Split data wtih spatial nestedness

[slevin75]

This PR includes the following updates:

• p4_nested_groups uses the add_nested_group_id function to determines which gages in the nested_gages matrix are nested based on the nested_threshold. Then it assigns a unique nested_group_id to every un-nested gage or groups of nested gages.
• split_data function was updated to ensure that any gage with the same nested_group_id are not split between training and testing datasets. If any nested gages were split up, it takes the gages in the testing dataset and adds them to the training dataset, then finds an equal number of un-nested gages in the training dataset to replace the moved gages in the testing data set.
• post_process_cvfold : this function post processes the vfold_cv function to ensure that any nested gages are not separated between the in / out of sample data using the same process as for the split_data function.
• Updated the Boruta_screening function calls and train_models_grid to work with new functions.

[slevin75]

OK - I think I got all the changes in there. I also split out part of the repeated lines of code that post process the split_function and cv function into a separate function. I reran the pipeline again and it went to completion.