/PySCES

AI-LSC-IVR is a code written in Python3 and used to perform ab initio nonadiabatic dynamics using linearized semiclassical initial value representation (LSC-IVR).

Primary LanguagePython

drawing

PySCES, Python code for Linearized Semi-Classical Dynamics with On-the-fly Electronic Structure, is a highly parallelized code for ab initio nonadiabatic molecular dynamics using linearized semiclassical initial value representation (LSC-IVR). As of now, on-the-fly updates of electronic structure variables can be performed either by GPU-assisted electronic structure package, TeraChem, or, as its original implementation was structured, GAMESS. The code is intended for computing electronic population correlation functions through one of three population estimators: Wigner, semiclassical, or spin mapping population estimator. See the following references for the original implementation and the application of the code:

  • K. Miyazaki and N. Ananth. "Nonadiabatic simulations of photoisomerization and dissociation in ethylene using ab initio classical trajectories," J. Chem. Phys. 159, 124110 (2023)

References for the most recent implementation will be added as they become available.

Disclaimer: The code contained in this package has been written, edited, and used by members of the Ananth group at Cornell University and the Isborn group at the University of California, Merced. It has not been formally reviewed, nor published, nor are there copyrights. Bugs and errors may be present.

Installation

The easiest way to get PySCES is to clone the repository from github. As the code is still in active development, this makes it convenient to get the latest changes as they are uploaded. PySCES can be installed with pip, and we recommend you make a new python environment with Conda prior to the installation:

    conda create -n pysces python
    conda activate pysces
    cd /dir/where/you/git/cloned/
    pip install -e .

The -e flag will tell pip not to copy over the code itself into your python environment. To get the latest updates, just go back to the location you cloned the repository, run a \texttt{git pull}, and your installation will also be updated. You do not need to run a pip install again.

Usage

For further details, including the various options that can control PySCES, please read the Manual in PDF form within the repository.

Acknowledgements

PySCES was initially written by Ken Miyazaki while at Cornell University. Christopher Myers extended the code to call TeraChem as the electronic structure driver at the University of California, Merced. The authors would also like to thank Thomas Trepl at University of Bayreuth, Germany, for his contributions to the code, including procedures for correcting nonadiabatic coupling sign-flips. The PySCES logo was designed by Fernanda Giongo Fernandes.