molprop.utils.analyze_mol_data

Functions

cross_check_data	Checks for overlaps between data sets If overlaps are found: returns pd.Series of overlapping entries between the two data sets Otherwise: returns None
extract_mol_features	Extracts all mol features for training, validation and test datasets.
get_mol_features	Creates dictionary containing atom and bond features.
get_mol_features_from_cfg
sanity_check_compounds
sanity_check_data	Checks for duplicates If duplicates are found: returns pd.Series of duplicates (input combinations that occur in duplicates) Otherwise: returns None --- check_conflicts checks for conflicting target values for duplicate input combinations (same input with different label) Note: can take some time if df is large
sanity_check_features	Checks for presence and variation of molecular features.
sanity_check_pipeline	Performs sanity checks for inputs and compund list.