CODEWITHSUNDEEP
python-2.7numpy
dranobob
dranobob

Reputation: 836

How to shift numpy elements to be between a set interval

I need to take a series of radian values and convert them to be in the interval of -pi to pi.

Here is the function for singular values

def angle_truncation(angle):
  while angle < 0.0:
    angle += np.pi * 2
  return ((angle + np.pi) % (np.pi * 2)) - np.pi

angle = 5.
print(angle_truncation(angle))

>> -1.28318530718

Because of the while loop though this won't work as is with numpy arrays. So would the above function be converted to work with numpy arrays using vectorization/broadcasting instead of just adding in a for loop?

ie

a = np.fill((3, 1), 5.)
print(angle_truncation(a))

>> [[-1.28318530718, -1.28318530718, -1.28318530718]]

Upvotes: 1

Views: 479

Answers (2)

wagnerpeer
wagnerpeer

Reputation: 947

How about calculating the fraction of pi and then adding the resulting multiple of it to the array.

import numpy as np

def angle_trunc(array) :
    below_pi = array < np.pi 
    fractions = np.abs(array[below_pi]) / (2 * np.pi) 
    array[below_zero] += np.ceil(fractions) * (2 * np.pi) 
    return (array % (2 * np.pi)) - np.pi

Upvotes: 1

llllllllll
llllllllll

Reputation: 16404

This while loop is needless, you only need to take the ceil() of the quotient:

angle_trunc = lambda a: (a+np.ceil(abs(a)/(2*np.pi))*2*np.pi+np.pi)%(np.pi*2) - np.pi

Upvotes: 1

Related Questions