def test_get_template():
"""Test nimare.utils.get_template."""
# 1mm template
img = utils.get_template(space="mni152_1mm", mask=None)
assert isinstance(img, nib.Nifti1Image)
assert not nib.is_proxy(img.dataobj)
img = utils.get_template(space="mni152_1mm", mask="brain")
assert isinstance(img, nib.Nifti1Image)
# 2mm template (default)
img = utils.get_template(space="mni152_2mm", mask=None)
assert isinstance(img, nib.Nifti1Image)
img = utils.get_template(space="mni152_2mm", mask="brain")
assert isinstance(img, nib.Nifti1Image)
assert not nib.is_proxy(img.dataobj)
# ALE template
img = utils.get_template(space="ale_2mm", mask=None)
assert isinstance(img, nib.Nifti1Image)
img = utils.get_template(space="ale_2mm", mask="brain")
assert isinstance(img, nib.Nifti1Image)
assert not nib.is_proxy(img.dataobj)
# Expect exceptions when incompatible spaces or masks are requested.
with pytest.raises(ValueError):
utils.get_template(space="something", mask=None)
with pytest.raises(ValueError):
utils.get_template(space="mni152_1mm", mask="gm")
with pytest.raises(ValueError):
utils.get_template(space="mni152_2mm", mask="gm")
with pytest.raises(ValueError):
utils.get_template(space="ale_2mm", mask="gm")
def printInfo(self):
print("Printing information for {0}".format(self.filename))
obj = self.img.dataobj
isArray = nib.is_proxy(obj)
print("Proxy: {0}".format(str(isArray)))
data = self.img.get_data()
print("Shape:\n{0}".format(str(data.shape)))
print("Array:\n{0}".format(data[:, :, 115]))
def img_data_dtype(niimg):
"""Determine type of data contained in image
Based on the information contained in ``niimg.dataobj``, determine the
dtype of ``np.array(niimg.dataobj).dtype``.
"""
dataobj = niimg.dataobj
# Neuroimages that scale data should be interpreted as floating point
if nibabel.is_proxy(dataobj) and (dataobj.slope, dataobj.inter) != (1.0, 0.0):
return np.float_
# ArrayProxy gained the dtype attribute in nibabel 2.2
if hasattr(dataobj, 'dtype'):
return dataobj.dtype
return niimg.get_data_dtype()
def img_data_dtype(niimg):
"""Determine type of data contained in image
Based on the information contained in ``niimg.dataobj``, determine the
dtype of ``np.array(niimg.dataobj).dtype``.
"""
dataobj = niimg.dataobj
# Neuroimages that scale data should be interpreted as floating point
if nibabel.is_proxy(dataobj) and (dataobj.slope, dataobj.inter) != (1.0,
0.0):
return np.float_
# ArrayProxy gained the dtype attribute in nibabel 2.2
if hasattr(dataobj, 'dtype'):
return dataobj.dtype
return niimg.get_data_dtype()
def check_img(img):
"""
Very basic checker that loads `img`` and ensures it's 3D/int
Parameters
--------
img : str or niimg-like
Filepath to or in-memory loaded image
Returns
-------
img : niimg-like
Loaded 3D/int image
"""
if isinstance(img, (str, os.PathLike)) and os.path.exists(img):
img = nib.load(img)
elif not isinstance(img, nib.spatialimages.SpatialImage):
raise TypeError('Provided image must be an existing filepath or a '
'pre-loaded niimg-like object')
# ensure 3D or squeezable to 3D
img = nib.funcs.squeeze_image(img)
if len(img.shape) != 3:
raise ValueError('Provided image must be 3D')
# check if atlas data is int or castable to int
# if image is arrayproxy convert it to an array for speed-up
data = np.asarray(img.dataobj)
cast = nib.is_proxy(img.dataobj)
if img.header.get_data_dtype().kind not in ['i', 'u']:
idata = data.astype('int32')
cast = np.allclose(idata, data)
data = idata
if not cast:
raise ValueError('Provided image should have integer values or '
'be safely castable to int without data loss')
if cast:
img = img.__class__(data, img.affine, header=img.header)
img.header.set_data_dtype(np.int32)
return img
def test_dataset_smoke():
"""Smoke test for nimare.dataset.Dataset initialization and get methods."""
