Decide on examples directory structure and provide an organized library of standard test cases

[tristanmontoya]

We should definitely organize the examples directory into subdirectories as this package becomes larger, but I am not sure if it makes sense to do so based on solver/mesh type the way Trixi.jl does. Since this package already provides several independent features relating to the physical aspects (i.e. the moist Euler equations) and geometrical aspects (i.e. solving PDEs on 2D manifolds in 3D) of atmospheric modelling, it is perhaps most natural to split the examples in such a way, for example, into examples/moist_euler/ and examples/spherical_advection/.

A separate question that we will need to address at some point is how to organize the subdirectories in src/ and whether a similar division makes sense there, or if we should continue to mimic the structure of Trixi.jl.

[benegee]

Absolutely agree! In general and also regarding the first level of our hierarchy: equation for moisture and PDEs on manifold are our current main features.

For the moist equations I would further like to keep information about the concrete set of equations (e.e.g authors) and the mesh type. This could be another level in a directory hierarchy or part of the file name.

As always we can and should revise this as soon as some directories get cluttered up. I would suggest the same for /src.

[tristanmontoya]

Following Oswald's suggestion at the TrixiAtmo meeting, I propose that we provide a library of standard test cases and reference values that can be used with the different models. My latest PR #31 begins this effort with the file equations/reference_data.jl which includes a Gaussian bell case for spherical advection, and defines some constants for reference geophysical parameters.