To disentangle the verifications from MRtrix3 as much as possible, the header transforms should be obtained without using mrinfo.
For NIfTI, nibabel might be the most unbiased.
Currently, header transform can be parsed from .mih simply. To do this from a .mif (which is currently used if optional metadata are exported to a JSON) would require parsing the END delimiter; or could just use .mih in both scenarios.
Lestropie
commented
4 days ago
To disentangle the verifications from MRtrix3 as much as possible, the header transforms should be obtained without using
mrinfo.nibabelmight be the most unbiased..mihsimply. To do this from a.mif(which is currently used if optional metadata are exported to a JSON) would require parsing theENDdelimiter; or could just use.mihin both scenarios.