I just updated the interpretation(https://github.com/beikov/jpa-treat-variations) to include what we discussed. Could you please review that and give me feedback?
I'll also add a test that asserts this specific behavior of treat in a condition context.