search

anthofflab / MimiFAIRv1_6_2.jl

MIT License
0 stars 1 forks source link

Add Testing Code #11

Open FrankErrickson opened 2 years ago

FrankErrickson commented 2 years ago

Ideally, we would run the Python code and directly compare the output to the Julia results under identical emissions and parameter settings (rather than just saving some CSVs of Python output).

@lrennels Think it'd be better to just place the FAIRv1.6.2 Python code in a testing folder here and run it that way, or do something fancier and try to call in Python code as a package from Julia?

lrennels commented 2 years ago

@FrankErrickson I've struggled with this, partially because I'm not as adept with the PyCall as I'd like to be ... maybe because I"m not as adept with Python as I'd like to be :).

So I agree with the ideal, but if it's too painful I think a reasonable happy medium is to put the scripts that run the Python within this folder, have them print CSVs, and then run that (once). Then for the regression testing etc. we can just run Julia code to compare to the CSVs, but it's very clear to us and future users how we got those CSVs. Thoughts?