def workflow(self):
images = self.images
rois = self.rois
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
toextract = {**dict_sequences['MR-RT'], **dict_sequences['OT']}
workflow = nipype.Workflow('features_extraction_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
for key in toextract:
session = toextract[key]
if session['scans'] is not None:
scans = session['scans']
reg_scans = [x for x in scans if x.endswith('_reg')]
segmented_masks = [x for x in scans if x in ['GTVPredicted',
'TumorPredicted',
'GTVPredicted-2modalities']]
add_scans = [x for x in scans if x in images]
add_masks = [x for x in scans if x in rois]
for image in reg_scans:
for roi in segmented_masks:
image_name = '{}_{}_reg'.format(key, image.split('_')[0])
roi_name = '{}_{}'.format(key, roi.split('.nii.gz')[0])
features = nipype.Node(
interface=FeatureExtraction(),
name='features_extraction_{}{}'.format(image_name, roi_name))
features.inputs.parameter_file = '/home/fsforazz/git/core/resources/Params_MR.yaml'
workflow.connect(datasource, image_name, features, 'input_image')
workflow.connect(datasource, roi_name, features, 'rois')
workflow.connect(features, 'feature_files', datasink,
'results.subid.{0}.@csv_file_{1}{2}'.format(
key, image_name, roi_name))
for image in add_scans:
for roi in add_masks:
image_name = '{}_{}'.format(key, image)
roi_name = '{}_{}'.format(key, roi.split('.nii.gz')[0])
features = nipype.Node(
interface=FeatureExtraction(),
name='features_extraction_{}{}'.format(image_name, roi_name))
features.inputs.parameter_file = '/home/fsforazz/git/core/resources/Params_MR.yaml'
workflow.connect(datasource, image_name, features, 'input_image')
workflow.connect(datasource, roi_name, features, 'rois')
workflow.connect(features, 'feature_files', datasink,
'results.subid.{0}.@csv_file_{1}{2}'.format(
key, image_name, roi_name))
datasink.inputs.substitutions = substitutions
return workflow
def workflow(self):
# self.datasource()
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
tobet = {**dict_sequences['MR-RT'], **dict_sequences['OT']}
workflow = nipype.Workflow('brain_extraction_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
substitutions += [('_preproc_corrected.', '_preproc.')]
datasink.inputs.substitutions = substitutions
for key in tobet:
files = []
# if tobet[key]['ref'] is not None:
# files.append(tobet[key]['ref'])
if tobet[key]['scans'] is not None:
files = files + tobet[key]['scans']
for el in files:
el = el.strip(self.extention)
node_name = '{0}_{1}'.format(key, el)
bet = nipype.Node(interface=HDBet(),
name='{}_bet'.format(node_name),
serial=True)
bet.inputs.save_mask = 1
bet.inputs.out_file = '{}_preproc'.format(el)
reorient = nipype.Node(interface=Reorient2Std(),
name='{}_reorient'.format(node_name))
if el in TON4:
n4 = nipype.Node(interface=N4BiasFieldCorrection(),
name='{}_n4'.format(node_name))
workflow.connect(bet, 'out_file', n4, 'input_image')
workflow.connect(bet, 'out_mask', n4, 'mask_image')
workflow.connect(
n4, 'output_image', datasink,
'results.subid.{0}.@{1}_preproc'.format(key, el))
else:
workflow.connect(
bet, 'out_file', datasink,
'results.subid.{0}.@{1}_preproc'.format(key, el))
workflow.connect(
bet, 'out_mask', datasink,
'results.subid.{0}.@{1}_preproc_mask'.format(key, el))
workflow.connect(reorient, 'out_file', bet, 'input_file')
workflow.connect(datasource, node_name, reorient, 'in_file')
return workflow
def workflow(self):
self.datasource()
datasource = self.data_source
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
regex = self.regex
roi_selection = self.roi_selection
if datasource is not None:
workflow = nipype.Workflow('rtstruct_extraction_workflow', base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir), "datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
ss_convert = nipype.MapNode(interface=RTStructureCoverter(),
iterfield=['reference_ct', 'input_ss'],
name='ss_convert')
mha_convert = nipype.MapNode(interface=MHA2NIIConverter(),
iterfield=['input_folder'],
name='mha_convert')
if roi_selection:
select = nipype.MapNode(interface=CheckRTStructures(),
iterfield=['rois', 'dose_file'],
name='select_gtv')
workflow.connect(mha_convert, 'out_files', select, 'rois')
workflow.connect(datasource, 'rt_dose', select, 'dose_file')
workflow.connect(select, 'checked_roi', datasink,
'results.subid.@masks')
else:
workflow.connect(mha_convert, 'out_files', datasink,
'results.subid.@masks')
for i, session in enumerate(self.rt['session']):
substitutions += [(('_select_gtv{}/'.format(i), session+'/'))]
substitutions += [(('_voxelizer{}/'.format(i), session+'/'))]
substitutions += [(('_mha_convert{}/'.format(i), session+'/'))]
datasink.inputs.substitutions =substitutions
workflow.connect(datasource, 'rtct_nifti', ss_convert, 'reference_ct')
workflow.connect(datasource, 'rts_dcm', ss_convert, 'input_ss')
workflow.connect(ss_convert, 'out_structures', mha_convert, 'input_folder')
workflow = self.datasink(workflow, datasink)
else:
workflow = nipype.Workflow('rtstruct_extraction_workflow', base_dir=nipype_cache)
return workflow
def brain_extraction(sub_id,
datasource,
sessions,
RESULT_DIR,
NIPYPE_CACHE,
reference,
t10=True):
bet = nipype.MapNode(interface=HDBet(),
iterfield=['input_file'],
name='bet')
bet.inputs.save_mask = 1
bet.inputs.out_file = 'T1_preproc'
if t10:
bet_t10 = nipype.Node(interface=HDBet(), name='t1_0_bet')
bet_t10.inputs.save_mask = 1
bet_t10.inputs.out_file = 'T1_0_bet'
datasink = nipype.Node(nipype.DataSink(base_directory=RESULT_DIR),
"datasink")
substitutions = [('subid', sub_id)]
for i, session in enumerate(sessions):
substitutions += [('_bet{}/'.format(i), session + '/')]
datasink.inputs.substitutions = substitutions
# Create Workflow
workflow = nipype.Workflow('brain_extraction_workflow',
base_dir=NIPYPE_CACHE)
workflow.connect(datasource, 't1', bet, 'input_file')
if t10:
workflow.connect(datasource, 't1_0', bet_t10, 'input_file')
workflow.connect(bet_t10, 'out_file', datasink,
'results.subid.T10.@T1_ref_bet')
workflow.connect(bet, 'out_file', datasink, 'results.subid.@T1_preproc')
workflow.connect(bet, 'out_mask', datasink, 'results.subid.@T1_mask')
workflow = datasink_base(datasink,
datasource,
workflow,
sessions,
reference,
t10=t10)
return workflow
def workflow(self):
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
toseg = {**dict_sequences['OT']}
workflow = nipype.Workflow('lung_segmentation_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
substitutions += [('_preproc_corrected.', '_preproc.')]
