feat(rust/sedona-geometry): Let CrsTransform handle M coordinate

[yutannihilation]

This is a followup of #606. Before addressing ST_Force3DM() and ST_Force4D(), I want to confirm if my approach is correct.

This pull request changes CrsTransform to have two additional methods to handle M coordinate (accordingly, transform_coord_3d() is renamed to transform_coord_xyz). Maybe I could create a separate transform_m() method, but I thought it's better to do the transformation in a single call.

• transform_xym()
• transform_xyzm()

This also requires transform_and_write_coords() to extract and fill M coordinate.

[paleolimbot]

Thank you!

I think this is correct...perhaps a good auxiliary check for this would be to ensure that ST_Translate has tests for the case where the XYM geometries are input and there is a deltaZ, ensuring that the M values are untouched (I think these tests already exist). I suppose we don't have a deltaM argument there, but maybe we should add one and test that Z values are untouched for XYZ and XYZM input (not for this PR since I see the tests here).

I also wonder if we've made the transform slower, but we have at least one optimization that we haven't implemented yet (transform arrays instead of scalar coordinates) and maybe that's the time to check in. Getting the default dimension math right is more important (we can always optimize easy cases later).

[yutannihilation]

Thanks!

I also wonder if we've made the transform slower

This was my main concern. But, I agree with you that we can optimize later.