thermoparser 3.3.0

Installation

ThermoParser can be installed from PyPI with pip:

pip install thermoparser

Note the package is imported as tp. To install the latest development version instead, or to work on ThermoParser itself, use git:

git clone https://github.com/smtg-bham/ThermoParser.git
cd ThermoParser
pip install .

After installing, you may want to copy tp/tprc.yaml to ~/.config/tprc.yaml, if you want to set your own default axis labels, unit conversions, default style sheets (two are provided), other aesthetic alterations and more! Its location in an installed copy can be found with:

python3 -c "import os, tp; print(os.path.join(os.path.dirname(tp.__file__), 'tprc.yaml'))"

Command Line Interface (CLI)

ThermoParser uses click, which has an easily navigable structure from the command-line, detailed in the tutorials. The most frequently useful commands are included in the CLI for maximum ease, including the tp get functions, which verbosely retrieve data from files which are normally tiresome and error-prone to navigate; and most of the simplest plot-types available through the Python interface.

Python Interface

ThermoParser is designed to have four main stages:

1. Axes:

   Pick an axis layout from tp.axes.

2. Load:

   Use the functions is tp.data.load to load the relevant data.

3. Add:

   Use functions in tp.plot modules to add graphs to the axes.

4. Save:

   Use plt.savefig or equivalent to produce the figure.

As ThermoParser is dependent on matplotlib, each stage can be substituted with bespoke code, e.g. using matplotlib.pyplot.subplots or matplotlib.axes.Axes.scatter. These can still be supplemented with ThermoParser helper functions, such as default labels which the user can set in tp.settings, colourmap generators in tp.plot.colour or legend helpers such as tp.axes.legend.alphabetise.

The best way to get a feel for ThermoParser is to see it in action: Take a look at our examples and tutorials. Currently supported codes are:

• Phononic properties:

  • Phonopy

  • Phono3py

• Electronic properties:

  • AMSET

  • BoltzTraP

Current plotting modes are split into four areas.

• tp.plot.phonons contains plots along a high-symmetry path, including phonon dispersions and plots which project other quantities onto these paths in various ways.

• tp.plot.frequency plots frequency on the x-axis, including density of states (DoS), cumulative kappa, “waterfall” and density plots. Each function has a main argument, which can be useful when plotting multiple quantities on the same set of axes; and an invert argument, which swaps the x and y axes to let you plot DoS-style next to a tp.plot.phonons plot.

• tp.plot.mfp contains a cumulative kappa against mean free path plot.

• tp.plot.heatmap contains a heatmap plotter, and wrappers which format appropriately for ZT against temperature and doping concentration; and one which plots the lattice thermal conductivity required to reach a target ZT, again against temperature and doping.

A set of example scripts is provided in the tp/examples folder and in our online examples, and there is documentation.

Contributing

We welcome any contributions, whether they be a feature request or a new piece of code (or anything else). Adding options is inteded to be straightforward; the modularity of the code means that each step is mostly independent of the others.

Bugs and feature requests can be submitted to the issue tracker, while contributions can be made using the fork and pull approach. Contributions should include comprehensive docstrings, and where appropriate examples, further documentation and tests are greatly appreciated. Documentation uses the sphinx package, and can be built from the docs directory with sphinx-build -b html src/ .. In order to build the docs, download the extra dependencies with, e.g., pip install .[docs] from the ThermoParser directory.

Testing

Tests use the unittest package, which is part of the Python standard library and so needs no installation. They can be run from the ThermoParser directory with python3 -m unittest.

Contributors

Many thanks to all those who contributed code or ideas to ThermoParser! Roughly chronologically, they are so far:

• Kieran B. Spooner

• Maud Einhorn

• David O. Scanlon

• Daniel W. Davies

• Bonan Zhu

• Sean R. Kavanagh

• Warda Rahim

• Katarina Brlec

• Joe Willis

• Dale Gaines II

Thanks also to the JOSS reviewers, Evan Walter Clarke Spotte-Smith, Enric Tomás Grau-Luque, and Francesco Nattino; and the editor Mojtaba Barzegari. An unintimidating and productive review process, which I would recommend if the opportunity arises!

License

ThermoParser is licensed under the GNU Affero General Public License v3 (AGPLv3).

Requirements

ThermoParser uses the following open-source packages:

• click

• h5py

• json

• matplotlib

• numpy

• pymatgen

• scipy

• sphinx

• yaml