Don't randomly generate auth_token in tests

In case something fails due to this, it's hard to debug it without
knowing the exact auth_token. Fuzzy testing should be done at another
place.

tests/conftest.py

@@ -2,8 +2,6 @@
 # SPDX-License-Identifier: AGPL-3.0-only
 # License-Filename: LICENSE.md
 
-from random import choice
-from string import ascii_letters, digits
 from unittest import mock
 
 import pytest
@@ -20,7 +18,7 @@ def api(auth_token):
 
 @pytest.fixture
 def auth_token():
-    return ''.join(choice(ascii_letters + digits) for i in range(33))
+    return 'NMDArEDjYlagINGd77WkKdxqDUAkYCWwh'
 
 
 @pytest.fixture