differential_geometry.dense_utils.dense_permute_tensor_indices#

differential_geometry.dense_utils.dense_permute_tensor_indices(tensor_field: ndarray, permutation: List[int], rank: int) → ndarray[source]#

Permutes the order of the tensor indices in the last rank dimensions of the tensor field.

Parameters:

  • tensor_field (numpy.ndarray) – Input tensor with shape (…, i₁, …, i_r), where the last rank axes are tensor indices.

  • permutation (list of int) – Permutation to apply to the tensor indices. Must be of length rank and contain a permutation of [0, …, rank-1].

  • rank (int) – The rank (number of tensor axes) of the tensor.

Returns:

Tensor field with permuted tensor indices.

Return type:

numpy.ndarray

Raises:

ValueError – If permutation is not valid.

Examples

>>> import numpy as np
>>> T = np.random.rand(4, 4, 3, 3)
>>> permute_tensor_indices(T, [1, 0], rank=2).shape
(4, 4, 3, 3)