def init_nodes(self):
"""Init the basic workflow and I/O nodes necessary before build."""
import nipype.interfaces.utility as nutil
import nipype.pipeline.engine as npe
if self.get_input_fields():
self._input_node = npe.Node(
name="Input",
interface=nutil.IdentityInterface(
fields=self.get_input_fields(), mandatory_inputs=False
),
)
else:
self._input_node = None
if self.get_output_fields():
self._output_node = npe.Node(
name="Output",
interface=nutil.IdentityInterface(
fields=self.get_output_fields(), mandatory_inputs=False
),
)
else:
self._output_node = None
Workflow.__init__(self, self._name, self.base_dir)
if self.input_node:
self.add_nodes([self.input_node])
if self.output_node:
self.add_nodes([self.output_node])
def __init__(self,
name='PermutationAnalysis',
session_dir='/export/data/mri/Neurometrics',
n=2000,
**kwargs):
Workflow.__init__(self, name=name, **kwargs)
self.permutation_workflows = [
Analysis1('p{0}'.format(i), session_dir, permute=True)
for i in range(n)
]
self.datasink = pe.Node(name='DataSink', interface=nio.DataSink())
self.datasink.inputs.base_directory = session_dir
self.datasink.inputs.container = 'ptest'
for pw in self.permutation_workflows:
self.connect(pw, 'output.frame_scores', self.datasink,
pw.name + '.frame')
self.connect(pw, 'output.block_scores', self.datasink,
pw.name + '.block')
self.connect(pw, 'output.halfrun_scores', self.datasink,
pw.name + '.halfrun')
self.connect(pw, 'output.run_scores', self.datasink,
pw.name + '.run')
def __init__(self, ct_file_name, tmp_dir, chest_regions=None):
Workflow.__init__(self, 'VesselParticlesWorkflow')
assert ct_file_name.rfind('.') != -1, "Unrecognized CT file name format"
self._tmp_dir = tmp_dir
self._cid = ct_file_name[max([ct_file_name.rfind('/'), 0])+1:\
ct_file_name.rfind('.')]
if ct_file_name.rfind('/') != -1:
self._dir = ct_file_name[0:ct_file_name.rfind('/')]
else:
self._dir = '.'
if vessel_seeds_mask_file_name is None:
self._vessel_seeds_mask_file_name = \
os.path.join(self._dir, self._cid + CM._vesselSeedsMask)
else:
self._vessel_seeds_mask_file_name = vessel_seeds_mask_file_name
generate_partial_lung_label_map = \
pe.Node(interface=cip.GeneratePartialLungLabelMap(),
name='generate_partial_lung_label_map')
generate_partial_lung_label_map.inputs.ct = ct_file_name
generate_partial_lung_label_map.inputs.
extract_chest_label_map = \
pe.Node(interface=cip.ExtractChestLabelMap(),
name='extract_chest_label_map')
extract_chest_label_map.inputs.outFileName =
extract_chest_label_map.inputs.
def __init__(self,name,input_fields=None,output_fields=None,**kwargs):
Workflow.__init__(self,name=name,**kwargs)
if input_fields:
self.input_node = pe.Node(name = 'input',
interface = util.IdentityInterface(fields=input_fields))
if output_fields:
self.output_node = pe.Node(name = 'output',
interface = util.IdentityInterface(fields=output_fields))
def __init__(self, name, input_fields=None, output_fields=None, **kwargs):
Workflow.__init__(self, name=name, **kwargs)
if input_fields:
self.input_node = pe.Node(
name='input',
interface=util.IdentityInterface(fields=input_fields))
if output_fields:
self.output_node = pe.Node(
name='output',
interface=util.IdentityInterface(fields=output_fields))
def __init__(self,name='PermutationAnalysis',session_dir='/export/data/mri/Neurometrics',n=2000,**kwargs):
Workflow.__init__(self,name=name,**kwargs)
self.permutation_workflows = [Analysis1('p{0}'.format(i),session_dir,permute=True) for i in range(n)]
self.datasink = pe.Node(name = 'DataSink',interface = nio.DataSink())
self.datasink.inputs.base_directory = session_dir
self.datasink.inputs.container = 'ptest'
for pw in self.permutation_workflows:
self.connect(pw,'output.frame_scores',self.datasink,pw.name+'.frame')
self.connect(pw,'output.block_scores',self.datasink,pw.name+'.block')
self.connect(pw,'output.halfrun_scores',self.datasink,pw.name+'.halfrun')
self.connect(pw,'output.run_scores',self.datasink,pw.name+'.run')
def __init__(self, bids_directory=None, caps_directory=None, tsv_file=None, name=None):
"""Inits a Pipeline object.
