euclidean distance tool

euclidean.py

import math
import numpy as np

def get_index_distance(index1, index2 ) :
    home_x, home_y = index1
    afar_x, afar_y = index2
    squares = abs(home_x-afar_x)**2 + abs(home_y-afar_y)**2
    return math.sqrt(squares)

def get_closest_index(array, index) :
    r,c = np.nonzero(array)
    min_idx = ((r - index[0])**2 + (c - index[1])**2).argmin()
    return r[min_idx], c[min_idx]

def euclidean_distance(arr, nodata=0):
    out = -np.ones(arr.shape)
    finders = np.where(arr == nodata)
    b = arr != nodata
    for finder  in zip(*finders):
        nearest = get_closest_index(b.astype(int), finder)
        out[finder] = get_index_distance(finder,nearest)
    out[out<0] = 0
    return out

if __name__ == '__main__':

    esri = np.zeros((6,6), dtype=int)
    esri[(0,0,1),(1,2,2)] = 1
    esri[(5,0)] = 4
    euclidean_distance(esri)

    l = np.zeros((7,7), dtype=int)
    l[(3,3)] = 1
    euclidean_distance(l)