README.rst
1hiPhive
2=======
3
4**hiPhive** is a tool for efficiently extracting high-order force constants
5from atomistic simulations, most commonly density functional theory
6calculations. A detailed description of the functionality provided as well as an
7extensive tutorial can be found in the
8`user guide <https://hiphive.materialsmodeling.org/>`_. Complete examples of
9using hiphive for force constants extaction can be found at `hiphive examples <https://gitlab.com/materials-modeling/hiphive-examples/>`_.
10
11**hiPhive** is written in Python, which allows
12easy integration with countless first-principles codes and analysis tools
13accessible in Python, and allows for a simple and intuitive user interface. For
14example using the following snippet one can train a force constant potential:
15
16.. code-block:: python
17
18 cs = ClusterSpace(primitive_cell, cutoffs)
19 sc = StructureContainer(cs, list_of_training_structure)
20 opt = Optimizer(sc.get_fit_data())
21 opt.train()
22 fcp = ForceConstantPotential(cs, opt.parameters)
23
24after wich it can be used in various ways, e.g., for generating phonon
25dispersions, computing phonon lifetimes, or running molecular dynamics
26simulations.
27
28
29Installation
30------------
31
32**hiPhive** can be installed via `pip`::
33
34 pip3 install hiphive
35
36If you want to get the absolutely latest (development) version you can clone
37the repo and then install **hiPhive** via::
38
39 git clone git@gitlab.com:materials-modeling/hiphive.git
40 cd hiphive
41 python3 setup.py install --user
42
43**hiPhive** requires Python3 and invokes functionality from
44several external libraries including the
45`atomic simulation environment <https://wiki.fysik.dtu.dk/ase>`_,
46`spglib <https://atztogo.github.io/spglib/>`_ and
47`SymPy <http://www.sympy.org/en/index.html>`_.
48Please note that the dependency on
49`scikit-learn <http://scikit-learn.org/>`_
50is not enforced during installation via `pip`.
51Please consult the
52`installation section of the user guide <https://hiphive.materialsmodeling.org/installation.html>`_
53for details.
54
55
56Credits
57-------
58
59* Fredrik Eriksson
60* Erik Fransson
61* Paul Erhart
62
63**hiPhive** has been developed at Chalmers University of Technology in
64Gothenburg, Sweden, in the
65`Materials and Surface Theory division <http://www.materialsmodeling.org>`_
66at the Department of Physics.
67
68When using **hiphive** in your research please cite the following paper:
69
70| Fredrik Eriksson, Erik Fransson, and Paul Erhart
71| *The Hiphive Package for the Extraction of High‐Order Force Constants by Machine Learning*
72| Adv. Theory. Sim., 1800184 (2019)
73| `doi: 10.1002/adts.201800184 <https://doi.org/10.1002/adts.201800184>`_
74
75Also consult the `Credits <https://hiphive.materialsmodeling.org/credits>`_
76page of the documentation for additional references.
77
78**hiphive** and its development are hosted on
79`gitlab <https://gitlab.com/materials-modeling/hiphive>`_.
80Bugs and feature requests are ideally submitted via the
81`gitlab issue tracker <https://gitlab.com/materials-modeling/hiphive/issues>`_.
82The development team can also be reached by email via
83hiphive@materialsmodeling.org.
84