Merge branch 'master' into meson

.codecov.yml

@@ -1,3 +1,6 @@
+github_checks:
+  annotations: false
+
 comment:
   layout: "reach, diff, files"
   behavior: default

dipy/align/streamlinear.py

@@ -1042,8 +1042,8 @@ def slr_with_qbx(static, moving,
     progressive : boolean, optional
             (default True)
 
-    rng : RandomState
-        If None creates RandomState in function.
+    rng : np.random.Generator
+        If None creates random generator in function.
 
     num_threads : int, optional
         Number of threads to be used for OpenMP parallelization. If None
@@ -1072,7 +1072,7 @@ def slr_with_qbx(static, moving,
 
     """
     if rng is None:
-        rng = np.random.RandomState()
+        rng = np.random.default_rng()
 
     if verbose:
         logger.info('Static streamlines size {}'.format(len(static)))
@@ -1207,8 +1207,8 @@ def groupwise_slr(bundles, x0='affine', tol=0, max_iter=20, qbx_thr=[4],
     verbose : bool, optional
         If True, logs information. Default: False.
 
-    rng : RandomState
-        If None, creates RandomState in function. Default: None.
+    rng : np.random.Generator
+        If None, creates random generator in function. Default: None.
 
     References
     ----------
@@ -1233,7 +1233,7 @@ def group_distance(bundles, n_bundle):
         return d
 
     if rng is None:
-        rng = np.random.RandomState()
+        rng = np.random.default_rng()
 
     metric = JointBundleMinDistanceMetric()
 

dipy/align/streamwarp.py

@@ -265,7 +265,9 @@ def bundlewarp_shape_analysis(moving_aligned, deformed_bundle, no_disks=10,
     indx = assignment_map(deformed_bundle, deformed_bundle, n)
     indx = np.array(indx)
 
-    colors = np.random.rand(n, 3)
+    rng = np.random.default_rng()
+
+    colors = rng.random((n, 3))
 
     disks_color = []
     for _, ind in enumerate(indx):

dipy/align/tests/test_api.py

@@ -23,6 +23,7 @@
 from dipy.io.stateful_tractogram import StatefulTractogram, Space
 from dipy.io.gradients import read_bvals_bvecs
 from dipy.io.image import load_nifti
+from dipy.testing.decorators import set_random_number_generator
 
 
 def setup_module():
@@ -246,11 +247,12 @@ def test_register_dwi_series_and_motion_correction():
         npt.assert_array_equal(reg_affines, reg_affines_2)
 
 
-def test_streamline_registration():
+@set_random_number_generator()
+def test_streamline_registration(rng):
     sl1 = [np.array([[0, 0, 0], [0, 0, 0.5], [0, 0, 1], [0, 0, 1.5]]),
            np.array([[0, 0, 0], [0, 0.5, 0.5], [0, 1, 1]])]
     affine_mat = np.eye(4)
-    affine_mat[:3, 3] = np.random.randn(3)
+    affine_mat[:3, 3] = rng.standard_normal(3)
     sl2 = list(transform_tracking_output(sl1, affine_mat))
     aligned, matrix = streamline_registration(sl2, sl1)
     npt.assert_almost_equal(matrix, np.linalg.inv(affine_mat))
@@ -259,7 +261,7 @@ def test_streamline_registration():
 
     # We assume the two tracks come from the same space, but it might have
     # some affine associated with it:
-    base_aff = np.eye(4) * np.random.rand()
+    base_aff = np.eye(4) * rng.random()
     base_aff[:3, 3] = np.array([1, 2, 3])
     base_aff[3, 3] = 1
 

dipy/align/tests/test_expectmax.py

@@ -309,7 +309,7 @@ def test_quantize_positive_2d(rng):
     # make sure additive noise doesn't change the quantization result
     noise_amplitude = np.min([delta / 4.0, min_positive / 4.0])
     sz = np.size(true_quantization)
-    noise = np.random.ranf(sz).reshape(img_shape) * noise_amplitude
+    noise = rng.random(sz).reshape(img_shape) * noise_amplitude
     noise = noise.astype(floating)
     input_image = np.ndarray(img_shape, dtype=floating)
     # assign intensities plus noise
@@ -372,7 +372,7 @@ def test_quantize_positive_3d(rng):
     # make sure additive noise doesn't change the quantization result
     noise_amplitude = np.min([delta / 4.0, min_positive / 4.0])
     sz = np.size(true_quantization)
-    noise = np.random.ranf(sz).reshape(img_shape) * noise_amplitude
+    noise = rng.random(sz).reshape(img_shape) * noise_amplitude
     noise = noise.astype(floating)
     input_image = np.ndarray(img_shape, dtype=floating)
     # assign intensities plus noise

dipy/align/tests/test_imaffine.py

@@ -175,7 +175,8 @@ def test_transform_origins_3d():
                     assert_array_almost_equal(actual.affine, expected)
 
