has_area({"xmin": 10, "ymin": 20, "xmax": 10, "ymax": 25})


def test_annotator_boxes_are_collapsed_includes_empty():
    # Empty and all-zero payloads are both unsafe to write over existing boxes
    # (Gradio often returns [] briefly on page turns).
    assert _annotator_boxes_are_collapsed(None) is True
    assert _annotator_boxes_are_collapsed([]) is True
    assert (
        _annotator_boxes_are_collapsed([{"xmin": 0, "ymin": 0, "xmax": 0, "ymax": 0}])
        is True
    )
    assert (
        _annotator_boxes_are_collapsed(
            [
                {"xmin": 0, "ymin": 0, "xmax": 0, "ymax": 0},
                {"xmin": 0.1, "ymin": 0.2, "xmax": 0.3, "ymax": 0.4},
            ]
        )
        is False
    )


def test_update_all_page_keeps_existing_when_incoming_boxes_collapsed():
    existing = {
        "image": "page_0.png",
        "boxes": [
            {
                "xmin": 0.1,
                "ymin": 0.2,
                "xmax": 0.3,
                "ymax": 0.4,
                "label": "PERSON",
                "id": "abc",
            }
        ],
    }
    collapsed = {
        "image": "page_0.png",
        "boxes": [
            {
                "xmin": 0,
                "ymin": 0,
                "xmax": 0,
                "ymax": 0,
                "label": "PERSON",
                "id": "abc",
            }
        ],
        "orientation": 0,
    }
    page_sizes = [
        {
            "page": 1,
            "image_path": "page_0.png",
            "image_width": 100,
            "image_height": 200,
        }
    ]

    updated, current, bottom = update_all_page_annotation_object_based_on_previous_page(
        collapsed,
        current_page=1,
        previous_page=1,
        all_image_annotations=[existing],
        page_sizes=page_sizes,
    )

    assert current == 1
    assert bottom == 1
    assert len(updated[0]["boxes"]) == 1
    assert updated[0]["boxes"][0]["xmin"] == 0.1
    assert updated[0]["boxes"][0]["ymax"] == 0.4


def test_update_all_page_keeps_existing_when_incoming_boxes_empty():
    existing = {
        "image": "page_0.png",
        "boxes": [
            {
                "xmin": 0.1,
                "ymin": 0.2,
                "xmax": 0.3,
                "ymax": 0.4,
                "label": "PERSON",
                "id": "abc",
            }
        ],
    }
    cleared = {"image": "page_0.png", "boxes": [], "orientation": 0}
    page_sizes = [
        {
            "page": 1,
            "image_path": "page_0.png",
            "image_width": 100,
            "image_height": 200,
        },
        {
            "page": 2,
            "image_path": "page_1.png",
            "image_width": 100,
            "image_height": 200,
        },
    ]

    updated, _, _ = update_all_page_annotation_object_based_on_previous_page(
        cleared,
        current_page=2,
        previous_page=1,
        all_image_annotations=[existing, {"image": "page_1.png", "boxes": []}],
        page_sizes=page_sizes,
    )

    assert len(updated[0]["boxes"]) == 1
    assert updated[0]["boxes"][0]["id"] == "abc"


def test_update_all_page_previous_page_zero_does_not_overwrite_last_page():
    state = [
        {
            "image": "doc_0.png",
            "boxes": [
                {
                    "xmin": 0.1,
                    "ymin": 0.2,
                    "xmax": 0.3,
                    "ymax": 0.4,
                    "label": "PERSON",
                    "id": "a",
                }
            ],
        },
        {
            "image": "doc_1.png",
            "boxes": [
                {
                    "xmin": 0.2,
                    "ymin": 0.3,
                    "xmax": 0.4,
                    "ymax": 0.5,
                    "label": "PERSON",
                    "id": "c",
                }
            ],
        },
    ]
    annotator = {
        "image": "doc_0.png",
        "boxes": [
            {
                "xmin": 100,
                "ymin": 280,
                "xmax": 300,
                "ymax": 560,
                "label": "PERSON",
                "id": "a",
            }
        ],
        "orientation": 0,
    }
    page_sizes = [
        {
            "page": 1,
            "image_path": "doc_0.png",
            "image_width": 1000,
            "image_height": 1400,
        },
        {
            "page": 2,
            "image_path": "doc_1.png",
            "image_width": 1000,
            "image_height": 1400,
        },
    ]

    updated, _, _ = update_all_page_annotation_object_based_on_previous_page(
        annotator,
        current_page=2,
        previous_page=0,  # initial State used to be 0
        all_image_annotations=state,
        page_sizes=page_sizes,
    )

    # previous_page=0 must coerce to page 1, not index -1 (last page)
    assert updated[1]["boxes"][0]["id"] == "c"
    assert updated[0]["boxes"][0]["id"] == "a"
    assert updated[0]["boxes"][0]["xmin"] == pytest.approx(0.1)