18091, "prashanth018", "Addition of NumPy like methods in Linear Algebra module.", "2021-07-21T19:13:58Z"
As part of this feature (implementation of onenormestimate for Matrix Exponentials), I have implemented few methods which could also be used publicly. Nevertheless, these methods are open for improvements (performance/readability). Also please do comment on what other generic numpy-like methods could be added to LA. Let this issue be used as tracking for suggestions or other discussions.
sparseDenseMatmul: Mutliplies Sparse and Dense Matrix sign_round_up: Rounds up the elements of the Matrix to 1.0 if element >= 0, -1.0 otherwise argsort: Given an array A, returns indices corresponding to the sorted ordering of A. elementary_vector: Returns an index's unit vector of size n. zeros: Given the Domain and element-type, the method returns a Matrix of zeros. ones: Given the Domain and element-type, the method returns a Matrix of ones. max: Gets max along specified axes. (Method for Sparse matrix) [currently only supports axes 0 and 1] absSum: Gets sums along specified axes. (Added Methods for both Sparse and Dense matrix) [currently only supports axes 0 and 1]