Citing MatMelt

If MatMelt contributes to work you publish, please cite it. The repository ships a Citation File Format file, CITATION.cff, which is the authoritative record of the title, authors and version — GitHub reads it directly and offers a Cite this repository button that converts it to BibTeX or APA on demand.

That conversion is deliberately not duplicated here: a hand-written BibTeX entry on this page would be a second copy of the authors and version, and the copy that drifts is always the one someone pastes into a paper.

Cite the version you ran

Results depend on the version that produced them. Record it from the machine that did the work:

$ matmelt --version

Screening settings travel with the data as well: each record.json carries an _mcli.screen.settings block naming the model, device, dtype, fmax and step limit that produced the verdict, so a reviewer can see what was actually run without taking your word for it. Get the results out covers getting those records out.

Cite the science too

MatMelt orchestrates other people’s scientific software; it does not replace it. A run that used the default funnel leaned on SMACT for charge-balanced compositions, PyXtal for symmetry-constrained structure generation, MACE and ASE for the relaxation and its verdict, and — if you enabled the optional stage — DFTB+ through AiiDA. Those projects ask to be cited in their own right, and the citation each of them prefers is documented on their own sites. What the stages actually compute sets out precisely which of them did what at each stage, which is also the honest basis for describing your method.

CITATION.cff

cff-version: 1.2.0
message: "If you use this software in work you publish, please cite it as below."
title: "MatMelt"
abstract: >-
  A config-driven command-line tool that runs a multi-stage materials-discovery
  funnel (compose, structure, screen, optional DFTB+ calculate, extract) on the
  researcher's own local and remote machines, with durable state and resumable runs.
type: software
authors:
  - name: "Quantum ARISE"
    website: "https://github.com/Quantum-ARISE-Acad"
  - family-names: "AMUZUGA"
    given-names: "Komi Djidoula Wisdom Sitou"
    alias: "ShellRuner"
    email: "sitouamu510@gmail.com"
    affiliation: "Quantum ARISE"
  - family-names: "LABBAH"
    given-names: "Elie"
    affiliation: "Quantum ARISE"
repository-code: "https://github.com/Quantum-ARISE-Acad/qa-materials-cli"
license: MIT
version: "0.2.0"
keywords:
  - materials-science
  - high-throughput screening
  - crystal structure prediction
  - machine-learning interatomic potentials
  - DFTB+
  - workflow