WanPy: A Wannier Tight-Binding Model Based Numerical Package

Requirements

WanPy requires the following packages to be installed:

numpy
scipy
pandas
sympy
h5py
mpi4py
fortio
spglib
matplotlib (optional)
phonopy (optional)

Installation

To install WanPy, create a Conda environment with the required packages:

$ conda create -n wanpy python numpy scipy pandas sympy h5py mpi4py fortio spglib matplotlib

Uncompress wanpy package in personal computer or clusters:

$ tar zxvf wanpy.tar.gz

$ git clone https://github.com/jincao2013/wanpy.git

Setup environment variables:

For debug in personal computer:

export PYTHONPATH="path_of_wanpy:$PYTHONPATH"
export PATH="path_of_wanpy/scripts:$PATH"
export WANPY_ROOT_DIR="your_work_dir"
export PYGUI="True"

For deployment in clusters:

export PYTHONPATH="path_of_wanpy:$PYTHONPATH"
export PATH="path_of_wanpy/scripts:$PATH"

Get start

# show usage
$ wanpy -h

# collect Wannier TB data to a single .h5 file
$ wanpy htb

# Collect Wannier TB data to a single .h5 file
$ wanpy wannier

# Write _hr.dat and _r.dat (optional) from .h5 file
$ wanpy write_dat

# Twist the order of .amn (uudd or udud) 
$ wanpy twist_amn

# compare Wannier and VASP band structure. support HSE type calculation. 
$ wpyplotband

# check job status
$ statistic_cores ...

Collect Wannier TB data

WanPy need to collect necessary Wannier TB data into a single .h5 file as the input for further calculations. This can be done by:

Important notice:

WanPy employs the uudd order of Wannier orbitals for its internal calculations, which is the default setting when using v1.2 of wannier_setup. If a higher version is used, one should utilize wanpy twist_amn to reorganize the .amn file into the uudd order and then proceed with the disentanglement process once more.

Method 1: from wanpy wannier, this will need POSCAR .nnkp, .wout, .chk, .eig, and WAVECAR if --spn. The nnkp file can be obtained by:

$ wannier90.x -pp wannier90.win

To start collecting dat:

# start to collect Wannier TB data 
$ wanpy wannier [options]

# For detail of options, see
$ wanpy wannier -h

Method 2: from wanpy htb, this will need at least POSCAR, .wout, _hr.dat . The files .nnkp, _r.dat, _wsvec.dat, and _spin.dat are optional.

# start to collect Wannier TB data 
$ wanpy htb [options]

# For detail of options, see
$ wanpy htb -h

Symmetric Wannier tight-binding models (SWTB)

The SWTB can be obtained during collecting Wannier TB data. Set symmetry related parameters in symmetry.in, a template file can also be produced by executing wanpy wannier --temp:

# Input file for building symmetric Wannier TB model

# Choose method, default is kspace
symmetric_method = rspace   # Options: {rspace, kspace}

# Choose between manually setting symmops or automatically detecting symmops from magmoms
parse_symmetry = man        # Options: {man, auto}

ngridR = 12 12 1            # used in kspace method, and use a slightly larger value than the original TB model 

# Parameters used when parse_symmetry = auto
symprec = 1e-5
&magmoms
0 0 1
0 0 -1
/

# Parameters used when parse_symmetry = man
&symmops
# TR  det  alpha  nx  ny  nz  taux  tauy  tauz
  0   1    0      0   0   1   0     0     0   # e
  0   1    180    0   0   1   0.5   0.5   0   # c2z
# Anti-unitary symmetry operations
  1   1    0      0   0   1   0.5   0.5   0   # T
  1   1    180    0   0   1   0     0     0   # Tc2z
/

Then start to get SWTB with an additional tag --symmetry

$ wanpy wannier --symmetry [options]

See Comput. Phys. Commun. 270, 108153 (2022) (arXiv:2012.08871) for detail.

License

WanPy is released under the GNU General Public License. See LICENSE for details.