fix: support for stricter checking with ants bsplinesyn

[satra]

@hjmjohnson - does this seem reasonable?

[coveralls]

Coverage remained the same when pulling 8c70c51 on satra:fix/antsbspline into 0928a3a on nipy:master.

[coveralls]

Coverage remained the same when pulling fda3eda on satra:fix/antsbspline into 0928a3a on nipy:master.

[hjmjohnson]

@satra Can this be merged? I thought this was ready for merging a few days ago.

[roelant001]

I'm still encountering this issue. I've installed the most recent version from github. Am I doing something wrong or is this "bug" still persisting? When manually removing the .0's from the command and running it, it runs successfully. Any suggestions?

relevant settings:

reg.inputs.transforms = ['Translation','Rigid', 'Affine', 'SyN', 'BSplineSyN']
reg.inputs.transform_parameters = [(0.1,),(0.1,), (0.1,), (0.2, 3.0, 0.0),(0.25,25,0)]

Description: itk::ERROR: CommandLineParser(0x183c7b0): ERROR: Parse error occured during command line argument processing
ERROR: Unable to convert '25.0' to type 'j' as unsigned int

[ntustison]

I accommodated this issue on the ANTs side with this commit

ANTsX/ANTs@05ed1ba

[satra]

@roelant001 - BSplineSyn needs 4 parameters, not 3, so the three parameter version is being cast to all floats.

see: https://github.com/nipy/nipype/blob/master/nipype/interfaces/ants/registration.py#L316

[roelant001]

Thanks!! @ntustison and @satra for this (very simple) explanation and solution.
I assumed the 4th parameter (splineOrder) would default when left empty. Assumptions...