Testing Class Material

extra_slides/packaging/noiser_project_final/noiser/tests/test_noise.py

@@ -0,0 +1,24 @@
+import numpy as np
+from numpy.testing import assert_allclose
+
+from noiser.noise import white_noise
+
+
+def test_white_noise():
+    n_images, height, width = 201, 101, 102
+    dtype = np.float32
+
+    # Create ``n_images`` identical image.
+    base_image = np.random.rand(1, height, width, 3).astype(dtype) - 0.5
+    images = np.repeat(base_image, n_images, axis=0)
+
+    std = 0.13
+    noisy = white_noise(images, std=std)
+
+    # dtype and shape are preserved.
+    assert noisy.dtype == dtype
+    assert noisy.shape == images.shape
+
+    # Mean and std of noisy image match expectations.
+    assert_allclose(images.mean(0), base_image[0], atol=1e-4)
+    assert np.isclose((noisy - images).std(), std, atol=1e-4)

extra_slides/packaging/noiser_project_final/noiser/tests/test_utils.py

@@ -0,0 +1,14 @@
+import numpy as np
+from numpy.testing import assert_array_equal
+
+from noiser.utils import copy_image
+
+
+def test_copy_image():
+    height, width = 101, 102
+    dtype = np.float32
+
+    image = np.random.rand(height, width, 3).astype(dtype)
+    copy = copy_image(image)
+    assert_array_equal(copy, image)
+