'''
ASCIIfy
Kristjan Kannike
Created: 2005-09-05
Changed: 2006-05-23

"ASCIIfy" a string.
'''

from unicodedata import decomposition
import re


def asciify(s):
    '''"ASCIIfy" a Unicode string by stripping all umlauts, tildes, etc.'''
    temp = u''
    for char in s:
        decomp = decomposition(char)
        if decomp:
            temp += unichr(int(decomp.split()[0], 16))
        else:
            temp += char
    return temp

def normalise(s):
    '''ASCIIfy a Unicode string, put it into lowercase, substitute '_' for ' ',
    and remove all non-alphanumeric characters.'''
    temp = asciify(s)
    temp = temp.lower()
    temp = re.sub(r'\s\s+', '', temp) # Remove excessive whitespace
    temp = temp.replace(' ', '_')
    temp = re.sub(r'\W', '', temp) # Remove all non-alphanumeric characters
    return temp