Unicode and urllib.open - python-2.7

Question

I am creating an application in python that can parse weather data from yr.no in Python. It works fine with regular ASCII strings, but fails when I use unicode.
def GetYRNOWeatherData(country, province, place):
#Parse the XML file
wtree = ET.parse(urllib.urlopen("http://www.yr.no/place/" + string.replace(country, ' ', '_').encode('utf-8') + "/" + string.replace(province, ' ', '_').encode('utf-8') + "/" + string.replace(place, ' ', '_').encode('utf-8') + "/forecast.xml"))
For example, when I try
GetYRNOWeatherData("France", "Île-de-France", "Paris")
I get this error
'charmap' codec can't encode character u'\xce' in position 0: character maps to <undefined>
Is it true that urllib doesn't handle unicode very well? Since I am using Tkinter as a frontend to this function, would that be the source of the problem (does the Tkinter Entry widget handle unicode well?)

Answer 1

You can handle this by keeping every string as a unicode right up until you actually make the urllib.urlopen request, at which point you encode to utf-8:
#!/usr/bin/python
# -*- coding: utf-8 -*-
# This import makes all literal strings in the file default to
# type 'unicode' rather than type 'str'. You don't need to use this,
# but you'd need to do u"France" instead of just "France" below, and
# everywhere else you have a string literal.
from __future__ import unicode_literals
import urllib
import xml.etree.ElementTree as ET
def do_format(*args):
ret = []
for arg in args:
ret.append(arg.replace(" ", "_"))
return ret
def GetYRNOWeatherData(country, province, place):
country, province, place = do_format(country, province, place)
url = "http://www.yr.no/place/{}/{}/{}/forecast.xml".format(country, province, place)
wtree = ET.parse(urllib.urlopen(url.encode('utf-8')))
return wtree
if __name__ == "__main__":
GetYRNOWeatherData("France", "Île-de-France", "Paris")