Utils

Warning

This page of the documentation is unfinished, and may significantly change in future versions.

Constraints and templates

class popcor.utils.constraint.Constraint(index=0, polyrow_a=None, polyrow_b=None, A_poly=None, A_sparse=None, b=None, a=None, a_full=None, b_full=None, mat_var_dict=None, mat_param_dict=None, known=False, template_idx=0)

This class serves the main purpose of not recomputing representations of constraints more than once.

Helpers for matrix and vector operations

popcor.utils.common.create_symmetric(vec, eps_sparse, correct=False, sparse=False)

Create a symmetric matrix from the vectorized elements of the upper half

popcor.utils.common.diag_indices(n)

Given the half kronecker product, return diagonal elements

popcor.utils.common.get_vec(mat, correct=True, sparse=False) → ndarray | csr_matrix | None

Convert NxN Symmetric matrix to (N+1)N/2 vectorized version that preserves inner product.

Parameters:

mat – (spmatrix or ndarray) symmetric matrix

Returns:

ndarray

popcor.utils.common.ravel_multi_index_triu(index_tuple, shape)

Equivalent of np.multi_index_triu, but using only the upper-triangular part of matrix.

popcor.utils.common.unravel_multi_index_triu(flat_indices, shape)

Equivalent of np.multi_index_triu, but using only the upper-triangular part of matrix.

popcor.utils.common.upper_triangular(p)

Given vector, get the half kronecker product.