The numpy.arctan2() method computes element-wise arc tangent of arr1/arr2 choosing the quadrant correctly. The quadrant is chosen so that arctan2(x1, x2) is the signed angle in radians between the ray ending at the origin and passing through the point (1, 0), and the ray ending at the origin and passing through the point (x2, x1). 
 

Syntax : numpy.arctan2(arr1, arr2, casting = 'same_kind', order = 'K', dtype = None, ufunc 'arctan') 
Parameters : 
arr1 : [array_like] real valued; y-coordinates 
arr2 : [array_like] real valued; x-coordinates. It must match shape of y-coordinates. 
out : [ndarray, array_like [OPTIONAL]] array of same shape as x. 
where : [array_like, optional] True value means to calculate the universal functions(ufunc) at that position, False value means to leave the value in the output alone.
Note : 
2pi Radians = 360 degrees 
The convention is to return the angle z whose real part lies in [-pi/2, pi/2].
Return : Element-wise arc tangent of arr1/arr2. The values are in the closed interval [-pi / 2, pi / 2]. 
 

  
Code #1 : Working 
 

# Python3 program explaining
# arctan2() function

import numpy as np

arr1 = [-1, +1, +1, -1]
arr2 = [-1, -1, +1, +1]

ans = np.arctan2(arr2, arr1) * 180 / np.pi

print ("x-coordinates : ", arr1)
print ("y-coordinates : ", arr2)

print ("\narctan2 values : \n", ans)

Output : 

x-coordinates :  [-1, 1, 1, -1]
y-coordinates :  [-1, -1, 1, 1]

arctan2 values : 
 [-135.  -45.   45.  135.]

  
Code #2 : Working 
 

# Python3 program showing
# of arctan2() function

import numpy as np

a = np.arctan2([0., 0., np.inf], [+0., -0., np.inf])

b = np.arctan2([1., -1.], [0., 0.])

print ("a : ", a)

print ("b : ", b)

Output : 
 

a :  [ 0.          3.14159265  0.78539816]
b :  [ 1.57079633 -1.57079633]

  
References : 
https://docs.scipy.org/doc/numpy-1.13.0/reference/generated/numpy.arctan2.html#numpy.arctan2 
.