Version 0.2

[zsunberg]

Added @provide to fix #1

Added the following functions

• valid_actions fixes valid_actions(env) #8
• clone fixes clone #2
• observations fixes observations(env) #10
• valid_action_mask fixes valid_actions_mask(env) #19
• render fixes render(env) #21

I did not include setstate! #9 , yet, because there is still a question about whether we want that or reset!(env, state).

I did not add observe #12 yet, because I don't understand the exact semantics of it. If someone can write a docstring for it that is unambiguous from both the environment writer and algorithm writer's perspective, then I think we could add it.

I did not set up Documenter.jl docs yet #24 , but I don't think this is necessary to be useful to release.

[zsunberg]

I'm requesting a couple of reviews, but if you want to review, please don't hesitate to jump in!

[findmyway]

In #20 @jonathan-laurent mentioned that some core common interfaces in single-agent and multi-agent environments can be contained here. So he may check it again in case some important functions are missing.

[jonathan-laurent]

I just made a proposal for supporting two-players, zero-sum games.

[jbrea]

I like this proposal!

[zsunberg]

OK, going to merge this but not going to register a new version. Will immediately start a PR with the changes discussed in #20 and #25