(technically pronounced peng-van, but some pronounce numpy as num-pee rather than num-pie, so who really knows?)
pengwann is a lightweight Python package for computing common descriptors of chemical bonding from Wannier functions (as output by Wannier90). Alternatively phrased: pengwann replicates the core functionality of LOBSTER, except that the local basis used to represent the Hamiltonian and the density matrix is comprised of Wannier functions rather than pre-defined atomic or pseudo-atomic orbitals. The primary advantage of this methodology is that (for energetically isolated bands) the spilling factor is strictly 0.
The latest tagged release of pengwann is pip-installable as:
pip install pengwann
Alternatively, to install the current development build:
pip install git+https://github.com/PatrickJTaylor/pengwann.git
If you think you have encountered a bug whilst using pengwann, please create an issue and let us know!
Contributions to pengwann via pull requests are also very welcome (see the contributions guide for more details).