While the files look like they are a mess, there is manner to the madness.
The first file to run is the R file 'create_hpc_experiment_files.R' These will build the python files for running an experiment, and will build the PBS files for running the experiments on the HPC. It will create a file called EXPERIMENT_LIST.csv, which contains a list of the experiemnts, the training and testing ranges, and the model used in that experiment.
You can download all the dataset from good as long as you have the google sheet id. You'll need a token.pickle and credntials.json to access GoogleSheet.
The resultant PBS files can be submitted to the HPC queue for processing
The __pyfile__01_build_models is the temptlate PY file. You can adjust TRAINING_RANGES, PREDICTION_RANGES, MODE_NAME, ANALYSIS_METHOD from lines 53-56 to have a custom and specific file to run.