Enhancements in CSRMatrix

[vickysharma0812]

Sparse info

Add following data to CSRMatrix_ field:

• max_nnz_in_row Maximum number of non zero entries in a row

• max_nnz_in_col Maximum number of non zero entries in a col

• Add a method to get the bandwidth of the sparse matrix, which is defined as:

$$ m_{l} = max( \left [(i-j); a(i,j) \ne 0 \right ]) $$

$$ m_{u} = max( (j-i); a(i,j) \ne 0 ) $$

where,

• $m_l$ is the bandwidth of the strict lower part of A.
• $m_u$ is the bandwidth of the strict upper part of A.

Get compact row

• Add methods to get compact row of a CSR matrix

Matrix vector multiplication:

Currently, we are using Sparsekit lib for matvec.

Instead Add following methods

• amux : A times a vector. Compressed Sparse Row (CSR) format.
• atmux : Transp(A) times a vector. CSR format.
• atmuxr: Transp(A) times a vector. CSR format. A rectangular.
• amuxms: A times a vector. Modified Compress Sparse Row format.
• B = scale * A by using Set(obj, value, scale)
• C = A + B
• C = A + scale * B

Others

• Implement GetDia method, currently we are using Sparsekit for this purpose.

Misc

• Make functions like scipy https://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.csr_matrix.html
• Make functions like scipy for COO https://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.coo_matrix.html
• Support for CSR, CSC, and CO
• https://docs.scipy.org/doc/scipy/reference/sparse.html