Random Number Generator in Python

random.py

# imports for abstract classes
from abc import ABCMeta, abstractmethod
import time


''' ==============================================================================================
Seed Classes
==============================================================================================='''
class Seed(object):
  """Abstract class for seeds"""

  __metaclass__ = ABCMeta

  @abstractmethod
  # Function that generates seed
  def generate_seed(self):
    pass

class TimeSeed(Seed):
  """ Generates seed from current time """

  def generate_seed(self):
    return time.time()

''' ==============================================================================================
End Seed Classes
==============================================================================================='''



''' ==================================================================================================
Randomizer Classes
==================================================================================================='''
class Randomizer(object):
  """Abstract class to generate random numbers"""

  __metaclass__ = ABCMeta

  @abstractmethod
  # Function that generates random numbers
  # 'Decorate' seed onto the random number generator
  def random(self, start, end, seed):
    pass


class XORShiftRandomizer(Randomizer):
  """XOR Shift Randomizer"""
  
  def random(self, start, end, seed):

    random_number = seed.generate_seed()

    # Get number after decimal point of seed because these are the numbers that actually vary
    random_number = random_number % 1
    random_number = str(random_number)

    # Check if the number is decimal first
    if random_number.find('.') != -1:
      random_number = random_number.split('.')[1]

    # Do not split if no decimal point and just take the integer as it is
    random_number = int(random_number)

    random_number ^= (random_number << 21);
    random_number ^= (random_number >> 35);
    random_number ^= (random_number << 4);

    # Convert the generated number to lie between start and end
    random_number = random_number % end
    if random_number < start:
      random_number = random_number + start

    return random_number


''' ==================================================================================================
End Randomizer Classes
==================================================================================================='''



if __name__ == "__main__":
  number_of_numbers = int(raw_input())
  number_of_digits = int(raw_input())

  xor_random = XORShiftRandomizer()

  if number_of_numbers == 0:
    exit(0)

  # Calculate start and end of the range from number of digits
  # start and end are starting as string because 0/9 will be appended to them and start and end will be generated. They will later be converted to integer
  start = "1"
  end = "9"
  if number_of_digits == 0:
    exit(0)
  elif number_of_digits == 1:
    start = 0
  else:
    while number_of_digits > 1:
      start = start + "0"
      end = end + "9"
      number_of_digits = number_of_digits - 1
  
  start = int(start)
  end = int(end)
  
  while number_of_numbers > 0:
    print xor_random.random(start, end, TimeSeed())
    number_of_numbers = number_of_numbers - 1

lines 64, 65 and 66 are the lines that actually perform the randomisation operation. They are actually the core of randomisation function.
I've used XORShift randomisation technique which performs an "OR" operation between the number with its 'shifts'. The numbers 21, 35 and 4 are known to provide best possible results.

Note: This is not cryptographically secure but generates a true random number which can be used for generating random ID's etc.

Just curious, what do you think about this article touting further speed improvements: http://xorshift.di.unimi.it/