fix fmri_spm_nested #1538

Author: oesteban

circle.yml

@@ -19,13 +19,13 @@ dependencies:
 test:
   override:
     - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch --entrypoint="/usr/bin/run_builddocs.sh" nipype/testbench
-    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench test_spm Linear /root/examples/ workflow3d workflow4d :
+    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench test_spm MultiProc /root/examples/ workflow3d workflow4d :
         timeout: 1600
-    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_fsl_feeds Linear /root/examples/ l1pipeline
-    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_spm_dartel Linear /root/examples/ level1 l2pipeline :
+    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_fsl_feeds MultiProc /root/examples/ l1pipeline
+    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_spm_dartel MultiProc /root/examples/ level1 l2pipeline :
         timeout: 1600
-    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_fsl_reuse Linear /root/examples/ level1_workflow
-    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_spm_nested Linear /root/examples/ level1 l2pipeline
+    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_fsl_reuse MultiProc /root/examples/ level1_workflow
+    - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch nipype/testbench fmri_spm_nested MultiProc /root/examples/ level1 l2pipeline
     - docker run -i -v /etc/localtime:/etc/localtime:ro -v ~/scratch:/scratch -w /scratch --entrypoint="/usr/bin/run_nosetests.sh" nipype/testbench :
         timeout: 2600
   post:

examples/fmri_spm_nested.py

@@ -17,16 +17,16 @@
 
 from __future__ import print_function
 from builtins import range
+import os.path as op                              # system functions
 
-import nipype.interfaces.io as nio           # Data i/o
-import nipype.interfaces.spm as spm          # spm
-import nipype.interfaces.matlab as mlab      # how to run matlab
-import nipype.interfaces.fsl as fsl          # fsl
-import nipype.interfaces.utility as niu      # utility
-import nipype.pipeline.engine as pe          # pypeline engine
-import nipype.algorithms.rapidart as ra      # artifact detection
-import nipype.algorithms.modelgen as model   # model specification
-import os                                    # system functions
+from nipype.interfaces import io as nio           # Data i/o
+from nipype.interfaces import spm as spm          # spm
+# from nipype.interfaces import matlab as mlab    # how to run matlab
+from nipype.interfaces import fsl as fsl          # fsl
+from nipype.interfaces import utility as niu      # utility
+from nipype.pipeline import engine as pe          # pypeline engine
+from nipype.algorithms import rapidart as ra      # artifact detection
+from nipype.algorithms import modelgen as model   # model specification
 
 """
 
@@ -65,19 +65,27 @@
 
 preproc = pe.Workflow(name='preproc')
 
+"""
+A node called :code:`inputnode` is set to designate the path in which input data
+are located:
+"""
+
+inputnode = pe.Node(niu.IdentityInterface(fields=['in_data']), name='inputnode')
+
+
 """Use :class:`nipype.interfaces.spm.Realign` for motion correction
 and register all images to the mean image.
 """
 
-realign = pe.Node(interface=spm.Realign(), name="realign")
+realign = pe.Node(spm.Realign(), name="realign")
 realign.inputs.register_to_mean = True
 
 """Use :class:`nipype.algorithms.rapidart` to determine which of the
 images in the functional series are outliers based on deviations in
 intensity or movement.
 """
 
-art = pe.Node(interface=ra.ArtifactDetect(), name="art")
+art = pe.Node(ra.ArtifactDetect(), name="art")
 art.inputs.use_differences = [True, False]
 art.inputs.use_norm = True
 art.inputs.norm_threshold = 1
@@ -89,14 +97,14 @@
 :class:`nipype.interfaces.fsl.BET`.
 """
 
-skullstrip = pe.Node(interface=fsl.BET(), name="skullstrip")
+skullstrip = pe.Node(fsl.BET(), name="skullstrip")
 skullstrip.inputs.mask = True
 
 """Use :class:`nipype.interfaces.spm.Coregister` to perform a rigid
 body registration of the functional data to the structural data.
 """
 
-coregister = pe.Node(interface=spm.Coregister(), name="coregister")
+coregister = pe.Node(spm.Coregister(), name="coregister")
 coregister.inputs.jobtype = 'estimate'
 
 
@@ -105,19 +113,19 @@
 includes the template image, T1.nii.
 """
 
-normalize = pe.Node(interface=spm.Normalize(), name="normalize")
-normalize.inputs.template = os.path.abspath('nipype-tutorial/data/T1.nii')
+normalize = pe.Node(spm.Normalize(), name="normalize")
 
 
 """Smooth the functional data using
 :class:`nipype.interfaces.spm.Smooth`.
 """
 
-smooth = pe.Node(interface=spm.Smooth(), name="smooth")
+smooth = pe.Node(spm.Smooth(), name="smooth")
 fwhmlist = [4]
 smooth.iterables = ('fwhm', fwhmlist)
 
