How to sort utf-8 string on Google App Engine in Python?

Here's a pure-Python approach:

alphabet = u'AĄBCĆDEĘFGHIJKLŁMNŃOÓPRSŚTUWYZŹŻaąbcćdeęfghijklłmnńoóprsśtuwyzźż'
dsort = dict((let, i) for i, let in enumerate(alphabet))

def key_utf8(utf8_string):
  s = utf8_string.decode('utf8')
  return map(dsort.get, s)

some_list_of_utf8_strings.sort(key=key_utf8)

You'd probably be best advised to keep list of unicode strings internally -- decoding utf8 input at once and encoding back into utf8 on output if needed -- but as long as you're happy to potentially pay the decoding computational cost/delay repeatedly, this pure-Python approach should work fine, in App Engine or anywhere else.

If you do follow the best practice of only ever keeping unicode strings internally (decoding on input, encoding if needed on output), then the sort could also use a key=lambda s: map(dsort.get, s) -- but I'd personally prefer using a named function (for clarity) instead of the somewhat-goofy lambda. Just a matter of style, really.