def test_exract_features(self):
"""
Assert exract_features returns a list
of 3 non-zero floats
"""
subimage = cv2.imread("test/assets/test_subimage_1.jpg")
clsfn = QRCodeClassification(self.test_config)
features = clsfn.exract_features(subimage)
self.assertEqual(len(features), 3)
for feat in features:
self.assertGreater(feat, 0.0)
self.assertIsInstance(feat, float)
def test_split_frames(self):
"""
assert returns correct number of subimages
"""
image = cv2.imread("test/assets/test_2592x1944_image.jpg")
height, width, ____ = image.shape
resolution = (height, width)
clsfn = QRCodeClassification(self.test_config)
sub_images = clsfn.split_frames(image)
num_images = (height / self.config.sub_image_dim[0]) * (
width / self.config.sub_image_dim[1])
# check correct number of
self.assertEquals(num_images, sub_images.shape[0])
def test_filter_contour_gt_min_lt_max_ht(self):
"""
positive case, inbetween the two bounds
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 4, and height 5
contour = np.asarray([[0, 1], [1, 4], [0, 0], [1, 0]])
clsfn = QRCodeClassification(self.test_config)
self.assertEquals(
clsfn.filter_contour(contour=contour,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len), contour)
def test_filter_contour_lt_min_height(self):
"""
Asserts that contour is less than min height
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 4, and height 1
contour = np.asarray([[0, 1], [1, 1], [0, 0], [1, 0]])
clsfn = QRCodeClassification(self.test_config)
self.assertEquals(
clsfn.filter_contour(contour=contour,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len), None)
def test_filter_countour_vertice_lt_min(self):
"""
negative case, vertices less than lower bound
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 1, and height 1
contour = np.asarray([[0, 1]])
clsfn = QRCodeClassification(self.test_config)
self.assertEquals(
clsfn.filter_contour(contour=contour,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len), None)
def test_filter_countour_gt_max_ht(self):
"""
negative case, exceeds upper bound
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 4, and height 100000
contour = np.asarray([[0, 1], [1, 100000], [0, 0], [1, 0]])
clsfn = QRCodeClassification(self.test_config)
self.assertEquals(
clsfn.filter_contour(contour=contour,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len), None)
def test_extract_contours(self):
"""
assert that correct datatype is returned
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 4, and height 5
contour = np.asarray([[0, 1], [1, 4], [0, 0], [1, 0]])
clsfn = QRCodeClassification(self.test_config)
self.assertEquals(
type(
clsfn.filter_contour(contour=contour,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len)), np.ndarray)
def test_get_contour_areas(self):
"""
correct areas are returned.
"""
max_height = 100
min_height = 2
max_len = 100
min_len = 2
# Square contour with length 4, and height 4
contour_1x4_len4 = np.asarray([[0, 1], [1, 4], [0, 0], [1, 0]])
contour_2x5_len4 = np.asarray([[0, 1], [2, 5], [0, 0], [2, 0]])
contour_1x1_len4 = np.asarray([[0, 1], [1, 1], [0, 0], [1, 0]])
test_countours = [contour_1x4_len4, contour_2x5_len4, contour_1x1_len4]
expected_areas = np.array([4.0, 10.0])
clsfn = QRCodeClassification(self.test_config)
# expect invalid areas to be filtered out
self.assertEquals(
clsfn.get_contour_areas(contours=test_countours,
min_height=min_height,
max_height=max_height,
max_len=max_len,
min_len=min_len).shape[0], expected_areas)
def test_invalid_dims(self):
"""
Set an invalid dim, and assert that invalid dimensions exception is thrown
"""
# both dims not divisible
self.test_config.sub_image_dim = (200, 200)
with self.assertRaises(InvalidSubImageDimError) as cm:
clsfn = QRCodeClassification(self.test_config)
clsfn.config = self.test_config
# first dim not divisible
self.test_config.sub_image_dim = (200, 216)
with self.assertRaises(InvalidSubImageDimError) as cm:
clsfn = QRCodeClassification(self.test_config)
clsfn.config = self.test_config
# second dim not divisible
self.test_config.sub_image_dim = (216, 200)
with self.assertRaises(InvalidSubImageDimError) as cm:
clsfn = QRCodeClassification(self.test_config)
clsfn.config = self.test_config