Bunch has no attribute insert (from neurostars)

[oesteban]

Just seen this in neurostars:

161023-16:13:46,511 workflow INFO:
         Running: antsApplyTransforms --default-value 0 --input /home/contematto/exp/fmri_pipelines/workdir/bids_preprocess/featpreproc/_subject_id_sub-sid000003_task_id_1back/addmean/mapflow/_addmean0/sub-sid000003_task-1back_run-01_bold_dtype_mcf_mask_smooth_mask_gms_tempfilt_maths.nii.gz --input-image-type 3 --interpolation Linear --output sub-sid000003_task-1back_run-01_bold_dtype_mcf_mask_smooth_mask_gms_tempfilt_maths_trans.nii.gz --reference-image /usr/share/fsl/5.0/data/standard/MNI152_T1_2mm.nii.gz --transform [ /home/contematto/exp/fmri_pipelines/workdir/bids_preprocess/registration/_subject_id_sub-sid000003_task_id_1back/antsRegister/output_Composite.h5, 0 ] --transform [ /home/contematto/exp/fmri_pipelines/workdir/bids_preprocess/registration/_subject_id_sub-sid000003_task_id_1back/convert2itk/affine.txt, 0 ]
161023-16:13:46,512 interface DEBUG:
         default_value_0.0
161023-16:13:46,512 interface DEBUG:
         input_image_/home/contematto/exp/fmri_pipelines/workdir/bids_preprocess/featpreproc/_subject_id_sub-sid000003_task_id_1back/addmean/mapflow/_addmean0/sub-sid000003_task-1back_run-01_bold_dtype_mcf_mask_smooth_mask_gms_tempfilt_maths.nii.gz
161023-16:13:46,512 interface DEBUG:
         input_image_type_3
161023-16:13:46,512 interface DEBUG:
         reference_image_/usr/share/fsl/5.0/data/standard/MNI152_T1_2mm.nii.gz
161023-16:13:46,513 interface DEBUG:
         {'DISPLAY': ':1'}
161023-16:15:40,447 interface DEBUG:
         Not a string literal
161023-16:15:40,511 interface DEBUG:
         Not a string literal
Traceback (most recent call last):
  File "preprocess.py", line 1129, in <module>
    wf.run(args.plugin)
  File "/home/contematto/exp/fmri_pipelines/venv2/local/lib/python2.7/site-packages/nipype/pipeline/engine/workflows.py", line 597, in run
    runner.run(execgraph, updatehash=updatehash, config=self.config)
  File "/home/contematto/exp/fmri_pipelines/venv2/local/lib/python2.7/site-packages/nipype/pipeline/plugins/linear.py", line 39, in run
    node.run(updatehash=updatehash)
  File "/home/contematto/exp/fmri_pipelines/venv2/local/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 394, in run
    self._run_interface()
  File "/home/contematto/exp/fmri_pipelines/venv2/local/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1306, in _run_interface
    updatehash=updatehash))
  File "/home/contematto/exp/fmri_pipelines/venv2/local/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1170, in _collate_results
    self._result.runtime.insert(i, None)
AttributeError: 'Bunch' object has no attribute 'insert'

[mvdoc]

@oesteban just following up here instead of on neurostars: if I disable the debug mode, the pipeline seems to execute that node

[mvdoc]

And for reference:

(venv2)contematto@head1 ~/exp/fmri_pipelines (master*) $ python -c "import nipype; print(nipype.__version__)"
0.12.1

[mattcieslak]

Any word on this? I'm also running into it when trying anything that uses a MapNode

[satra]

@mvdoc - any chance you could test this with current master?

[mattcieslak]

I get the same error on 0.13.0-g5e0597b.dev

[satra]

@mattcieslak - thanks for the update. it would really help if you can create a simple example? because we do run many pipelines with mapnodes in our test suite, so we are trying to figure out why we don't see this error.

