Correct sign for BgB convectivity

[theo-brown]

See discussion #446.

[jcitrin]

First of all, I see that unittests aren't launched when you push, which is odd and something we need to look at. Probably with the way github actions is set up.

In any case, I looked deeper at the case, and realized that v_face_el is near-zero (1e-5 level). I don't see any bug in the formula compared to Tholerus et al. It's a bit odd. Maybe worthwhile comparing to JETTO values? Thus I wouldn't expect this sign change, while correct, to actually significantly the ITER density peaking in the comparison with QLKNN you've made. However, maybe you can check that and post the plot with pre and post BgB bugfix to this PR.

For TORAX, it may be helpful to sync to head and use the new plotting tools, whose default plot_config has more extensive information including D and V. See https://torax.readthedocs.io/en/latest/plotting.html

[theo-brown]

Regarding unit tests, might it be because the on: pull-request flag isn't set in the Github actions workflow pytest.yml? I think because I don't have write permissions on this repo, I have to fork it and make my changes in a fork, then PR from the fork to this repo. This means I don't trigger the on: push that sets off pytest.yml.

I think perhaps "Pass unit tests" should be a check, alongside the import/copybara and cla/google checks.

[jcitrin]

Thanks for the tips. We'll take a look and fix permissions / configuration. Ideally the unittests should run at every commit to a branch, and every PR. Ideally you also wouldn't need to fork, and could push to a branch. Will investigate.

[jcitrin]

@theo-brown , merged with 35ce9b0