Merge new variable structure and simple Python testing

[Arcadia197]

New variable structure.

• A variable can now simply be created with constructing
• CmmVar2D Var(NX, NY, NZ, BoundsPointer, Type)
• This class contains informations about the grid, sizes and the actual data on the device
• The same was done for particles (forwarded or advected)

Python testing

• wrote small tests built in Python, those run full simulations and later compare the output
• Included first tests for initializing the variables, testing time-stepping, remapping, as well as one broad one for euler2D and vlasov1D
• those tests rely on new debug settings: increasing verbose to 4 or higher disables all saving of variables but instead saves global attributes of the variables: Min, Max, Avg, L2, Hash
• the saving of the fields as well as remapping has been valiated with existing results for euler2D. I mirrored all changes to Vlasov1D but did not extensively validate it

Additional comments can be found in the individual commits. I hope that this clarifies the general structure and helps with streamlining the simulation variables

[Arcadia197]

I will do a test on Anticythere or Idris and once this succeeds I will merge it

[Arcadia197]

First test looked okay. The differences need to be adjusted, as they can range up to 10^-10. One variable bug was not yet applied for reference data. Particle P2 for remapping showed higher difference. I should check that

[Arcadia197]

I noticed issues with saving maps due to OOM error for the MapStack on Idris production run. This needs to be investigated

[Arcadia197]

I noticed issues with saving maps due to OOM error for the MapStack on Idris production run. This needs to be investigated

Those OOM errors were not related to the changes but rather slurm workload management not allocating enough memory.

[Arcadia197]

I consider this enough tested to be merged.

[Arcadia197]

Oh, I accidentely closed without merging