add BIDS conversion script

[loj]

No description provided.

[yarikoptic]

I now stronger believe that some of those .txt could and should be renamed to .tsv. I would also check on naming of output files for motion correction etc produced by fmriprep which I believe is pretty much the guide for BIDS derivatives layout support.

[yarikoptic]

A bit controversial, feel free to ignore I guess:
FWIW, since renaming anyways (it is sad that they couldn't be modified), might be worth looking into changing extension at least for some of them? I see there .csv and may be some .tsv (although without headers) files:

$> head -n 2 Results/rfMRI_REST1_LR/*.txt             
... stripped since might be sensitive data ;)

==> Results/rfMRI_REST1_LR/ReclassifyAsNoise.txt <==

==> Results/rfMRI_REST1_LR/ReclassifyAsSignal.txt <==

==> Results/rfMRI_REST1_LR/rfMRI_REST1_LR_Atlas_stats.txt <==
TCSName,NumSignal,NumNoise,NumTotal,MEAN,STD,COV,TSNR,OrigVar,HighPassVar,MotionVar,StructNoiseVar,BOLDVar,UnstructNoiseVar,TotalSharedVar,CNR,OrigMGTVar,HighPassMGTVar,PostMotionMGTVar,CleanedMGTVar,OrigMGTbeta,HighPassMGTbeta,PostMotionMGTbeta,CleanedMGTbeta,HighPassVarRatio,MotionVarRatio,StructNoiseVarRatio,BOLDVarRatio,UnstructNoiseVarRatio,OrigMGTVarRatio,HighPassMGTVarRatio,PostMotionMGTVarRatio,CleanedMGTVarRatio,CleanedMGTVarVsBOLDVarRatio,OrigMGTrsq,HighPassMGTrsq,PostMotionMGTrsq,CleanedMGTrsq,UnstructNoiseMGTbeta,UnstructNoiseMGTVar,UnstructNoiseMGTVarRatio,UnstructNoiseMGTrsq,NoiseMGTbeta,NoiseMGTVar,NoiseMGTVarRatio,NoiseMGTrsq,WMbeta,WMVar,WMVarVsBOLDVarRatio,WMCleanedBOLDVar,WMCleanedMGTbeta,WMCleanedMGTVar,WMCleanedMGTVarVsWMCleanedBOLDVarRatio,CSFbeta,CSFVar,CSFVarVsBOLDVarRatio,CSFCleanedBOLDVar,CSFCleanedMGTbeta,CSFCleanedMGTVar,CSFCleanedMGTVarVsCSFCleanedBOLDVarRatio,WMCSFbetaWM,WMCSFbetaCSF,WMCSFVar,WMCSFVarVsBOLDVarRatio,WMCSFCleanedBOLDVar,WMCSFCleanedMGTbeta,WMCSFCleanedMGTVar,WMCSFCleanedMGTVarVsWMCSFCleanedBOLDVarRatio
... and there is more

I am not sure if there is a convention for an extension to indicate that file is just a single entry (thus could be .tsv .csv etc but without header).

Why asking, is because in general indication of format or structure in the file is recommended. and that is why BIDS atm has all .tsv's. (there are some .txt "leaked in" for MEG but that is about it... I put some flame into the fire where it is due).

[yarikoptic]

might be worth checking how current version fmriprep stores any such related derivatives, and follow along.

[yarikoptic]

checked -- this one is a .tsv (no header). BIDS likes .tsv.
Interestingly, BIDS seems to demands them to be .gz'ed! https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/06-physiological-and-other-continuous-recordings.html

But I think this should be sensible to ask to rectify: bids-standard/bids-specification#472

[yarikoptic]

anyone (@loj @adswa) would be interested to become a BIDS contributor and propose a PR which would make .gz optional for these files? There seems to be no negative votes so far in the BIDS issue ;)

[yarikoptic]

why not to

    for lr in LR RL; do
....

and avoid duplication -- would simplify adjusting it as well.

[yarikoptic]

here is a "fuller" paradigm which IMHO would also make review easier

# nice concise specific differences on what ot rename into what
echo "\
Movement_RelativeRMS.txt _desc-MovementRelativeRMS.txt
other another" \
| while read fr_ to_; do
    for lr in LR RL; do
        # one command to rule them all
        git mv git mv -fk ${subdir}/MNINonLinear/Results/rfMRI_REST1_${lr}/${fr_} ${newsub}/ses-3T/func/${newsub}_ses-3T_task-rest_dir-${lr}_run-1_${to_}
    done
done