Introduce enable/disable experiment cohorting

[robin-phung]

What problem does this solve?

Disabling Cohorting: Allow existing registered participants time to convert while no longer accepting new participants into the experiment. New participants will be given the control will be given the control and will not be recorded as being apart of the experiment.

Why is this useful?

• Gives implementers the ability to soft close experiments, allowing shifting new participants into other experiments before concluding the result of the experiment (selecting a winner).
• Allows for better accuracy of data, ~> allows participants the chance to convert if they engaged late before the end of the experiment.

How does this solve it?

• Adds enable/disable button to the dashboard
• Provides a brief alert description describing the resulting behaviour when performed.
• Adds functions to the experiment object to check and mutate the flag stored in redis.
• Deletion of the experiment will result in the deletion of the flag (resetting it back to its default state = cohorting enabled)
• Continues to perform appropriate callbacks if the user is already an existing participant to the experiment.
• Performs no callbacks if the participant is first entering the experiment for the first time when cohorting is disabled.
• existing overrides and disables take precedence

[andrehjr]

Thanks for the Pull Request @robin-phung ✨ I wanna do some testing just to make sure it's all good before merging.

The Trial#choose! method is a bit complex as the way it is right now(before your PR), I have some plans to refactor it.

[robin-phung]

@andrehjr Let me know what your thoughts are on refactoring as I'm also interested (and may be able to assist/collab) in a refactor of the Helper#ab_test ~> Trial#choose!.

[andrehjr]

Just a few changes and we should be good to merge. 🚀 Sorry for taking so long.

Being able to 'soft close' an experiment does look useful to give more control over experiments.

[andrehjr]

@andrehjr Let me know what your thoughts are on refactoring as I'm also interested (and may be able to assist/collab) in a refactor of the Helper#ab_test ~> Trial#choose!.

Sure! Help is always welcome. I've started the other day a few spikes starting with the ab_test method in #623 in order to start preparing things.

[robin-phung]

Updated PR with the following changes:

• Memoizes cohorting_disabled to prevent repeated calls to redis when calling #cohorting_disabled?
• Places enable/disable cohorting buttons within the else condition of #has_winner? ~> only displays the buttons when the experiment does not have a winner.
• Places the condition logic of new_participant && @experiment.cohorting_disabled? into a variable new_participant_and_cohorting_disabled

[andrehjr]

@robin-phung can you rebase these changes? 18c0514 is already on master 648576a

And that's it! 🚀

[robin-phung]

My rebase went out of whack, it should be resolved now.