thearn

import pprocess
import time
import numpy as np

# Define a function to parallelize:
def takeuptime(x):
    """A function to waste CPU cycles"""
    time.sleep(0.1)
    return x[-1]
    

thearn

import pickle
import gzip

def save(object, filename, protocol = 0):
        """Saves a compressed object to disk
        """
        file = gzip.GzipFile(filename, 'wb')
        file.write(pickle.dumps(object, protocol))
        file.close()

thearn

import numpy as np


def low_rank_approx(SVD=None, A=None, r=1):
    """
    Computes an r-rank approximation of a matrix
    given the component u, s, and v of it's SVD

    Requires: numpy
    """

thearn

from numpy.fft import fft, ifft, fft2, ifft2, fftshift
import numpy as np

def fft_convolve2d(x,y):
    """ 2D convolution, using FFT"""
    fr = fft2(x)
    fr2 = fft2(np.flipud(np.fliplr(y)))
    m,n = fr.shape
    cc = np.real(ifft2(fr*fr2))
    cc = np.roll(cc, -m/2+1,axis=0)