TopologyArray#

class sigmaepsilon.mesh.topoarray.TopologyArray(*topo, cuts: Iterable | None = None, force_numpy: bool = True)[source]#

A class to handle complex topologies. It is a subclass of the JaggedArray class from the Neumann library and is compatible with Numpy universal functions.

Parameters:

*topo (Iterable) – One or more 2d arrays definig topologies for polygonal cells.
cuts (Iterable, Optional) – An iterable that tells how to unflatten an 1d array into a 2d jagged shape. Only if topology is provided as a 1d array. Default is None.
force_numpy (bool, Optional) – Forces dense inputs to be NumPy arrays in the background. Default is True.

Examples

The following could be the definiton for a mesh consisting from two line cells, one with 3 nodes and another with 2:

>>> import numpy as np
>>> from sigmaepsilon.mesh import TopologyArray
>>> data = np.array([0, 1, 2, 3, 4])
>>> TopologyArray(data, cuts=[3, 2])
TopologyArray([[0, 1, 2], [3, 4]])

The same mesh defined in another way:

>>> topo1 = np.array([0, 1, 2])
>>> topo2 = np.array([3, 4])
>>> TopologyArray(topo1, topo2)
TopologyArray([[0, 1, 2], [3, 4]])

Let assume we have two 4-noded quadrilaterals as well:

>>> topo3 = np.array([[5, 6, 7, 8],[6, 7, 9, 10]])
>>> TopologyArray(topo1, topo2, topo3)
TopologyArray([[0, 1, 2], [3, 4], [5, 6, 7, 8], [6, 7, 9, 10]])

Since the TopologyArray class is a subclass of JaggedArray, we can easily transform it to a CSR matrix, or an Awkward array:

>>> TopologyArray(topo1, topo2, topo3).to_csr()
>>> TopologyArray(topo1, topo2, topo3).to_ak()
<Array [[0, 1, 2], [3, 4, ... [6, 7, 9, 10]] type='4 * var * int32'>

To get the unique indices in a mesh, you can simply use NumPy:

>>> t = TopologyArray(topo1, topo2, topo3)
>>> np.unique(t)
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10])

You can also combine different topologies into one object by stacking:

>>> t1 = TopologyArray(topo1, topo2)
>>> t2 = TopologyArray(topo3)
>>> np.vstack([t1, t2])
TopologyArray([[0, 1, 2], [3, 4], [5, 6, 7, 8], [6, 7, 9, 10]])