PythTB is a Python library for constructing and analyzing tight-binding models, built for modern topological band theory applications. It provides a streamlined path from model specification to physical interpretation, making it useful for both learning electronic structure and conducting research-level studies. With only a few lines of code, you can define lattice models, build tight-binding Hamiltonians, and compute electronic properties.

PythTB provides tools for:

• Band structures and density of states
• Berry phases and Berry curvature
• Chern numbers, Wilson loops, and related invariants
• Quantum geometric tensors and local Chern markers
• Chern-Simons axion angle
• Maximally localized Wannier functions
• Wannier-based tight-binding models generated through Wannier90

• Documentation: https://pythtb.readthedocs.io/en/latest/
• Contributing: https://pythtb.readthedocs.io/en/latest/development.html
• Tutorials: https://pythtb.readthedocs.io/en/latest/tutorials.html
• Formalism: https://pythtb.readthedocs.io/en/latest/formalism.html
• Source: https://github.com/pythtb/pythtb
• Report Issues: https://github.com/pythtb/pythtb/issues

PythTB is available through conda-forge (recommended) and PyPI.

# Conda (pythtb >= 1.8.0)
conda install -c conda-forge pythtb

# pip
pip install pythtb

To install from source in editable mode:

git clone https://github.com/pythtb/pythtb.git
cd pythtb
pip install -e .

PythTB ≥ 2.0.0 requires Python ≥ 3.12 and the core dependencies:

• numpy ≥ 2.0
• matplotlib ≥ 3.9

Optional extras can be installed via pip install .[group]:

• [plotting]: Plotly for interactive visualization
• [speedup]: TensorFlow for GPU-accelerated routines
• [notebooks]: Jupyter support (IPython ≥ 8.17, ipykernel, notebook, jupyter, jupyterlab)
• [docs]: Sphinx toolchain for documentation
• [tests]: pytest
• [dev]: developer tools (pytest, ruff, pre-commit)

For more detailed instructions, see:

• Full Installation Guide
• Wiki

If you use the code in your paper, please cite us

@software{Cole_Python_Tight_Binding_2025,
author = {Cole, Trey and Coh, Sinisa and Vanderbilt, David},
doi = {10.5281/zenodo.12721315},
license = {GPL-3.0-or-later},
month = nov,
title = {{Python Tight Binding (PythTB)}},
url = {https://zenodo.org/records/12721315},
version = {2.0.0},
year = {2025}
}