datasink.inputs.substitutions = substitutions
for key in toseg:
files = []
# if tobet[key]['ref'] is not None:
# files.append(tobet[key]['ref'])
if toseg[key]['scans'] is not None:
files = files + toseg[key]['scans']
for el in files:
el = el.strip(self.extention)
node_name = '{0}_{1}'.format(key, el)
preproc = nipype.Node(interface=LungSegmentationPreproc(),
name='{}_ls_preproc'.format(node_name))
preproc.inputs.new_spacing = self.new_spacing
lung_seg = nipype.Node(interface=LungSegmentationInference(),
name='{}_ls'.format(node_name))
lung_seg.inputs.weights = self.network_weights
workflow.connect(datasource, node_name, preproc, 'in_file')
workflow.connect(preproc, 'tensor', lung_seg, 'tensor')
workflow.connect(preproc, 'image_info', lung_seg, 'image_info')
workflow.connect(
lung_seg, 'segmented_lungs', datasink,
'results.subid.{0}.@{1}_segmented_lungs'.format(key, el))
return workflow
rf_1 = nipype.MapNode(interface=fsl.RobustFOV(),
iterfield=['in_file'],
name='rf_1')
rf_ref = nipype.Node(interface=fsl.RobustFOV(), name='rf_ref')
bet_1 = nipype.MapNode(interface=HDBet(),
iterfield=['input_file'],
name='bet_1')
bet_1.inputs.save_mask = 1
bet_1.inputs.out_file = '{}_bet'.format(contrast)
bet_ref = nipype.Node(interface=HDBet(), name='bet_ref')
bet_ref.inputs.save_mask = 1
bet_ref.inputs.out_file = '{}_bet'.format(contrast)
datasink = nipype.Node(nipype.DataSink(base_directory=RESULT_DIR),
"datasink")
substitutions = [('contrast', contrast),
('sub', sub.split('/')[-1]),
('session', ref_tp + '_reference_tp')]
for i, session in enumerate(sessions):
substitutions += [('_bet_1{}/'.format(i), session + '/')]
datasink.inputs.substitutions = substitutions
workflow = nipype.Workflow('temporal_analysis_preproc_workflow',
base_dir=os.path.join(
CACHE_DIR,
sub_name + '_' + contrast))
workflow.connect(datasource, 'reference', rf_ref, 'in_file')
workflow.connect(datasource, 'to_reg', rf_1, 'in_file')
workflow.connect(rf_1, 'out_roi', bet_1, 'input_file')
datasource.inputs.field_template = dict(reference='%s/%s/%sCT.nii.gz',
to_reg='%s/%s/%s.nii.gz')
datasource.inputs.template_args = dict(to_reg=[['sub_id', 'sessions', 'contrasts']],
reference=[['sub_id', 'ref_tp', '']])
datasource.inputs.raise_on_empty = False
datasource.inputs.contrasts = contrast
datasource.inputs.sub_id = sub
datasource.inputs.sessions = sessions
datasource.inputs.ref_tp = ref_tp
reg = nipype.MapNode(interface=AntsRegSyn(), iterfield=['input_file'], name='ants_reg')
reg.inputs.transformation = 'r'
reg.inputs.num_dimensions = 3
reg.inputs.num_threads = 4
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir), "datasink")
substitutions = [('contrast', contrast), ('sub', sub)]
for i, session in enumerate(sessions):
substitutions += [('_ants_reg{}/'.format(i), session+'/')]
datasink.inputs.substitutions =substitutions
workflow = nipype.Workflow('registration_workflow', base_dir=cache_dir)
workflow.connect(datasource, 'reference', reg, 'ref_file')
workflow.connect(datasource, 'to_reg', reg, 'input_file')
workflow.connect(reg, 'reg_file', datasink, 'registration.contrast.sub.@reg_image')
workflow.connect(reg, 'regmat', datasink, 'registration.contrast.sub.@affine_mat')
workflow.connect(datasource, 'reference', datasink,
'registration.contrast.sub.@reference')
# workflow.run()
workflow.run('MultiProc', plugin_args={'n_procs': 4})
def workflow(self):
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
toreg = {**dict_sequences['MR-RT'], **dict_sequences['OT']}
workflow = nipype.Workflow('registration_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
mr_rt_ref = None
rtct = None
if dict_sequences['MR-RT'] and self.normilize_mr_rt:
ref_session = list(dict_sequences['MR-RT'].keys())[0]
ref_scans = dict_sequences['MR-RT'][ref_session]['scans']
for pr in POSSIBLE_REF:
for scan in ref_scans:
if pr in scan.split('_')[0]:
mr_rt_ref = '{0}_{1}_preproc'.format(
ref_session,
scan.split('_')[0])
mr_rt_ref_name = '{}_preproc'.format(
scan.split('_')[0])
break
else:
continue
break
if dict_sequences['RT'] and self.normilize_rtct:
rt_session = list(dict_sequences['RT'].keys())[0]
ct_name = dict_sequences['RT'][rt_session]['rtct']
if ct_name is not None and mr_rt_ref is not None:
rtct = '{0}_rtct'.format(rt_session, ct_name)
reg_mr2ct = nipype.Node(interface=AntsRegSyn(),
name='{}_lin_reg'.format(rt_session))
reg_mr2ct.inputs.transformation = 'r'
reg_mr2ct.inputs.num_dimensions = 3
reg_mr2ct.inputs.num_threads = 4
reg_mr2ct.inputs.out_prefix = '{}_reg2RTCT'.format(
mr_rt_ref_name)
reg_mr2ct.inputs.interpolation = 'BSpline'
workflow.connect(datasource, mr_rt_ref, reg_mr2ct,
'input_file')
workflow.connect(datasource, rtct, reg_mr2ct, 'ref_file')
workflow.connect(
reg_mr2ct, 'regmat', datasink,
'results.subid.{0}.@{1}_reg2RTCT_mat'.format(
ref_session, mr_rt_ref_name))
workflow.connect(
reg_mr2ct, 'reg_file', datasink,
'results.subid.{0}.@{1}_reg2RTCT'.format(
ref_session, mr_rt_ref_name))
substitutions += [
('{}_reg2RTCTWarped.nii.gz'.format(mr_rt_ref_name),
'{}_reg2RTCT.nii.gz'.format(mr_rt_ref_name))
]
substitutions += [
('{}_reg2RTCT0GenericAffine.mat'.format(mr_rt_ref_name),
'{}_reg2RTCT_linear_mat.mat'.format(mr_rt_ref_name))
]
for key in toreg:
session = toreg[key]
if session['scans'] is not None:
scans = session['scans']
scans = [x for x in scans if 'mask' not in x]
ref = None
for pr in POSSIBLE_REF:
for scan in scans:
if pr in scan:
ref = '{0}_{1}_preproc'.format(
key,
scan.split('_')[0])
scans.remove('{}_preproc'.format(
scan.split('_')[0]))
ref_name = scan.split('_')[0]
workflow.connect(
datasource, ref, datasink,
'results.subid.{0}.@{1}_reg'.format(
key, ref_name))
substitutions += [
('{}_preproc'.format(scan.split('_')[0]),
'{}_reg'.format(scan.split('_')[0]))
]
break
else:
continue
break
if ref is not None:
if mr_rt_ref is not None and key != ref_session:
reg_mr_rt = nipype.Node(interface=AntsRegSyn(),
name='{}_def_reg'.format(key))
reg_mr_rt.inputs.transformation = 's'
reg_mr_rt.inputs.num_dimensions = 3
reg_mr_rt.inputs.num_threads = 6
reg_mr_rt.inputs.out_prefix = '{}_reg2MR_RT'.format(
ref_name)
workflow.connect(datasource, ref, reg_mr_rt,
'input_file')
workflow.connect(datasource, mr_rt_ref, reg_mr_rt,
'ref_file')
workflow.connect(
reg_mr_rt, 'regmat', datasink,
'results.subid.{0}.@{1}_reg2MR_RT_linear_mat'.
