This repo is for model testing. It provides a test model for testing Dojo's model registration workflow.
Install with pip install -r requirements.txt.
Next, edit the rainfall setting in parameters.json to a float between -1 and 1 (e.g. 0.7), where 0 is no percentage perturbation to rainfall (baseline).
Next, edit the color_hue setting in fakeDataType.json to a str that is one of the following colors:"blue","red","yellow","green","purple","orange".
Then run with python3 main.py --temp 1.2 where the value passed to --temp is the percentage perturbation to temperature (0 is baseline).
Outputs are stored to output/output_{{ rainfall }}_{{ temp }}.csv (based on the model parameters). For example, output/output_0.7_1.2.csv.
Now, this model also produces an Excel file, Geotiff, and NetCDF. The Geotiff and NetCDF files are based on an interpolated grid.
docker build -t test-model .
then set parameters.json with the value of your choosing and run:
mkdir output
docker run -v $PWD/configFiles:/model/configFiles -v $PWD/output:/model/output --entrypoint python3 test-model /model/main.py --temp=0.5