distances.py

#!/usr/bin/python
# -*- coding: utf-8 -*-
#
# Author        :   Vitor Rodrigues Di Toro
# E-Mail        :   vitorrditoro@gmail.com
# Date          :   14/03/2018
# Last Update   :   19/03/2018
#

import unittest
   


def euclideanDistance(p1, p2):
    '''
      EuclideanDistance Distance implementation
      See: https://en.wikipedia.org/wiki/Euclidean_distance
    
      Returns the EuclideanDistance distance between two vectors.
      
      - p1 : first array
      - p2 : second array
    '''
    
    sum = 0
    
    for i in range(len(p1)):
        sum += (p1[i]-p2[i])**2
   
    
    return sum**(0.5)
    

    
def manhattanDistance(p1, p2):
    '''
      ManhattanDistance Distance implementation
      See: https://en.wikipedia.org/wiki/Taxicab_geometry
     
      Returns the ManhattanDistance distance between two vectors.
      
      - p1 : first array
      - p2 : second array
    '''
    
    sum = 0
    
    for i in range(len(p1)):
        sum += abs(p1[i]-p2[i])
        
    return sum



def minkowskiDistance(p, q, n):
    '''
      Minkowski Distance implementation
      See: http://en.wikipedia.org/wiki/Minkowski_distance
     
      Returns the Minkowski distance between two vectors.
      
      - p : first array
      - q : second array
      - n : distance order
    '''
    
    sum = 0
    
    for x,y in zip(p,q):

        sum += abs(x-y)**n

    return sum**(1 / n)
    


class TestDistances(unittest.TestCase):
    
    def test_euclideanDistance(self):
        #ref: http://calculator.vhex.net/calculator/distance/euclidean-distance
        
        p1 = [0, 0, 0]
        p2 = [2, 2, 2]
        self.assertEqual(euclideanDistance(p1,p2), (2*(3) ** (0.5)))
        
        p1 = [0, 5, -8,  9, 3]
        p2 = [2, 7, -9, -1, 3]
        self.assertEqual(round(euclideanDistance(p1,p2),6), 10.440307)
        
        p1 = [0.500, 5.30, -8.2,  9.54, 3.134]
        p2 = [2.123, 7.43, -9.2, -1.50, 3.000]
        self.assertEqual(round(euclideanDistance(p1,p2),6), 11.404849)       
        
    def test_manhattanDistance(self):
        #ref: http://calculator.vhex.net/calculator/distance/manhattan-distance
        
        p1 = [0, 0, 0]
        p2 = [2, 2, 2]
        self.assertEqual(manhattanDistance(p1,p2), 6)
        
        p1 = [0, 5, -8,  9, 3]
        p2 = [2, 7, -9, -1, 3]
        self.assertEqual(manhattanDistance(p1,p2), 15)
        
        p1 = [0.500, 5.30, -8.2,  9.54, 3.134]
        p2 = [2.123, 7.43, -9.2, -1.50, 3.000]
        self.assertEqual(manhattanDistance(p1,p2), 15.927)
        
    def test_minkowskiDistance(self):     
        #ref: http://people.revoledu.com/kardi/tutorial/Similarity/MinkowskiDistance.html
        
        p1 = [0, 3, 4,  5]
        p2 = [7, 6, 3, -1]
        n  = 3
        self.assertEqual(round(minkowskiDistance(p1,p2,n),3), 8.373)
        
        p1 = [0.5 , 3.9 , -4.37,  5.5]
        p2 = [7.72, 6.36,  3.27, -1.98]
        n  = -0.25
        self.assertEqual(round(minkowskiDistance(p1,p2,n),5), 0.02139)
        
        p1 = [0.5 , 3.9 , -4.37,  5.5]
        p2 = [7.72, 6.36,  3.27, -1.98]
        n  = 4
        self.assertEqual(round(minkowskiDistance(p1,p2,n),3), 9.818)
    
    

if __name__ == "__main__":
    unittest.main()