bug: failing API tests when run in isolation

[Flix6x]

In case anyone is having trouble running a flexmeasures API test in isolation due to some UNAUTHORIZED error (e.g. with pytest -k test_trigger_and_get_schedule)? There appears to be some unwanted dependency between API tests. A silly but effective workaround is to rename the test you want to run in isolation by prepending test_auth_token and running pytest -k test_auth_token instead. This then runs both the test_auth_token and the actual test you wanted to run in isolation. For example, update def test_trigger_and_get_schedule to def test_auth_token_test_trigger_and_get_schedule. There appears to be something in test_auth_token that other tests need.

The test suite is unaffected by this bug.