format(key, ref_name))
workflow.connect(
reg_mr_rt, 'reg_file', datasink,
'results.subid.{0}.@{1}_reg2MR_RT'.format(
key, ref_name))
workflow.connect(
reg_mr_rt, 'warp_file', datasink,
'results.subid.{0}.@{1}_reg2MR_RT_warp'.format(
key, ref_name))
substitutions += [
('{}_reg2MR_RT0GenericAffine.mat'.format(ref_name),
'{}_reg2MR_RT_linear_mat.mat'.format(ref_name))
]
substitutions += [
('{}_reg2MR_RT1Warp.nii.gz'.format(ref_name),
'{}_reg2MR_RT_warp.nii.gz'.format(ref_name))
]
substitutions += [
('{}_reg2MR_RTWarped.nii.gz'.format(ref_name),
'{}_reg2MR_RT.nii.gz'.format(ref_name))
]
if rtct is not None and key != ref_session:
apply_ts_rt_ref = nipype.Node(
interface=ApplyTransforms(),
name='{}_norm2RT'.format(ref_name))
apply_ts_rt_ref.inputs.output_image = (
'{}_reg2RTCT.nii.gz'.format(ref_name))
workflow.connect(datasource, ref, apply_ts_rt_ref,
'input_image')
workflow.connect(datasource, rtct, apply_ts_rt_ref,
'reference_image')
workflow.connect(
apply_ts_rt_ref, 'output_image', datasink,
'results.subid.{0}.@{1}_reg2RTCT'.format(
key, ref_name))
merge_rt_ref = nipype.Node(
interface=Merge(4),
name='{}_merge_rt'.format(ref_name))
merge_rt_ref.inputs.ravel_inputs = True
workflow.connect(reg_mr2ct, 'regmat', merge_rt_ref,
'in1')
workflow.connect(reg_mr_rt, 'regmat', merge_rt_ref,
'in3')
workflow.connect(reg_mr_rt, 'warp_file', merge_rt_ref,
'in2')
workflow.connect(merge_rt_ref, 'out', apply_ts_rt_ref,
'transforms')
for el in scans:
el = el.strip(self.extention)
el_name = el.split('_')[0]
node_name = '{0}_{1}'.format(key, el)
reg = nipype.Node(interface=AntsRegSyn(),
name='{}_lin_reg'.format(node_name))
reg.inputs.transformation = 'r'
reg.inputs.num_dimensions = 3
reg.inputs.num_threads = 4
reg.inputs.interpolation = 'BSpline'
reg.inputs.out_prefix = '{}_reg'.format(el_name)
workflow.connect(datasource, node_name, reg,
'input_file')
workflow.connect(datasource, ref, reg, 'ref_file')
workflow.connect(
reg, 'reg_file', datasink,
'results.subid.{0}.@{1}_reg'.format(key, el_name))
workflow.connect(
reg, 'regmat', datasink,
'results.subid.{0}.@{1}_regmat'.format(
key, el_name))
substitutions += [
('{}_regWarped.nii.gz'.format(el_name),
'{}_reg.nii.gz'.format(el_name))
]
substitutions += [
('{}_reg0GenericAffine.mat'.format(el_name),
'{}_linear_regmat.mat'.format(el_name))
]
if mr_rt_ref is not None and key != ref_session:
merge = nipype.Node(
interface=Merge(3),
name='{}_merge_MR_RT'.format(node_name))
merge.inputs.ravel_inputs = True
workflow.connect(reg, 'regmat', merge, 'in3')
workflow.connect(reg_mr_rt, 'regmat', merge, 'in2')
workflow.connect(reg_mr_rt, 'warp_file', merge,
'in1')
apply_ts = nipype.Node(
interface=ApplyTransforms(),
name='{}_norm2MR_RT'.format(node_name))
apply_ts.inputs.output_image = '{}_reg2MR_RT.nii.gz'.format(
el_name)
workflow.connect(merge, 'out', apply_ts,
'transforms')
workflow.connect(datasource, node_name, apply_ts,
'input_image')
workflow.connect(datasource, mr_rt_ref, apply_ts,
'reference_image')
workflow.connect(
apply_ts, 'output_image', datasink,
'results.subid.{0}.@{1}_reg2MR_RT'.format(
key, el_name))
if rtct is not None:
apply_ts_rt = nipype.Node(
interface=ApplyTransforms(),
name='{}_norm2RT'.format(node_name))
apply_ts_rt.inputs.output_image = '{}_reg2RTCT.nii.gz'.format(
el_name)
workflow.connect(datasource, node_name,
apply_ts_rt, 'input_image')
workflow.connect(datasource, rtct, apply_ts_rt,
'reference_image')
workflow.connect(
apply_ts_rt, 'output_image', datasink,
'results.subid.{0}.@{1}_reg2RTCT'.format(
key, el_name))
if key != ref_session:
merge_rt = nipype.Node(
interface=Merge(4),
name='{}_merge_rt'.format(node_name))
merge_rt.inputs.ravel_inputs = True
workflow.connect(reg_mr2ct, 'regmat', merge_rt,
'in1')
workflow.connect(reg, 'regmat', merge_rt,
'in4')
workflow.connect(reg_mr_rt, 'regmat', merge_rt,
'in3')
workflow.connect(reg_mr_rt, 'warp_file',
merge_rt, 'in2')
workflow.connect(merge_rt, 'out', apply_ts_rt,
'transforms')
else:
merge_rt = nipype.Node(
interface=Merge(2),
name='{}_merge_rt'.format(node_name))
merge_rt.inputs.ravel_inputs = True
workflow.connect(reg_mr2ct, 'regmat', merge_rt,
'in1')
workflow.connect(reg, 'regmat', merge_rt,
'in2')
workflow.connect(merge_rt, 'out', apply_ts_rt,
'transforms')
datasink.inputs.substitutions = substitutions
return workflow
def workflow(self):
# self.datasource()
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
regex = self.regex
roi_selection = self.roi_selection
workflow = nipype.Workflow('rtstruct_extraction_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
substitutions += [('_mha_convert/', '/')]
rt_sessions = dict_sequences['RT']
for key in rt_sessions:
rt_files = rt_sessions[key]
if rt_files['phy_dose'] is not None:
dose_name = '{0}_phy_dose'.format(key)
elif rt_files['rbe_dose'] is not None:
dose_name = '{0}_rbe_dose'.format(key)
elif rt_files['ot_dose'] is not None:
dose_name = '{0}_ot_dose'.format(key)
else:
roi_selection = False
if rt_files['rtct'] is not None and rt_files[
'rtstruct'] is not None:
ss_convert = nipype.Node(interface=RTStructureCoverter(),
name='ss_convert')
mha_convert = nipype.Node(interface=MHA2NIIConverter(),
name='mha_convert')
if roi_selection:
select = nipype.Node(interface=CheckRTStructures(),
name='select_gtv')
workflow.connect(mha_convert, 'out_files', select, 'rois')
workflow.connect(datasource, dose_name, select,
'dose_file')
workflow.connect(select, 'checked_roi', datasink,
'results.subid.{}.@masks'.format(key))
else:
workflow.connect(mha_convert, 'out_files', datasink,
'results.subid.{}.@masks'.format(key))
datasink.inputs.substitutions = substitutions
workflow.connect(datasource, '{0}_rtct'.format(key),
ss_convert, 'reference_ct')
workflow.connect(datasource, '{0}_rtstruct'.format(key),
ss_convert, 'input_ss')
workflow.connect(ss_convert, 'out_structures', mha_convert,
'input_folder')
else:
print(
'NO RTCT OR RTSTRUCT!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!')