-preproc.connect([(realign, coregister, [('mean_image', 'source'),
+preproc.connect([(inputnode, normalize, [(('in_data', _template_path), 'template')]),
+                 (realign, coregister, [('mean_image', 'source'),
                                         ('realigned_files', 'apply_to_files')]),
                  (coregister, normalize, [('coregistered_files', 'apply_to_files')]),
                  (normalize, smooth, [('normalized_files', 'in_files')]),
@@ -140,40 +148,40 @@
 :class:`nipype.interfaces.spm.SpecifyModel`.
 """
 
-modelspec = pe.Node(interface=model.SpecifySPMModel(), name="modelspec")
+modelspec = pe.Node(model.SpecifySPMModel(), name="modelspec")
 modelspec.inputs.concatenate_runs = True
 
 """Generate a first level SPM.mat file for analysis
 :class:`nipype.interfaces.spm.Level1Design`.
 """
 
-level1design = pe.Node(interface=spm.Level1Design(), name="level1design")
+level1design = pe.Node(spm.Level1Design(), name="level1design")
 level1design.inputs.bases = {'hrf': {'derivs': [0, 0]}}
 
 """Use :class:`nipype.interfaces.spm.EstimateModel` to determine the
 parameters of the model.
 """
 
-level1estimate = pe.Node(interface=spm.EstimateModel(), name="level1estimate")
+level1estimate = pe.Node(spm.EstimateModel(), name="level1estimate")
 level1estimate.inputs.estimation_method = {'Classical': 1}
 
 """Use :class:`nipype.interfaces.spm.EstimateContrast` to estimate the
 first level contrasts specified in a few steps above.
 """
 
-contrastestimate = pe.Node(interface=spm.EstimateContrast(), name="contrastestimate")
+contrastestimate = pe.Node(spm.EstimateContrast(), name="contrastestimate")
 
 """Use :class: `nipype.interfaces.utility.Select` to select each contrast for
 reporting.
 """
 
-selectcontrast = pe.Node(interface=niu.Select(), name="selectcontrast")
+selectcontrast = pe.Node(niu.Select(), name="selectcontrast")
 
 """Use :class:`nipype.interfaces.fsl.Overlay` to combine the statistical output of
 the contrast estimate and a background image into one volume.
 """
 
-overlaystats = pe.Node(interface=fsl.Overlay(), name="overlaystats")
+overlaystats = pe.Node(fsl.Overlay(), name="overlaystats")
 overlaystats.inputs.stat_thresh = (3, 10)
 overlaystats.inputs.show_negative_stats = True
 overlaystats.inputs.auto_thresh_bg = True
@@ -182,7 +190,7 @@
 statistical volumes for a report of the first-level results.
 """
 
-slicestats = pe.Node(interface=fsl.Slicer(), name="slicestats")
+slicestats = pe.Node(fsl.Slicer(), name="slicestats")
 slicestats.inputs.all_axial = True
 slicestats.inputs.image_width = 750
 
@@ -243,8 +251,7 @@
 info = dict(func=[['subject_id', ['f3', 'f5', 'f7', 'f10']]],
             struct=[['subject_id', 'struct']])
 
-inputnode = pe.Node(niu.IdentityInterface(fields=['in_data']), name='inputnode')
-infosource = pe.Node(interface=niu.IdentityInterface(fields=['subject_id']), name="infosource")
+infosource = pe.Node(niu.IdentityInterface(fields=['subject_id']), name="infosource")
 
 """Here we set up iteration over all the subjects. The following line
 is a particular example of the flexibility of the system.  The
@@ -265,7 +272,7 @@
 functionality.
 """
 
-datasource = pe.Node(interface=nio.DataGrabber(infields=['subject_id'],
+datasource = pe.Node(nio.DataGrabber(infields=['subject_id'],
                                                outfields=['func', 'struct']),
                      name='datasource')
 datasource.inputs.template = 'nipype-tutorial/data/%s/%s.nii'
@@ -357,7 +364,7 @@ def subjectinfo(subject_id):
 """
 
 level1 = pe.Workflow(name="level1")
-level1.base_dir = os.path.abspath('spm_tutorial2/workingdir')
+level1.base_dir = op.abspath('spm_tutorial2/workingdir')
 
 level1.connect([(inputnode, datasource,  [('in_data', 'base_directory')]),
                 (infosource, datasource, [('subject_id', 'subject_id')]),
@@ -389,16 +396,16 @@ def subjectinfo(subject_id):
 the mean image would be copied to that directory.
 """
 
