Code hardwired to Atari and emulator environments

[captify-alapite]

In many places in the code, there's a hardwired expectation that the environment will either be from ALE, or else it will be from some other type of video game. This limits the usefulness of the RL algorithms, even though there's nothing in their definition which stops them being useful in other contexts. In particular, there's no reason why an environment has to be something that can be rendered to a screen, e.g. the new GuessingGame-v0 environment provided with OpenAI Gym.

[steveKapturowski]

'render' should only be called if you set the flag '-v 1', so that isn't a strict dependency, but I did notice when trying out GuessingGame-v0 that I'm not properly handling Discrete input spaces so I'll submit a fix for that shortly since I'd like the code to be compatible with all gym environments.

I agree that the code shouldn't be coupled too closely with OpenAI gym so it would be straightforward to use the algorithms in conjunction with e.g. DeepMind Lab, Universe, custom environments, etc. so I'll try to clean up the interfaces to make integration with other environments easier