search

TIGRLab / SCanD_project

base repo for the SCanD codebase - meant to be forked/cloned for every SCanD dataset
MIT License
1 stars 11 forks source link

[BUG] qsirecon 2 (i.e. enigma dti participant) not robust to many study schemes #56

Closed edickie closed 4 months ago

edickie commented 4 months ago

In my test run of qsirecon 2, it failed because the dwi filenames had "run" in them.

I also suspect (from reading the script) that the script would fail for cross sectional studies (with no session id).

We might need to recode the file finding in python using pybids?

GhazalehManj commented 4 months ago

Why did you change qsiprep container to enigma in qsirecon step2?

edickie commented 4 months ago

FSL's dtifit is present in both the qsiprep container and the tbss container - so either container could work for that step.

Although I think we need a python wrapper for DTIFIT that finds the preprocessed DWI files - because the bash code for finding DWI files (as currently written is failing for some projects.

GhazalehManj commented 4 months ago

I see, I'll edit the code.

GhazalehManj commented 4 months ago

Done. I added multiple conditions to capture "run" and "acquisition". About session, all scans (longitudinal, cross-sectional) have session part, right?

edickie commented 4 months ago

In the BIDS spec, cross-sectional projects can exist without sessions - (i,e. func and anat folder appear directly under teh subjects folder) (an example here: https://openneuro.org/datasets/ds001132/versions/1.0.0 Most of the labs data is longitudinal (or we add session for no reason). But our pipelines should be able to work with either version.

GhazalehManj commented 4 months ago

Done!