Cut cache_generator

[nickdrozd]

Is cache_generator doing anything? Cutting it doesn't break anything
as far as I can tell, and it Pylint even seems to go a little faster
without it.

[nickdrozd]

Someone else should try this on a large repo and see what happens.

[PCManticore]

It's possible cache_generator is not having the performance impact we expected it to have, but having some tests with pylint on top of libraries like numpy or pandas could prove that.

[kodonnell]

As per here:

Allow partial caching. Currently an inference value is only cached if the inference is completely exhausted. Since there are many instances of only retrieving the first value of the inference, caching is ignored because there is no partial caching.

So - probably want to fix it (easy?) as opposed to getting rid of it - and hopefully that'll actually improve performance.

[PCManticore]

Yeah, that makes sense, it is possible that addressing some of its flaws could result in a performance improvement albeit minor.