Dealing with intendedFor and "file dependencies"

[Remi-Gau]

So I would not necessarily think this is needed for events but definitely, we need a way to deal with the intendedFor for fieldmaps, especially that it is being reused for other data types as well (ASL, PET...) .

@nbeliy

The new PR for ASL seems to incorporate the idea you had for "file dependencies" as aspects of ASL includes the "intendedFor" field

See PR #127 and also the code here: https://github.com/ExploreASL/bids-matlab/blob/6cba2ee686fc26fb9b182f6a50a99ae15a0a11fa/%2Bbids/layout.m#L481

I would suggest that we start small and tackle first those "explicit" dependencies covered by the specification before we move on to those that are more use case dependent.

What do you think?

Originally posted by @Remi-Gau in #122 (comment)

[Remi-Gau]

from @nbeliy

I would suggest that we start small and tackle first those "explicit" dependencies covered by the specification before we move on to those that are more use case dependent.

In my idea dependency is to revert IntendedFor. A fmap image do not have a value by itself, and it is not that important to know for what scan it was taken. But for a given scan it is important to have an associated fmap. Actually it would be nice to have it in BIDS itself, but I guess is too late.

From technical point of view, it can be done at the end of layout function. Just scanning all images for IntendedFor field, and adding found files to corresponding scans.

[Remi-Gau]

I agree that this makes more sense when using the data indeed.

@HenkMutsaerts what do you think?