Sampling With Replacement Using Weights in Python

Here is the Python function corresponding to sample() call in R. We based it on the code here; only changed it so that the inputs use seperate weight and value vectors instead of one vector that has tuples of weight, value pairs.

import random

items = [(10, "low"),
       (100, "mid"),
       (890, "large")]

w = [10, 100, 890]
v = ["low", "mid", "large"]

def weighted_sample(ws, vs, n):
  total = float(sum(w for w in ws))
  i = 0
  w = ws[0]
  v = vs[0]
  while n:
      x = total * (1 - random.random() ** (1.0 / n))
      total -= x
      while x > w:
          x -= w
          i += 1
          w = ws[i]
          v = vs[i]
      w -= x
      yield v
      n -= 1

for i in weighted_sample(w, v, 500):
  print i