Explicit "intendedfor" annotation

[yarikoptic]

Absence of "IntendedFor" handling is a notable shortcoming. Recent issue within ReproIn about it was #36 (comment) by @dlevitas.

Although it might be possible to deduce it automagically (nipy/heudiconv#148 (comment)), and by default in case of a single fmap (or pair of se scans with differing _dir) to assign to all matching bold/dwi; I wonder if in ReproIn context we should better proceed with explicit convention. E.g. annotating both fmap and target scans with some _key-<index> (e.g. _fmap-1) to bind those together. With solutions for nipy/heudiconv#148 it would also be possible to verify that what was intended to be the fmap is compatible with the target scan.

ideas/feedback?

[satra]

sure, if there is an explicit match in named relation

annotating both fmap and target scans with some _key- (e.g. _fmap-1) to bind those together.

sounds reasonable. what happens for each of the different types of fmaps (pepolar, mag/phase, built in -such as pepolar diffusion scans)

[dlevitas]

I don't see anything pertaining to IntendedFor in the ReproIn code, so I'm assuming specifying a key to link fmap to corresponding func/dwi acquisitions wouldn't work at this point?

Not sure if this could also be lumped in with the heuristic-options

[yarikoptic]

see my initial review within nipy/heudiconv#482 .