return a function with partial argument embedded

Question

I have a function which takes two numpy arrays, x and y, as arguments. I want to be able to pass in only one argument, y. How can I return a function that accepts x, with y embedded in original function?

For example:

def z(x: [x1, x2, x3], y: [y1, y2, y3]):
    return x1 * y1 + x2 * y2 + x3 * y3

I want z(y=[2, 3, 4]) to return x1 * 2 + x2 * 3 + x3 * 4 as a new function.

That's not the correct syntax for deconstructing the arguments, by the way. — chepner, Jul 12 '22 at 13:12
Python 2 allowed you to write `def z((x1, x2, x3), (y1, y2, y3]): ...` (providing a tuple of names for the argument to unpack into *instead* of a single name). I miss that syntax sometimes. — chepner, Jul 12 '22 at 13:19

score 3 · Answer 1 · answered Jul 12 '22 at 13:11

3

Use functools.partial:

>>>  from functools import partial
>>> z_with_y = partial(z, y=[2,3,4])
>>> z_with_y([5,6,7])
56
>>> z([5,6,7], [2,3,4])
56

answered Jul 12 '22 at 13:11

chepner

497,756
71
530
681

Mad Physicist · Answer 2 · 2022-07-12T13:13:29.257

3

Given z = np.dot, as in your example, or any other 2-arg function, you can make a partial fucuntion

def z1(y):
    return lambda x: z(x, y)

Without using lambda:

def z1(y):
    def inner(x):
        return z(x, y)
    return inner

edited Jul 12 '22 at 13:13

answered Jul 12 '22 at 13:12

Mad Physicist

107,652
25
181
264

I think you need to treat `z` as a black box, not knowing that it's computing a dot product. – chepner Jul 12 '22 at 13:12
@chepner. Good call. Not sure why I put `np.dot` in the return instead of `z`. Fixed now – Mad Physicist Jul 12 '22 at 13:13

return a function with partial argument embedded

2 Answers2