QCANT

Utilities for near-term applications of quantum computing in chemistry and materials science.

This repository currently contains a lightweight, template-derived QCANT package. The public API is small and intended to grow as project modules are added.

Install

You can install QCANT from PyPI:

pip install QCANT

QCANT requires scientific Python dependencies (installed automatically when you pip install QCANT):

• numpy<2, scipy<2
• pennylane
• pyscf
• autoray<0.7

For development (recommended: conda env for the full stack):

conda env create -f devtools/conda-envs/qcant.yaml
conda activate qcant
pip install -e . --no-deps

For development (pip/venv):

pip install -e .

To enable the optional Qulacs simulator backend:

pip install -e ".[qulacs]"

For users (once QCANT is published to PyPI):

pip install QCANT

Release (PyPI)

This repo is configured to publish to PyPI from GitHub Actions using Trusted Publishing.

1. Create the project on PyPI and enable "Trusted Publishing" for:
   • Owner: srivathsanps-quantum
   • Repo: QCANT
   • Workflow: publish-pypi.yml
2. Tag a release (tag must start with a digit, e.g. 1.0.0) and push the tag:

git tag 1.0.0
git push origin 1.0.0

That tag push triggers the publish workflow.

Quickstart

import QCANT

print(QCANT.canvas())

Qulacs Backend

QCANT also exposes optional Qulacs-backed exact-state routines for the simulator-heavy algorithms:

• QCANT.qkud_qulacs
• QCANT.qrte_qulacs
• QCANT.qrte_pmte_qulacs
• QCANT.adapt_vqe_qulacs
• QCANT.cvqe_qulacs

The Qulacs path is aimed at reducing simulator overhead rather than replacing the chemistry stack. PySCF and PennyLane are still used for Hamiltonian and active-space construction, while Qulacs handles the state evolution and expectation-value hot loops.

The current speed-focused implementation does three main things:

• compiles PennyLane chemistry gates once into reusable Qulacs parametric circuits
• uses Qulacs backprop for native ADAPT-VQE gradients
• reduces CVQE inner optimization by solving the selected-determinant coefficients in the current reference space and optimizing only the ansatz parameters

For larger exact-state runs, prefer evolution_mode="trotter" in the Qulacs QRTE/QKUD routines. The "sparse" mode still materializes the Hamiltonian matrix and is not the scalable choice near the 16-20 qubit range.

Documentation

Hosted documentation:

• https://asthanalab.github.io/QCANT/

The documentation lives in docs/ and is built with Sphinx:

cd docs
make html

The output will be in docs/_build/html.

Copyright

Copyright (c) 2025, Asthana Lab

Acknowledgements

Project based on the Computational Molecular Science Python Cookiecutter version 1.11.