West London Hack Night 14/08/14 - Soundex Algorithm

hacknight.py

import sys
import soundex
import random
import pickle

soundalikes = {}

def load_dict():
    with open('/usr/share/dict/words') as f:
        for line in f:
            for word in line.split():
                key = soundex.soundex(word)
                soundalikes.setdefault(key, []).append(word)

def process_text(line):
    for word in line.split():
        if len(word) < 5:
            yield word
        else:
            key = soundex.soundex(word)
            choices = soundalikes.get(key, [word])
            yield random.choice(choices)

if __name__ == '__main__':
    try:
        with open('dict', 'r') as infile:
            soundalikes = pickle.load(infile)
    except IOError:
        load_dict()
        with open('dict', 'w') as outfile:
            pickle.dump(soundalikes, outfile)
    for line in sys.stdin:
        print ' '.join(list(process_text(line)))

output

Load up on gynoecia bromacetanilide your frantically
It's fun to lose and to pratingly
Sosia over breth and self assortedness
Oh no, I know a derride word

hoyle hulu hello how low? [x3]
hele Halawi helluo

With the liquidogenic out, it's less dangerless
Here we are now, entermete us
I feel stafette and conducive
Here we are now, endermically us
A mullid
An albumosuria
A megawatt
My libbet
Yay! [x3]

I'm warehouse at what I do best
And for this gift I feel ballstock
Our lately garava has Alaki been
And Allasch will undelivery the end

haole hale hale how low? [x3]
hill hala hollo

With the leucotic out, it's less dunger
Here we are now, enterable us
I feel stuffed and candescence
Here we are now, enterochirurgia us
A militia
An Alphonso
A muscot
My lupoid
Yay! [x3]

And I forestem just why I touchwood
Oh yoe I gecko it Maycock me smally
I font it Horatio it's hard to find
Oh whilly wetbird nabber mind

Hal hewel hala how low? [x3]
Holly howel hala

With the liquidizer out, it's less denigration
Here we are now, endearing us
I feel stapedial and contise
Here we are now, enteradenography us
A moloid
An albinistic
A muckweed
My loppet

A Daniele !! [x9]

soundex.py

import re
import string
import sys
nonword = re.compile('[^a-z]+')
repetitions = re.compile(r'([0-9])([hw]?\1)+')
cruft = re.compile('[hwaeiouy]+')
numbers = string.maketrans('bfpvcgjkqsxzdtlmnr', '111122222222334556')

def soundex(word):
    result = nonword.sub('', word.lower())

    if result:
        numbered = result.translate(numbers)
        norepetitions = repetitions.sub(lambda match: match.group(1), numbered)
        nocruft = cruft.sub('', norepetitions[1:])
        result = result[0] + nocruft
    missing = max(4 - len(result), 0)
    result = (result + '0' *  missing)[:4]
    return result

if __name__ == '__main__':
    print [soundex(w) for w in sys.argv[1:]]