Open carlosmunozmoncayo opened 2 months ago
We did have something that would effectively call a Riemann solver in a mode that would be an actual Riemann problem. Something like that would be useful although I am not sure how hard it would be to hook up generally to the Fortran Riemann solvers that use things like modules. A Fortran test harness maybe appropriate even in that case.
It could be useful to have some testing workflow for the Riemann solvers. I was thinking about writing some Pytest-based regression tests for the Python and f2py-compiled Fortran solvers. The idea would be to compare, for a set of problems, the output of each solver (waves, speeds, normal and transverse fluctuations) with some reference values stored in dictionaries, instead of focusing on the Riemann solution itself. Would something like this make sense?