matador-db 0.9.7

Installation

In the simplest case (e.g. you already have Python 3.6+ set up), pip install matador-db is sufficient to get up and running, preferably in a fresh virtual environment. More detailed instructions can be found in the Installation instructions.

Tests can be run on your local machine with python -m unittest discover -v -b or simply with py.test (required pytest to be installed); in order to test CASTEP-running functionality, the tests will look for an MPI-enabled executable named castep on your $PATH.

Upgrading to the latest version should be as simple as pip install -U matador-db.

Usage

matador is primarily a Python library that can be used inside Python scripts/modules to create a custom workflow. There are, however, several command-line scripts bundled with matador itself. All of these scripts are listed under CLI Usage.

For basic command-line usage, please explore the help system for command. Common workflows can be found inside examples/ and in the online docs.

Please consult the full Python API documentation for programmatic usage.

Core functionality

The API has many features that can be explored in the examples and API documentation. As a summary, matador can be used for:

• Scraping of CASTEP (and Quantum Espresso) input/output files into flexible Python dictionaries/models.

• The creation and curation of MongoDB collections of geometry optimisation calculations, with a powerful querying CLI/API.

• Customisable, publication-ready plots for all models, e.g. phase diagrams, PDF, PXRD, voltage profiles, electronic/vibrational bandstructures etc.

• High-throughput geometry optimisations, electronic and vibrational properties using CASTEP (and Quantum Espresso) with run3. Tested on several supercomputers. run3 is designed primarily for simple workflows and offers little in the way of tools for creating complex workflows; if this is your use case, then check out some of the other codes listed below.

• Creation of phase diagrams and electrochemical voltage profiles from the results of DFT calculations.

This functionality is achieved by interfacing with much of the standard scientific Python stack (NumPy, SciPy, matplotlib), some more materials-specific packages (spglib, SeeK-path, periodictable) and other general packages (pymongo, python-ternary, numba).

Similar packages

This package is by no means unique in its functionality or goals. Below is a list of similar packages and an overview of where they overlap with matador:

• ASE: manipulation of structures, parsing and exporting files, running jobs and local databases with ase-db. An interface is provided to ASE’s Atoms object.

• pymatgen: similar to ASE, with a focus on the VASP DFT code. An interface is provided to pymatgen’s Structure object. Local databases can be constructed with the pymatgen-db add-on and high-throughput workflows are achieved with Fireworks.

• AiiDA: high-throughput job running, provenance tracking and database storage for many simulation codes.

• sumo publication quality plotting, primarily for VASP but also with support for other codes.

If you think this list is outdated, incorrect or simply incomplete, then please raise an issue!

Citing matador

If you use matador in your work, we kindly ask that you cite the source code archive on Zenodo:

Matthew Evans, ml-evs/matador (2020), Zenodo. http://doi.org/10.5281/zenodo.3908574