[SciPy-User] Creating meshgrid from meshgrid

[Pauli Virtanen]

Florian Lindner kirjoitti 06.09.2017 klo 06:31:
[clip]
>      def fun(x):
>          return x
> 
>      mgrid = np.meshgrid( a,b )
>      A = fun( np.abs(mgrid[0] - mgrid[1] ) )
> 
> Result A is of size N x N:
> 
>      array([[ 9,  8,  7],
>             [19, 18, 17],
>             [29, 28, 27]])

Don't use meshgrid, use broadcasting:

from numpy import newaxis

ax = np.array([0, 1, 2])
ay = np.array([4, 5])
bx = np.array([7, 8, 9, 10])
by = np.array([11, 12, 13, 14])

def fun(x1, y1, x2, y2):
     return abs(x1 - x2) + abs(y1 - y2)

d = fun(ax[:,newaxis,newaxis,newaxis],
         ay[newaxis,:,newaxis,newaxis],
         bx[newaxis,newaxis,:,newaxis],
         by[newaxis,newaxis,newaxis,:])

or

p1 = np.array([[0,1], [1,3], [5,6], [7,9]])
p2 = np.array([[3,6], [2,7], [9,2], [1,3]])

def fun(a, b):
     return np.hypot(a[...,0] - b[...,0], a[...,1] - b[...,1])

d = fun(p1[:,newaxis,:], p2[newaxis,:,:])

depending on what it is that you actually want to evaluate.

-- 
Pauli Virtanen