Fix missing __init__ for wheels

[ericl]

Why are these changes needed?

Related issue number

Checks

• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[ericl]

Build seems sane so merging as hotfix

[waleedkadous]

@ericl Can you help me understand how this passed tests?

Can we write a test to prevent this happening in future?

[ericl]

@simon-mo the issue as I understand is that it was a problem in the wheel upload process. Is it possible to add some sanity check to the wheel build job that tests you can install / ray.init() a wheel?

[simon-mo]

They are actually tested https://github.com/ray- project/ray/blob/master/ci/travis/test-wheels.sh#L24-L25

On March 11, 2021, GitHub ***@***.***> wrote: @simon-mo <https://github.com/simon-mo> the issue as I understand is that it was a problem in the wheel upload process. Is it possible to add some sanity check to the wheel build job that tests you can install / ray.init() a wheel?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/ray- project/ray/pull/14615#issuecomment-796959197>, or unsubscribe <https://github.com/notifications/unsubscribe- auth/AFBD7AZHEC26BXTUPDSSVNLTDD6S7ANCNFSM4Y7HUY6Q>.

[ericl]

Got it, so it was just my fault not checking the wheel status here. It infrequently breaks so I tend to ignore it :)

Separately, I think downstream consumers of nightly should run sanity checks (import ray; ray.init()) before auto ingesting new nightly builds.

[waleedkadous]

Can we communicate when it fails more effectively, e.g. have it send an e-mail / page when it happens? I can't see from the code where buildkite failures are surfaced.