ABMPTools (ABINIT-MP Tools)
Copyright 2022-2026 Koji Okuwaki

This product includes software developed by Koji Okuwaki and contributors.

Licensed under the Apache License, Version 2.0 (see LICENSE file for the
full text of the license).

================================================================================

Citation request
----------------

If you use ABMPTools in academic work, please cite the project. Citation
metadata is provided in CITATION.cff (machine-readable, used by GitHub's
"Cite this repository" button) and in the README.md "How to cite" section.

Until a peer-reviewed publication is released, please cite as:

    Okuwaki, K. (2026). ABMPTools: a Python toolkit for ABINIT-MP
    Fragment Molecular Orbital pre/post-processing.
    https://github.com/kojioku/abmptools

A Zenodo DOI will be issued upon the first GitHub release tag and added
to CITATION.cff.

================================================================================

Third-party software
--------------------

ABMPTools depends on, but does not bundle or modify, several open-source
projects (numpy, pandas, ase, rdkit, OpenMM, OpenFF Toolkit, GROMACS,
AmberTools, vermouth, insane, GENESIS, acpype, ABINIT-MP, etc.).

A complete inventory with license details is provided in
docs/licenses_third_party.md. All listed dependencies are used either
via Python import (with their own license terms preserved separately)
or via subprocess invocation (mere aggregation, no linking, no source
modification).

When redistributing ABMPTools, recipients must retain this NOTICE file
verbatim per Apache-2.0 §4(d).

================================================================================

Re-implementation attribution
-----------------------------

The `abmptools.cg.dpd` subpackage (v1.26.0+) re-implements the DPD input
generation logic of DPDgen (https://github.com/kojioku/dpdgen) for the
Cognac DPD UDF (R1) and J-OCTA dpm patch (R2) routes. DPDgen is authored
by the same author (Koji Okuwaki), so the porting is in-house. No DPDgen
source code is bundled; `abmptools.cg.dpd` provides an independent
implementation that consumes the same fcews-unified aij.dat format
(Python dict script: `aij = [[seg_i, seg_j, value], ...]` or `chi = [...]`).
For reference DPDgen samples, see /home/okuwaki/repos/dpdgen (private).
