Dev varRBE robopt world coordinate system

[amitantony]

This PR should replace the old world coordinate system PR #704

[amitantony]

comments for reviewers

1. Ive checked the examples run through , please let me know if there are new parts of code not covered in the examples that ive missed
2. would appreciate Reviewers with more MC experience to test out those parts of the code, results for Example 12 were quite close to that of the clean dev varRBErobOpt but Im unsure of how close, close should be
   cheers 🍻

[wahln]

Can you quickly write a summary about

• the different coordinates system now used and where there (default) origin is? I see the world system, which probably is given by ct.x/y/z and is defaulted like it was before. And I see the cube system which uses the i,j,k and translates directly to coordinates.
• Which entities are living in which coordinate system? Isocenter always in world? doseGrid.x/y/z also in world? When is the cube system used (only internal functions)?
• Maybe it would make sense to already consider the orientation of the system kn the functions somehow. Such that in the future, we could also transform easily between different medical systems / anatomical orientations.

[wahln]

Good work, but I think some stuff could be a bit cleaner. Also, there seems to be some regular transformations which could deserve their own functions and are not directly clear why they are done this way (maybe add comments).

[amitantony]

Summary of the different coordinate systems :

1. voxel coordinates ( i, j, k) in cube
   used for all handling within the cube and GUI

2. world coordinates (x, y, z) [mm]
   where the isocenter is defined, therefore consistent with all dicom patient positioning

3. raytracing coordinates, same as world but shifted so that isocenter is at origin (x, y, z)
   this is used by raytracing and all the dosecalc Functions

[JenHardt]

comments for reviewers

1. Ive checked the examples run through , please let me know if there are new parts of code not covered in the examples that ive missed
2. would appreciate Reviewers with more MC experience to test out those parts of the code, results for Example 12 were quite close to that of the clean dev varRBErobOpt but Im unsure of how close, close should be
   cheers 🍻

i think from a MC point of view you are good to go

[wahln]

Also there's a huge number of tests failing:
https://github.com/e0404/matRad/actions/runs/10060022215
Unfortunately currently printingthe failures do not work on all PRs, but it will as soon as the new pipeline is on the master-branch.

[wahln]

Looks good now!
However, I wonder how you manually tested the alignment of the grids? Just by looking at the dosimetric results (equivalence) or did you debug and check that similar values are generated?

[wahln]

Will take a look on why the tests don't run through.

[wahln]

Will take a look on why the tests don't run through.

My bad, it's only the failing coverage report in PRs.