CliMA / OceanParameterizations.jl

Machine learning (and uncertainty quantification?) of climate model parameterizations using differentiable (and probabilistic?) programming.
MIT License
21 stars 2 forks source link

Bit of cleanup #27

Closed ali-ramadhan closed 3 years ago

ali-ramadhan commented 3 years ago

This PR just creates a OceanParameterizations.PhysicalParameterizations submodule and makes coarse_grain work again.

It also gets rid of the OceanTurb existing identifier conflict warnings (since both OceanTurb and Oceananigans define Cell and Face):

WARNING: using OceanTurb.Cell in module OceanParameterizations conflicts with an existing identifier.
WARNING: using OceanTurb.Face in module OceanParameterizations conflicts with an existing identifier.
ali-ramadhan commented 3 years ago

This PR now also moves my old stuff from ocean_convection.jl into a local FreeConvection submodule. This means OceanParameterizations.jl itself doesn't need to depend on heavy packages like DifferentialEquations.jl so it precompiles more quickly now.

@adelinehillier I also added using Statistics to NeuralNetworks.jl so it might fix #23.

ali-ramadhan commented 3 years ago

No worries! Can always work on other branches so no rush on reviewing PRs!