cogito-dft 0.1.0

Crystal Orbital Guided Iteration To atomic-Orbitals (COGITO) finds an adaptive atomic orbital basis for a given density functional theory (DFT) run, and then decodes the atom and bond contributions to electronic structure.

COGITO - Crystal Orbital Guided Iteration To atomic-Orbitals

This repository hosts the code for COGITO, a tool for obtaining quantum chemistry from plane wave DFT calculations. The main codes modules are COGITO, COGITOpost, COGITOanalyze, and COGITOico.

Installation

pip install --upgrade pip
pip install cogito-dft

To install optional dependences (scikit-image, dash, dash-ag-grid) that are used in some COGITOpost functions:

pip install "cogito-dft[plot]"

To avoid thread oversubscription and possible stalls (especially on HPC), set:

export OMP_NUM_THREADS=1

Quick Start

Visit the live website: cogito-dft.readthedocs.io

Section	Description
Tutorial	Step-by-step analysis workflow
File Guide	Interactive guide of inputs and outputs
API Docs	Complete function reference

Basic Workflow:

1. Run VASP - Static calculation with saved wavefunctions and high NBANDS
2. Generate COGITO model - Creates COGITO basis functions and tight binding parameters
3. Verify quality - Check convergence, charge spilling, orbital mixing, and band interpolation
4. Analyze chemistry - COHP, bonding, charge analysis

# Run VASP 
vasp_std

# See 'COGITO --help' for variable options
COGITO --dir './'

# Check that quality metrics are within range (check band interpolation after COGITOpost)
COGITOanalyze --dir './'

# Generate atom and bond partition of charge / band energies, make bond plots, and more
COGITOpost --dir './'
# Customize a runTBmodel.py file to get specific plots
python runTBmodel.py