def test_generate_one_layer_of_perfectly_fit_domains(self):
actual = generate_domains(10, 15, 5, 1)
expected = [
Block(5, 5, (0, 0)),
Block(5, 5, (5, 0)),
Block(5, 5, (0, 5)),
Block(5, 5, (5, 5)),
Block(5, 5, (0, 10)),
Block(5, 5, (5, 10))
]
self.assertEqual(expected, actual)
def test_generate_three_layers_of_imperfectly_fit_domains(self):
actual = generate_domains(17, 11, 6, 3)
expected = [
Block(6, 6, (0, 0)),
Block(5, 6, (6, 0)),
Block(6, 6, (11, 0)),
Block(6, 5, (0, 6)),
Block(5, 5, (6, 6)),
Block(6, 5, (11, 6)),
Block(8, 11, (0, 0)),
Block(9, 11, (8, 0)),
Block(17, 11, (0, 0))
]
self.assertEqual(expected, actual)
def test_generate_one_layer_of_imperfectly_fit_domains(self):
actual = generate_domains(16, 27, 7, 1)
expected = [
Block(5, 7, (0, 0)),
Block(6, 7, (5, 0)),
Block(5, 7, (11, 0)),
Block(5, 7, (0, 7)),
Block(6, 7, (5, 7)),
Block(5, 7, (11, 7)),
Block(5, 6, (0, 14)),
Block(6, 6, (5, 14)),
Block(5, 6, (11, 14)),
Block(5, 7, (0, 20)),
Block(6, 7, (5, 20)),
Block(5, 7, (11, 20))
]
self.assertEqual(expected, actual)
def encode(image):
"""
:param image: np.ndarray
:param width: number
:param height: number
:param metadata: dict
:return models.EncodedImage.EncodedImage
"""
width = image.shape[1]
height = image.shape[0]
range_width = get_range_width(width, height)
domain_width = range_width * DOMAIN_TO_RANGE_SIZE_RATIO
ranges = generate_ranges(width, height, range_width)
domains = generate_domains(width, height, domain_width, DOMAINS_DEPTH)
print(f'Encoding {len(ranges)} ranges')
return EncodedImage(
[find_best_transformation(image, matrix_from_block(image, range_block), domains, index) for index, range_block in enumerate(ranges)],
width,
height
)
def test_generate_one_domain(self):
self.assertEqual([Block(1, 1, (0, 0))], generate_domains(1, 1, 1, 1))