MRIConvert interface mismatch output name from freesurfer mri_convert ouputs

[vferat]

Summary

The behavior of list outputs from Nipype MRIConvert interface is different from the one of freesurfer mri_convert function at least for spm (img + hdr) files. The interface ouput file as an index but freesurfer files doesn't. Problem may comes from this line.

Actual behavior

traits.trait_errors.TraitError: The trait 'out_file' of a MRIConvertOutputSpec instance is an existing file name, but the path '/home/vferat/aparc+aseg_out001.img' does not exist.

Nipype interface list output like

• [input file name] + _out001.img

while freesurfer outputs files named:

• [input file name] + _out.img
• [input file name] + _out.hdr

Expected behavior

Filename from freesurfer output and Nipype MRIConvert interface should be the same

How to replicate the behavior

from nipype.interfaces import freesurfer
mc = freesurfer.MRIConvert()
mc.inputs.in_file = "aparc+aseg.mgz"
mc.inputs.out_type = 'spm'
mc.cmdline
mc.run()

Platform details:

freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c

commit_source': 'installation', 'commit_hash': 'a0f756d1c', 'nipype_version': '1.2.0', 'sys_version': '3.6.8 (default, Jan 14 2019, 11:02:34) \n[GCC 8.0.1 20180414 (experimental) [trunk revision 259383]]', 'sys_executable': '/usr/bin/python3', 'sys_platform': 'linux', 'numpy_version': '1.17.0', 'scipy_version': '1.3.0', 'networkx_version': '2.3', 'nibabel_version': '2.5.0', 'traits_version': '5.1.2'}

Execution environment

• My python environment outside container : Python 3.6.8 (default, Jan 14 2019, 11:02:34)

[effigies]

Ah, it looks like this is a case where you have a 4D image with dim[4] == 1. If this is the case, then we should add a check for that where you identified it. Would you care to submit a pull request?

[vferat]

I checked my file and the image is 3D

size = load(self.inputs.in_file).shape
print(size)
(256, 256, 256)

Interface works fine for other output types.
Maybe mri_convert add a suffix only if the input_file is a 4D image, but unfortunatly I do not have such a file to test it.

If it's the case, editing this lines should work (it does for my file)

if self.inputs.out_type in ['spm', 'analyze']:
                # generate all outputs
                size = load(self.inputs.in_file).shape
                if len(size) == 3:
                    outputs['out_file'] = self._get_outfilename()
                else:
                    tp = size[-1]
                    # have to take care of all the frame manipulations
                    raise Exception('Not taking frame manipulations into account- please warn the developers')

I can open a PR but I'm not sure this works for all input formats and I have not much experience with MRI data format, so I prefer to have your feedbacks first