 
-def test_affreg_all_transforms():
+@set_random_number_generator(202311)
+def test_affreg_all_transforms(rng):
     # Test affine registration using all transforms with typical settings
 
     # Make sure dictionary entries are processed in the same order regardless
@@ -198,7 +199,8 @@ def test_affreg_all_transforms():
                                                                       trans,
                                                                       factor,
                                                                       nslices,
-                                                                      1.0)
+                                                                      1.0,
+                                                                      rng=rng)
         # Sum of absolute differences
         start_sad = np.abs(static - moving).sum()
         metric = imaffine.MutualInformationMetric(32, sampling_pc)
@@ -243,7 +245,8 @@ def test_affreg_all_transforms():
                               np.zeros_like(smask), np.zeros_like(mmask))
 
 
-def test_affreg_defaults():
+@set_random_number_generator(202311)
+def test_affreg_defaults(rng):
     # Test all default arguments with an arbitrary transform
     # Select an arbitrary transform (all of them are already tested
     # in test_affreg_all_transforms)
@@ -260,7 +263,7 @@ def test_affreg_defaults():
         factor = factors[ttype][0]
         transform = regtransforms[ttype]
         static, moving, static_grid2world, moving_grid2world, smask, mmask, T = \
-            setup_random_transform(transform, factor, nslices, 1.0)
+            setup_random_transform(transform, factor, nslices, 1.0, rng=rng)
         # Sum of absolute differences
         start_sad = np.abs(static - moving).sum()
 
@@ -297,7 +300,7 @@ def test_affreg_defaults():
             end_sad = np.abs(moving - transformed_inv).sum()
             reduction = 1 - end_sad / start_sad
             print("%s>>%f" % (ttype, reduction))
-            assert(reduction > 0.9)
+            assert(reduction > 0.89)
 
 
 @set_random_number_generator(2022966)
@@ -326,7 +329,7 @@ def test_mi_gradient(rng):
             start.param_to_matrix(0.25 * rng.standard_normal(nrot))
         # Get data (pair of images related to each other by an known transform)
         static, moving, static_g2w, moving_g2w, smask, mmask, M = \
-            setup_random_transform(transform, factor, nslices, 2.0)
+            setup_random_transform(transform, factor, nslices, 2.0, rng=rng)
 
         # Prepare a MutualInformationMetric instance
         mi_metric = imaffine.MutualInformationMetric(32, sampling_proportion)
@@ -608,10 +611,11 @@ def test_affine_map(rng):
             assert_raises(AffineInversionError, AffineMap, mat_large_dim)
 
 
-def test_MIMetric_invalid_params():
+@set_random_number_generator()
+def test_MIMetric_invalid_params(rng):
     transform = regtransforms[('AFFINE', 3)]
-    static = np.random.rand(20, 20, 20)
-    moving = np.random.rand(20, 20, 20)
+    static = rng.random((20, 20, 20))
+    moving = rng.random((20, 20, 20))
     n = transform.get_number_of_parameters()
     sampling_proportion = 0.3
     theta_sing = np.zeros(n)

dipy/align/tests/test_parzenhist.py

@@ -31,7 +31,7 @@
            ('AFFINE', 3): 0.1}
 
 
-def create_random_image_pair(sh, nvals, seed):
+def create_random_image_pair(sh, nvals, rng=None):
     r""" Create a pair of images with an arbitrary, non-uniform joint PDF
 
     Parameters
@@ -42,6 +42,9 @@ def create_random_image_pair(sh, nvals, seed):
         maximum number of different values in the generated 2D images.
         The voxel intensities of the returned images will be in
         {0, 1, ..., nvals-1}
+    rng : numpy.random.generator
+        numpy's random generator. If None, it is set with a random seed.
+        Default is None
 
