1568 Enhance tests to use self random state (Project-MONAI#1599)

* [DLMED] change tests to use self random state

Signed-off-by: Nic Ma <nma@nvidia.com>

* [MONAI] python code formatting

Signed-off-by: monai-bot <monai.miccai2019@gmail.com>

* [DLMED] fix CI tests

Signed-off-by: Nic Ma <nma@nvidia.com>

Co-authored-by: monai-bot <monai.miccai2019@gmail.com>

Author: Nic-Ma

tests/test_rand_rotate.py

@@ -52,7 +52,7 @@ def test_correct_results(self, degrees, keep_size, mode, padding_mode, align_cor
             self.imt[0, 0], -np.rad2deg(angle), (0, 1), not keep_size, order=_order, mode=_mode, prefilter=False
         )
         expected = np.stack(expected).astype(np.float32)
-        np.testing.assert_allclose(expected, rotated[0])
+        np.testing.assert_allclose(expected, rotated[0], rtol=1e-2, atol=1)
 
 
 class TestRandRotate3D(NumpyImageTestCase3D):

tests/test_rand_rotated.py

@@ -54,7 +54,7 @@ def test_correct_results(self, degrees, keep_size, mode, padding_mode, align_cor
             self.imt[0, 0], -np.rad2deg(angle), (0, 1), not keep_size, order=_order, mode=_mode, prefilter=False
         )
         expected = np.stack(expected).astype(np.float32)
-        self.assertTrue(np.allclose(expected, rotated["img"][0]))
+        self.assertTrue(np.allclose(expected, rotated["img"][0], rtol=1e-2, atol=1))
 
 
 class TestRandRotated3D(NumpyImageTestCase3D):

tests/utils.py

@@ -30,6 +30,7 @@
 
 from monai.config.deviceconfig import USE_COMPILED
 from monai.data import create_test_image_2d, create_test_image_3d
+from monai.transforms import Randomizable
 from monai.utils import ensure_tuple, optional_import, set_determinism
 from monai.utils.module import get_torch_version_tuple
 
@@ -434,14 +435,25 @@ def _call_original_func(name, module, *args, **kwargs):
     return f(*args, **kwargs)
 
 
-class NumpyImageTestCase2D(unittest.TestCase):
+class NumpyImageTestCase2D(unittest.TestCase, Randomizable):
     im_shape = (128, 64)
     input_channels = 1
     output_channels = 4
     num_classes = 3
 
+    def randomize(self, data=None):
+        return create_test_image_2d(
+            width=self.im_shape[0],
+            height=self.im_shape[1],
+            num_objs=4,
+            rad_max=20,
+            noise_max=0,
+            num_seg_classes=self.num_classes,
+            random_state=self.R,
+        )
+
     def setUp(self):
-        im, msk = create_test_image_2d(self.im_shape[0], self.im_shape[1], 4, 20, 0, self.num_classes)
+        im, msk = self.randomize()
 
         self.imt = im[None, None]
         self.seg1 = (msk[None, None] > 0).astype(np.float32)
@@ -456,14 +468,26 @@ def setUp(self):
         self.segn = torch.tensor(self.segn)
 
 
-class NumpyImageTestCase3D(unittest.TestCase):
+class NumpyImageTestCase3D(unittest.TestCase, Randomizable):
     im_shape = (64, 48, 80)
     input_channels = 1
     output_channels = 4
     num_classes = 3
 
+    def randomize(self, data=None):
+        return create_test_image_3d(
+            height=self.im_shape[0],
+            width=self.im_shape[1],
+            depth=self.im_shape[2],
+            num_objs=4,
+            rad_max=20,
+            noise_max=0,
+            num_seg_classes=self.num_classes,
+            random_state=self.R,
+        )
+
     def setUp(self):
-        im, msk = create_test_image_3d(self.im_shape[0], self.im_shape[1], self.im_shape[2], 4, 20, 0, self.num_classes)
+        im, msk = self.randomize()
 
         self.imt = im[None, None]
         self.seg1 = (msk[None, None] > 0).astype(np.float32)