Python Random Lines

randomlines.py

import os.path
import random

class RandomLines(object):
	def __init__(self, input_file, cache_index=True):
		if isinstance(input_file, basestring):
			self.source_file = open(input_file,'rb')
			filename = input_file
		else:
			self.source_file = input_file
			filename = input_file.name
		self.index = []

		if not os.path.isfile(filename+'.lineindex'):
			bytes_counter = 0
			for line in self.source_file:
				bytes_counter += len(line)
				if len(line.strip()):
					self.index.add(bytes_counter-len(line))
			if cache_index:
				open(filename+'.lineindex','w').write('\n'.join(str(i) for i in self.index))
		else:
			for line in open(filename+'.lineindex'):
				self.index.append(int(line.strip()))

	def __iter__(self):
		return self

	def next(self):
		while len(self.index):
			offset = self.index.pop(random.randrange(0,len(self.index)))
			self.source_file.seek(offset,0)
			return self.source_file.readline().strip()
		raise StopIteration