Make it all to be a BIDS study dataset type?

[yarikoptic]

Congrats on the great work/product!

BIDS allows to structure entire studies following BIDS as for placement of its source data and derivatives. That simplifies then review etc across studies. This repository could have been like that and become a "proper" BIDS dataset. Pretty much it would just entail not centering around paper (? if that the center since unclear) and thus hiding datasets somewhere under data/ :

hyperface-data-paper/
├── data/                           # BIDS dataset (OpenNeuro ds007329)
│   ├── stimuli/                    # 707 video clips (4s each)
│   └── derivatives/
│       ├── fmriprep/               # fMRIPrep 25.1.4 (OpenNeuro ds007384)
│       ├── freesurfer/             # FreeSurfer 7.3.2 (OpenNeuro ds007378)
│       ├── qa/                     # QA outputs (generated by this repo)
│       ├── stimuli/                # Stimulus descriptors (embeddings, RDMs, ratings)
│       └── jiahui-2023/            # Data from Jiahui et al. (2023) PNAS
├── src/hyperface/                  # Python analysis package
└── scripts/qa/                     # QA pipeline scripts

but smth like what you would get at large after

mkdir -p sourcedata/
git mv data/ sourcedata/rawbids  # or alike
mkdir derivatives/
# re-populate those subdatasets hidden within rawbids/derivatives right here
mkdir code
git mv src/* scripts/qa code/

? paper materials could go e.g. under docs/ if desired etc. Then anyone with BIDS awareness who would get to it would know where to find everything without digging.

WDYT?

[mvdoc]

Yarik, my head is spinning...I haven't thought of that, now I see what you meant.
My mental model was that this repository would be an analysis code repository that installs datasets/derivatives. You're suggesting that the code/repository instead is just a submodule of the BIDS dataset. So the BIDS dataset is the main entity and everything else is dependent.

I think I can do that. I'll wait for the reviews to come back to make the change after the reviewers have looked at this repo first.

Thanks for the feedback!!

[yarikoptic]

yes -- BIDS "study" dataset, not BIDS "raw" dataset (as you made it somewhat). I am slowly brewing collection of those across OpenNeuro and will stage yours to come "earlier" than later ;) -- may be we could add another twist there -- to join forces kinda on this study dataset. Have a look at https://github.com/OpenNeuroStudies/OpenNeuroStudies and if you did not use yet https://www.visidata.org/ -- pip install it right away to checkout those studies.tsv and studies+derivatives.tsv for starters ;)