organize output of FFX and RFX in separate derivatives folder

[Remi-Gau]

Currently we have:

└── SPM12_CPPL
    ├── JOBS
    ├── group
    ├── sub-pilot001
    │   ├── ses-001
    │   └── stats
    └── sub-pilot002
        ├── ses-001
        └── stats

This makes it hard to export the stats data of all the subjects without taking the preprocessed data as well

What we at least want is:

└── SPM12_CPPL
    ├── JOBS
    ├── group
    ├── stats
    │   ├── sub-pilot001
    │   └── sub-pilot002
    ├── sub-pilot001
    │   └── ses-001
    └── sub-pilot002
        └── ses-001

What we ideally want is:

├── SPM12_CPPL - preprocess
│    ├── JOBS
│    ├── sub-pilot001
│    │   └── ses-001
│    └── sub-pilot002
│        └── ses-001
├── SPM12_CPPL - stats
│    ├── JOBS
│    ├── sub-pilot001
│    └── sub-pilot002
└── SPM12_CPPL - group
     ├── JOBS
     ├── sub-pilot001
     └── sub-pilot002

[Remi-Gau]

Another BIDS compliant option is one with nesting: https://bids-specification.readthedocs.io/en/stable/02-common-principles.html#source-vs-raw-vs-derived-data

<dataset>/derivatives/spm-preproc/sub-0001
<dataset>/derivatives/spm-preproc/derivatives/spm-stats/sub-0001

I am not a big fan because it makes it hard to travel up and down the hierarchy of files.

[Remi-Gau]

• needs to be merged after the next release

[Remi-Gau]

I also opened an issue in the Datalad handbook to see what type of data structure would be the most Datalad friendly.

datalad-handbook/book#618

I think this it the kind of thing that I would like to have a "democratic" decision on after discussing the pros and cons of it (and personal preference).