Skip to content

TEST: Only use stable argsorts in PARREC tests #1234

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Jun 27, 2023
Merged

TEST: Only use stable argsorts in PARREC tests #1234

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
57a3add
TEST: Remove potentially unstable argsort from parrec tests
effigies Jun 27, 2023
c385a53
FIX: Use stable argsort
effigies Jun 27, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 8 additions & 14 deletions nibabel/tests/test_parrec.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -192,15 +192,6 @@
]


def _shuffle(arr):
"""Return a copy of the array with entries shuffled.

Needed to avoid a bug in np.random.shuffle for numpy 1.7.
see: numpy/numpy#4286
"""
return arr[np.argsort(np.random.randn(len(arr)))]


def test_top_level_load():
# Test PARREC images can be loaded from nib.load
img = top_load(EG_PAR)
Expand Down Expand Up @@ -332,7 +323,7 @@ def test_sorting_dual_echo_T1():
t1_hdr = PARRECHeader.from_fileobj(fobj, strict_sort=True)

# should get the correct order even if we randomly shuffle the order
t1_hdr.image_defs = _shuffle(t1_hdr.image_defs)
np.random.shuffle(t1_hdr.image_defs)

sorted_indices = t1_hdr.get_sorted_slice_indices()
sorted_echos = t1_hdr.image_defs['echo number'][sorted_indices]
Expand Down Expand Up @@ -363,7 +354,7 @@ def test_sorting_multiple_echos_and_contrasts():
t1_hdr = PARRECHeader.from_fileobj(fobj, strict_sort=True)

# should get the correct order even if we randomly shuffle the order
t1_hdr.image_defs = _shuffle(t1_hdr.image_defs)
np.random.shuffle(t1_hdr.image_defs)

sorted_indices = t1_hdr.get_sorted_slice_indices()
sorted_slices = t1_hdr.image_defs['slice number'][sorted_indices]
Expand Down Expand Up @@ -402,7 +393,7 @@ def test_sorting_multiecho_ASL():
asl_hdr = PARRECHeader.from_fileobj(fobj, strict_sort=True)

# should get the correct order even if we randomly shuffle the order
asl_hdr.image_defs = _shuffle(asl_hdr.image_defs)
np.random.shuffle(asl_hdr.image_defs)

sorted_indices = asl_hdr.get_sorted_slice_indices()
sorted_slices = asl_hdr.image_defs['slice number'][sorted_indices]
Expand Down Expand Up @@ -524,7 +515,7 @@ def test_diffusion_parameters_strict_sort():
dti_hdr = PARRECHeader.from_fileobj(fobj, strict_sort=True)

# should get the correct order even if we randomly shuffle the order
dti_hdr.image_defs = _shuffle(dti_hdr.image_defs)
np.random.shuffle(dti_hdr.image_defs)

assert dti_hdr.get_data_shape() == (80, 80, 10, 8)
assert dti_hdr.general_info['diffusion'] == 1
Expand All @@ -533,7 +524,10 @@ def test_diffusion_parameters_strict_sort():
# DTI_PAR_BVECS gives bvecs copied from first slice each vol in DTI.PAR
# Permute to match bvec directions to acquisition directions
# note that bval sorting occurs prior to bvec sorting
assert_almost_equal(bvecs, DTI_PAR_BVECS[np.ix_(np.argsort(DTI_PAR_BVALS), [2, 0, 1])])
assert_almost_equal(
bvecs,
DTI_PAR_BVECS[np.ix_(np.argsort(DTI_PAR_BVALS, kind='stable'), [2, 0, 1])],
)
# Check q vectors
assert_almost_equal(dti_hdr.get_q_vectors(), bvals[:, None] * bvecs)

Expand Down
2 changes: 1 addition & 1 deletion nibabel/tests/test_scripts.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -418,7 +418,7 @@ def test_parrec2nii_with_data():
assert_almost_equal(np.loadtxt('DTI.bvals'), np.sort(DTI_PAR_BVALS))
img = load('DTI.nii')
data_sorted = img.get_fdata()
assert_almost_equal(data[..., np.argsort(DTI_PAR_BVALS)], data_sorted)
assert_almost_equal(data[..., np.argsort(DTI_PAR_BVALS, kind='stable')], data_sorted)
del img

# Writes .ordering.csv if requested
Expand Down