ponychen123 / cryspy

Website and document: https://tomoki-yamashita.github.io/CrySPY

CrySPY

CrySPY is a crystal structure prediction tool written in Python.

Latest version

version 0.8.0 (2020 February 16)

Important changes

• version 0.8.0
  • Migrated to Python 3
  • Several variable names
  • Several data formats
  • Unit of energy in output: eV/cell --> eV/atom
  • No. of working directories corresponds to structure ID
• version 0.7.0
  • Evolutionary algorithm is now available
• version 0.6.2
  • LAMMPS can be used in CrySPY
• version 0.6.0
  • LAQA is now available
  • Changed the data format of init_struc_data and opt_struc_data from list to dict

System requirements

Python

• Python 3.x.x
• COMBO
• numpy
• pandas
• pymatgen

Structure optimizer

At least one optimizer is required.

• VASP (tested with version 5.4.1)
• Quantum ESPRESSO (tested with version 6.1, version 5.x does not work)
• soiap (tested with version 0.2.2)
• LAMMPS

Others

• find_wy: find_wy can randomly select a combination of Wyckoff positions for a given chemical composition and space group.

Document

CrySPY document

Google group

Google gruop of CrySPY

Tutorial (written in Japanese)

チュートリアルと解説

Reference

Bayesian optimization

• T. Yamashita, N. Sato, H. Kino, T. Miyake, K. Tsuda, and T. Oguchi, Phys. Rev. Mater. 2, 013803 (2018).
  • https://link.aps.org/doi/10.1103/PhysRevMaterials.2.013803

LAQA

• K.Terayama, T. Yamashita, T. Oguchi, and K. Tsuda, npj Comput. Mater. 4, 32 (2018).
  • https://www.nature.com/articles/s41524-018-0090-y

License

CrySPY is distributed under the MIT License.
Copyright (c) 2018 CrySPY Development Team