readability.py

"""Readability

[Wiki](http://en.wikipedia.org/wiki/Flesch%E2%80%93Kincaid_readability_test)

The Flesch Reading Ease Score (FRES) is
206.835 - 1.015 * (#words / #sentences) - 84.6 * (#syllables / #words)

Example:
"The Australian platypus is seemingly a hybrid of a mammal and reptilian
creature."

This (one) sentence has thirteen words and twenty four syllables. So its score
is
206.835 - 1.015 * (words / sentences) - 84.6 * (syllables / words)

>>> text = "The Australian platypus is seemingly a hybrid of a mammal and " + \
           "reptilian creature."
>>> readability(text)
37.455384615384645
"""

from __future__ import division

import sys

from nltk.corpus import cmudict
from nltk.tokenize import word_tokenize, sent_tokenize


_syll_map = cmudict.dict()

def num_syllables(word):
    """Return the number of syllables in `word`."""

    if word in _syll_map:
        return max(len([phone for phone in phones if phone[-1] in '0123456789'])
                   for phones in _syll_map[word])
    else:
        return None


def readability(text):
    """Return the Flesch Reading Ease Score of `text."""

    words = sentences = syllables = 0
    for sent in sent_tokenize(text):
        sentences += 1

        for word in word_tokenize(sent):
            word = word.lower()
            num_sylls = num_syllables(word)

            if num_sylls is not None:
                words += 1
                syllables += num_sylls

    if syllables == 0:
        return None

    score = (206.835 -
             1.015 * (words / sentences) -
             84.6 * (syllables / words))

    return score


def main(args):
    while True:
        try:
            text = raw_input('Text: ')
        except EOFError:
            print
            return 0
        print readability(text)


if __name__ == '__main__':
    exit(main(sys.argv[1:]))