Currently there are N tags associated with python tests in our evergreen.yml file:

• "unit_test"
• "unit_test_long"
• "python"
• "unit_test_extra_long"

Depending on which tag you include in your variant you will get, any or all of the task. For this reason the "python" tag is not useful. If you include it tasks like "unit-test-macos" will be selected regardless of which is variant is selected.

We also have the "unit-test" task which in theory should be identical in terms of tests to ".unit_test" which seems unclear.