ANTS interface: incorrect parameter passing in WarpTimeSeriesImageMultiTransform

[piloubazin]

Summary

when using WarpTimeSeriesImageMultiTransform.inputs.invert_affine = [1], the parameter passed to the corresponding ANTS command is ['-i'], and should instead be '-i' (as in WarpImageMutiTransfrom) in resampling.py

Actual behavior

The command fails to run or to write with .cmdline
(TypeError: sequence item 0: expected string or Unicode, list found)

Expected behavior

In [8]: wimt.cmdline
Out[8]: u'WarpTimeSeriesImageMultiTransform 4 my_fixed_time_series.nii.gz my_fixed_time_series_warped.nii.gz -R my_moving_image.nii.gz -i synAffine.txt synInverseWarp.nii.gz'

How to replicate the behavior

just run WarpTimeSeriesImageMultiTransform with .inputs.invert_affine = [1]

Script/Workflow details

l.89: series += ['-i'], -> series += '-i',

Platform details:

{'nibabel_version': '2.1.0', 'sys_executable': '/usr/bin/python', 'networkx_version': '1.11', 'numpy_version': '1.11.1rc1', 'sys_platform': 'linux2', 'sys_version': '2.7.12+ (default, Sep 17 2016, 12:08:02) \n[GCC 6.2.0 20160914]', 'commit_source': u'repository', 'commit_hash': '3f8c8b1', 'pkg_path': '/mnt/drive/pilou/Code/github/src/nipype/nipype', 'nipype_version': u'1.0.0-dev+g3f8c8b1', 'traits_version': '4.6.0', 'scipy_version': '0.17.1'}
1.0.0-dev+g3f8c8b1

Execution environment

Choose one
My python environment outside container

[djarecka]

Hi, thanks for reporting. I'm reading the code and I believe that changing ['-i'] to -i in line 89 shouldn't help, since line 91 would add a space between - and i.

I might be wrong, never run this interface. Would you mind sharing the code for wimt so I can test it? The simplest examples I run didn't give me an error.

What exactly do you get when you ask for wimt.cmdline?

[djarecka]

@chrisfilo : having problem with repeating the error, but wanted to ask if you really wanted to have series += '-i' (not series += ['-i']) in WarpImageMultiTransform, see 198

[chrisgorgo]

I'm not 100% sure, but series += ['-i'] does seem to make more sense.

[djarecka]

I was wrong. I missed the comma at the end, so python treats it as adding a list, I unified it in #2229.

The reason why I didn't get any error earlier was that I provided wrong (i.e. out of range) index in invert_affine (see test_WarpImageMultiTransform_invaffine_wrong and test_WarpTimeSeriesImageMultiTransform_invaffine_wrong)