Demonstrations of fast approximate variable-width broadening

Abstract

We have developed a fast approximate method of applying variable-width broadening to spectral data. This dataset contains demonstration scripts which generate the plots for a related academic manuscript. They also reproduce the parametrisation of polynomial fits to the relationship between the approximation error and the scale factor between broadening kernels used in the method. Some of the scripts use the implementation in the Euphonic library: to reproduce published results use Euphonic v1.3.0

https://pace-neutrons.github.io/Euphonic/versions/v1.3.0.html https://github.com/pace-neutrons/Euphonic/releases/tag/v1.3.0

Most of the scripts are written in Python. One Bash shell script is also included which needs to call the euphonic-powder-map program included in the Euphonic package. A Makefile is provided so that all plots can be generated with "make plots" using GNU Make. .yaml datafiles were derived from data in the Kyoto phonon database ("phonondb") under a CC-by license: the Phonopy calculation directories were downloaded and Phonopy was run with =--include-all= and =--nac= options to create .yaml files containing all the information needed by Euphonic.