Make it easier to change the surface normal in Bragg2DPtycho.from_dataset

[yoshikisd]

Makes the surface normal vector a parameter in Bragg2DPtycho.from_dataset to make it easier for folks to define what the surface normal is.

[allevitan]

This is really nice, seems like this is a super helpful feature!

Right now, if I'm reading the code right, it's written so that if there is an 'orientation' defined in dataset.sample_info, then that will override the provided surface_normal argument, and the same if a 'scattering_mode' argument is set.

I think it would make the most sense if surface_normal > scattering_mode > dataset.sample_info['orientation'], i.e. an explicitly defined surface_normal overrides everything, a scattering_mode overrides the info in dataset.sample_info, and dataset.sample_info serves as a fallback. Does that make sense?

[yoshikisd]

Makes sense to me, and I think that sounds like a good idea. The latest push has this hierarchy implemented.

Currently, it's set so that surface_normal > scattering_mode > dataset.sample_info['orientation'] > np.array([0, 0, 1]) (default transmission mode). There is a RuntimeError implemented to prevent surface_normal from being anything other than a 3-element numpy vector. However, issues with scattering_mode or dataset.sample_info['orientation'] may silently cause a default transmission geometry to be used for the surface_normal. Maybe it might be a good idea to introduce guards for that as well, perhaps in a different PR?

[allevitan]

I just added a few suggested tweaks to the branch, let me know if they look good:

• I made the scattering_vector argument work with anything that can cast to a numpy array with np.asarray
• I broadened the check for length 3
• I stopped it from silently falling back to default if an invalid scattering_mode is given
• I slightly broadened the allowed strings for 'r' and 't' mode.

If it looks good, let me know and I can clear my comments and merge it.

[yoshikisd]

Thanks for making these changes; they all look good to me. Would you do the honors of merging?