runnerase-core

Shared core library for the Python interface to RuNNer, the Ruhr university Neural Network energy representation. Every other runnerase package (runnerase-prediction, runnerase-training, runnerase-workflows, runnerase-feature-calculation) builds on top of runnerase-core.

Provides

• SymmetryFunction, SymmetryFunctionSet, CutoffFunction: symmetry function definitions used to describe atomic environments to RuNNer.
• read_inputnn / write_inputnn: read and write RuNNer's input.nn configuration file.
• generate_default_symmetry_functions, generate_symmetryfunctions_sftype2, generate_symmetryfunctions_sftype3, prune_feature_maps: routines for generating and reducing symmetry function sets.
• Utilities for working with ASE Atoms objects (enforce_total_charge, get_minimum_distances, get_triplets, extract_environment, ...) and evaluating fit quality (calc_mae, calc_rmse).

Usage

from runnerase_core import read_inputnn, generate_default_symmetry_functions

parameters = read_inputnn("input.nn")
sfs = generate_default_symmetry_functions(elements=["H", "O"])

Documentation

Full documentation, including tutorials on configuring symmetry functions and reading/writing input.nn: https://runner.theochem2.rub.de/runnerase-docs/latest/

License

GPL-3.0-or-later. Part of the runnerase project.