Exercice Sam & Max

# encoding: utf-8
from __future__ import unicode_literals, absolute_import, print_function

import sys
import re
import unicodedata
from collections import defaultdict

from pprint import pprint

def main():

    with open(sys.argv[1]) as f:
        text = f.read().decode('utf-8').lower()

    re_punctuation = re.compile('[^a-z0-9]')

    text = ''.join(c for c in unicodedata.normalize('NFD', text)
                  if unicodedata.category(c) != 'Mn')
    text = text.replace("œ", "oe")
    text = re_punctuation.sub(' ', text).strip()

    words = re.split(r' +', text)

    df = defaultdict(list)
    for pos, word in enumerate(words):
        df[word].append(pos)

    df = sorted(df.items(), key = lambda (word,positions):(len(positions), sorted(positions)))

    pprint(df)


if __name__ == "__main__":
    main()