Project Overview

PAOFLOW is an open-source Python package for post-processing Density Functional Theory (DFT) calculations. It takes Kohn–Sham eigenstates from Quantum ESPRESSO or VASP and constructs tight-binding Hamiltonians in a Projected Atomic Orbital (PAO) basis — giving you a compact, ab initio starting point for computing electronic, transport, and spectroscopic properties like band structures, density of states, and Berry curvature effects, all without empirical parameters.

This section is the developer-facing side of the documentation: installation notes, contribution guidelines, module overviews, and internal workflow descriptions.

Design Philosophy

A few core ideas shape how PAOFLOW is built and maintained:

PAO basis construction. Everything starts from a converged DFT calculation. PAOFLOW projects Bloch wavefunctions onto atomic orbital bases to produce compact, ab initio tight-binding Hamiltonians that are efficient enough for large-scale property calculations.

Single source of truth. Each module owns one well-defined responsibility. Shared state travels through the DataController dictionary, and every function that touches it documents exactly which keys it reads or modifies — so you always know where data comes from and where it ends up.

Minimal external coupling. The driver (rank-0 Python process) and MPI worker processes talk to each other exclusively through pickle files, keeping the orchestration layer simple and free of MPI dependencies.

Testability first. New functionality comes with integration tests. We keep them fast by using low-accuracy settings (reduced grids and k-points) and staging large inputs as release assets rather than committing them to the repository. See Testing for the full picture.