Use SDF inputs

By default presto reads SMILES strings under param_settings.molecules. To use one or more SDF files instead, set molecule_input_type: sdf.

CLI form

presto train \
    --param-settings.molecule-input-type sdf \
    --param-settings.molecules input_molecule.sdf

For multiple SDF files, repeat the --param-settings.molecules flag, or use the YAML form.

YAML form

param_settings:
    molecule_input_type: sdf
    molecules:
        - ligand_a.sdf
        - ligand_b.sdf

Multi-molecule SDF behaviour

Each .sdf may contain one or more molecules. All molecules across all SDF files are loaded into the same list and fitted simultaneously (see Fit a congeneric series for the shared-parameter recipe). Molecule indices in output filenames (*_mol0, *_mol1, …) match the load order.