# if datasource is not None:
#
# workflow = nipype.Workflow('rtstruct_extraction_workflow', base_dir=nipype_cache)
#
# datasink = nipype.Node(nipype.DataSink(base_directory=result_dir), "datasink")
# substitutions = [('subid', sub_id)]
# substitutions += [('results/', '{}/'.format(self.workflow_name))]
#
# ss_convert = nipype.MapNode(interface=RTStructureCoverter(),
# iterfield=['reference_ct', 'input_ss'],
# name='ss_convert')
# mha_convert = nipype.MapNode(interface=MHA2NIIConverter(),
# iterfield=['input_folder'],
# name='mha_convert')
#
# if roi_selection:
# select = nipype.MapNode(interface=CheckRTStructures(),
# iterfield=['rois', 'dose_file'],
# name='select_gtv')
# workflow.connect(mha_convert, 'out_files', select, 'rois')
# workflow.connect(datasource, 'rt_dose', select, 'dose_file')
# workflow.connect(select, 'checked_roi', datasink,
# 'results.subid.@masks')
# else:
# workflow.connect(mha_convert, 'out_files', datasink,
# 'results.subid.@masks')
#
# for i, session in enumerate(self.rt['session']):
# substitutions += [(('_select_gtv{}/'.format(i), session+'/'))]
# substitutions += [(('_voxelizer{}/'.format(i), session+'/'))]
# substitutions += [(('_mha_convert{}/'.format(i), session+'/'))]
#
# datasink.inputs.substitutions =substitutions
#
# workflow.connect(datasource, 'rtct_nifti', ss_convert, 'reference_ct')
# workflow.connect(datasource, 'rts_dcm', ss_convert, 'input_ss')
# workflow.connect(ss_convert, 'out_structures', mha_convert, 'input_folder')
#
# workflow = self.datasink(workflow, datasink)
# else:
# workflow = nipype.Workflow('rtstruct_extraction_workflow', base_dir=nipype_cache)
return workflow
def tumor_segmentation(datasource,
sub_id,
sessions,
gtv_model,
tumor_model,
result_dir,
nipype_cache,
reference,
reg_workflow=None,
bet_workflow=None):
if reg_workflow is None:
if reference:
iterfields_t1 = ['in1', 'in2', 'in3']
if_0 = 2
else:
iterfields_t1 = ['in1', 'in2']
if_0 = 1
merge_ts_t1 = nipype.MapNode(interface=Merge(len(iterfields_t1)),
iterfield=iterfields_t1,
name='merge_t1')
merge_ts_t1.inputs.ravel_inputs = True
apply_ts_gtv = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_gtv')
apply_ts_gtv.inputs.interpolation = 'NearestNeighbor'
apply_ts_tumor = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_tumor')
apply_ts_tumor.inputs.interpolation = 'NearestNeighbor'
apply_ts_tumor1 = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_tumor1')
apply_ts_tumor1.inputs.interpolation = 'NearestNeighbor'
if reference:
merge_ts_t1ref = nipype.MapNode(interface=Merge(len(iterfields_t1)),
iterfield=['in1', 'in2'],
name='merge_t1ref')
merge_ts_t1ref.inputs.ravel_inputs = True
apply_ts_gtv_t1ref = nipype.MapNode(
interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_gtv_t1ref')
apply_ts_gtv_t1ref.inputs.interpolation = 'NearestNeighbor'
apply_ts_tumor_t1ref = nipype.MapNode(
interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_tumor_t1ref')
apply_ts_tumor_t1ref.inputs.interpolation = 'NearestNeighbor'
apply_ts_tumor1_t1ref = nipype.MapNode(
interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_tumor1_t1ref')
apply_ts_tumor1_t1ref.inputs.interpolation = 'NearestNeighbor'
outname = 'reg2CT'
else:
outname = 'reg2T1ref'
tumor_seg = nipype.MapNode(interface=HDGlioPredict(),
iterfield=['t1', 'ct1', 't2', 'flair'],
name='tumor_segmentation')
tumor_seg.inputs.out_file = 'segmentation'
mi = nipype.MapNode(Merge(2), iterfield=['in1', 'in2'], name='merge')
gtv_seg_data_prep = nipype.MapNode(interface=NNUnetPreparation(),
iterfield=['images'],
name='gtv_seg_data_prep')
gtv_seg = nipype.MapNode(interface=NNUnetInference(),
iterfield=['input_folder'],
name='gtv_segmentation')
gtv_seg.inputs.model_folder = gtv_model
tumor_seg_2mods = nipype.MapNode(interface=NNUnetInference(),
iterfield=['input_folder'],
name='tumor_seg_2mods')
tumor_seg_2mods.inputs.model_folder = tumor_model
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('/segmentation.nii.gz', '/Tumor_predicted.nii.gz')]
substitutions += [('subid', sub_id)]
for i, session in enumerate(sessions):
substitutions += [('_tumor_segmentation{}/'.format(i), session + '/')]
substitutions += [('_gtv_segmentation{}/subject1'.format(i),
session + '/GTV_predicted')]
substitutions += [('_tumor_seg_2mods{}/subject1'.format(i),
session + '/Tumor_predicted_2modalities')]
substitutions += [
('_apply_ts_gtv{}/subject1_trans.nii.gz'.format(i),
session + '/' + 'GTV_predicted_{}.nii.gz'.format(outname))
]
substitutions += [
('_apply_ts_tumor1{}/subject1_trans.nii.gz'.format(i), session +
'/' + 'Tumor_predicted_2modalities_{}.nii.gz'.format(outname))
]
substitutions += [
('_apply_ts_tumor{}/segmentation_trans.nii.gz'.format(i),
session + '/' + 'Tumor_predicted_{}.nii.gz'.format(outname))
]
substitutions += [
('_apply_ts_gtv_t1ref{}/subject1_trans.nii.gz'.format(i),
session + '/' + 'GTV_predicted_reg2T1ref.nii.gz')
]
substitutions += [
('_apply_ts_tumor1_t1ref{}/subject1_trans.nii.gz'.format(i),
session + '/' + 'Tumor_predicted_2modalities_reg2T1ref.nii.gz')
]
substitutions += [
('_apply_ts_tumor_t1ref{}/segmentation_trans.nii.gz'.format(i),
session + '/' + 'Tumor_predicted_reg2T1ref.nii.gz')
]
datasink.inputs.substitutions = substitutions
# Create Workflow
workflow = nipype.Workflow('tumor_segmentation_workflow',
base_dir=nipype_cache)
# Connect from registration workflow, if provided
if reg_workflow is not None:
workflow.connect(reg_workflow, 'masking0.out_file', mi, 'in1')
workflow.connect(reg_workflow, 'masking2.out_file', mi, 'in2')
workflow.connect(reg_workflow, 'masking0.out_file', tumor_seg, 'ct1')
workflow.connect(reg_workflow, 'masking1.out_file', tumor_seg, 't2')
workflow.connect(reg_workflow, 'masking2.out_file', tumor_seg, 'flair')
workflow.connect(bet_workflow, 'bet.out_file', tumor_seg, 't1')
workflow.connect(reg_workflow, 'merge_t1.out', apply_ts_tumor,
'transforms')
workflow.connect(reg_workflow, 'merge_t1.out', apply_ts_gtv,
'transforms')
workflow.connect(reg_workflow, 'merge_t1.out', apply_ts_tumor1,
'transforms')
if reference:
workflow.connect(reg_workflow, 'reg2T1.regmat', merge_ts_t1ref,
'in2')
workflow.connect(reg_workflow, 'reg2T1.warp_file', merge_ts_t1ref,
'in1')
else:
# for i in range(len(sessions)):
# workflow.connect(datasource, 't12ct_mat', fake_merge,
# 'in{}'.format(i+1))
workflow.connect(datasource, 'reg2t1_mat', merge_ts_t1,
