uf3.data.geometry

This module provides functions for generating supercells for machine learning while avoiding the need for minimum-image convention.

Functions

generate_displacements_from_forces	WIP implementation of data augmentation as introduced in https://doi.org/10.1038/s41524-020-0323-8
get_supercell	Generate supercell, centered on original unit cell, with sufficient number of images along each lattice vector direction to ensure that atoms within the unit cell may interact with neighbors (in periodic images) at distances of r_cut or more.
get_supercell_factors	Identify minimum number of replicas along each lattice vector direction to ensure that atoms within the unit cell may interact with neighbors (in periodic images) at distances of r_cut or more.
sort_image_indices	Sort image indices based on distance to origin.