Simplifying structure of scripts for environmental forecasting usage

[JimCircadian]

The environmental forecasting pipeline changes some of the scripts significantly, so am using this as an opportunity to rationalise the scripts so that they're a bit more user friendly:

• Prepare data get the data assets output
• [x] Prepare datasets from configuration for training prep_training_data.sh
• [x] Prepare dataset from CLI date for prediction (e.g. for daily forecasting) prep_prediction_data.sh
• Run model get the model / prediction assets to the results folder
• [x] Train a single model run_training.sh
• [x] Train an ensemble run_train_ensemble.sh
• [x] Predict via a single model run_prediction.sh
• [x] Predict via an ensemble run_predict_ensemble.sh
• Other changes
• [x] Rename relevant post processing scripts to process_...<dot>sh or plot_...<dot>sh
• [x] Rename dataset utilities to dataset_...<dot>sh

Note much of this is in place already, but there is a clearer delineation between data and ML operators, as well as other processing type activities. This should make documenting and navigating the repository easier.

Thoughts @bnubald?

[bnubald]

So, pre-process scripts would be denoted by prep, and post-process by process (except for plotting scripts)?

If so, LGTM.