autoplex.data.common.utils

Utility functions for training data jobs.

Functions

boltz	Calculate the Boltzmann factor for a given energy.
boltzhist_cur	Sample atoms from a list according to boltzmann energy weighting and CUR diversity.
check_distances	Take in a pymatgen Structure object and check minimum distances between atoms using minimum image convention.
convexhull_cur	Sample atoms from a list according to Boltzmann energy weighting relative to convex hull and CUR diversity.
create_soap_descriptor	Generate a SOAP descriptor string based on the given parameters.
cur_select	Perform CUR selection on a set of atoms to get representative SOAP descriptors.
data_distillation	For data distillation.
energy_plot	Plot the distribution of energy per atom on the output vs the input.
extract_base_name	Extract the base of a file name to easier manipulate other file names.
filter_outlier_energy	Filter data outliers per energy criteria and write them into files.
filter_outlier_forces	Filter data outliers per force criteria and write them into files.
flatten	Flatten an iterable fully, but excluding Atoms objects.
force_plot	Plot the distribution of force components per atom on the output vs the input.
mc_rattle	Take in a pymatgen Structure object and generates rattled structures.
parallel_calc_descriptor_vec	Calculate the SOAP descriptor vector for a given atom and hypers in parallel.
plot_energy_forces	Plot energy and forces of the data.
random_vary_angle	Take in a pymatgen Structure object and generates angle-distorted structures.
rms_dict	Compute RMSE and standard deviation of predictions with reference data.
scale_cell	Take in a pymatgen Structure object and generates stretched or compressed structures.
std_rattle	Take in a pymatgen Structure object and generates rattled structures.
stratified_dataset_split	Split the dataset.
to_ase_trajectory	Convert to an ASE .Trajectory.

Classes

ElementCollection	A class to handle different species operations for a collection of atoms.
GPa	Convert a string or number to a floating point number, if possible.
TYPE_CHECKING	bool(x) -> bool