convert fLoc/rois to a dataset and add as a submodule

[MarieStLaurent]

From local repo copy on beluga, convert fLoc/rois directory and its entire content into a cneuromod datalad dataset called floc.rois

Local path on beluga: projects/rrg-pbellec/mstlaure/cneuromod-things/fLoc/rois

Description: fLoc vision localizer derivatives and documentation, including contrast maps and binary ROI masks.

IMPORTANT: remove files downloaded from the Kanwisher group's website from the dataset? (or should we keep them to garante version control?). If so, remove fLoc/rois/standard_masks/kanwisher_parcels/cvs/cvs_*_parcels folders from final dset (currently present in my repo on beluga).

Then, install floc.rois as a submodule called rois directly under cneuromod-things/fLoc to replace the placeholder folder (not a repo). Update .gitignore file to track the submodule's content.