def test_correctly_chooses_parallel():
# forcing 1 core should always return False
assert multiprocessing_necessary((100, 10, 10), cores=1) is False
# shapes less than 10 should return false
assert multiprocessing_necessary((10, 10, 10), cores=12) is False
assert multiprocessing_necessary(10, cores=12) is False
# shapes over 10 should return True
assert multiprocessing_necessary((11, 10, 10), cores=12) is True
assert multiprocessing_necessary(11, cores=12) is True
def filter_func(images: Images,
diff=None,
radius=_default_radius,
axis=0,
cores=None,
progress=None):
"""
:param images: Input data
:param diff: Pixel value difference above which to crop bright pixels
:param radius: Size of the median filter to apply
:param cores: The number of cores that will be used to process the data.
:return: The processed 3D numpy.ndarray
"""
if not utility.multiprocessing_necessary(images.data.shape, cores):
cores = 1
if diff and radius and diff > 0 and radius > 0:
data = images.projections if axis == 0 else images.sinograms
func = psm.create_partial(OutliersISISFilter._execute,
psm.return_fwd_func,
diff=diff,
radius=radius)
psm.execute(
data,
func,
cores,
progress=progress,
msg=f"Outliers with threshold {diff} and kernel {radius}")
return images
def filter_func(data,
diff=None,
radius=_default_radius,
mode=_default_mode,
axis=0,
type=_default_dim,
cores=None,
progress=None):
"""
Requires tomopy to be available.
:param data: Input data as a 3D numpy.ndarray
:param diff: Pixel value difference above which to crop bright pixels
:param radius: Size of the median filter to apply
:param mode: Spot brightness to remove.
Either 'bright' or 'dark'.
:param cores: The number of cores that will be used to process the data.
:return: The processed 3D numpy.ndarray
"""
progress = Progress.ensure_instance(progress,
task_name='Outliers',
num_steps=3)
if not utility.multiprocessing_necessary(data.data.shape, cores):
cores = 1
if diff and radius and diff > 0 and radius > 0:
with progress:
progress.update(
msg="Applying outliers with threshold: {0} and "
"radius {1}".format(diff, radius))
sample = data.data
# By default tomopy only clears bright outliers.
# As a workaround inverting the image makes the dark outliers the brightest
if mode == OUTLIERS_DARK:
np.negative(sample, out=sample)
tomopy = importer.do_importing('tomopy')
if type == "2D":
tomopy.misc.corr.remove_outlier(sample,
diff,
radius,
axis,
ncore=cores,
out=sample)
else:
tomopy.misc.corr.remove_outlier1d(sample,
diff,
radius,
axis,
ncore=cores,
out=sample)
progress.update()
# reverse the inversion
if mode == OUTLIERS_DARK:
np.negative(sample, out=sample)
return data
def test_correctly_chooses_parallel(shape: Union[int, List, Tuple[int, int,
int]],
cores: int, should_be_parallel: bool):
assert multiprocessing_necessary(shape, cores) is should_be_parallel