test.unit.test_loss_label.test_cauchy_kernel1d() - Code Metrics - Inspection of "Separable convolutions for LNCC loss" - DeepRegNet/DeepReg - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — main (#605)

unknown

created 2021-01-26 11:23 UTC

test.unit.test_loss_label.test_cauchy_kernel1d() A

↳ Parent: test.unit.test_loss_label.TestMultiScaleLoss.test_get_config()

Complexity

Conditions

Size

Total Lines	7
Code Lines	7

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
eloc	7
dl	0
loc	7
rs	10
c	0
b	0
f	0
cc	1
nop	1

# coding=utf-8

"""
Tests for deepreg/model/loss/label.py in
pytest style
"""

from test.unit.util import is_equal_tf

import numpy as np
import pytest
import tensorflow as tf

import deepreg.loss.label as label


class TestMultiScaleLoss:

    def test_call(self):

        loss = label.MultiScaleLoss()
        with pytest.raises(NotImplementedError):
            loss.call(0, 0)

    def test_get_config(self):

        loss = label.MultiScaleLoss()
        got = loss.get_config()
        expected = dict(
            scales=None,
            kernel="gaussian",
            reduction=tf.keras.losses.Reduction.SUM,
            name="MultiScaleLoss",
        )
        assert got == expected


class TestDiceScore:

    shape = (3, 3, 3, 3)

    @pytest.fixture()
    def y_true(self):

        return np.ones(shape=self.shape) * 0.6

    @pytest.fixture()
    def y_pred(self):

        return np.ones(shape=self.shape) * 0.3

    @pytest.mark.parametrize(

        "binary,neg_weight,scales,expected",
        [
            (True, 0.0, None, 0.0),
            (False, 0.0, None, 0.4),
            (False, 0.2, None, 0.4 / 0.94),
            (False, 0.2, [0, 0], 0.4 / 0.94),
            (False, 0.2, [0, 1], 0.46030036),
        ],

    )
    def test_call(self, y_true, y_pred, binary, neg_weight, scales, expected):
        expected = np.array([expected] * self.shape[0])  # call returns (batch, )
        got = label.DiceScore(binary=binary, neg_weight=neg_weight, scales=scales).call(
            y_true=y_true, y_pred=y_pred
        )
        assert is_equal_tf(got, expected)
        got = label.DiceLoss(binary=binary, neg_weight=neg_weight, scales=scales).call(
            y_true=y_true, y_pred=y_pred
        )
        assert is_equal_tf(got, -expected)

    def test_get_config(self):

        got = label.DiceScore().get_config()
        expected = dict(
            binary=False,
            neg_weight=0.0,
            scales=None,
            kernel="gaussian",
            reduction=tf.keras.losses.Reduction.SUM,
            name="DiceScore",
        )
        assert got == expected


class TestCrossEntropy:

    shape = (3, 3, 3, 3)

    @pytest.fixture()
    def y_true(self):

        return np.ones(shape=self.shape) * 0.6

    @pytest.fixture()
    def y_pred(self):

        return np.ones(shape=self.shape) * 0.3

    @pytest.mark.parametrize(
        "binary,neg_weight,scales,expected",
        [
            (True, 0.0, None, -np.log(1.0e-7)),
            (False, 0.0, None, -0.6 * np.log(0.3)),
            (False, 0.2, None, -0.48 * np.log(0.3) - 0.08 * np.log(0.7)),
            (False, 0.2, [0, 0], -0.48 * np.log(0.3) - 0.08 * np.log(0.7)),
            (False, 0.2, [0, 1], 0.5239637),
        ],

    )
    def test_call(self, y_true, y_pred, binary, neg_weight, scales, expected):
        expected = np.array([expected] * self.shape[0])  # call returns (batch, )
        got = label.CrossEntropy(
            binary=binary, neg_weight=neg_weight, scales=scales
        ).call(y_true=y_true, y_pred=y_pred)
        assert is_equal_tf(got, expected)

    def test_get_config(self):

        got = label.CrossEntropy().get_config()
        expected = dict(
            binary=False,
            neg_weight=0.0,
            scales=None,
            kernel="gaussian",
            reduction=tf.keras.losses.Reduction.SUM,
            name="CrossEntropy",
        )
        assert got == expected


class TestJaccardIndex:

    shape = (3, 3, 3, 3)

