GK checks on right handed coordinate system

[manauref]

Add check to ensure that the GK coordinate system is right handed.

Most importantly, this PR changes the definition of y from y=-alpha to y=alpha, and we change the signs in the definition of the binormal coordinate y (and the size of B_Z) in IWL mapc2p input files.

NOTE: the IWL input files don't pass the consistency checks. We've tried to investigate this some, even completely rewriting the mapc2p for those input files, but have not figured out why they fail. We leave it as failing at the moment, but at least now we know. It prints messages like

Running command: srun -u -n 4 ./gkeyll_20251111_180358 -g -M -c 1 -d 1 -e 4  -o max_run_time=21600
Creating app object ...
calc_metric.c: B and mapc2p are inconsistent (hat{b} and e_3 are not parallel).

Unfortunately, because we need every part of the domain to perform these checks and alert us if they fail, every MPI process is printing this warning, so the log file will have a lot of these at t=0.

Below we do a comparison of a TCV IWL 3x2v simulation ran with this branch vs. one ran in main.

IWL test

Here we compare a TCV IWL 3x2v with fixed source ran in main vs in this righthanded_checks branch. Up to nearly 0.5 ms the moments look basically the same.

[tnbernard]

When is this check called? For any input file with mapc2p?

[manauref]

When is this check called? For any input file with mapc2p?

Yes, any input file with mapc2p or with tokamak numerical geo.

[manauref]

Re-opening:
The shift in twistshift BCs was wrong previously. With the correct shift there are issues due to production of high-k structures, but we need to do additional work to address that. For now, if someone wants to run as before, they can just decrease the shift (e.g. divide it by 2).