When I run mypy main.py --strict on

# main.py
import numpy as np
import numpy.typing as npt

def foo(x: npt.NDArray[np.float64]) -> npt.NDArray[np.float64]:
    return np.square(x)

I get Success: no issues found in 1 source file

But when I try to calculate the np.mean I get an error

import numpy as np
import numpy.typing as npt

def foo(x: npt.NDArray[np.float64]) -> np.float64:
    return np.mean(np.square(x))

error: Returning Any from function declared to return "floating[_64Bit]"  [no-any-return]
Found 1 error in 1 file (checked 1 source file)

What am I doing wrong? Is it because it doesn’t know the shape of x, therefore np.mean could return an array if x was more than one-dimensional? In my case, x is always one-dimensional therefore I always will return a float and I want the type-hint to reflect that the user should expect a single float returned.

Related – scalar/array multiplication return Any type