Provide agent base class, agnostic to Pac-Man simulator

[matheusportela]

Currently, ClientAgent is the most basic agent class, but it's simply too coupled with the Pac-Man simulator. For instance, it inherits from BerkeleyGameAgent, implements the getAction method, create state message, among others.

Let's make a base class that is agnostic to Pac-Man simulator, such as the BaseLearningAlgorithm class. We can draw ideas from PyBrain, RL Glue, OpenAI Gym, and Maja on how to create an agent class that can be reused with different scenarios.

[matheusportela]

Things a good BaseAgent class must have:

• Connection via ZMQ with the Multiagent RL controller
• Easy setup of learning algorithms, such as via constructor parameter
• Standard interface with reinforcement learning environments, such as get_action, receive_state, and receive_reward methods
• Simple toggle from learning to testing mode
• Methods to inform when an episode has started or finished, as well as a method to reset the agent state to be used across episodes

[matheusportela]

I just find out three modules that are a first attempt of decoupling agents from the Pac-Man simulator:

• pac_agents.py: Contains a AgentBase (that should be renamed to BaseAgent) with init, setup, and cleanup methods, and provide other interfaces, such as BehaviorInterface, LearningInterface and QLearningInterface. Also attempts to implement some agents using these classes
• pac_experiment.py: Implements PacmanExperiment based on Experiment from pac_utils.py
• pac_utils.py: Provide Logger class (that must be replaced by logging and Experiment, an abstraction to a learning experiment

[matheusportela]

As implemented in #67, now we have three classes: BaseAgent, BaseController and BaseExperiment, but they aren't really useful yet because all code is still implement in Pac-Man specific concrete classes. Now, we need to refactor out as much code as possible so they become reusable in other environments.

[matheusportela]

Now that we have BaseAdapterAgent, we may move most of the logic in BerkeleyAdapter to Experiment considering that most experiment will require a basic setup:

• Instantiate agents specific to the experiment
• Start the experiment and execute the agents' start_experiment method
• Start the game and execute the agents' start_game method
• Execute the game
• Bookkeeping after executing the game
• Finish the game and execute the agents' finish_game method
• Finish the experiment and execute the agents' finish_experiment method
• Bookkeeping after executing the experiment

[matheusportela]

Moved most reusable logic from experiments at acf899297a1d7ad9221d68bb0a3b199e75e6427a. I believe most refactoring from the adapter, in order to create a base agent class, has been done. I'll close this issue and open a new one specific to refactoring the controller.