[mvdoc]

@satra I found something interesting and subtle while writing up a small example. It seems that if there is an error in the pipeline, and config.enable_debug_mode() is set, the pipeline fails with AttributeError: 'Bunch' object has no attribute 'insert'. But if the pipeline exits fine, no such error is thrown. Example below:

This is the small working example

from nipype import MapNode, config
from nipype.interfaces import Function
config.enable_debug_mode()


def myfunction(string):
    return string + 'meh'

node = MapNode(Function(input_names=['string'],
                        output_names=['newstring'],
                        function=myfunction),
               iterfield=['string'],
               name='myfunc')

node.inputs.string = ['string' + str(i) for i in range(3)]
node.run()

And this is the output, running fine with enable_debug_mode set

contematto@talete ~/Documents/experiments_work/pipelines (master*) $ python example-mapnode.py
161107-19:43:37,85 workflow INFO:
         Executing node myfunc in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp43HUxH/myfunc
161107-19:43:37,97 workflow INFO:
         Executing node _myfunc0 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp43HUxH/myfunc/mapflow/_myfunc0
161107-19:43:37,123 workflow INFO:
         Runtime memory and threads stats unavailable
161107-19:43:37,125 workflow INFO:
         Executing node _myfunc1 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp43HUxH/myfunc/mapflow/_myfunc1
161107-19:43:37,139 workflow INFO:
         Runtime memory and threads stats unavailable
161107-19:43:37,141 workflow INFO:
         Executing node _myfunc2 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp43HUxH/myfunc/mapflow/_myfunc2
161107-19:43:37,154 workflow INFO:
         Runtime memory and threads stats unavailable

Now if I introduce a small error in the code, and run the following code (note the input_names=['WRONG']), it results in the AttributeError:

from nipype import MapNode, config
from nipype.interfaces import Function
config.enable_debug_mode()


def myfunction(string):
    return string + 'meh'

node = MapNode(Function(input_names=['WRONG'],
                        output_names=['newstring'],
                        function=myfunction),
               iterfield=['WRONG'],
               name='myfunc')

node.inputs.WRONG = ['string' + str(i) for i in range(3)]
node.run()

here's the output

contematto@talete ~/Documents/experiments_work/pipelines (master*) $ python example-mapnode.py
161107-19:46:02,754 workflow INFO:
         Executing node myfunc in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmpZUoT2V/myfunc
161107-19:46:02,767 workflow INFO:
         Executing node _myfunc0 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmpZUoT2V/myfunc/mapflow/_myfunc0
Traceback (most recent call last):
  File "example-mapnode.py", line 16, in <module>
    node.run()
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 394, in run
    self._run_interface()
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1306, in _run_interface
    updatehash=updatehash))
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1170, in _collate_results
    self._result.runtime.insert(i, None)
AttributeError: 'Bunch' object has no attribute 'insert'

If I disable the enable_debug_mode(), it pukes and that's it:

contematto@talete ~/Documents/experiments_work/pipelines (master*) $ python example-mapnode.py
161107-19:47:19,859 workflow INFO:
         Executing node myfunc in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp5fBpWk/myfunc
161107-19:47:19,876 workflow INFO:
         Executing node _myfunc0 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp5fBpWk/myfunc/mapflow/_myfunc0
161107-19:47:19,909 workflow INFO:
         Executing node _myfunc1 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp5fBpWk/myfunc/mapflow/_myfunc1
161107-19:47:19,919 workflow INFO:
         Executing node _myfunc2 in dir: /var/folders/gs/4v1cn8mj4n52qrr9r6xv4ckm0000gn/T/tmp5fBpWk/myfunc/mapflow/_myfunc2
Traceback (most recent call last):
  File "example-mapnode.py", line 16, in <module>
    node.run()
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 394, in run
    self._run_interface()
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1306, in _run_interface
    updatehash=updatehash))
  File "/Users/contematto/anaconda/lib/python2.7/site-packages/nipype/pipeline/engine/nodes.py", line 1211, in _collate_results
    (self.name, '\n'.join(msg)))
Exception: Subnodes of node: myfunc failed:
Subnode 0 failed
Error:
myfunction() got an unexpected keyword argument 'WRONG'
Interface Function failed to run. 
Subnode 1 failed
Error:
myfunction() got an unexpected keyword argument 'WRONG'
Interface Function failed to run. 
Subnode 2 failed
Error:
myfunction() got an unexpected keyword argument 'WRONG'
Interface Function failed to run. 

And for reference, this is with 0.12.1

contematto@talete ~/Documents/experiments_work/pipelines (master*) $ python -c "import nipype; print(nipype.__version__)"
0.12.1

[satra]

@mvdoc - awesome that helps a lot - we need to create a pipeline+mapnode test with an error :) - i'll add this as a test. should be a relatively easy fix.

[mvdoc]

Just run with master and got the same error

contematto@talete ~/Documents/experiments_work/pipelines (master*) $ python -c "import nipype; print(nipype.__version__)"
0.13.0-dev