Question

I would like to be able to add attributes to characters within a str or unicode string, chop the string and move pieces around, and have the attributes still be present at the end.

Would this be possible (and best) with a subclass of str, unicode or possibly basestring?

Example:

s = u"hello world"
s[6].foo = u'bar'
s2 = s.split(' ')
assert(s2 == u'world')
assert(s2[0].foo == u'bar')

Thanks for thoughts!

Answer 1

You can try using that as a starter:

class Mine(unicode):
    #
    def __init__(self, *args, **kwargs):
        super(Mine, self).__init__(*args, **kwargs)
    #
    def __setattr__(self, attr, value):
        try:
            super(Mine, self).__setattr__(attr, value)
        except AttributeError:
            self.__dict__[attr] = value
    #
    def __getattr__(self, attr):
        try:
            super(Mine, self).__getattr__(attr)
        except AttributeError:
            try:
                return self.__dict__[attr]
            except KeyError:
                raise AttributeError

    def __getitem__(self, item):
        obj = Mine(super(Mine, self).__getitem__(item))
        obj.__dict__ = self.__dict__
        return obj

Of course, you’ll have to create a specific split method, where each item of the output list will be a Mine object:

def split(self, arg=' '):
    result = []
    for item in super(Mine, self).split(arg):
        i = Mine(item)
        i.__dict__ = self.__dict__
        result.append(i)
    return result

The generic idea is to overload every single method of the parent class (at least, the ones you’re really interested in) so that it returns an instance of your class and inherits the __dict__ of the caller… That can be a lot of work.