FRUST Documentation

FRUST is a Python workflow package for building and screening frustrated Lewis pair substrate structures. It connects substrate inputs to calculation outputs: template-based structure generation, conformer handling, staged xTB and ORCA runs, UMA and g-xTB integrations, and parquet-backed result tables.

FRUST is active research software. The workflows are useful now, but APIs and defaults may still change as the project evolves.

Start Here

• Installation covers the basic Python package setup.
• External tool setup covers .env, xTB, ORCA, OET, UMA prerequisites, and g-xTB paths.
• Quickstart shows the shortest practical high-level and Stepper workflows.
• Workflow overview explains when to use the high-level pipeline functions, staged cluster pipelines, or direct Stepper calls.
• TS guess generation explains how templates, reactive positions, and conformers become TS inputs.
• Inspecting results gives the practical checks to run before using a barrier or scope prediction.
• DataFrames and results documents the main input and output conventions used across workflows.
• Troubleshooting failed calculations starts from *-NT, *-error, and saved backend files.

External Tools

FRUST does not install quantum chemistry engines for you. These integrations must be configured in the environment where calculations run:

• ORCA and xTB for standard calculations.
• ORCA-External-Tools for UMA and ORCA-driven g-xTB.
• A g-xTB-capable xtb executable for direct and ORCA-driven g-xTB.

Start with External tool setup when using UMA or ORCA-driven g-xTB. Then use the UMA and g-xTB pages for method-specific FRUST usage.

API Reference

The API reference focuses on the public surface most users touch directly: Stepper, frust.pipes, cluster submission helpers, dataframe schema helpers, and selected external-tool utilities.