while aggregating task- files I guess we should gather records and see if we should add/extend/update records in dataset_description.json . Might be tricky to do reliably but we should try.
Trickier would be to annotate the heuristic since we do not have consistent "naming" or "versioning" mechanism for them if they are not coming from heudiconv and later to reassess their identity/version from a copy under .heudiconv/. So may be we should add some versioning way etc.
ATM we are recording ad-hoc fields for heudiconv and dcm2niix does the same. Since released in BIDS 1.9.0
Trickier would be to annotate the heuristic since we do not have consistent "naming" or "versioning" mechanism for them if they are not coming from heudiconv and later to reassess their identity/version from a copy under
.heudiconv/
. So may be we should add some versioning way etc.