'in{}'.format(if_0 + 1))
workflow.connect(datasource, 'reg2t1_warp', merge_ts_t1,
'in{}'.format(if_0))
if reference:
workflow.connect(datasource, 't12ct_mat', merge_ts_t1, 'in1')
workflow.connect(datasource, 'reg2t1_mat', merge_ts_t1ref, 'in1')
workflow.connect(datasource, 'reg2t1_warp', merge_ts_t1ref, 'in2')
workflow.connect(merge_ts_t1, 'out', apply_ts_tumor, 'transforms')
workflow.connect(merge_ts_t1, 'out', apply_ts_gtv, 'transforms')
workflow.connect(merge_ts_t1, 'out', apply_ts_tumor1, 'transforms')
workflow.connect(datasource, 'ct1_preproc', mi, 'in1')
workflow.connect(datasource, 'flair_preproc', mi, 'in2')
workflow.connect(datasource, 'ct1_preproc', tumor_seg, 'ct1')
workflow.connect(datasource, 't2_preproc', tumor_seg, 't2')
workflow.connect(datasource, 'flair_preproc', tumor_seg, 'flair')
workflow.connect(datasource, 't1_preproc', tumor_seg, 't1')
# Connect from datasource
if reference:
workflow.connect(merge_ts_t1ref, 'out', apply_ts_tumor_t1ref,
'transforms')
workflow.connect(merge_ts_t1ref, 'out', apply_ts_gtv_t1ref,
'transforms')
workflow.connect(merge_ts_t1ref, 'out', apply_ts_tumor1_t1ref,
'transforms')
workflow.connect(datasource, 'reference', apply_ts_gtv,
'reference_image')
workflow.connect(datasource, 'reference', apply_ts_tumor1,
'reference_image')
workflow.connect(datasource, 'reference', apply_ts_tumor,
'reference_image')
workflow.connect(datasource, 't1_0', apply_ts_gtv_t1ref,
'reference_image')
workflow.connect(datasource, 't1_0', apply_ts_tumor1_t1ref,
'reference_image')
workflow.connect(datasource, 't1_0', apply_ts_tumor_t1ref,
'reference_image')
else:
workflow.connect(datasource, 't1_0', apply_ts_gtv, 'reference_image')
workflow.connect(datasource, 't1_0', apply_ts_tumor1,
'reference_image')
workflow.connect(datasource, 't1_0', apply_ts_tumor, 'reference_image')
# Connect other nodes
# Nodes to prepare the data before nnUNet inference
workflow.connect(mi, 'out', gtv_seg_data_prep, 'images')
# Nodes to segment GTV and tumor using nnUNet
workflow.connect(gtv_seg_data_prep, 'output_folder', gtv_seg,
'input_folder')
workflow.connect(gtv_seg_data_prep, 'output_folder', tumor_seg_2mods,
'input_folder')
# Nodes to normalize segmentations to CT space
workflow.connect(gtv_seg, 'output_file', apply_ts_gtv, 'input_image')
workflow.connect(tumor_seg_2mods, 'output_file', apply_ts_tumor1,
'input_image')
workflow.connect(tumor_seg, 'out_file', apply_ts_tumor, 'input_image')
# Connect datasink nodes to save outputs
workflow.connect(tumor_seg, 'out_file', datasink,
'results.subid.@tumor_seg')
workflow.connect(gtv_seg, 'output_file', datasink,
'results.subid.@gtv_seg')
workflow.connect(tumor_seg_2mods, 'output_file', datasink,
'results.subid.@tumor_seg_2mods')
workflow.connect(apply_ts_gtv, 'output_image', datasink,
'results.subid.@gtv_reg2CT')
workflow.connect(apply_ts_tumor, 'output_image', datasink,
'results.subid.@tumor_reg2CT')
workflow.connect(apply_ts_tumor1, 'output_image', datasink,
'results.subid.@tumor1_reg2CT')
if reference:
workflow.connect(tumor_seg_2mods, 'output_file', apply_ts_tumor1_t1ref,
'input_image')
workflow.connect(tumor_seg, 'out_file', apply_ts_tumor_t1ref,
'input_image')
workflow.connect(gtv_seg, 'output_file', apply_ts_gtv_t1ref,
'input_image')
workflow.connect(apply_ts_gtv_t1ref, 'output_image', datasink,
'results.subid.@gtv_reg2T1ref')
workflow.connect(apply_ts_tumor_t1ref, 'output_image', datasink,
'results.subid.@tumor_reg2T1ref')
workflow.connect(apply_ts_tumor1_t1ref, 'output_image', datasink,
'results.subid.@tumor1_reg2T1ref')
workflow = datasink_base(datasink, datasource, workflow, sessions,
reference)
return workflow
def longitudinal_registration(sub_id,
datasource,
sessions,
reference,
result_dir,
nipype_cache,
bet_workflow=None):
"""
This is a workflow to register multi-modalities MR (T2, T1KM, FLAIR) to their
reference T1 image, in multiple time-points cohort. In particular, for each
subject, this workflow will register the MR images in each time-point (tp)
to the corresponding T1, then it will register all the T1 images to a reference T1
(the one that is the closest in time to the radiotherapy session), and finally the
reference T1 to the BPLCT. At the end, all the MR images will be saved both in T1 space
(for each tp) and in CT space.
"""
reg2T1 = nipype.MapNode(interface=AntsRegSyn(),
iterfield=['input_file'],
name='reg2T1')
reg2T1.inputs.transformation = 's'
reg2T1.inputs.num_dimensions = 3
reg2T1.inputs.num_threads = 6
if reference:
regT12CT = nipype.MapNode(interface=AntsRegSyn(),
iterfield=['input_file'],
name='regT12CT')
regT12CT.inputs.transformation = 'r'
regT12CT.inputs.num_dimensions = 3
regT12CT.inputs.num_threads = 4
reg_nodes = []
for i in range(3):
reg = nipype.MapNode(interface=AntsRegSyn(),
iterfield=['input_file', 'ref_file'],
name='ants_reg{}'.format(i))
reg.inputs.transformation = 'r'
reg.inputs.num_dimensions = 3
reg.inputs.num_threads = 4
reg.inputs.interpolation = 'BSpline'
reg_nodes.append(reg)
apply_mask_nodes = []
for i in range(3):
masking = nipype.MapNode(interface=ApplyMask(),
iterfield=['in_file', 'mask_file'],
name='masking{}'.format(i))
apply_mask_nodes.append(masking)
apply_ts_nodes = []
for i in range(3):
apply_ts = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts{}'.format(i))
apply_ts_nodes.append(apply_ts)
# Apply ts nodes for T1_ref normalization
apply_ts_nodes1 = []
for i in range(3):
apply_ts = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts1{}'.format(i))
apply_ts_nodes1.append(apply_ts)
split_ds_nodes = []
for i in range(4):
split_ds = nipype.Node(interface=Split(), name='split_ds{}'.format(i))
split_ds.inputs.splits = [1] * len(sessions)
split_ds_nodes.append(split_ds)
apply_ts_t1 = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_t1')
merge_nodes = []
if reference:
iterfields = ['in1', 'in2', 'in3', 'in4']
iterfields_t1 = ['in1', 'in2', 'in3']
if_0 = 2
else:
iterfields = ['in1', 'in2', 'in3']
iterfields_t1 = ['in1', 'in2']
if_0 = 1
for i in range(3):
merge = nipype.MapNode(interface=Merge(len(iterfields)),
iterfield=iterfields,
name='merge{}'.format(i))
merge.inputs.ravel_inputs = True
merge_nodes.append(merge)
# Merging transforms for normalization to T1_ref
merge_nodes1 = []
for i in range(3):
