try to fix bugs of ActionTransformedEnv

[peterchen96]

When I want to get the action space of the player 1 with an ActionTransformedEnv, it will raise errors like the following:

julia> using ReinforcementLearning
julia> env = KuhnPokerEnv();
julia> wrapped_env = ActionTransformedEnv(env);
julia> action_space(wrapped_env, 1)
ERROR: MethodError: no method matching identity(::Base.OneTo{Int64}, ::Int64)

Maybe we should move args... into the action_space function...

[findmyway]

@all-contributors please add @peterchen96 for code, doc

[allcontributors]

@findmyway

I've put up a pull request to add @peterchen96! 🎉

[findmyway]

Sorry, it seems I forgot to add you in the README.