Support pepolar fmaps

[yarikoptic]

Not yet sure if might be reproin heuristic specific. ATM for SE sequences intended to be pepolar fieldmaps https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/01-magnetic-resonance-imaging-data.html#case-4-multiple-phase-encoded-directions-pepolar we are getting files with _magnitude suffix, whenever it should be _epi.

Attn @amanelis17

[pvelasco]

Hi @yarikoptic,

Yes, it looks like the problem is with the reproin heuristic file. The pepolar fieldmaps are magnitude images, so they have the M label in the s.image_type and reproin will assign the seqtype_label magnitude.
What we do in our heuristic file is we also check the s.sequence_name: epse2d is for the pepolar fieldmaps, while the traditional GRE fieldmaps s.sequence_name is fm2d. (NB: This is for a Siemens Prisma)
So, I would do something like this:

if seqtype == 'fmap' and not seqtype_label:
    if 'fm2d' in s.sequence_name:
        if not dcm_image_iod_spec:
            raise ValueError("Do not know image data type yet to make decision")
        seqtype_label = {
            'M': 'magnitude',  # might want explicit {file_index}  ?
            'P': 'phasediff',
            'DIFFUSION': 'epi', # according to KODI those DWI are the EPIs we need
        }[dcm_image_iod_spec]
    else:
        seqtype_label = 'epi'     # (or whichever label you want to give it).

For other scanner vendors, you will need to check the sequence_name for the GRE fieldmaps and include them in the new if statement.

I hope this helps.

[yarikoptic]

THANK YOU @pvelasco ! I think in our (reproin) case we could judge based on having _dir- field present, which is defined/makes sense only for pepolar ones and not for generic fieldmaps