Deleting the CCPP subdirectory

[nbren12]

Would there be any objections to deleting the CCPP subdirectory since we don't actually use this for any builds? I think this has been proposed in the past by e.g @ofuhrer.

It's a potential source of conflicts with upstream, and it seems like a good opportunity to delete some effectively dead code.

I ran into this when comparing our code to the version Alexie and Vladimir are using for emulation: https://github.com/AlexBelochitski-NOAA/fv3atm_full_physics_nn_emulator.

[ofuhrer]

Not sure if I suggested it, but I tend to agree with this. This directory introduces a lot of duplicate and dead code which is not helpful when searching, editing, grepping, ... throughout the code base. Adding it back would be trivial in case needed.

[nbren12]

@frodre you were using this at some point I recall, but needed to compile from the top-level UFS repo?

[frodre]

@nbren12, sorry I forgot to reply to this. I'm fine with removal too. We were compiling UFS at one point, but I don't plan or think we have plans for using that anytime soon.

[nbren12]

Thoughts @mcgibbon?

[mcgibbon]

I reference this directory periodically but I can go to the CCPP repo directly instead. The CCPP subdirectory is the main place I go when I want to understand what a part of the physics is actually doing. The main risk of looking elsewhere for this info is that upstream CCPP may receive updates our code will not.

[nbren12]

Ok. I think I'll remove it. I agree the CCPP package is easier to follow, but there are no guarantees that the CCPP code corresponds to the gfsphysics (although in practice this is true for some configs).