Should we create a `--group-label` parameter?

[tsalo]

Should we add a --group-name/--group-label parameter to write out summary figures, reports, and aggregated TSVs for a given group-level run to a unique set of filenames?

[effigies]

--group-label would make sense as long as there's a group entity. How are groups determined? I would expect a column in a participants.tsv file, which you could then subselect with --group-label, analogous to --participant-label.

[tsalo]

I guess I saw it as determining the output directory more than determining which participants to use, but I like your idea too.

In my proposal, I saw it as something like this:

app \
    /dset/derivatives/preproc \
    /dset/derivatives/postproc \
    group \
    --participant-label a b c \
    --group-level subset01

That would produce the following in `/dset/derivatives/postproc:

grp-subset01.html
grp-subset01_seg-Schaefer_stat-pearsoncorrelation_relmat.tsv

With your idea, I imagine it would basically be the same, except it wouldn't require --participant-label a b c, right?

[effigies]

I see. Yeah, I suppose you could say --group-label and look it up in participants.tsv, and if that fails, interpret that as assigning a label to the selected participants (default all).