DScribe
========

.. image:: https://travis-ci.org/SINGROUP/dscribe.svg?branch=master
    :target: https://travis-ci.org/SINGROUP/dscribe

.. image:: https://coveralls.io/repos/github/SINGROUP/dscribe/badge.svg?branch=master
    :target: https://coveralls.io/github/SINGROUP/dscribe?branch=master

DScribe is a Python package for transforming atomic structures into fixed-size
numerical fingerprints. These fingerprints are often called "descriptors" and
they can be used in various tasks, including machine learning, visualization,
similarity analysis, etc. To get started you can check the :doc:`basic tutorial
<tutorials/basics>`.

.. note::
   Python 3.8 is supported starting from version **0.3.5**.

.. note::
   Creating SOAP output for large systems has has been greatly improved
   starting in version **0.3.2**. Updating is recommended if working with
   larger systems.


Capabilities at a Glance
========================

DScribe currently includes the following descriptors:

  - :doc:`Coulomb matrix <tutorials/coulomb_matrix>`
  - :doc:`Sine matrix <tutorials/sine_matrix>`
  - :doc:`Ewald sum matrix <tutorials/ewald_sum_matrix>`
  - :doc:`Atom-centered Symmetry Functions (ACSF) <tutorials/acsf>`
  - :doc:`Smooth Overlap of Atomic Positions (SOAP) <tutorials/soap>`
  - :doc:`Many-body Tensor Representation (MBTR) <tutorials/mbtr>`
  - :doc:`Local Many-body Tensor Representation (LMBTR) <tutorials/lmbtr>`

Check the tutorials for more information.

Go Deeper
=========
You can find more details in our open-access article: `DScribe: Library of
descriptors for machine learning in materials science
<https://doi.org/10.1016/j.cpc.2019.106949>`_

Documentation for the source code :doc:`can be found here <doc/modules>`. The
full source code with examples and regression tests can be explored at `github
<https://github.com/SINGROUP/dscribe>`_.

.. toctree::
    :hidden:

    install
    tutorials/tutorials
    API <doc/modules>
    contributing
    publications
    citing
    about