ReaxKit

ReaxKit is a modular, extensible Python toolkit for pre‑processing, post‑processing, and analysis of ReaxFF molecular dynamics simulations. It provides a clean separation between file parsing, analysis routines, and reproducible workflows, with Python APIs, a CLI interface, and a Dash-based GUI.

ReaxKit is designed for researchers who want a transparent, scriptable bridge between raw ReaxFF data and quantitative, publication‑ready results.

Key Capabilities

Engine

• Shared engine abstractions for consistent parsing and generating input/output files across 3 different ReaxFF simulation engines, namely ams, lammps, reaxff
• Example IO and generator modules under the reaxff engine:
  • ReaxFF input and output files such as xmolout, fort.7, fort.13, molfra, and more
  • Input generators for control, geo, eregime, tregime, and related files

Analysis

• Engine-agnostic analyzers built on ReaxKit's domain data models
• Analyzer tasks organized around separated request, result, and task objects
• Analysis components separated from engine IO so the same computations can be reused across scripts, workflows, web UI, and presentation modules

Workflows

• Reproducible workflows organized by data handling, file tools, meta orchestration, presentation, and study design
• Automation paths for common ReaxFF pre-processing, analysis, post-processing, and presentation tasks

Web UI

• Dash-based interface components for interactive ReaxFF data inspection
• Backend, UI, and presentation layers for browser-driven workflows

Presentation

• Publication-ready plotting utilities:
  • 2D plots, dual-axis plots, tornado plots, 3D scatter, and heatmaps
• Video generators

Utilities

• Shared data, media, and numerical utilities
• Common infrastructure used by analysis, workflows, web UI, and presentation modules

See the full documentation (API reference, tutorials, examples) on ReaxKit Site.

Project Layout

src/reaxkit/
├── analysis/        # Engine-agnostic analysis tasks
├── cli/             # Command-line entry points
├── core/            # Registries and shared core infrastructure
├── data/            # Packaged reference data and resources
├── domain/          # Central data models, requests, and results
├── engine/          # IO handlers and input generators
├── help/            # Introspection and help system
├── presentation/    # Plotting, active-site views, and media presentation
├── utils/           # Shared data, media, and numerical utilities
├── webui/           # Dash/web interface backend, UI, and presentation layers
└── workflows/       # Workflow orchestration and automation

Testing

Run unit tests with:

pytest -s tests/

to test the package and get the timing for their execution.

Citation

If you use ReaxKit in your work, please cite:

Dinani, A. M., van Duin, A., Shin, Y. K., & Sepehrinezhad, A. (2025).
ReaxKit: A modular Python toolkit for ReaxFF simulation analysis.
Zenodo. https://doi.org/10.5281/zenodo.18485384

Source code: https://github.com/ali-m-dinani/reaxkit

Future Directions

• Implement machine learning-based surrogate models for rapid property prediction and active-site identification.
• Develop a plugin system for user-contributed analysis routines and visualization components.
• Integrate with Jupyter notebooks for interactive analysis and visualization.

If you have any feature request, you can submit it through the ReaxKit's GitHub page, or directly sending an email to Dinani@psu.edu.

Additional resources:

• AUTHORS.md — Full credits and acknowledgments.
• LICENSE — Full license terms under the MIT License