merge = nipype.MapNode(interface=Merge(3),
iterfield=['in1', 'in2', 'in3'],
name='merge1{}'.format(i))
merge.inputs.ravel_inputs = True
merge_nodes1.append(merge)
merge_ts_t1 = nipype.MapNode(interface=Merge(len(iterfields_t1)),
iterfield=iterfields_t1,
name='merge_t1')
merge_ts_t1.inputs.ravel_inputs = True
# have to create a fake merge of the transformation from t10 to CT in order
# to have the same number if matrices as input in mapnode
fake_merge = nipype.Node(interface=Merge(len(sessions)), name='fake_merge')
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
for i, session in enumerate(sessions):
substitutions += [('session'.format(i), session)]
substitutions += [('_masking0{}/antsregWarped_masked.nii.gz'.format(i),
session + '/' + 'CT1_preproc.nii.gz')]
substitutions += [('_reg2T1{}/antsreg0GenericAffine.mat'.format(i),
session + '/' + 'reg2T1_ref.mat')]
substitutions += [('_reg2T1{}/antsreg1Warp.nii.gz'.format(i),
session + '/' + 'reg2T1_ref_warp.nii.gz')]
substitutions += [('_reg2T1{}/antsregWarped.nii.gz'.format(i),
session + '/' + 'T1_reg2T1_ref.nii.gz')]
substitutions += [('_regT12CT{}/antsreg0GenericAffine.mat'.format(i),
'/regT1_ref2CT.mat')]
substitutions += [('_masking1{}/antsregWarped_masked.nii.gz'.format(i),
session + '/' + 'T2_preproc.nii.gz')]
substitutions += [('_masking2{}/antsregWarped_masked.nii.gz'.format(i),
session + '/' + 'FLAIR_preproc.nii.gz')]
substitutions += [('_apply_ts0{}/CT1_trans.nii.gz'.format(i),
session + '/' + 'CT1_reg2CT.nii.gz')]
substitutions += [('_apply_ts1{}/T2_trans.nii.gz'.format(i),
session + '/' + 'T2_reg2CT.nii.gz')]
substitutions += [('_apply_ts2{}/FLAIR_trans.nii.gz'.format(i),
session + '/' + 'FLAIR_reg2CT.nii.gz')]
substitutions += [('_apply_ts_t1{}/T1_trans.nii.gz'.format(i),
session + '/' + 'T1_reg2CT.nii.gz')]
substitutions += [('_apply_ts10{}/CT1_trans.nii.gz'.format(i),
session + '/' + 'CT1_reg2T1_ref.nii.gz')]
substitutions += [('_apply_ts11{}/T2_trans.nii.gz'.format(i),
session + '/' + 'T2_reg2T1_ref.nii.gz')]
substitutions += [('_apply_ts12{}/FLAIR_trans.nii.gz'.format(i),
session + '/' + 'FLAIR_reg2T1_ref.nii.gz')]
datasink.inputs.substitutions = substitutions
# Create Workflow
workflow = nipype.Workflow('registration_workflow', base_dir=nipype_cache)
for i, reg in enumerate(reg_nodes):
workflow.connect(datasource, SEQUENCES[i + 1], reg, 'input_file')
workflow.connect(datasource, SEQUENCES[0], reg, 'ref_file')
# bring every MR in CT space
for i, node in enumerate(apply_ts_nodes):
workflow.connect(datasource, SEQUENCES[i + 1], node, 'input_image')
if reference:
workflow.connect(datasource, 'reference', node, 'reference_image')
else:
workflow.connect(datasource, 't1_0', node, 'reference_image')
workflow.connect(merge_nodes[i], 'out', node, 'transforms')
workflow.connect(node, 'output_image', datasink,
'results.subid.@{}_reg2CT'.format(SEQUENCES[i + 1]))
# bring every MR in T1_ref space
for i, node in enumerate(apply_ts_nodes1):
workflow.connect(datasource, SEQUENCES[i + 1], node, 'input_image')
workflow.connect(datasource, 't1_0', node, 'reference_image')
workflow.connect(merge_nodes1[i], 'out', node, 'transforms')
workflow.connect(
node, 'output_image', datasink,
'results.subid.@{}_reg2T1_ref'.format(SEQUENCES[i + 1]))
for i, node in enumerate(merge_nodes):
workflow.connect(reg_nodes[i], 'regmat', node, 'in{}'.format(if_0 + 2))
workflow.connect(reg2T1, 'regmat', node, 'in{}'.format(if_0 + 1))
workflow.connect(reg2T1, 'warp_file', node, 'in{}'.format(if_0))
if reference:
workflow.connect(fake_merge, 'out', node, 'in1')
for i, node in enumerate(merge_nodes1):
workflow.connect(reg_nodes[i], 'regmat', node, 'in3')
workflow.connect(reg2T1, 'regmat', node, 'in2')
workflow.connect(reg2T1, 'warp_file', node, 'in1')
for i, mask in enumerate(apply_mask_nodes):
workflow.connect(reg_nodes[i], 'reg_file', mask, 'in_file')
if bet_workflow is not None:
workflow.connect(bet_workflow, 'bet.out_mask', mask, 'mask_file')
else:
workflow.connect(datasource, 't1_mask', mask, 'mask_file')
workflow.connect(mask, 'out_file', datasink,
'results.subid.@{}_preproc'.format(SEQUENCES[i + 1]))
if bet_workflow is not None:
workflow.connect(bet_workflow, 'bet.out_file', reg2T1, 'input_file')
workflow.connect(bet_workflow, 't1_0_bet.out_file', reg2T1, 'ref_file')
else:
workflow.connect(datasource, 't1_bet', reg2T1, 'input_file')
workflow.connect(datasource, 't1_0_bet', reg2T1, 'ref_file')
if reference:
for i, sess in enumerate(sessions):
workflow.connect(regT12CT, 'regmat', fake_merge,
'in{}'.format(i + 1))
workflow.connect(regT12CT, 'regmat', datasink,
'results.subid.{0}.@regT12CT_mat'.format(sess))
workflow.connect(datasource, 'reference', regT12CT, 'ref_file')
workflow.connect(datasource, 't1_0', regT12CT, 'input_file')
workflow.connect(fake_merge, 'out', merge_ts_t1, 'in1')
workflow.connect(datasource, 'reference', apply_ts_t1,
'reference_image')
else:
workflow.connect(datasource, 't1_0', apply_ts_t1, 'reference_image')
workflow.connect(datasource, 't1', apply_ts_t1, 'input_image')
workflow.connect(merge_ts_t1, 'out', apply_ts_t1, 'transforms')
workflow.connect(reg2T1, 'regmat', merge_ts_t1, 'in{}'.format(if_0 + 1))
workflow.connect(reg2T1, 'warp_file', merge_ts_t1, 'in{}'.format(if_0))
workflow.connect(reg2T1, 'warp_file', datasink,
'results.subid.@reg2CT_warp')
workflow.connect(reg2T1, 'regmat', datasink, 'results.subid.@reg2CT_mat')
workflow.connect(reg2T1, 'reg_file', datasink, 'results.subid.@T12T1_ref')
workflow.connect(apply_ts_t1, 'output_image', datasink,
'results.subid.@T1_reg2CT')
if bet_workflow is not None:
workflow = datasink_base(datasink, datasource, workflow, sessions,
reference)
else:
workflow = datasink_base(datasink,
datasource,
workflow,
sessions,
reference,
extra_nodes=['t1_bet'])
return workflow
def single_tp_registration(sub_id,
datasource,
session,
reference,
result_dir,
nipype_cache,
bet_workflow=None):
"""
This is a workflow to register multi-modalities MR (T2, T1KM, FLAIR) to their
reference T1 image, in one single time-point cohort. In particular, for each
subject, this workflow will register the MR images in the provided time-point (tp)
to the corresponding T1, then it will register the T1 image to the BPLCT (if present)'
'. At the end, all the MR images will be saved both in T1 space and in CT space.
