Mrtrix3 `dwi2response` - bad algorithm argument position

[jguillon]

Summary

Th Mrtrix3 dwi2response CL wrapper generates the following runtime error:

dwi2response:
mrinfo: [ERROR] no diffusion encoding information found in image "<DWI_FILE>"
dwi2response: [ERROR] Script requires diffusion gradient table: either in image header, or using -grad / -fslgrad option

It turns out that the command generated by nipype does not respect (my version of) the Mrtrix3 CL format.

Actual behavior

Generated command (not runnable):

dwi2response -fslgrad <BVEC_FILE> <BVAL_FILE> -mask <MASK_FILE> tournier <WM_FILE>

Expected behavior

Runnable command:

dwi2response tournier -fslgrad <BVEC_FILE> <BVAL_FILE> -mask <MASK_FILE> <WM_FILE>

Environment

• MRtrix 3.0_RC2-117-gf098f097 dwi2response bin version: 3.0_RC2-117-gf098f097
• Python 2.7.12
• nipype v1.0.0

Quick and dirty solution

I'm really not sure how clean it is, but it worked for me; in the ResponseSDInputSpec class, I changed position=-6 to position=1 in the algorithm traits.

[mgxd]

@jguillon yeah - looks like if any of the inherited inputs:

nipype/nipype/interfaces/mrtrix3/base.py

Lines 28 to 50 in 2c35800

	grad_file = File(
	exists=True,
	argstr='-grad %s',
	desc='dw gradient scheme (MRTrix format')
	grad_fsl = traits.Tuple(
	File(exists=True),
	File(exists=True),
	argstr='-fslgrad %s %s',
	desc='(bvecs, bvals) dw gradient scheme (FSL format')
	bval_scale = traits.Enum(
	'yes',
	'no',
	argstr='-bvalue_scaling %s',
	desc='specifies whether the b - values should be scaled by the square'
	' of the corresponding DW gradient norm, as often required for '
	'multishell or DSI DW acquisition schemes. The default action '
	'can also be set in the MRtrix config file, under the '
	'BValueScaling entry. Valid choices are yes / no, true / '
	'false, 0 / 1 (default: true).')
	in_bvec = File(
	exists=True, argstr='-fslgrad %s %s', desc='bvecs file in FSL format')
	in_bval = File(exists=True, desc='bvals file in FSL format')

are defined, the position of algorithm is shifted - would you mind submitting a PR with this fix?