Spezial whitespace strip...

[jens]

Mal wieder was zum optimieren, wer lust hat. Denn so richtig gut finde ich meine Lösung nicht:

Code: Alles auswählen

import re

space_re = re.compile(r"^(\s*)(.*?)(\s*)$", re.DOTALL)
def clean_whitespace(txt):
    """
    >>> clean_whitespace(u"\\n\\nfoo bar\\n\\n")
    u'\\nfoo bar\\n'
    
    >>> clean_whitespace(u"   foo bar  \\n  \\n")
    u' foo bar\\n'

    >>> clean_whitespace(u" \\n \\n  foo bar   ")
    u'\\nfoo bar '
    
    >>> clean_whitespace(u"foo   bar")
    u'foo   bar'
    """
    def striped(txt):
        if "\n" in txt:
            return "\n"
        elif " " in txt:
            return " "
        else:
            return ""
        
    def cleanup(match):
        #print "all:", repr(match.group(0))
        start, txt, end = match.groups()
        return striped(start) + txt + striped(end)   
            
    return space_re.sub(cleanup, txt)

if __name__ == '__main__':
    import doctest
    doctest.testmod()