# -*- coding: utf-8 -*-
#!/usr/bin/env python
from collection import Collection
from BeautifulSoup import UnicodeDammit
from math import sqrt
from string import punctuation

def depunctuate(text):
    puncs = list(punctuation)
    for p in puncs: text = text.replace(p,'')
    return text

def freq(seq):
    fq = {}
    for e in seq:
      if e not in fq: fq[e] = 1
      else: fq[e] += 1
    return fq 

def normalize(seq): 
    self.maximum = max(seq)
    return [ int(float(n)/maximum * 100) for n in seq]    

def scalar(vec):
    total = 0 
    for elem in vec:
        total += vec[elem] * vec[elem]
    return sqrt(total)

def sim(v, w):
    total = 0 
    for elem in v:
        if elem in w:
            total += v[elem] * w[elem]
    return float(total) / (scalar(v) * scalar(w))


def _test():
   import doctest
   import align
   myoptionflags = doctest.ELLIPSIS|doctest.NORMALIZE_WHITESPACE
   doctest.testmod(optionflags=myoptionflags)

if __name__ == "__main__":
   _test()