Restructure Repository

CAM_interface/README.md

@@ -17,6 +17,7 @@ require revisions to these files in future.
 In addition to the files in `[/NN_Module](/NN_Module)` containing the parameterisation,
 the additional files here are required:
 
+- `nn_interface_CAM.F90` - The interface for performing conversion from CAM variables and grid into the variables/grid expected by the YOG parameterisation.
 - `yog_intr.F90` - The interface between the CAM model and the YOG parameterisation.
 
 Changes also have to be made to:
@@ -27,3 +28,9 @@ Changes also have to be made to:
 - `/src/physics/cam/physpkg.F90` - To call the new parameteristion routines detailed in `yog_intr.F90`.
 
 Full details of these can be seen in the specific implementation in CAM linked above.
+
+
+## Tests
+
+There are test routines associated with this code in `/tests/test_CAM_interface/`.
+Guidance on running these can be found in the main README.

README.md

@@ -27,31 +27,58 @@ Long term developments of this project will seek to re-deploy more complex ML pa
 ## Repository structure
 
 ```
-├── NN_module
+├── YOG_convection
 │   └── ...
 ├── torch_nets
 │   └── ...
-└── CAM_interface
+├── CAM_interface
+│   └── ...
+└── tests
     └── ...
-```
-
 
+```
 
 ### Contents
 
-### `NN_module/`
-This folder contains the fortran neural net extracted from the [code referenced above](https://github.com/yaniyuval/Neural_nework_parameterization/tree/v.1.0.3), along with any dependencies, that may be compiled as a standalone fortran module.
-
-Currently there is code that can be built on CSD3 using the included shell script.
-This now needs cleaning up, testing, and a proper makefile creating (see open issues #9 and #10).
+### `YOG_convection/`
+This directory contains the fortran code for the YOG convection parameterisation, along with any dependencies. 
+It can be included as a standalone fortran module in a larger model.
 
 ### ``torch_nets/``
-The directory contains the PyTorch versions of the neural networks we are interested in.
+This directory contains the PyTorch versions of the neural networks used in the YOG convection parameterisation.
 
 ### ``CAM_interface/``
-The directory contains the additional files or details to interface the code with the CAM atmospheric model
+The directory contains the additional files or details to interface the YOG code with the CAM atmospheric model
 as part of the CESM model suite. It also includes a link to an implementation in a fork of CAM.
 
+### ``tests/``
+
+There are some tests for the NN_module and interface code in the `test/`
+subdirectory.
+These require a Fortran compiler, `netcdf` and
+`netcdf-fortran` libraries, and CMake to build.
+
+They can be built and run with CMake using the following commands:
+```
+cd tests
+mkdir build
+cd build
+cmake ..
+cmake --build .
+```
+This will create executables in the `build/` subdirectory which can be
+run to execute the tests using:
+```
+./test_YOG_convection
+./test_CAM_interface
+```
+with output printed to the console.
+
+Note: To build only a specific subset of tests instead of all of them use:
+```
+cmake --build . --target <name_of_test_executable>
+```
+
 
 ## Contributing
 

NN_module/README.md → YOG_convection/README.md

@@ -10,7 +10,6 @@ Files:
 
 - `nn_cf_net.f90` - The core neural net routines
 - `nn_convection_flux.f90` - wrapper around the neural net routines to perform physics-based operations
-- `nn_interface_SAM.f90` - wrapper around the parameterisation routines to interface with the SAM model
 - `test.f90` - simple smoke tests for parameterisation routines
 - `NN_weights_YOG_convection.nc` - NetCDF file with weights for the neural net
 - Makefile - Makefile to compile these files
@@ -26,11 +25,7 @@ python sounding_to_netcdf.py
 deactivate
 ```
 
-## Running tests
+## Tests
 
-To run the tests (requires `ifort` or `gfortran` compiler and and `netcdf` and
-`netcdf-fortran` libraries) edit `Makefile` to select the appropriate compiler
-(`ifort` or `gfortran`), set flags, and select files, and then build with:
-```
-make
-```
+There are test routines associated with this code in `/tests/test_NN_module/`.
+Guidance on running these can be found in the main README.