[nipype]define individual interface inputs

[Juliane]

Hi everyone,

I am working on my own interface. Is it possible to create new interface inputs when i wrapped it in a node in the pipeline script? and how would you define the input spec for the interface then?
And is it possible to create dynamic outputs in the _run_interface()-function?

Thank you! 

[Juliane]

And is it possible to create dynamic outputs in the _run_interface()-function?

More information: I would like to use dicts instead of lists in the pipeline and i would like to have an interface that takes the dict for example from the convert pipeline, looks up what kind of keys are there, like 'epi', 't1'  and creates new outputs depended on the number of key it has. So that i have as many specific output as i have keys for connection with further pipelines. hope someone understands that.

[Satrajit Ghosh]

hi juliane,

I am working on my own interface. Is it possible to create new interface inputs when i wrapped it in a node in the pipeline script? and how would you define the input spec for the interface then?

it depends on the underlying interface. see examples below.

 

And is it possible to create dynamic outputs in the _run_interface()-function?

here is an example:

Function (allows inputs/outputs to be specified via init)

https://github.com/nipy/nipype/blob/ef9122ec82239979400703f33569aa27f44fa4f0/nipype/interfaces/utility.py#L320

DataSink (creates inputs on the fly)

https://github.com/nipy/nipype/blob/ef9122ec82239979400703f33569aa27f44fa4f0/nipype/interfaces/io.py#L151

many of the interfaces in interfaces.io or interfaces.utility have dynamic inputs or outputs.

cheers,

satra

[Satrajit Ghosh]

hi juliane,

And is it possible to create dynamic outputs in the _run_interface()-function?

More information: I would like to use dicts instead of lists in the pipeline and i would like to have an interface that takes the dict for example from the convert pipeline, looks up what kind of keys are there, like 'epi', 't1'  and creates new outputs depended on the number of key it has. So that i have as many specific output as i have keys for connection with further pipelines. hope someone understands that.

this is kind of what datagrabber/finder does. 

cheers,

satra