sparse matrix support about graspologic

[bdpedigo]

https://scikit-learn.org/stable/modules/generated/sklearn.utils.extmath.safe_sparse_dot.html#sklearn.utils.extmath.safe_sparse_dot

[j1c]

Never really understood the point of that function.

def safe_sparse_dot(a, b, dense_output=False):
    """Dot product that handle the sparse matrix case correctly
    Uses BLAS GEMM as replacement for numpy.dot where possible
    to avoid unnecessary copies.
    Parameters
    ----------
    a : array or sparse matrix
    b : array or sparse matrix
    dense_output : boolean, default False
        When False, either ``a`` or ``b`` being sparse will yield sparse
        output. When True, output will always be an array.
    Returns
    -------
    dot_product : array or sparse matrix
        sparse if ``a`` or ``b`` is sparse and ``dense_output=False``.
    """
    if sparse.issparse(a) or sparse.issparse(b):
        ret = a * b
        if dense_output and hasattr(ret, "toarray"):
            ret = ret.toarray()
        return ret
    else:
        return np.dot(a, b)

Just does * instead of @, which I think is equivalent in later versions of python.

import numpy as np
from scipy.sparse import csr_matrix

X = csr_matrix(np.random.normal(size=(100, 4)))
U = X * X.T
V = X @ X.T

(U != V).nnz == 0
>>> True

[j1c]

TODO:

• ASE should require no changes
• LSE might require changes to to_laplace
• Omni requires change in omnibus matrix creation
• cmds requires change for matrix pairwise distance computation, but no change for vector pairwise distance computation

[j1c]

More TODO:

• Change PTR to output sparse matrix if input is sparse
• Changes to simulations to output sparse matrix similar to link
  • This is lots of work
• Need to check the following if output is sparse if input is sparse
  • LCC functions
  • symmetrize
  • augment_diagonal
  • remove_loops