db_file = op.join(get_test_data_path(), "neurosynth_dset.json")
dset = dataset.Dataset(db_file)
dset.update_path(get_test_data_path())
assert isinstance(dset, nimare.dataset.Dataset)
# Test that Dataset.masker is portable
assert not nib.is_proxy(dset.masker.mask_img_.dataobj)
methods = [
dset.get_images, dset.get_labels, dset.get_metadata, dset.get_texts
]
for method in methods:
assert isinstance(method(), list)
assert isinstance(method(ids=dset.ids[:5]), list)
assert isinstance(method(ids=dset.ids[0]), list)
assert isinstance(dset.get_images(imtype="beta"), list)
assert isinstance(dset.get_metadata(field="sample_sizes"), list)
assert isinstance(dset.get_studies_by_label("cogat_cognitive_control"),
list)
assert isinstance(dset.get_studies_by_coordinate(np.array([[20, 20, 20]])),
list)
# If label is not available, raise ValueError
with pytest.raises(ValueError):
dset.get_studies_by_label("dog")
mask_data = np.zeros(dset.masker.mask_img.shape, int)
mask_data[40, 40, 40] = 1
mask_img = nib.Nifti1Image(mask_data, dset.masker.mask_img.affine)
assert isinstance(dset.get_studies_by_mask(mask_img), list)
dset1 = dset.slice(dset.ids[:5])
dset2 = dset.slice(dset.ids[5:])
assert isinstance(dset1, dataset.Dataset)
dset_merged = dset1.merge(dset2)
assert isinstance(dset_merged, dataset.Dataset)
img.affine img.header header = img.header print(header) print(header.get_data_shape()) # Use: get_data_dtype() to get the numpy data type in which the image data is stored (or will be stored if you save the image): print(header.get_data_dtype()) #Use: get_zooms() to get the voxel sizes in millimeters: print(header.get_zooms()) img.dataobj nib.is_proxy(img.dataobj) array_data = np.arange(24, dtype=np.int16).reshape((2, 3, 4)) affine = np.diag([1, 2, 3, 1]) array_img = nib.Nifti1Image(array_data, affine) array_img.dataobj array_img.dataobj is array_data nib.is_proxy(array_img.dataobj) image_data = array_img.get_data() image_data.shape image_data is array_data image_data = img.get_data() image_data.shape img.dataobj data_again = img.get_data()
cosa = np.cos(alpha)
sina = np.sin(alpha)
Trotate = np.array([[1, 0, 0, 0],
[0, cosa, -sina, 0],
[0, sina, cosa, 0],
[0, 0, 0, 1]])
Tsclrot = Tscale.dot(Trotate)
Ttranslate = np.array([[1, 0, 0, -78],
[0, 1, 0, -76],
[0, 0, 1, -64],
[0, 0, 0, 1]])
Twhole = Ttranslate.dot(Tsclrot)
# epi image proxy (i.e. points to epi image without loading it from disk yet):
nib.is_proxy(epi.dataobj)
# Creating a nifti image from an array using nibable:
mat = np.arange(24, dtype=np.int16).reshape((2, 3, 4))
affine = np.diag([1, 2, 3, 1])
newnifti = nib.Nifti1Image(mat, affine)
'''
Here the newnifti data is already a numpy array,
and there is no version of the array on disk.
The dataobj property of newnifti is the array itself
rather than a proxy for the array:
'''
print(newnifti.dataobj)
print(newnifti.dataobj is mat)
print(nib.is_proxy(newnifti.dataobj))
