pyacs.lib.glinalg.odot

Scale block rows of G by vector a.

pyacs.lib.glinalg.odot.odot(a, G)

Scale block rows of G by vector a (element-wise then reshape).

G is treated as a stack of submatrices G_1,…,G_n; result is [a_1*G_1; …; a_n*G_n]. Implemented with numpy broadcasting. If G.shape = (n, m) and a.shape = (l,), then each block has n/l rows.

Parameters:

• a (numpy.ndarray) – 1D array of scalars (multipliers).

• G (numpy.ndarray) – 2D array; number of rows must be divisible by len(a).

Returns:

Matrix with same shape as G; each block of rows scaled by corresponding a[i].

Return type:

numpy.ndarray