We are working on #428 which will automatically order components. It would be useful to have integrated tests with lots of components to catch any instances of problematic ordering.
Because of this, I thought it would be a good opportunity to review which components are currently not covered by integrated tests.
Here is the list after https://github.com/boutproject/hermes-3/pulls is merged. These come in two categories - niche components and components linked to 3D turbulence:
I think some could be lumped together into a few tests (e.g. you can add into a 1D case). Some others are more tricky. Some should really be deleted (the solkit ones).
In an ideal world, all of the above would be covered by integrated tests that expose all of their settings combinations to ensure that the component sort algorithm works correctly. However, it would be quite a task to do this and we don't have the resources for it at the moment.
Maybe we could add a couple of tests specifically aiming to capture component order by featuring as many components as possible. For example, a 1D test based on 1D-dthe with classical_diffusion, fixed_fraction_ions, isothermal, detachment_controller, simple_pump and set_temperature. Maybe we can do something similar for 2D/3D turbulence?
We are working on #428 which will automatically order components. It would be useful to have integrated tests with lots of components to catch any instances of problematic ordering.
Because of this, I thought it would be a good opportunity to review which components are currently not covered by integrated tests.
Here is the list after https://github.com/boutproject/hermes-3/pulls is merged. These come in two categories - niche components and components linked to 3D turbulence:
I think some could be lumped together into a few tests (e.g. you can add into a 1D case). Some others are more tricky. Some should really be deleted (the
solkitones).In an ideal world, all of the above would be covered by integrated tests that expose all of their settings combinations to ensure that the component sort algorithm works correctly. However, it would be quite a task to do this and we don't have the resources for it at the moment.
Maybe we could add a couple of tests specifically aiming to capture component order by featuring as many components as possible. For example, a 1D test based on
1D-dthewithclassical_diffusion,fixed_fraction_ions,isothermal,detachment_controller,simple_pumpandset_temperature. Maybe we can do something similar for 2D/3D turbulence?