csr_matrix.py

Define the CSRmatrix class.

class openmdao.matrices.csr_matrix.CSRMatrix(submats)

Bases: COOMatrix

Sparse matrix in Compressed Row Storage format.

Parameters:

submatsdict

Dictionary of sub-jacobian data keyed by (row_name, col_name).

Attributes:

_coo_to_csr_mapndarray of int

Single array of size nnz. Entry i holds the CSR data index for COO entry i. Multiple COO entries may map to the same CSR index when duplicate (row, col) pairs exist across subjacs.

_has_within_subjac_duplicatesdict

Maps subjac key to bool. True if that subjac has duplicate (row, col) entries within its own COO slice, requiring np.add.at instead of +=.

Methods

dump(msginfo)	Dump the matrix to stdout.
todense()	Return a dense version of the matrix.
transpose()	Transpose the matrix.

__init__(submats)

Initialize all attributes.

dump(msginfo)

Dump the matrix to stdout.

Parameters:

msginfostr

Message info.

todense()

Return a dense version of the matrix.

Returns:

ndarray

Dense array representation of the matrix.

transpose()

Transpose the matrix.

Returns:

csc_matrix

Transposed matrix.