Args:
bids_directory (optional): Path to a BIDS directory.
caps_directory (optional): Path to a CAPS directory.
tsv_file (optional): Path to a subjects-sessions `.tsv` file.
name (optional): A pipelines name.
"""
import nipype.interfaces.utility as nutil
import nipype.pipeline.engine as npe
import inspect
import os
self._is_built = False
self._bids_directory = bids_directory
self._caps_directory = caps_directory
self._verbosity = 'debug'
self._tsv_file = tsv_file
self._info_file = os.path.join(
os.path.dirname(os.path.abspath(inspect.getfile(self.__class__))),
'info.json')
self._info = {}
if name:
self._name = name
else:
self._name = self.__class__.__name__
self._parameters = {}
if self._bids_directory is None:
if self._caps_directory is None:
raise IOError('%s does not contain BIDS nor CAPS directory' %
self._name)
self._sessions, self._subjects = get_subject_session_list(
input_dir=self._caps_directory,
ss_file=self._tsv_file,
is_bids_dir=False
)
else:
self._sessions, self._subjects = get_subject_session_list(
input_dir=self._bids_directory,
ss_file=self._tsv_file,
is_bids_dir=True
)
if self.get_input_fields():
self._input_node = npe.Node(name="Input",
interface=nutil.IdentityInterface(
fields=self.get_input_fields(),
mandatory_inputs=False))
else:
self._input_node = None
if self.get_output_fields():
self._output_node = npe.Node(name="Output",
interface=nutil.IdentityInterface(
fields=self.get_output_fields(),
mandatory_inputs=False))
else:
self._output_node = None
Workflow.__init__(self, self._name)
if self.input_node:
self.add_nodes([self.input_node])
if self.output_node:
self.add_nodes([self.output_node])
def __init__(self, ct_file_name, label_map_file_name,
tmp_dir, vessel_seeds_mask_file_name=None):
Workflow.__init__(self, 'VesselParticlesMaskWorkflow')
assert ct_file_name.rfind('.') != -1, "Unrecognized CT file name format"
self._tmp_dir = tmp_dir
self._cid = ct_file_name[max([ct_file_name.rfind('/'), 0])+1:\
ct_file_name.rfind('.')]
if ct_file_name.rfind('/') != -1:
self._dir = ct_file_name[0:ct_file_name.rfind('/')]
else:
self._dir = '.'
if vessel_seeds_mask_file_name is None:
self._vessel_seeds_mask_file_name = \
os.path.join(self._dir, self._cid + CM._vesselSeedsMask)
else:
self._vessel_seeds_mask_file_name = vessel_seeds_mask_file_name
# Params for feature strength computation
self._ct_file_name = ct_file_name
self._label_map_file_name = label_map_file_name
self._distance_map_file_name = \
os.path.join(self._tmp_dir, self._cid + '_distanceMap.nhdr')
self._feature_mask_file_name = \
os.path.join(self._tmp_dir, self._cid + '_featureMask.nhdr')
self._masked_strength_file_name = \
os.path.join(self._tmp_dir, self._cid + '_maskedStrength.nhdr')
self._equalized_strength_file_name = \
os.path.join(self._tmp_dir, self._cid + '_equalized.nhdr')
self._thresholded_equalized_file_name = \
os.path.join(self._tmp_dir, self._cid + '_thresholded.nhdr')
self._converted_thresholded_equalized_file_name = \
os.path.join(self._tmp_dir, self._cid + '_converted.nhdr')
self._strength_file_name = \
os.path.join(self._tmp_dir, self._cid + '_strength.nhdr')
self._scale_file_name = \
os.path.join(self._tmp_dir, self._cid + '_scale.nhdr')
self._thinned_file_name = \
os.path.join(self._tmp_dir, self._cid + '_thinned.nhdr')
self._sigma_step_method = 1
self._rescale = False
self._threads = 0
self._method = 'Frangi'
self._alpha = 0.63 # In [0, 1]
self._beta = 0.51 # In [0. 1]
self._C = 245 # In [0, 300]
self._alphase = 0.25 # In [0, 1]
self._nu = 0 # In [-1, 0.5]
self._kappa = 0.5 # In [0.01, 2]
self._betase = 0.1 # In [0.01, 2]
self._sigma = 1.0
self._sigma_min = 0.7
self._sigma_max = 4.0
self._num_steps = 7
self._gaussianStd = [self._sigma_min, self._sigma_max, self._num_steps]
# Params for histogram equalization (unu heq node)
self._bin = 10000
self._amount = 0.96
self._smart = 5
# Param for thresholding the histogram-equalized strength image
self._vesselness_th = 0.5
# Params for 'unu_2op_lt' node
self._distance_from_wall = -2.0