     Returns
     -------
@@ -50,7 +53,8 @@ def create_random_image_pair(sh, nvals, seed):
     moving : array, shape=sh
         second image in the image pair
     """
-    rng = np.random.default_rng(seed)
+    if rng is None:
+        rng = np.random.default_rng()
     sz = reduce(mul, sh, 1)
     sh = tuple(sh)
     static = rng.integers(0, nvals, sz).reshape(sh)
@@ -173,10 +177,10 @@ def test_parzen_joint_histogram():
                     assert_equal(index, P.padding + i)
 
 
-def test_parzen_densities():
+@set_random_number_generator(1246592)
+def test_parzen_densities(rng):
     # Test the computation of the joint intensity distribution
     # using a dense and a sparse set of values
-    seed = 1246592
     nbins = 32
     nr = 30
     nc = 35
@@ -186,10 +190,10 @@ def test_parzen_densities():
     for dim in [2, 3]:
         if dim == 2:
             shape = (nr, nc)
-            static, moving = create_random_image_pair(shape, nvals, seed)
+            static, moving = create_random_image_pair(shape, nvals, rng)
         else:
             shape = (ns, nr, nc)
-            static, moving = create_random_image_pair(shape, nvals, seed)
+            static, moving = create_random_image_pair(shape, nvals, rng)
 
         # Initialize
         parzen_hist = ParzenJointHistogram(nbins)
@@ -257,7 +261,6 @@ def test_parzen_densities():
                                   expected_smarginal_sparse)
 
 
-@set_random_number_generator(3147702)
 def setup_random_transform(transform, rfactor, nslices=45, sigma=1, rng=None):
     r""" Creates a pair of images related to each other by an affine transform
 
@@ -276,7 +279,14 @@ def setup_random_transform(transform, rfactor, nslices=45, sigma=1, rng=None):
         added to the identity parameters to create the random transform
     nslices: int
         number of slices to be stacked to form the volumes
+    sigma: float
+        standard deviation of the gaussian filter
+    rng: np.random.Generator
+        numpy's random generator. If None, it is set with a random seed.
+        Default is None
     """
+    if rng is None:
+        rng = np.random.default_rng()
     dim = 2 if nslices == 1 else 3
     if transform.get_dim() != dim:
         raise ValueError("Transform and requested volume have different dims.")
@@ -316,7 +326,8 @@ def setup_random_transform(transform, rfactor, nslices=45, sigma=1, rng=None):
     return static, moving, static_g2w, moving_g2w, smask, mmask, M
 
 
-def test_joint_pdf_gradients_dense():
+@set_random_number_generator(3147702)
+def test_joint_pdf_gradients_dense(rng):
     # Compare the analytical and numerical (finite differences) gradient of
     # the joint distribution (i.e. derivatives of each histogram cell) w.r.t.
     # the transform parameters. Since the histograms are discrete partitions
@@ -351,7 +362,8 @@ def test_joint_pdf_gradients_dense():
         theta = transform.get_identity_parameters()
 
         static, moving, static_g2w, moving_g2w, smask, mmask, M = \
-            setup_random_transform(transform, factor, nslices, 5.0)
+            setup_random_transform(transform, factor, nslices, 5.0,
+                                   rng=rng)
         parzen_hist = ParzenJointHistogram(32)
         parzen_hist.setup(static, moving, smask, mmask)
 
@@ -412,7 +424,8 @@ def test_joint_pdf_gradients_dense():
         assert(std_cosine < 0.25)
 
 
-def test_joint_pdf_gradients_sparse():
+@set_random_number_generator(3147702)
+def test_joint_pdf_gradients_sparse(rng):
     h = 1e-4
 
     # Make sure dictionary entries are processed in the same order regardless
@@ -435,14 +448,14 @@ def test_joint_pdf_gradients_sparse():
         theta = transform.get_identity_parameters()
 
         static, moving, static_g2w, moving_g2w, smask, mmask, M = \
-            setup_random_transform(transform, factor, nslices, 5.0)
+            setup_random_transform(transform, factor, nslices, 5.0,
+                                   rng=rng)
         parzen_hist = ParzenJointHistogram(32)
         parzen_hist.setup(static, moving, smask, mmask)
 
         # Sample the fixed-image domain
         k = 3
         sigma = 0.25
-        rng = np.random.default_rng(1234)
         shape = np.array(static.shape, dtype=np.int32)
         samples = sample_domain_regular(k, shape, static_g2w, sigma, rng)
         samples = np.array(samples)