Using np.where in cython as index

Question

I am trying to use np.where in a cython program to select where values are within a radius. Ususally this would work with numpy, but with cython I get

TypeError: 'tuple' object cannot be interpreted as an integer

def test(double[:,::1] array, double[:] point, double radius):
    array = np.asarray(array)
    idx = np.where(np.logical_and(np.greater_equal(array[:, 0], point[0] - radius), np.less_equal(array[:, 0], point[0] + radius)))
    a = array[idx]

I tried doing this but got error

only integer scalar arrays can be converted to a scalar index

a = array[idx[0].astype(int)]

Answer 1

奇怪的是，即使我已经做了 np.asarray，它也能做到这一点

a = np.array(array)[idx[0]]