Two desc-<value> pairs in qsirecon output

[chiuhoward]

Summary

• QSIPrep version: 0.20.1.dev0+geda84d5.d20240112
• Docker version: NA
• Singularity version: apptainer version 1.2.5-1.el7

What were you trying to do?

I was trying to run pyAFQ with qsirecon output as input for pyAFQ.

What did you expect to happen?

I expected pyAFQ to pick up the ifod2.tck files.

What actually happened?

pyAFQ ignored the files which were not compliant as there were two desc- pairs qsirecon_errors.txt .

Reproducing the bug

[mattcieslak]

This is fixed in the next release. If you're interested in trying it out it's available under in the pennbbl/qsiprep:unstable image

[tsalo]

It looks like the PyAFQ outputs still have duplicate entities, like two descs and two spaces in sub-ABCD_acq-10per000_space-T1w_desc-preproc_dwimap/ROIs/sub-ABCD_acq-10per000_space-T1w_desc-preproc_dwi_space-subject_desc-ARCL1end_mask.nii.gz

[chiuhoward]

Hi @tsalo, what version of pyAFQ are you running? the ARC abbreviation in the ROI mask leads me to believe that this is pyAFQ<1.3

[tsalo]

That's just one of the outputs from the integration tests in https://github.com/PennLINC/qsiprep/pull/763, so it's whatever PyAFQ version is bundled into the current unstable version of QSIPrep.

[chiuhoward]

@tsalo this should have been fixed in the current release of pyAFQ, if i’m reading the code correctly…

anecdotally, my outputs are also BIDS compliant

[arokem]

Looks like qsiprep is pinning to <=1.3.1. Is there a reason to do that? Current pyAFQ is 1.3.2, which may fix this.

[tsalo]

Setting max versions makes it less likely that backwards-incompatible dependency releases will be merged in. I imagine @mattcieslak just hasn't gotten around to updating the pinned PyAFQ version.

[arokem]

OK - let's give it a try: https://github.com/PennLINC/qsiprep/pull/764