You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
is not saying that smoothparam is a keyword to use on the command line -- it is just a "description" for that positional field. So if I wanted to just change the optional-laplacian-tolerance to 0.999 (and assuming that all those =value correspond to defaults I would need to call
ATM, Nipype interface, if I just specify \.opt_tolerance=0.999 (and none of the other prior optional positional arguments) then it would produce (do not have a nipype test, only an integration test result observation):
LaplacianThickness WM.nii GM.nii Out.nii 0.999
which would place that 0.999 into where ANTs expects (and consumes) the smoothparam.
Expected behavior
If nipype to not carry the defaults for all those fields (which might change etc within ANTs, so not a good idea I guess), then at least I expect it to blow stating that specification of some later positional argument requires specification (or a default value) for all preceding it arguments.
deftest_LaplacianThickness_wrongargs(change_dir, create_lt):
lt=create_ltlt.inputs.prior_thickness=5.9# 500 must not be placed as smooth_paramassertlt.cmdline!='LaplacianThickness functional.nii diffusion_weighted.nii functional_thickness.nii 5.9'# probably should have just raised an exception that "smooth_param"# should also be defined
which currently would fail!
The text was updated successfully, but these errors were encountered:
Summary
This is a supplement to #2846 fixing up ANTs LaplacianSmoothness interface. The main issue is that the cmdline call description
is not saying that
smoothparam
is a keyword to use on the command line -- it is just a "description" for that positional field. So if I wanted to just change theoptional-laplacian-tolerance
to 0.999 (and assuming that all those=value
correspond to defaults I would need to callActual behavior
ATM, Nipype interface, if I just specify
\.opt_tolerance=0.999
(and none of the other prior optional positional arguments) then it would produce (do not have a nipype test, only an integration test result observation):which would place that 0.999 into where ANTs expects (and consumes) the
smoothparam
.Expected behavior
If nipype to not carry the defaults for all those fields (which might change etc within ANTs, so not a good idea I guess), then at least I expect it to blow stating that specification of some later positional argument requires specification (or a default value) for all preceding it arguments.
Here is the unittest on top of #2846
which currently would fail!
The text was updated successfully, but these errors were encountered: