nipype

nipype copied to clipboard

In our efforts to port more of Clinica to nipype, we are faced with a challenge regarding our use of mri_surf2surf. We are using mri_surf2surf in a similar way as example 4 in the docs.

I am not sure how this would translate to the SurfaceSmooth interface. Semanticly spaking, it sounds more like a surface resampling operation rather than smoothing. Should we extend the SurfaceSmooth interface to support the additional set of parameters? Or make a new one dedicated to this use case?

I guess that fits in a wider topic of how to efficiently wrap tools which have a single entrypoint (without subcommands) for multiple distinct use cases.

I think it would make sense to copy SurfaceSmooth and just call it Surf2Surf and provide all of the options. The reason it's here as it is is that it matched somebody's use case. Some contributors have been very thorough to document all options, while others possibly didn't want to encode ones they're less familiar with to not imply support.

In any event, nipype is pretty promiscuous in its acceptance, and it's fine to have multiple interfaces to the same tool.

Some contributors have been very thorough to document all options, while others possibly didn't want to encode ones they're less familiar with to not imply support.

To be honest, I can't blame whoever contributed this interface considering these options are not listed in the help string and you have got to scroll all the way down to the example section to learn about them.

I think it would make sense to copy SurfaceSmooth and just call it Surf2Surf and provide all of the options.

Thank you for confirming. Do you think we should be doing the same for Vol2Surf (and still provide SampleToSurface as an alias)? That would also reduce the cognitive work to map which interface belongs to which CLI tool down the line. I always end up searching for the name of the actual FreeSurfer command to find the correct interface :sweat_smile:

Sounds good to me.

Hi @ghisvail - I have the same use case as yourself (I'd like to resample surfaces into the subject's native diffusion space).

Did you manage to make any progress on a Surf2Surf interface?

Did you manage to make any progress on a Surf2Surf interface?

It's on my ever growing todo list :sweat_smile:

Feel free to have a go at it if you've got time.