"""
session = session[0]
if reference:
regT12CT = nipype.MapNode(interface=AntsRegSyn(),
iterfield=['input_file'],
name='regT12CT')
regT12CT.inputs.transformation = 'r'
regT12CT.inputs.num_dimensions = 3
regT12CT.inputs.num_threads = 4
reg_nodes = []
for i in range(3):
reg = nipype.MapNode(interface=AntsRegSyn(),
iterfield=['input_file', 'ref_file'],
name='ants_reg{}'.format(i))
reg.inputs.transformation = 'r'
reg.inputs.num_dimensions = 3
reg.inputs.num_threads = 4
reg.inputs.interpolation = 'BSpline'
reg_nodes.append(reg)
apply_mask_nodes = []
for i in range(3):
masking = nipype.MapNode(interface=ApplyMask(),
iterfield=['in_file', 'mask_file'],
name='masking{}'.format(i))
apply_mask_nodes.append(masking)
if reference:
apply_ts_nodes = []
for i in range(3):
apply_ts = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts{}'.format(i))
apply_ts_nodes.append(apply_ts)
apply_ts_t1 = nipype.MapNode(interface=ApplyTransforms(),
iterfield=['input_image', 'transforms'],
name='apply_ts_t1')
merge_nodes = []
for i in range(3):
merge = nipype.MapNode(interface=Merge(2),
iterfield=['in1', 'in2'],
name='merge{}'.format(i))
merge.inputs.ravel_inputs = True
merge_nodes.append(merge)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('session', session)]
substitutions += [('_regT12CT0/antsreg0GenericAffine.mat',
'/reg2T1_ref.mat')]
substitutions += [('_masking00/antsregWarped_masked.nii.gz',
session + '/' + 'CT1_preproc.nii.gz')]
substitutions += [('_regT12CT/antsreg0GenericAffine.mat',
'/regT1_ref2CT.mat')]
substitutions += [('_masking10/antsregWarped_masked.nii.gz',
session + '/' + 'T2_preproc.nii.gz')]
substitutions += [('_masking20/antsregWarped_masked.nii.gz',
session + '/' + 'FLAIR_preproc.nii.gz')]
substitutions += [('_apply_ts00/antsregWarped_masked_trans.nii.gz',
session + '/' + 'CT1_reg2CT.nii.gz')]
substitutions += [('_apply_ts10/antsregWarped_masked_trans.nii.gz',
session + '/' + 'T2_reg2CT.nii.gz')]
substitutions += [('_apply_ts20/antsregWarped_masked_trans.nii.gz',
session + '/' + 'FLAIR_reg2CT.nii.gz')]
substitutions += [('_apply_ts_t10/T1_preproc_trans.nii.gz',
session + '/' + 'T1_reg2CT.nii.gz')]
datasink.inputs.substitutions = substitutions
# Create Workflow
workflow = nipype.Workflow('registration_workflow', base_dir=nipype_cache)
for i, reg in enumerate(reg_nodes):
workflow.connect(datasource, SEQUENCES[i + 1], reg, 'input_file')
workflow.connect(datasource, SEQUENCES[0], reg, 'ref_file')
# bring every MR in CT space
if reference:
for i, node in enumerate(merge_nodes):
workflow.connect(reg_nodes[i], 'regmat', node, 'in2')
workflow.connect(regT12CT, 'regmat', node, 'in1')
for i, node in enumerate(apply_ts_nodes):
workflow.connect(apply_mask_nodes[i], 'out_file', node,
'input_image')
workflow.connect(datasource, 'reference', node, 'reference_image')
workflow.connect(regT12CT, 'regmat', node, 'transforms')
workflow.connect(
node, 'output_image', datasink,
'results.subid.@{}_reg2CT'.format(SEQUENCES[i + 1]))
workflow.connect(regT12CT, 'regmat', datasink,
'results.subid.{0}.@regT12CT_mat'.format(session))
workflow.connect(datasource, 'reference', regT12CT, 'ref_file')
workflow.connect(datasource, 't1', regT12CT, 'input_file')
if bet_workflow is not None:
workflow.connect(bet_workflow, 'bet.out_file', apply_ts_t1,
'input_image')
else:
workflow.connect(datasource, 't1_bet', apply_ts_t1, 'input_image')
workflow.connect(datasource, 'reference', apply_ts_t1,
'reference_image')
workflow.connect(apply_ts_t1, 'output_image', datasink,
'results.subid.@T1_reg2CT')
workflow.connect(regT12CT, 'regmat', apply_ts_t1, 'transforms')
for i, mask in enumerate(apply_mask_nodes):
workflow.connect(reg_nodes[i], 'reg_file', mask, 'in_file')
if bet_workflow is not None:
workflow.connect(bet_workflow, 'bet.out_mask', mask, 'mask_file')
else:
workflow.connect(datasource, 't1_mask', mask, 'mask_file')
workflow.connect(mask, 'out_file', datasink,
'results.subid.@{}_preproc'.format(SEQUENCES[i + 1]))
if bet_workflow is not None:
workflow = datasink_base(datasink,
datasource,
workflow, [session],
reference,
t10=False)
else:
workflow = datasink_base(datasink,
datasource,
workflow, [session],
reference,
extra_nodes=['t1_bet'],
t10=False)
return workflow
def convertion_workflow(self):
self.datasource()
datasource = self.data_source
ref_sequence = self.ref_sequence
t10 = self.t10
sub_id = self.sub_id
result_dir = self.result_dir
nipype_cache = self.nipype_cache
sequences = self.sequences
reference = self.reference
rt_data = self.rt
if rt_data is not None:
rt_session = rt_data['session']
workflow = nipype.Workflow('data_convertion_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
if type(ref_sequence) == list:
to_convert = sequences + ref_sequence
else:
to_convert = sequences + [ref_sequence]
if rt_data is not None:
rt_sequences = [
x for x in rt_data.keys()
if rt_data[x] and x != 'session' and x != 'labels'
]
workflow.connect(datasource, 'rt', datasink, 'results.subid.@rt')
to_convert = to_convert + rt_sequences
else:
rt_sequences = []
if reference:
to_convert.append('reference')
if t10:
to_convert.append('t1_0')
if self.ct_sessions:
to_convert.append('ct')
for seq in to_convert:
if seq not in rt_sequences:
dc = nipype.MapNode(interface=DicomCheck(),
iterfield=['dicom_dir'],
name='dc{}'.format(seq))
workflow.connect(datasource, seq, dc, 'dicom_dir')
converter = nipype.MapNode(
interface=Dcm2niix(),
iterfield=['source_dir', 'out_filename'],
name='converter{}'.format(seq))
converter.inputs.compress = 'y'
converter.inputs.philips_float = False
if seq == 'reference' or seq == 'ct':
converter.inputs.merge_imgs = True
else:
converter.inputs.merge_imgs = False
check = nipype.MapNode(interface=ConversionCheck(),
iterfield=['in_file', 'file_name'],
name='check_conversion{}'.format(seq))
workflow.connect(dc, 'outdir', converter, 'source_dir')
workflow.connect(dc, 'scan_name', converter, 'out_filename')
workflow.connect(dc, 'scan_name', check, 'file_name')
workflow.connect(converter, 'converted_files', check,
'in_file')
if seq == 'reference':
workflow.connect(
check, 'out_file', datasink,
'results.subid.REF.@{}_converted'.format(seq))
elif seq == 't1_0':
workflow.connect(
check, 'out_file', datasink,
'results.subid.T10.@{}_converted'.format(seq))
else:
workflow.connect(check, 'out_file', datasink,
'results.subid.@{}_converted'.format(seq))
for i, session in enumerate(self.session_names[seq]):
substitutions += [(('_converter{0}{1}/'.format(seq, i),
session + '/'))]
else:
if seq != 'rtstruct':
if seq == 'rtct':
converter = nipype.MapNode(
interface=Dcm2niix(),
iterfield=['source_dir', 'out_filename'],
name='converter{}'.format(seq))
converter.inputs.compress = 'y'
converter.inputs.philips_float = False
converter.inputs.merge_imgs = True
else:
converter = nipype.MapNode(
interface=DoseConverter(),
iterfield=['input_dose', 'out_name'],
name='converter{}'.format(seq))
if seq == 'doses':
converter = nipype.MapNode(
interface=DoseConverter(),
iterfield=['input_dose'],
name='converter{}'.format(seq))
get_dose = nipype.MapNode(interface=GetRefRTDose(),
iterfield=['doses'],
name='get_doses')
workflow.connect(datasource, 'doses', get_dose,
'doses')
workflow.connect(get_dose, 'dose_file', converter,
'input_dose')
converter.inputs.out_name = 'Unused_RTDOSE.nii.gz'
workflow.connect(
converter, 'out_file', datasink,
'results.subid.@{}_converted'.format(seq))
else:
dc = nipype.MapNode(interface=DicomCheck(),
iterfield=['dicom_dir'],
name='dc{}'.format(seq))
workflow.connect(datasource, seq, dc, 'dicom_dir')
if seq == 'rtct':
check = nipype.MapNode(
interface=ConversionCheck(),
iterfield=['in_file', 'file_name'],
name='check_conversion{}'.format(seq))
workflow.connect(dc, 'outdir', converter,
'source_dir')
workflow.connect(dc, 'scan_name', converter,
'out_filename')
workflow.connect(dc, 'scan_name', check,
'file_name')
workflow.connect(converter, 'converted_files',
check, 'in_file')
workflow.connect(
check, 'out_file', datasink,
'results.subid.@{}_converted'.format(seq))
else:
workflow.connect(dc, 'dose_file', converter,
'input_dose')
workflow.connect(dc, 'scan_name', converter,
'out_name')
workflow.connect(
converter, 'out_file', datasink,
'results.subid.@{}_converted'.format(seq))
else:
dc = nipype.MapNode(interface=DicomCheck(),
iterfield=['dicom_dir'],
name='dc{}'.format(seq))
workflow.connect(datasource, seq, dc, 'dicom_dir')
workflow.connect(dc, 'outdir', datasink,
'results.subid.@rtstruct')
for i, session in enumerate(rt_session):
substitutions += [
(('_dc{0}{1}/checked_dicoms'.format(seq, i),
session + '/RTSTRUCT_used'))
]
for i, session in enumerate(rt_session):
substitutions += [(('_converter{0}{1}/'.format(seq, i),
session + '/'))]
substitutions += [('_converterreference0/', '')]
substitutions += [('_convertert1_00/', '')]
datasink.inputs.substitutions = substitutions
return workflow
def sorting_workflow(self,
subject_name_position=-3,
renaming=False,
mr_classiffication=True,
checkpoints=None,
sub_checkpoints=None):
nipype_cache = os.path.join(self.nipype_cache, 'data_sorting')
result_dir = self.result_dir
workflow = nipype.Workflow('sorting_workflow', base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
# prep = nipype.Node(interface=FolderPreparation(), name='prep')
# prep.inputs.input_dir = self.base_dir
# create_list = nipype.Node(interface=CreateSubjectsList(), name='cl')
# create_list.inputs.input_dir = self.base_dir
file_check = nipype.Node(interface=FileCheck(), name='fc')
file_check.inputs.input_dir = self.base_dir
file_check.inputs.subject_name_position = subject_name_position
file_check.inputs.renaming = renaming
prep = nipype.MapNode(interface=FolderPreparation(),
name='prep',
iterfield=['input_list'])
sort = nipype.MapNode(interface=FolderSorting(),
name='sort',
iterfield=['input_dir'])
mr_rt_merge = nipype.MapNode(interface=Merge(2),
name='mr_rt_merge',
iterfield=['in1', 'in2'])
mr_rt_merge.inputs.ravel_inputs = True
merging = nipype.Node(interface=FolderMerge(), name='merge')
if mr_classiffication:
if checkpoints is None or sub_checkpoints is None:
raise Exception('MRClass weights were not provided, MR image '
'classification cannot be performed!')
mrclass = nipype.MapNode(interface=MRClass(),
name='mrclass',
iterfield=['mr_images'])
mrclass.inputs.checkpoints = checkpoints
mrclass.inputs.sub_checkpoints = sub_checkpoints
else:
mr_rt_merge.inputs.in1 = None
rt_sorting = nipype.MapNode(interface=RTDataSorting(),
name='rt_sorting',
iterfield=['input_dir'])
# workflow.connect(create_list, 'file_list', file_check, 'input_file')
workflow.connect(file_check, 'out_list', prep, 'input_list')
workflow.connect(prep, 'out_folder', sort, 'input_dir')
workflow.connect(sort, 'out_folder', rt_sorting, 'input_dir')
if mr_classiffication:
workflow.connect(sort, 'mr_images', mrclass, 'mr_images')
workflow.connect(mrclass, 'out_folder', mr_rt_merge, 'in1')
workflow.connect(rt_sorting, 'out_folder', mr_rt_merge, 'in2')
workflow.connect(mr_rt_merge, 'out', merging, 'input_list')
workflow.connect(merging, 'out_folder', datasink, '@rt_sorted')
else:
workflow.connect(rt_sorting, 'out_folder', datasink, '@rt_sorted')
substitutions = [('_rt_sorting\d+/', '')]
datasink.inputs.regexp_substitutions = substitutions
return workflow
def convertion_workflow(self):
self.datasource()
datasource = self.data_source
dict_sequences = self.dict_sequences
nipype_cache = self.nipype_cache
result_dir = self.result_dir
sub_id = self.sub_id
toprocess = {**dict_sequences['MR-RT'], **dict_sequences['OT']}
workflow = nipype.Workflow('data_convertion_workflow',
base_dir=nipype_cache)
datasink = nipype.Node(nipype.DataSink(base_directory=result_dir),
"datasink")
substitutions = [('subid', sub_id)]
substitutions += [('results/', '{}/'.format(self.workflow_name))]
substitutions += [('checked_dicoms', 'RTSTRUCT_used')]
datasink.inputs.substitutions = substitutions
for key in toprocess:
files = []
if toprocess[key]['scans'] is not None:
files = files + toprocess[key]['scans']
for el in files:
el = el.strip(self.extention)
node_name = '{0}_{1}'.format(key, el)
dc = nipype.Node(interface=DicomCheck(),
name='{}_dc'.format(node_name))
workflow.connect(datasource, node_name, dc, 'dicom_dir')
converter = nipype.Node(interface=Dcm2niix(),
name='{}_convert'.format(node_name))
converter.inputs.compress = 'y'
converter.inputs.philips_float = False
if el == 'CT':
converter.inputs.merge_imgs = True
else:
converter.inputs.merge_imgs = False
check = nipype.Node(interface=ConversionCheck(),
name='{}_cc'.format(node_name))
workflow.connect(dc, 'outdir', converter, 'source_dir')
workflow.connect(dc, 'scan_name', converter, 'out_filename')
workflow.connect(dc, 'scan_name', check, 'file_name')
workflow.connect(converter, 'converted_files', check,
'in_file')
workflow.connect(
check, 'out_file', datasink,
'results.subid.{0}.@{1}_converted'.format(key, el))
check = nipype.Node(interface=ConversionCheck(),
name='{}_cc'.format(node_name))
for key in dict_sequences['RT']:
doses = []
if dict_sequences['RT'][key]['phy_dose'] is not None:
doses.append('{}_phy_dose'.format(key))
if dict_sequences['RT'][key]['rbe_dose'] is not None:
doses.append('{}_rbe_dose'.format(key))
for el in doses:
el = el.strip(self.extention)
node_name = el.strip(self.extention)
converter = nipype.Node(interface=DoseConverter(),
name='{}_dose_conv'.format(node_name))
dc = nipype.Node(interface=DicomCheck(),
name='{}_dc'.format(node_name))
workflow.connect(datasource, node_name, dc, 'dicom_dir')
workflow.connect(dc, 'dose_file', converter, 'input_dose')
workflow.connect(dc, 'scan_name', converter, 'out_name')
workflow.connect(
converter, 'out_file', datasink,
'results.subid.{0}.@{1}_converted'.format(key, el))
if dict_sequences['RT'][key]['ot_dose'] is not None:
el = '{}_ot_dose'.format(key)
node_name = el.strip(self.extention)
converter = nipype.Node(interface=DoseConverter(),
name='{}_convert'.format(node_name))
get_dose = nipype.Node(interface=GetRefRTDose(),
name='{}_get_dose'.format(node_name))
workflow.connect(datasource, node_name, get_dose, 'doses')
workflow.connect(get_dose, 'dose_file', converter,
'input_dose')
converter.inputs.out_name = 'Unused_RTDOSE.nii.gz'
workflow.connect(
converter, 'out_file', datasink,
'results.subid.{0}.@{1}_converted'.format(key, el))
if dict_sequences['RT'][key]['rtct'] is not None:
el = '{}_rtct'.format(key)
node_name = el.strip(self.extention)
converter = nipype.Node(interface=Dcm2niix(),
name='{}_convert'.format(node_name))
converter.inputs.compress = 'y'
converter.inputs.philips_float = False
converter.inputs.merge_imgs = True
dc = nipype.Node(interface=DicomCheck(),
name='{}_dc'.format(node_name))
workflow.connect(datasource, node_name, dc, 'dicom_dir')
check = nipype.Node(interface=ConversionCheck(),
name='{}_cc'.format(node_name))
workflow.connect(dc, 'outdir', converter, 'source_dir')
workflow.connect(dc, 'scan_name', converter, 'out_filename')
workflow.connect(dc, 'scan_name', check, 'file_name')
workflow.connect(converter, 'converted_files', check,
'in_file')
workflow.connect(
check, 'out_file', datasink,
'results.subid.{0}.@{1}_converted'.format(key, el))
if dict_sequences['RT'][key]['rtstruct'] is not None:
el = '{}_rtstruct'.format(key)
node_name = el.strip(self.extention)
dc = nipype.Node(interface=DicomCheck(),
name='{}_dc'.format(node_name))
workflow.connect(datasource, node_name, dc, 'dicom_dir')
workflow.connect(dc, 'outdir', datasink,
'results.subid.{0}.@rtstruct'.format(key, el))
return workflow
