def test_parse_username_unicode_handling(self):
unparsed_strings = [u'sès', u'مو', u'èm']
expected_strings = ['S\xc3\xa8s', '\xd9\x85\xd9\x88', '\xc3\x88m']
for unparsed, expected in zip(unparsed_strings, expected_strings):
username = parse_username(unparsed)
assert_equal(username, expected)
def test_validate_username(self):
# this username has a few problems that the normalize call should handle
# 1. normal ascii space in front
# 2. lowercase
# 3. nasty trailing unicode space (the reason this file has coding:utf-8)
problem_username = '******'
parsed_user = parse_username(problem_username)
assert_equal(parsed_user, 'Editor test-specific-0')
def validate_users(wikiusers, project, validate_as_user_ids):
"""
Parameters
wikiusers : the wikiusers with a candidate mediawiki_username
project : the project these wikiusers should belong to
validate_as_user_ids : if True, records will be checked against user_id
if False, records are checked against user_name
"""
session = db.get_mw_session(project)
users_dict = {wu.mediawiki_username: wu for wu in wikiusers}
try:
# validate
if validate_as_user_ids:
keys_as_ints = [int(k) for k in users_dict.keys() if k.isdigit()]
clause = MediawikiUser.user_id.in_(keys_as_ints)
else:
clause = MediawikiUser.user_name.in_(users_dict.keys())
matches = session.query(MediawikiUser).filter(clause).all()
# update results
for match in matches:
if validate_as_user_ids:
key = str(match.user_id)
else:
key = parse_username(match.user_name)
users_dict[key].mediawiki_username = match.user_name
users_dict[key].mediawiki_userid = match.user_id
users_dict[key].valid = True
users_dict[key].reason_invalid = None
# remove valid matches
users_dict.pop(key)
# mark the rest invalid
for key in users_dict.keys():
if validate_as_user_ids:
users_dict[key].reason_invalid = 'invalid user_id: {0}'.format(key)
else:
users_dict[key].reason_invalid = 'invalid user_name: {0}'.format(key)
users_dict[key].valid = False
except Exception, e:
# clear out the dictionary in case of an exception, and raise the exception
for key in users_dict.keys():
users_dict.pop(key)
raise e