# Create distance map node. We want to isolate a region that is
# not too close to the lung periphery (particles can pick up noise in
# that region)
compute_distance_transform = \
pe.Node(interface=cip.ComputeDistanceMap(),
name='compute_distance_transform')
compute_distance_transform.inputs.labelMap = self._label_map_file_name
compute_distance_transform.inputs.distanceMap = \
self._distance_map_file_name
# Create node for thresholding the distance map
unu_2op_lt = pe.Node(interface=unu.unu_2op(), name='unu_2op_lt')
unu_2op_lt.inputs.operator = 'lt'
unu_2op_lt.inputs.type = 'short'
unu_2op_lt.inputs.in2_scalar = self._distance_from_wall
unu_2op_lt.inputs.output = self._feature_mask_file_name
# Create node for generating the vesselness feature strength image
compute_feature_strength = \
pe.Node(interface=cip.ComputeFeatureStrength(),
name='compute_feature_strength')
compute_feature_strength.inputs.inFileName = self._ct_file_name
compute_feature_strength.inputs.outFileName = self._strength_file_name
compute_feature_strength.inputs.ssm = self._sigma_step_method
compute_feature_strength.inputs.rescale = self._rescale
compute_feature_strength.inputs.threads = self._threads
compute_feature_strength.inputs.method = self._method
compute_feature_strength.inputs.feature = 'RidgeLine'
compute_feature_strength.inputs.alpha = self._alpha
compute_feature_strength.inputs.beta = self._beta
compute_feature_strength.inputs.C = self._C
compute_feature_strength.inputs.alphase = self._alphase
compute_feature_strength.inputs.nu = self._nu
compute_feature_strength.inputs.kappa = self._kappa
compute_feature_strength.inputs.betase = self._betase
compute_feature_strength.inputs.std = self._gaussianStd
# Create node for isolating the strength values within the mask
unu_2op_x_iso = pe.Node(interface=unu.unu_2op(), name='unu_2op_x_iso')
unu_2op_x_iso.inputs.operator = 'x'
unu_2op_x_iso.inputs.type = 'float'
unu_2op_x_iso.inputs.output = self._masked_strength_file_name
# Create node for histogram equalization
unu_heq = pe.Node(interface=unu.unu_heq(), name='unu_heq')
unu_heq.inputs.bin = self._bin
unu_heq.inputs.amount = self._amount
unu_heq.inputs.smart = self._smart
unu_heq.inputs.output = self._equalized_strength_file_name
# Set up thresholder to isolate all histogram-equalized voxels that
# are above a pre-defined threshold
unu_2op_gt = pe.Node(interface=unu.unu_2op(), name='unu_2op_gt')
unu_2op_gt.inputs.operator = 'gt'
unu_2op_gt.inputs.in2_scalar = self._vesselness_th = 0.5
unu_2op_gt.inputs.output = self._thresholded_equalized_file_name
# Now convert the equalized, thresholded image to short type
unu_convert = pe.Node(interface=unu.unu_convert(), name='unu_convert')
unu_convert.inputs.type = 'short'
unu_convert.inputs.output = \
self._converted_thresholded_equalized_file_name
# Thin the thresholded image
generate_binary_thinning_3d = \
pe.Node(interface=cip.GenerateBinaryThinning3D(),
name='generate_binary_thinning_3d')
generate_binary_thinning_3d.inputs.out = \
self._thinned_file_name
# Create node for isolating the strength values within the mask
unu_2op_x_ves = pe.Node(interface=unu.unu_2op(), name='unu_2op_x_ves')
unu_2op_x_ves.inputs.operator = 'x'
unu_2op_x_ves.inputs.output = self._vessel_seeds_mask_file_name
unu_2op_x_ves.inputs.in1_scalar = 768
# Set up the workflow connections
self.connect(compute_distance_transform, 'distanceMap',
unu_2op_lt, 'in1_file')
self.connect(unu_2op_lt, 'output', unu_2op_x_iso, 'in1_file')
self.connect(compute_feature_strength, 'outFileName',
unu_2op_x_iso, 'in2_file')
self.connect(unu_2op_x_iso, 'output', unu_heq, 'input')
self.connect(unu_heq, 'output', unu_2op_gt, 'in1_file')
self.connect(unu_2op_gt, 'output', unu_convert, 'input')
self.connect(unu_convert, 'output',
generate_binary_thinning_3d, 'opt_in')
self.connect(generate_binary_thinning_3d, 'out',
unu_2op_x_ves, 'in2_file')
self.config['execution'] = {'remove_unnecessary_outputs': 'False'}
