pass RNG instead of seeds

[costachris]

Closes #260

Pass RNG objects instead of seeds to make randomness more traceable.

[odunbar]

@costachris Could you rebase with main, before I review, hopefully then all tests will pass! Cheers

[costachris]

@odunbar The EKP test is failing because we're doing two sequential calls to construct_initial_ensemble, passing the same rng object without reseeding, and checking that they're equal. This test made sense when we were passing the seed to construct_initial_ensemble, but that's no longer the case. The simple solution is to reseed the rng object right before the second call to construct_initial_ensemble (outside of the function) in the test, but just want to double-check that's the behavior we want?

[odunbar]

Looks good, Thank you! Regarding your comment. I think we should just remove that test.

As you say, the reproducibility should not occur unless the user does something external now - by explicitly copying/reseeding the rng, or the GLOBAL_RNG outside the function. This is then something already tested in Random/Distributions etc. so not our concern any more.

Once this is done, please merge!

[costachris]

bors r+

[bors]

Build succeeded!

The publicly hosted instance of bors-ng is deprecated and will go away soon.

If you want to self-host your own instance, instructions are here.
For more help, visit the forum.

If you want to switch to GitHub's built-in merge queue, visit their help page.

• docs-build
• format
• test_success