.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        Click :ref:`here <sphx_glr_download_examples_rotation3d.py>`     to download the full example code
    .. rst-class:: sphx-glr-example-title

    .. _sphx_glr_examples_rotation3d.py:


Rotation in 3D Space
====================

This tutorial demonstrates how rotate a group of points in 3D space.

To begin with, we must understand that rotation is a linear transformation.
The set of all rotations forms the group SO(3). It is a Lie group that each
group element is described by an orthogonal 3x3 matrix.

That is, if you have two points :math:`\vec{r}_1` and :math:`\vec{r}_2`, and
you want to rotate the two points along the same axis for the same number of
degrees, then there is a single orthogonal matrix :math:`R` that no matter
the value of :math:`\vec{r}_1` and :math:`\vec{r}_2`, their rotation is always
:math:`R\cdot\vec{r}_1` and :math:`R\cdot\vec{r}_2`.

Let's import libraries first


.. code-block:: default

    import math
    import torch
    import pytest
    import sys
    import nnp.so3 as so3








Let's first take a look at a special case: rotating the three unit vectors
``(1, 0, 0)``, ``(0, 1, 0)`` and ``(0, 0, 1)`` along the diagonal for 120
degree, this should permute these points:


.. code-block:: default

    rx = torch.tensor([1, 0, 0])
    ry = torch.tensor([0, 1, 0])
    rz = torch.tensor([0, 0, 1])
    points = torch.stack([rx, ry, rz])








Now let's compute the rotation matrix


.. code-block:: default

    axis = torch.ones(3) / math.sqrt(3) * (math.pi * 2 / 3)
    R = so3.rotate_along(axis)








Note that we need to do matrix-vector product for all these unit vectors
so we need to do a transpose in order to use ``@`` operator.


.. code-block:: default

    rotated = (R @ points.float().t()).t()
    print(rotated)






.. rst-class:: sphx-glr-script-out

 Out:

 .. code-block:: none

    tensor([[ 0.0000e+00,  1.0000e+00, -2.1865e-08],
            [ 0.0000e+00,  0.0000e+00,  1.0000e+00],
            [ 1.0000e+00,  0.0000e+00,  0.0000e+00]])




After this rotation, the three vector permutes: rx->ry, ry->rz, rz->rx. Let's
programmically check that. This check will be run by pytest later.


.. code-block:: default

    def test_rotated_unit_vectors():
        expected = torch.stack([ry, rz, rx]).float()
        assert torch.allclose(rotated, expected, atol=1e-5)









Now let's run all the tests


.. code-block:: default

    if __name__ == '__main__':
        pytest.main([sys.argv[0], '-v'])




.. rst-class:: sphx-glr-script-out

 Out:

 .. code-block:: none

    ============================= test session starts ==============================
    platform linux -- Python 3.7.7, pytest-5.4.3, py-1.8.2, pluggy-0.13.1 -- /opt/hostedtoolcache/Python/3.7.7/x64/bin/python
    cachedir: .pytest_cache
    rootdir: /home/runner/work/nnp/nnp, inifile: setup.cfg
    collecting ... collected 1 item

    rotation3d.py::test_rotated_unit_vectors PASSED                          [100%]

    ============================== 1 passed in 0.01s ===============================





.. rst-class:: sphx-glr-timing

   **Total running time of the script:** ( 0 minutes  0.643 seconds)


.. _sphx_glr_download_examples_rotation3d.py:


.. only :: html

 .. container:: sphx-glr-footer
    :class: sphx-glr-footer-example



  .. container:: sphx-glr-download sphx-glr-download-python

     :download:`Download Python source code: rotation3d.py <rotation3d.py>`



  .. container:: sphx-glr-download sphx-glr-download-jupyter

     :download:`Download Jupyter notebook: rotation3d.ipynb <rotation3d.ipynb>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_