changed constants to comply with sklearn type_of_target
changed dataset_property flag for few keys like X_train, train_indices as these will be added by the AutoML class after performing splits on the dataset.
added interpret_dataset_properties to tabular dataset
Changes made: