Framework for random structure search using polynomial machine learning potentials

rsspolymlp is a Python code designed to conduct random structure searches (RSS) using polynomial machine learning potentials (MLPs). The code provides functionalities for generating initial random structures, performing geometry optimization using polynomial MLPs, and eliminating duplicate structures. The formulation and development of polynomial MLPs are detailed in pypolymlp.

Installation

Required libraries and python modules

• python >= 3.10
• scikit-learn
• joblib
• pypolymlp
• spglib
• symfc

[Optional]

• matplotlib, seaborn
• pymatgen

How to install

• Install from conda-forge

Name	Downloads	Version	Platforms

conda create -n rsspolymlp
conda activate rsspolymlp
conda install -c conda-forge rsspolymlp

• Install from PyPI

conda create -n rsspolymlp
conda activate rsspolymlp
conda install -c conda-forge scikit-learn joblib pypolymlp spglib symfc
pip install rsspolymlp

How to use rsspolymlp

• Workflow of RSS with polynomial MLPs
  • Initial structure generation
  • Global RSS with polynomial MLPs
  • Unique structure identification and RSS result summarization
  • Phase stability analysis
• Development kit for polynomial MLPs
  • MLP dataset generation
  • DFT dataset division
  • Polynomial MLP development
  • Pareto-optimal MLP selection
• Python API
  • RSS workflow
  • VASP calculation utilities
    • Single-point calculation
    • Local geometry optimization
  • Matplotlib utilities

Citation of rsspolymlp

If you use rsspolymlp in your study, please cite the following articles.

“Systematic global structure search of bismuth-based binary systems under pressure using machine learning potentials”, Phys. Rev. Materials 10, 093801 (2026)

@article{Phys.Rev.Mater.10.093801,
  title = {Systematic global structure search of bismuth-based binary systems under pressure using machine learning potentials},
  author = {Wakai, Hayato and Ishiwata, Shintaro and Seko, Atsuto},
  journal = {Phys. Rev. Mater.},
  volume = {10},
  issue = {9},
  pages = {093801},
  numpages = {28},
  year = {2026},
  month = {Sep},
  publisher = {American Physical Society},
  doi = {10.1103/k9sg-v7hl},
  url = {https://link.aps.org/doi/10.1103/k9sg-v7hl}
}