redflag
redflag copied to clipboard
`is_standard_normal` test is flaky
Relax confidence level? Or could seed it.
Weirdly only seems flaky on Python 3.8.
Argh, still flaky. Clearly this test is not as robust as I hoped, maybe need to use something other than (or in addition to?) K-S.
Temporarily fixed this byusing default_rng with a seed -- but would rather have a test that a decent sample passes on.
Much better now but failed again on 3.9 here https://github.com/scienxlab/redflag/actions/runs/6385283773