molprop.utils.utils.perform_train_test_split

molprop.utils.utils.perform_train_test_split(overall_dataset_path, model_name, split_scenario, smiles_columheader=None, temperature_columnheader=None) → None

Function to create train test splits, takes file path of original dataset and stores train and test set respectively.

Parameters:

• overall_dataset_path – Complete File path to overall dataset

• model_name – Name of model to be trained, needs to exist as folder in 03_model_input

• split_scenario – either random, tempExt (Temperature extrapolate) or molExt (Molecule extrapolate)