Add guide for using GauXC from C and Fortran

[awvwgk]

See wavefunction91/GauXC#172
See wavefunction91/GauXC#174

Preview at https://awvwgk.github.io/skala/gauxc/c-library.html and https://awvwgk.github.io/skala/gauxc/fortran-library.html

[kjh-giesbertz]

Some questions on the GauXC in C instructions

• It feels weird to start adding MPI guards when we ask in the cmake config to set MPI off. Perhaps one could move that to the end in a subsection specifying what changes are needed for MPI?
• I needed to comment the last 3 lines of cmake/skala-gauxc.cmake out to make the generation of build-files work, so the fix if the gau2grid::gg is not found
• Why is the Default on the enum GauXC_AtomicGridSizeDefault?

[awvwgk]

• I needed to comment the last 3 lines of cmake/skala-gauxc.cmake out to make the generation of build-files work, so the fix if the gau2grid::gg is not found

Good point, now this all handled in GauXC since wavefunction91/GauXC#169 was merged.

[awvwgk]

• It feels weird to start adding MPI guards when we ask in the cmake config to set MPI off. Perhaps one could move that to the end in a subsection specifying what changes are needed for MPI?

Since GauXC has MPI support I would like to highlight it as well. Also, it does impact the API for the runtime environment so it is something that can't be easily ignored. We are testing all examples with and without MPI to make sure they actually work correctly.

[awvwgk]

• Why is the Default on the enum GauXC_AtomicGridSizeDefault?

I follow the naming used in GauXC: include/gauxc/enums.hpp#L28-L39

[kjh-giesbertz]

my compiler complains it cannot find a matching argument pattern. Removing model fixed the problem in my case.

Suggested change

	call gauxc_eval_exc_vxc(status, integrator, p_s, p_z, model, exc, vxc_s, vxc_z)
	call gauxc_eval_exc_vxc(status, integrator, p_s, p_z, exc, vxc_s, vxc_z)

[awvwgk]

Model should be passed here to dispatch to Skala interface, if you can't find a missing interface you might not have Skala enabled in GauXC.

[kjh-giesbertz]

The mpi communicator error is now the same variable as the error in cli. I am not sure wether this in causing conflicts, but giving separate variables didn't solve the mpi issue in my case.