    @pytest.fixture()
    def y_true(self):

        return np.ones(shape=self.shape) * 0.6

    @pytest.fixture()
    def y_pred(self):

        return np.ones(shape=self.shape) * 0.3

    @pytest.mark.parametrize(

        "binary,scales,expected",
        [
            (True, None, 0),
            (False, None, 0.25),
            (False, [0, 0], 0.25),
            (False, [0, 1], 0.17484076),
        ],

    )
    def test_call(self, y_true, y_pred, binary, scales, expected):
        expected = np.array([expected] * self.shape[0])  # call returns (batch, )
        got = label.JaccardIndex(binary=binary, scales=scales).call(
            y_true=y_true, y_pred=y_pred
        )
        assert is_equal_tf(got, expected)
        got = label.JaccardLoss(binary=binary, scales=scales).call(
            y_true=y_true, y_pred=y_pred
        )
        assert is_equal_tf(got, -expected)

    def test_get_config(self):

        got = label.JaccardIndex().get_config()
        expected = dict(
            binary=False,
            scales=None,
            kernel="gaussian",
            reduction=tf.keras.losses.Reduction.SUM,
            name="JaccardIndex",
        )
        assert got == expected


def test_foreground_prop_binary():
    """
    Test foreground function with a
    tensor of zeros with some ones, asserting
    equal to known precomputed tensor.
    Testing with binary case.
    """
    array_eye = np.identity(3, dtype=np.float32)
    tensor_eye = np.zeros((3, 3, 3, 3), dtype=np.float32)
    tensor_eye[:, :, 0:3, 0:3] = array_eye
    expect = tf.convert_to_tensor([1.0 / 3, 1.0 / 3, 1.0 / 3], dtype=tf.float32)
    get = label.foreground_proportion(tensor_eye)
    assert is_equal_tf(get, expect)


def test_foreground_prop_simple():
    """
    Test foreground functions with a tensor
    of zeros with some ones and some values below
    one to assert the thresholding works.
    """
    array_eye = np.identity(3, dtype=np.float32)
    tensor_eye = np.zeros((3, 3, 3, 3), dtype=np.float32)
    tensor_eye[:, 0, :, :] = 0.4 * array_eye  # 0
    tensor_eye[:, 1, :, :] = array_eye
    tensor_eye[:, 2, :, :] = array_eye
    tensor_eye = tf.convert_to_tensor(tensor_eye, dtype=tf.float32)
    expect = [54 / (27 * 9), 54 / (27 * 9), 54 / (27 * 9)]
    get = label.foreground_proportion(tensor_eye)
    assert is_equal_tf(get, expect)


def test_compute_centroid():
    """
    Testing compute centroid function
    and comparing to expected values.
    """
    tensor_mask = np.zeros((3, 2, 2, 2))
    tensor_mask[0, :, :, :] = np.ones((2, 2, 2))
    tensor_mask = tf.constant(tensor_mask, dtype=tf.float32)

    tensor_grid = np.ones((2, 2, 2, 3))
    tensor_grid[:, :, :, 1] *= 2
    tensor_grid[:, :, :, 2] *= 3
    tensor_grid = tf.constant(tensor_grid, dtype=tf.float32)

    expected = np.ones((3, 3))  # use 1 because 0/0 ~= (0+eps)/(0+eps) = 1
    expected[0, :] = [1, 2, 3]
    got = label.compute_centroid(tensor_mask, tensor_grid)
    assert is_equal_tf(got, expected)


def test_compute_centroid_d():
    """
    Testing compute centroid distance between equal
    tensors returns 0s.
    """
    array_ones = np.ones((2, 2))
    tensor_mask = np.zeros((3, 2, 2, 2))
    tensor_mask[0, :, :, :] = array_ones
    tensor_mask = tf.convert_to_tensor(tensor_mask, dtype=tf.float32)

    tensor_grid = np.zeros((2, 2, 2, 3))
    tensor_grid[:, :, :, 0] = array_ones
    tensor_grid = tf.convert_to_tensor(tensor_grid, dtype=tf.float32)

    get = label.compute_centroid_distance(tensor_mask, tensor_mask, tensor_grid)
    expect = np.zeros((3))
    assert is_equal_tf(get, expect)


