test/test_module.py

import unittest
import src.SegmentAnything2Assist.SegmentAnything2Assist as SegmentAnything2Assist
import cv2

import numpy


class TestSegmentAnything2Assist(unittest.TestCase):
    def setUp(self) -> None:
        return super().setUp()

    def tearDown(self) -> None:
        return super().tearDown()

    def _loading_all_sam_model_types(self):
        # Test loading all types of SAM2 models.
        all_sam_models_type = [
            "sam2_hiera_tiny",
            "sam2_hiera_small",
            "sam2_hiera_base_plus",
            "sam2_hiera_large",
        ]
        for sam_model_type in all_sam_models_type:
            sam_model = SegmentAnything2Assist.SegmentAnything2Assist(
                sam_model_name=sam_model_type, download=True, device="cpu"
            )
            self.assertEqual(sam_model.is_model_available(), True)

            sam_model = SegmentAnything2Assist.SegmentAnything2Assist(
                sam_model_name=sam_model_type,
                download=False,
                model_path=f".tmp/checkpoints/{sam_model_type}.pth",
                device="cpu",
            )

            with self.assertRaises(Exception):
                sam_model = SegmentAnything2Assist.SegmentAnything2Assist(
                    sam_model_name=sam_model_type,
                    download=False,
                    model_path=".",
                    device="cpu",
                )

    def _generate_automatic_mask(self):
        image = cv2.imread("test/assets/liberty.jpg")

        sam_model = SegmentAnything2Assist.SegmentAnything2Assist(
            sam_model_name="sam2_hiera_tiny", download=True, device="cpu"
        )

        segmentation_masks, bboxes, predicted_iou, stability_score = (
            sam_model.generate_automatic_masks(image)
        )

        self.assertEqual(len(segmentation_masks.shape), 4)
        self.assertEqual(segmentation_masks[0].shape, image.shape)
        self.assertEqual(segmentation_masks.shape[3], 3)
        self.assertEqual(type(segmentation_masks[0][0][0][0]), numpy.uint8)
        self.assertEqual(len(bboxes.shape), 2)
        self.assertEqual(bboxes[0].shape, (4,))
        self.assertEqual(type(bboxes[0][0]), numpy.uint32)
        self.assertEqual(len(predicted_iou.shape), 1)
        self.assertEqual(type(predicted_iou[0]), numpy.float32)
        self.assertEqual(len(stability_score.shape), 1)
        self.assertEqual(type(stability_score[0]), numpy.float32)

        for segmentation_mask in segmentation_masks:
            self.assertEqual(segmentation_mask.shape, image.shape)

    def test_generate_masks_from_image(self):
        image = cv2.imread("test/assets/liberty.jpg")

        sam_model = SegmentAnything2Assist.SegmentAnything2Assist(
            sam_model_name="sam2_hiera_tiny", download=True, device="cpu"
        )

        mask_chw, mask_iou = sam_model.generate_masks_from_image(
            image, None, None, None
        )

        self.assertEqual(len(mask_chw.shape), 3)
        self.assertEqual(mask_chw[0].shape, image.shape)
        self.assertEqual(mask_chw.shape[0], 1)

        self.assertEqual(len(mask_iou.shape), 1)
        self.assertEqual(mask_iou.shape[0], 1)