FieldRotator#

class discretisedfield.FieldRotator(field)#

Rotate a field.

This class can be used to rotate a field object. During rotation a new region and mesh are constructed and field values are interpolated onto the new mesh. Multiple consecutive rotations are possible without additional numerical errors. Rotation always starts from the initial unrotated field.

Periodic boundary conditions have no effect.

Parameters:

field (discretisedfield.Field) – Field to rotate.

Examples

>>> import discretisedfield as df
>>> from math import pi

Create a field to rotate.

>>> region = df.Region(p1=(0, 0, 0), p2=(20, 10, 2))
>>> mesh = df.Mesh(region=region, cell=(1, 1, 1))
>>> field = df.Field(mesh, nvdim=3, value=(0, 0, 1))
>>> field.mesh.n
array([20, 10,  2])

Create a FieldRotator object for the field.

>>> field_rotator = df.FieldRotator(field)

Rotate the field.

>>> field_rotator.rotate('from_euler', seq='x', angles=pi/2)

Access the rotated field.

>>> field_rotator.field
Field(...)
>>> field_rotator.field.mesh.n
array([20,  2, 10])

Apply a second rotation.

>>> field_rotator.rotate('from_euler', seq='z', angles=pi/2)
>>> field_rotator.field.mesh.n
array([ 2, 20, 10])

Methods

__dir__

Default dir() implementation.

__eq__

Return self==value.

__repr__

Representation string.

clear_rotation

Remove all rotations.

rotate

Rotate the field.

Properties

field

Rotated field.
__repr__()#

Representation string.

Internally self._repr_html_() is called and all html tags are removed from this string.

Returns:

Representation string.

Return type:

str

Example

  1. Getting representation string.

>>> import discretisedfield as df
...
>>> p1 = (0, 0, 0)
>>> p2 = (2, 2, 1)
>>> cell = (1, 1, 1)
>>> mesh = df.Mesh(p1=p1, p2=p2, cell=cell)
...
>>> field = df.Field(mesh, nvdim=1, value=1)
>>> rotator = df.FieldRotator(field)
>>> rotator
FieldRotator(...)
clear_rotation()#

Remove all rotations.

rotate(method, /, *args, n=None, **kwargs)#

Rotate the field.

Rotate the field using the given method. The definition of the rotation is based on scipy.spatial.transform.Rotation. Additional parameters required for the different possible rotation methods must be specified following to the scipy documentation. These are passed directly to the relevant scipy function. For a detailed explanation and required arguments of the different methods please refer directly to the scipy documentation.

The only method that differs from scipy is align_vector. This method expects two keyword arguments initial and final (array-like, 3). This method rotates the vector initial to the vector final, the cross product is kept fixed, i.e. it defines the rotation vector.

The rotation of the field consists of three steps, rotation of the region, remeshing, and rotation + interpolation of the field values. Rotation of the region produces as new, larger region for the new field. If n is not specified remeshing is done automatically and the cell volume is kept mostly constant. Interpolation of the field values is done using linear interpolation. For more details on the rotation process please refer to the detailed documentation notebook.

Parameters:

  • method (str) – Rotation method. One of 'from_quat', 'from_matrix', 'from_rotvec', 'from_mpr', 'from_euler', or 'align_vector'.

  • args – Additional positional arguments for the rotation method.

  • n (array-like, 3, optional) – Number of cells in the new mesh. If not specified n is chosen automatically to keep the cell volume mostly constant. Defaults to None.

  • kwargs – Additional keyword arguments for the rotation method.

Examples

>>> import discretisedfield as df
>>> from math import pi
>>> region = df.Region(p1=(0, 0, 0), p2=(20, 10, 2))
>>> mesh = df.Mesh(region=region, cell=(1, 1, 1))
>>> field = df.Field(mesh, nvdim=3, value=(0, 0, 1))
>>> field_rotator = df.FieldRotator(field)
>>> field_rotator.rotate('from_euler', seq='x', angles=pi/2)
property field#

Rotated field.