1		# coding=utf-8
2
3		"""
4		Tests for deepreg/model/loss/label.py in
5		pytest style
6		"""
7
8		from test.unit.util import is_equal_tf
9
10		import numpy as np
11		import pytest
12		import tensorflow as tf
13
14		import deepreg.loss.label as label
15
16
17		class TestMultiScaleLoss:
		0 ignored issues – show introduced 2021-01-26 11:27 UTC by Report Bug Copy Issue Report Missing class docstring Loading history...
18		def test_call(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
19		loss = label.MultiScaleLoss()
20		with pytest.raises(NotImplementedError):
21		loss.call(0, 0)
22
23		def test_get_config(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
24		loss = label.MultiScaleLoss()
25		got = loss.get_config()
26		expected = dict(
27		scales=None,
28		kernel="gaussian",
29		reduction=tf.keras.losses.Reduction.SUM,
30		name="MultiScaleLoss",
31		)
32		assert got == expected
33
34
35		class TestDiceScore:
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing class docstring Loading history...
36		shape = (3, 3, 3, 3)
37
38		@pytest.fixture()
39		def y_true(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
40		return np.ones(shape=self.shape) * 0.6
41
42		@pytest.fixture()
43		def y_pred(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
44		return np.ones(shape=self.shape) * 0.3
45
46	View Code Duplication	@pytest.mark.parametrize(
		0 ignored issues – show Duplication introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report This code seems to be duplicated in your project. Loading history...
47		"binary,neg_weight,scales,expected",
48		[
49		(True, 0.0, None, 0.0),
50		(False, 0.0, None, 0.4),
51		(False, 0.2, None, 0.4 / 0.94),
52		(False, 0.2, [0, 0], 0.4 / 0.94),
53		(False, 0.2, [0, 1], 0.46030036),
54		],
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
55		)
56		def test_call(self, y_true, y_pred, binary, neg_weight, scales, expected):
57		expected = np.array([expected] * self.shape[0]) # call returns (batch, )
58		got = label.DiceScore(binary=binary, neg_weight=neg_weight, scales=scales).call(
59		y_true=y_true, y_pred=y_pred
60		)
61		assert is_equal_tf(got, expected)
62		got = label.DiceLoss(binary=binary, neg_weight=neg_weight, scales=scales).call(
63		y_true=y_true, y_pred=y_pred
64		)
65		assert is_equal_tf(got, -expected)
66
67		def test_get_config(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
68		got = label.DiceScore().get_config()
69		expected = dict(
70		binary=False,
71		neg_weight=0.0,
72		scales=None,
73		kernel="gaussian",
74		reduction=tf.keras.losses.Reduction.SUM,
75		name="DiceScore",
76		)
77		assert got == expected
78
79
80		class TestCrossEntropy:
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing class docstring Loading history...
81		shape = (3, 3, 3, 3)
82
83		@pytest.fixture()
84		def y_true(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
85		return np.ones(shape=self.shape) * 0.6
86
87		@pytest.fixture()
88		def y_pred(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
89		return np.ones(shape=self.shape) * 0.3
90
91		@pytest.mark.parametrize(
92		"binary,neg_weight,scales,expected",
93		[
94		(True, 0.0, None, -np.log(1.0e-7)),
95		(False, 0.0, None, -0.6 * np.log(0.3)),
96		(False, 0.2, None, -0.48 * np.log(0.3) - 0.08 * np.log(0.7)),
97		(False, 0.2, [0, 0], -0.48 * np.log(0.3) - 0.08 * np.log(0.7)),
98		(False, 0.2, [0, 1], 0.5239637),
99		],
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
100		)
101		def test_call(self, y_true, y_pred, binary, neg_weight, scales, expected):
102		expected = np.array([expected] * self.shape[0]) # call returns (batch, )
103		got = label.CrossEntropy(
104		binary=binary, neg_weight=neg_weight, scales=scales
105		).call(y_true=y_true, y_pred=y_pred)
106		assert is_equal_tf(got, expected)
107
108		def test_get_config(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
109		got = label.CrossEntropy().get_config()
110		expected = dict(
111		binary=False,
112		neg_weight=0.0,
113		scales=None,
114		kernel="gaussian",
115		reduction=tf.keras.losses.Reduction.SUM,
116		name="CrossEntropy",
117		)
118		assert got == expected
119
120
121		class TestJaccardIndex:
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing class docstring Loading history...
122		shape = (3, 3, 3, 3)
