If no CSV supplied, read task directly from .sdf with mol.getProp()

      yield load_pickle_from_disk(input_file)

def load_sdf_files(input_files, clean_mols):

def load_sdf_files(input_files, clean_mols, tasks=[]):

  """Load SDF file into dataframe."""

  dataframes = []

  for input_file in input_files:

    # Tasks are stored in .sdf.csv file

    raw_df = next(load_csv_files([input_file + ".csv"], shard_size=None))

    # Tasks are either in .sdf.csv file or in the .sdf file itself

    has_csv = os.path.isfile(input_file + ".csv")

    # Structures are stored in .sdf file

    print("Reading structures from %s." % input_file)

    suppl = Chem.SDMolSupplier(str(input_file), clean_mols, False, False)

    for ind, mol in enumerate(suppl):

      if mol is not None:

        smiles = Chem.MolToSmiles(mol)

        df_rows.append([ind, smiles, mol])

    mol_df = pd.DataFrame(df_rows, columns=('mol_id', 'smiles', 'mol'))

        df_row = [ind, smiles, mol]

        for task in tasks:

          df_row.append(mol.GetProp(str(task)))

        df_rows.append(df_row)

    mol_df = pd.DataFrame(

        df_rows, columns=('mol_id', 'smiles', 'mol') + tuple(tasks))

    # Tasks are stored either in .sdf.csv file

    if has_csv:

      raw_df = next(load_csv_files([input_file + ".csv"], shard_size=None))

      dataframes.append(pd.concat([mol_df, raw_df], axis=1, join='inner'))

    else:

      dataframes.append(mol_df)

  return dataframes