BatchDosFeaturizer

BatchDosFeaturizer#

class lobsterpy.featurize.batch.BatchDosFeaturizer(path_to_lobster_calcs, add_element_dos_moments=False, fingerprint_type='summed_pdos', normalize=True, n_bins=56, e_range=[-15.0, 0.0], n_jobs=4, use_lso_dos=True)[source]#

Bases: object

BatchFeaturizer to generate Lobster DOS moment features and fingerprints.

Parameters:

  • path_to_lobster_calcs (str | Path) – path to root directory consisting of all lobster calc

  • add_element_dos_moments (bool) – add element dos moment features alongside orbital dos

  • normalize (bool) – bool to state to normalize the fingerprint data

  • n_bins (int) – sets number for bins for fingerprint objects

  • e_range (list[float]) – range of energy relative to fermi for which moment features needs to be computed

  • n_jobs – number of parallel processes to run

  • fingerprint_type (Literal['s', 'p', 'd', 'f', 'summed_pdos', 'tdos']) – Specify fingerprint type to compute, can accept {s/p/d/f/}summed_{pdos/tdos} (default is summed_pdos)

  • use_lso_dos (bool) – Will force featurizer to use DOSCAR.LSO.lobster instead of DOSCAR.lobster

get_df()[source]#

Generate a pandas dataframe with all moment features.

Moment features are PDOS (optional: element dos) center, width, skewness, kurtosis and upper band edge.

Returns:

A pandas dataframe with moment features

Return type:

DataFrame

get_fingerprints_df()[source]#

Generate a pandas dataframe with DOS fingerprints.

Returns:

A pandas dataframe with fingerprint objects

Return type:

DataFrame