How Does Frompyfunc Iterate Over Arrays?

I am working with large arrays representing a maze routing grid, each element is a Cell object with x,y attributes. I am trying to use numpyfunc to initialize the coordinates in ea

Solution 1:

From your comments and suppose that Cell objects have x, y attributes and some other default attribute that doesn't come to play:

classCell:def__init__(self, x,y):
        self.x = x
        self.y = y
    ...

Suppose you want a 100*100 array, initiate your array like this:

CellList = [[Cell(x,y) for y in range(100)] for x in range(100)]
# Optional translate it into np.array for easier indexingCellArr = np.array(CellList)

This will return your 100*100 Cell array that has correct Cell elements. To verify:

CellArr[1,2].x
>>>1

Note that numpy can't actually speed up your array much because Cell cannot actually go through C code when vectorizing. It could only be used for better indexing.

Vectorizing does not actually help your speed:

%%timeit
CellList = [[Cell(x,y) for y inrange(100)] for x inrange(100)]
# Optional translate it into np.array for easier indexing
CellArr = np.array(CellList)

>>> 24.2 ms ± 542 µs per loop

Vectorizing functions:

def vecX(c, x):c= Cell(x,0)returnc

def vecY(c, y):
    c.y = y
    returnc

vec = np.vectorize(vecX)
vey = np.vectorize(vecY)

results:

%%timeit
l = []
n = np.zeros((100,100))
for i in range(len(n)):
    l.append(vec(n[i, :], i))
CellArr = np.vstack(l)
for j in range(len(CellArr)):
    vey(CellArr[:, j], j)

>>> 23.5 ms ± 5 ms per loop