Change coordinates to reflectometry coordinate system

[arm61]

Currently an internal (or maybe NeXus) standard is used for what is x, y, and z. This doesn't work for reflectometry where there is already an accepted standard that depends on the sample orientation (see https://orsopy.readthedocs.io/en/latest/orsopy.fileio.base.html#orsopy.fileio.base.ValueVector). The Amor code should be modified to adhere to this reflectometry standard.

[SimonHeybrock]

@jokasimr Is this still relevant with the recent rewrite?

[jokasimr]

Yes it is still relevant. But I'm not sure if we want to do anything about it. It would make using the coordinate transformation code in scippneutron more complicated.

Essreflectometry uses a global coordinate system similar to the rest of the instrument workflows. That is, sample is origin, z is the 'beam direction' (this becomes fuzzy when it comes to instruments with focusing like amor and estia, but it's basically the vector from sample to the the detector, projected to the horizontal plane), y is up, and x is whatever is left.

[SimonHeybrock]

Is there a way to avoid referring to x, y, z at all, e.g., by callings things up, sample_normal, etc.?

[jl-wynen]

This can mask some differences. But there is a fundamental difference between reflectometry and the other techniques. Normally, we define

• one vector along beam
• one vector along gravity
• one vector in terms of first three to make a right-handed orthogonal system

But in reflectometry it seems that people use

• one vector along beam
• one vector normal to the sample
• one vector based on the other two In this setup, you don't get an orthogonal coordinate system.