FIX: Correctly check SDCFlows registry of IntendedFor files

[oesteban]

Follows up on #140

EDIT 12/17/2020 --

Summary

This PR explores the new SDCFlows API to apply fieldmaps. Although the control over the available fieldmaps is almost none and many issues are still to be solved, this is a second step towards the integration.

Now, the CircleCI reports contain SDC results reportlets for ds001771 - one of them looks decent, the other needs some investigation.

This dataset contains two fake _fieldmap images - these are the two with the weird mask (the problem is that SDCFlows expects a magnitude image, but receives an EPI, because these are fake fieldmaps derived from the TOPUP solution).

[oesteban]

Failing for the same root cause of nipreps/sdcflows#158 (broken dependency of scikit-image).

[josephmje]

This dataset contains two fake _fieldmap images - these are the two with the weird mask
(the problem is that SDCFlows expects a magnitude image, but receives an EPI, because these
are fake fieldmaps derived from the TOPUP solution).

Ha! I tested this out earlier and was wondering why the fieldmaps looked the same as the TOPUP solution.

[oesteban]

Ha! I tested this out earlier and was wondering why the fieldmaps looked the same as the TOPUP solution.

Wonder no more. I think this is something that @edickie spotted a long while ago.

[josephmje]

Looks good to me! I just made a few small suggestions for the wording of the reportlet. Waiting for the tests to run to see what you mean by the pre-processing of the other fieldmap.