-datasink = pe.Node(interface=nio.DataSink(), name="datasink")
-datasink.inputs.base_directory = os.path.abspath('spm_tutorial2/l1output')
-report = pe.Node(interface=nio.DataSink(), name='report')
-report.inputs.base_directory = os.path.abspath('spm_tutorial2/report')
+datasink = pe.Node(nio.DataSink(), name="datasink")
+datasink.inputs.base_directory = op.abspath('spm_tutorial2/l1output')
+report = pe.Node(nio.DataSink(), name='report')
+report.inputs.base_directory = op.abspath('spm_tutorial2/report')
 report.inputs.parameterization = False
 
 
 def getstripdir(subject_id):
-    import os
-    return os.path.join(os.path.abspath('spm_tutorial2/workingdir'), '_subject_id_%s' % subject_id)
+    import os.path as op
+    return op.join(op.abspath('spm_tutorial2/workingdir'), '_subject_id_%s' % subject_id)
 
 # store relevant outputs from various stages of the 1st level analysis
 level1.connect([(infosource, datasink, [('subject_id', 'container'),
@@ -440,7 +447,7 @@ def getstripdir(subject_id):
 contrast_ids = list(range(1, len(contrasts) + 1))
 l2source = pe.Node(nio.DataGrabber(infields=['fwhm', 'con']), name="l2source")
 # we use .*i* to capture both .img (SPM8) and .nii (SPM12)
-l2source.inputs.template = os.path.abspath('spm_tutorial2/l1output/*/con*/*/_fwhm_%d/con_%04d.*i*')
+l2source.inputs.template = op.abspath('spm_tutorial2/l1output/*/con*/*/_fwhm_%d/con_%04d.*i*')
 # iterate over all contrast images
 l2source.iterables = [('fwhm', fwhmlist),
                       ('con', contrast_ids)]
@@ -453,10 +460,10 @@ def getstripdir(subject_id):
 """
 
 # setup a 1-sample t-test node
-onesamplettestdes = pe.Node(interface=spm.OneSampleTTestDesign(), name="onesampttestdes")
-l2estimate = pe.Node(interface=spm.EstimateModel(), name="level2estimate")
+onesamplettestdes = pe.Node(spm.OneSampleTTestDesign(), name="onesampttestdes")
+l2estimate = pe.Node(spm.EstimateModel(), name="level2estimate")
 l2estimate.inputs.estimation_method = {'Classical': 1}
-l2conestimate = pe.Node(interface=spm.EstimateContrast(), name="level2conestimate")
+l2conestimate = pe.Node(spm.EstimateContrast(), name="level2conestimate")
 cont1 = ('Group', 'T', ['mean'], [1])
 l2conestimate.inputs.contrasts = [cont1]
 l2conestimate.inputs.group_contrast = True
@@ -467,14 +474,22 @@ def getstripdir(subject_id):
 """
 
 l2pipeline = pe.Workflow(name="level2")
-l2pipeline.base_dir = os.path.abspath('spm_tutorial2/l2output')
+l2pipeline.base_dir = op.abspath('spm_tutorial2/l2output')
 l2pipeline.connect([(l2source, onesamplettestdes, [('outfiles', 'in_files')]),
                     (onesamplettestdes, l2estimate, [('spm_mat_file', 'spm_mat_file')]),
                     (l2estimate, l2conestimate, [('spm_mat_file', 'spm_mat_file'),
                                                  ('beta_images', 'beta_images'),
                                                  ('residual_image', 'residual_image')]),
                     ])
 
+"""
+Inlining functions in connect()
+-------------------------------
+"""
+def _template_path(in_data):
+    return op.abspath(op.join(in_data, 'nipype-tutorial/data/T1.nii'))
+
+
 """
 Execute the second level pipeline
 ---------------------------------

tools/run_examples.py

@@ -2,16 +2,25 @@
 import os
 import sys
 from shutil import rmtree
+from multiprocessing import cpu_count
 
 
-def run_examples(example, pipelines, data_path, plugin):
-    print('running example: %s with plugin: %s' % (example, plugin))
+def run_examples(example, pipelines, data_path, plugin=None):
     from nipype import config
+    from nipype.interfaces.base import CommandLine
+
+    if plugin is None:
+        plugin = 'MultiProc'
+
+    print('running example: %s with plugin: %s' % (example, plugin))
     config.enable_debug_mode()
     config.enable_provenance()
-    from nipype.interfaces.base import CommandLine
     CommandLine.set_default_terminal_output("stream")
 
+    plugin_args = {}
+    if plugin == 'MultiProc':
+        plugin_args['n_procs'] = cpu_count()
+
     __import__(example)
     for pipeline in pipelines:
         wf = getattr(sys.modules[example], pipeline)
@@ -22,7 +31,7 @@ def run_examples(example, pipelines, data_path, plugin):
                                    'stop_on_first_rerun': 'true',
                                    'write_provenance': 'true'}}
         wf.inputs.inputnode.in_data = os.path.abspath(data_path)
-        wf.run(plugin=plugin, plugin_args={'n_procs': 4})
+        wf.run(plugin=plugin, plugin_args=plugin_args)
         # run twice to check if nothing is rerunning
         wf.run(plugin=plugin)