fix: clean accent characters from username (#28157)

Also cleaned username for Authn MFE while registering using SSO/SAML.

VAN-483

common/djangoapps/third_party_auth/models.py

@@ -68,7 +68,7 @@ def clean_json(value, of_type):
 
 def clean_username(username=''):
     """ Simple helper method to ensure a username is compatible with our system requirements. """
-    return re.sub(r'[^-\w]+', '_', username)[:USERNAME_MAX_LENGTH]
+    return ('_').join(re.findall(r'[a-zA-Z0-9\-]+', username))[:USERNAME_MAX_LENGTH]
 
 
 class AuthNotConfigured(SocialAuthBaseException):

common/djangoapps/third_party_auth/tests/test_pipeline.py

@@ -75,7 +75,7 @@ class PipelineOverridesTest(SamlIntegrationTestUtilities, IntegrationTestMixin,
         ('S', 'S9fe2', False),
         ('S', 'S9fe2', True),
         ('S.K', 'S_K', False),
-        ('S.K.', 'S_K_', False),
+        ('S.K.', 'S_K', False),
         ('S.K.', 'S_K_9fe2', True),
         ('usernamewithcharacterlengthofmorethan30chars', 'usernamewithcharacterlengthofm', False),
         ('usernamewithcharacterlengthofmorethan30chars', 'usernamewithcharacterlengt9fe2', True),

openedx/core/djangoapps/user_authn/views/utils.py

@@ -8,6 +8,7 @@ from ipware.ip import get_client_ip
 
 from common.djangoapps import third_party_auth
 from common.djangoapps.third_party_auth import pipeline
+from common.djangoapps.third_party_auth.models import clean_username
 from openedx.core.djangoapps.site_configuration import helpers as configuration_helpers
 from openedx.core.djangoapps.geoinfo.api import country_code_from_ip
 
@@ -70,6 +71,9 @@ def third_party_auth_context(request, redirect_to, tpa_hint=None):
             current_provider = third_party_auth.provider.Registry.get_from_pipeline(running_pipeline)
             user_details = running_pipeline['kwargs']['details']
             if user_details:
+                username = running_pipeline['kwargs'].get('username') or user_details.get('username')
+                if username:
+                    user_details['username'] = clean_username(username)
                 context['pipeline_user_details'] = user_details
 
             if current_provider is not None: