FIX: Load ITK fields from H5 correctly

[effigies]

Ports nipreps/fmriprep#3300 to nitransforms.

I've gotten ANTs to generate a warp field with a non-cardinal rotation matrix and anistropic spacing:

Need to verify that we're raveling the fixed parameters in the correct order. Numpy defaults to C.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.79%. Comparing base (d6170fb) to head (1c74669).
Report is 36 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #211   +/-   ##
=======================================
  Coverage   95.79%   95.79%           
=======================================
  Files          14       14           
  Lines        1307     1307           
  Branches      259      259           
=======================================
  Hits         1252     1252           
  Misses         52       52           
  Partials        3        3           

Flag	Coverage Δ
unittests	95.79% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[effigies]

Oh no.

[effigies]

In [37]: np.allclose(np.diag([-1, -1, 1, 1]) @ targetrot, nb.affines.from_matvec(np.reshape(params, (3, 3))))
Out[37]: True