FIX: BET raising "No image files match: ..." with very long file names

[ZviBaratz]

Trying to fix #3211.
@satra or @oesteban, I'll try a couple of things and see if it works, but I'm not confident if I completely understood:

2. make the output filename be local (no absolute path) for generation, but not for list_outputs.

If you could help clarify how exactly this would be best implemented, I would be happy to update the PR and hopefully resolve this issue.

• (Mandatory) I acknowledge that this contribution will be available under the Apache 2 license.

[codecov]

Codecov Report

Merging #3309 (489a48d) into master (47fe00b) will increase coverage by 0.00%.
The diff coverage is 77.77%.

@@           Coverage Diff           @@
##           master    #3309   +/-   ##
=======================================
  Coverage   64.70%   64.70%           
=======================================
  Files         302      302           
  Lines       39869    39875    +6     
  Branches     5288     5289    +1     
=======================================
+ Hits        25796    25801    +5     
- Misses      12984    12985    +1     
  Partials     1089     1089           

Flag	Coverage Δ
unittests	64.70% <77.77%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
nipype/interfaces/fsl/preprocess.py	71.50% <77.77%> (+0.09%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 47fe00b...489a48d. Read the comment docs.

[ZviBaratz]

I changed _gen_outfilename() and _list_outputs() to the best of my understanding, but it doesn't seem to solve the problem (at least not in my use case).

[ZviBaratz]

Just for context: I encountered this issue trying to run MRIQC. It seems even running BET manually with the MRIQC workflow's relative work/ paths raises an error.

[satra]

@ZviBaratz - thank you for taking a stab at this. could you please tell me what the BET command (the cmdline that get's executed) looks like before it crashes?

[oesteban]

@satra #3211 (comment)

[ZviBaratz]

@satra This is the command generated by MRIQC when using nipype on this PR's branch:

bet /home/flavus/Projects/mriqc/work/mriqc_wf/anatMRIQC/HeadMaskWorkflow/_in_file_..media..veracrypt1..media..MRI..NIfTI..sub-772..ses-201905310935..anat..sub-772_acq-54363corrected_T1w.nii.gz/fsl_bet/sub-772_acq-54363corrected_T1w_conformed_corrected.nii.gz sub-772_acq-54363corrected_T1w_conformed_corrected_brain.nii.gz -A

For readability's sake:

in_file = "/home/flavus/Projects/mriqc/work/mriqc_wf/anatMRIQC/HeadMaskWorkflow/_in_file_..media..veracrypt1..media..MRI..NIfTI..sub-772..ses-201905310935..anat..sub-772_acq-54363corrected_T1w.nii.gz/fsl_bet/sub-772_acq-54363corrected_T1w_conformed_corrected.nii.gz"
out_file = "sub-772_acq-54363corrected_T1w_conformed_corrected_brain.nii.gz"

Unfortunately, unlike @oesteban case, even trying to run the generated command manually raises the same exception.
I realize my BIDS labels are slightly long, but I would imagine it to be an acceptable use case. What do you think?

[satra]

@ZviBaratz - the command line doesn't quite look right. the issue is that it's still reflecting the relative path. in this particular case you would do the same thing for in_file that you have done for out_file but by overriding the function format_arg

here is an example from FAST, which you can adapt for BET

nipype/nipype/interfaces/fsl/preprocess.py

Line 381 in e66bc41

def _format_arg(self, name, spec, value):

if you do this for in_file, it should reduce the path length.

[ZviBaratz]

@satra Thank you for clarifying this. I implemented what you suggested, and I am happy to report it indeed solves the issue (at least in my use case).

[effigies]

Looks sensible. Need to get CI back up and running, but I don't think we need to hold this up.