transforms new_capi version of gvec_to_xy does not use optional arguments bmat and v_inv

[donald-e-boyce]

The arguments bmat and v_inv to gvec_to_xy are accepted but not used in the new_capi version of the transforms module. The xf_numpy also has those arguments but handles them.

The bmat matrix converts HKLs to the components of the normal vector in the crystallographic reference frame. If this is not None then the input gvec_c are expected to be HKLs. The v_inv argument accounts for the grain stretch in fitting.

However, I don't see any place where these arguments are actually used in calls to gvec_to_xy.

[donald-e-boyce]

Maybe we should just raise a NotImplementedError if those arguments are used.

[donald-e-boyce]

I don't think you're talking about the matrix bmat, which uses the lattice spacing to convert HKLs to crystal components. In xy_to_gvec, the argument rmat_b refers to the beam reference frame, in which the 3-direction is the beam and eta is the azimuthal angle calculated from (x, y) and the 1-direction vector (eta = 0). It is true the middle column is not used here.

[kevindlewis23]

You're right, deleted my comment

[psavery]

@donald-e-boyce Oscar had mentioned that there were some C functions that exist where Joel wanted to add new parameters/features, so he rewrote them in Python and added the new parameter/feature, and this resulted in an inconsistency between the C and Python versions (and the Python version would have to be used when those parameters were required).

Perhaps this is one of those cases?