runnerase-workflows 0.3.1

Collection of routines for applying the Ruhr University Neural Network energy representation (RuNNer).

runnerase-workflows

High-level workflows and command-line entry points built on top of the other runnerase packages — for users who want a ready-made procedure rather than assembling one from runnerase-core/runnerase-prediction/runnerase-training primitives.

Provides

• ActiveLearning, DisagreementFunctions, Settings: active-learning loop for iteratively improving a potential based on prediction disagreement.
• RuNNer-LAMMPS interface helpers (setup_runner_context, run_md, execute_runner_simulation, format_cmd_pair_style, ...) for running MD simulations with RuNNer potentials in LAMMPS.
• prepare_prediction_from_trained_potential: set up a prediction run from an already-trained potential.
• open_ovito: visualize structures and trajectories with OVITO (requires the optional ovito extra).

Command-line tools

This package also installs several CLI entry points:

• active-learning
• dataset-reduction
• prepare_prediction_from_trained_potential
• enforce-total-charge
• open_ovito
• prune_feature_maps

Installation

uv pip install runnerase-workflows
# with OVITO visualization support:
uv pip install "runnerase-workflows[ovito]"

Documentation

Tutorials on active learning, structure visualization, and LAMMPS-based simulations: https://runner.theochem2.rub.de/runnerase-docs/latest/

License

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