araby.py 文件源码

def strip_harakat(text):
    """Strip Harakat from arabic word except Shadda.
    The striped marks are :
        - FATHA,  DAMMA,  KASRA
        - SUKUN
        - FATHATAN,  DAMMATAN,  KASRATAN,  ,  ,  .
    Example:
        >>> text = u"?????????????"
        >>> stripTashkeel(text)
        ????????
    @param text: arabic text.
    @type text: unicode.
    @return: return a striped text.
    @rtype: unicode.
    """
    # if text:
        # return  re.sub(HARAKAT_PATTERN, u'', text)
    # return text
    if not text:
        return text
    elif is_vocalized(text): 
        for char in HARAKAT:
            text  = text.replace(char, '')
    return text