Tessero\Sparse\CsrMatrix

fromDense(mixed $dense): self

fromTriplets(Tessero\NDArray|array $row, Tessero\NDArray|array $col, Tessero\NDArray|array $values, array $shape): self

From coordinate triplets (scipy.sparse.coo_matrix(...).tocsr()); duplicates are summed.

fromArrays(mixed $indptr, mixed $indices, mixed $data, array $shape): self

Wrap existing CSR arrays (validated).

identity(int $n): self

shape(): array

nnz(): int

indptr(): Tessero\NDArray

indices(): Tessero\NDArray

data(): Tessero\NDArray

toDense(): Tessero\NDArray

dot(mixed $x): Tessero\NDArray

A @ x for a dense vector (n,) or matrix (n, k).

transpose(): self

scale(float $alpha): self

diagonal(): Tessero\NDArray

Main diagonal as a dense vector.

cg(mixed $b, mixed $x0 = null, float $rtol = 1.0E-5, float $atol = 0.0, ?int $maxiter = null, bool $throw = false): array

Conjugate gradients (scipy.sparse.linalg.cg) with Jacobi preconditioning; A must be SPD.

bicgstab(mixed $b, mixed $x0 = null, float $rtol = 1.0E-5, float $atol = 0.0, ?int $maxiter = null, bool $throw = false): array

BiCGSTAB for general square systems (scipy.sparse.linalg.bicgstab).