FCP-INDI / C-PAC

Configurable Pipeline for the Analysis of Connectomes
https://fcp-indi.github.io/
GNU Lesser General Public License v3.0
64 stars 41 forks source link

Functional Preprocessing Nipype Errors #436

Open ebridge2 opened 9 years ago

ebridge2 commented 9 years ago

These errors refer to a functional preprocessing pipeline, created using the command:

preproc=create_func_preproc(use_bet=True)

1) Erode_one_voxel is lacking an input traited spec. All of the other functional preprocessing methods have one.

2) Deoblique is missing an out_file location in its traited spec.

3) Several of the functions do not use the out_file spec, even when it exists in the function and has been user defined, when determining the location of their outputs. I found that 3 of the functions did not place the output in the location I specify in the out_file (I cannot remember which exactly, as I removed the folder when I realized a ton of the outputs were missing). I can find out for you by running another test in the morning if you are curious.

That's pretty much all I've found thus far when going through the module. As a side note, I could not find scan parameters as traited specs for the preprocessing module. I tried using the phrases scan_params (what the comments in func_preproc.py says they are called), scan_parameters, and parameters in every combination as a spec (ie preproc.scan_params) and as a spec within the inputs spec (ie preproc.inputs.scan_params). I am not sure if this is by accident or if you guys found that they don't make a significant impact on the preprocessing or something.

@jovo

anibalsolon commented 5 years ago

Thank you @ebridge2 for the feedback.

Although an old issue, I will check for the problems you were having.

The scan parameters are used earlier in the pipeline, for slice timing correction. Maybe slice timing correction should be folded into create_func_preproc, but I need to check the impact of this change.