123
124		@pytest.fixture()
125		def y_true(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
126		return np.ones(shape=self.shape) * 0.6
127
128		@pytest.fixture()
129		def y_pred(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
130		return np.ones(shape=self.shape) * 0.3
131
132	View Code Duplication	@pytest.mark.parametrize(
		0 ignored issues – show Duplication introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report This code seems to be duplicated in your project. Loading history...
133		"binary,scales,expected",
134		[
135		(True, None, 0),
136		(False, None, 0.25),
137		(False, [0, 0], 0.25),
138		(False, [0, 1], 0.17484076),
139		],
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
140		)
141		def test_call(self, y_true, y_pred, binary, scales, expected):
142		expected = np.array([expected] * self.shape[0]) # call returns (batch, )
143		got = label.JaccardIndex(binary=binary, scales=scales).call(
144		y_true=y_true, y_pred=y_pred
145		)
146		assert is_equal_tf(got, expected)
147		got = label.JaccardLoss(binary=binary, scales=scales).call(
148		y_true=y_true, y_pred=y_pred
149		)
150		assert is_equal_tf(got, -expected)
151
152		def test_get_config(self):
		0 ignored issues – show introduced 2021-01-26 01:41 UTC by Report Bug Copy Issue Report Missing function or method docstring Loading history...
153		got = label.JaccardIndex().get_config()
154		expected = dict(
155		binary=False,
156		scales=None,
157		kernel="gaussian",
158		reduction=tf.keras.losses.Reduction.SUM,
159		name="JaccardIndex",
160		)
161		assert got == expected
162
163
164		def test_foreground_prop_binary():
165		"""
166		Test foreground function with a
167		tensor of zeros with some ones, asserting
168		equal to known precomputed tensor.
169		Testing with binary case.
170		"""
171		array_eye = np.identity(3, dtype=np.float32)
172		tensor_eye = np.zeros((3, 3, 3, 3), dtype=np.float32)
173		tensor_eye[:, :, 0:3, 0:3] = array_eye
174		expect = tf.convert_to_tensor([1.0 / 3, 1.0 / 3, 1.0 / 3], dtype=tf.float32)
175		get = label.foreground_proportion(tensor_eye)
176		assert is_equal_tf(get, expect)
177
178
179		def test_foreground_prop_simple():
180		"""
181		Test foreground functions with a tensor
182		of zeros with some ones and some values below
183		one to assert the thresholding works.
184		"""
185		array_eye = np.identity(3, dtype=np.float32)
186		tensor_eye = np.zeros((3, 3, 3, 3), dtype=np.float32)
187		tensor_eye[:, 0, :, :] = 0.4 * array_eye # 0
188		tensor_eye[:, 1, :, :] = array_eye
189		tensor_eye[:, 2, :, :] = array_eye
190		tensor_eye = tf.convert_to_tensor(tensor_eye, dtype=tf.float32)
191		expect = [54 / (27 * 9), 54 / (27 * 9), 54 / (27 * 9)]
192		get = label.foreground_proportion(tensor_eye)
193		assert is_equal_tf(get, expect)
194
195
196		def test_compute_centroid():
197		"""
198		Testing compute centroid function
199		and comparing to expected values.
200		"""
201		tensor_mask = np.zeros((3, 2, 2, 2))
202		tensor_mask[0, :, :, :] = np.ones((2, 2, 2))
203		tensor_mask = tf.constant(tensor_mask, dtype=tf.float32)
204
205		tensor_grid = np.ones((2, 2, 2, 3))
206		tensor_grid[:, :, :, 1] *= 2
207		tensor_grid[:, :, :, 2] *= 3
208		tensor_grid = tf.constant(tensor_grid, dtype=tf.float32)
209
210		expected = np.ones((3, 3)) # use 1 because 0/0 ~= (0+eps)/(0+eps) = 1
211		expected[0, :] = [1, 2, 3]
212		got = label.compute_centroid(tensor_mask, tensor_grid)
213		assert is_equal_tf(got, expected)
214
215
216		def test_compute_centroid_d():
217		"""
218		Testing compute centroid distance between equal
219		tensors returns 0s.
220		"""
221		array_ones = np.ones((2, 2))
222		tensor_mask = np.zeros((3, 2, 2, 2))
223		tensor_mask[0, :, :, :] = array_ones
224		tensor_mask = tf.convert_to_tensor(tensor_mask, dtype=tf.float32)
225
226		tensor_grid = np.zeros((2, 2, 2, 3))
227		tensor_grid[:, :, :, 0] = array_ones
228		tensor_grid = tf.convert_to_tensor(tensor_grid, dtype=tf.float32)
229
230		get = label.compute_centroid_distance(tensor_mask, tensor_mask, tensor_grid)
231		expect = np.zeros((3))
232		assert is_equal_tf(get, expect)
233

DeepRegNet / DeepReg

Pull Request — main (#605)

test.unit.test_loss_label.test_cauchy_kernel1d() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like