Buffalo-gym

Added

- TiredBuffalo -- FatigueBandit

Changed

- Minor fixes for warnings in pytest from gym checker

Added

- BounlessBuffalo -- InfiniteArmedBandit
- DuelingBuffalo -- DuelingBandit
- SymbolicState

Changed

- BuffaloTrail more closely matches the notion of a bandit. It was previously unconstrained on two axes
compared to a traditional bandit (control over future states and dependence on history). Control over future
states was removed as dependence on history is interesting in its own right.

Added

- Utility which might help in verifying what an agent is learning
- Citation information

Changed

- Added repeatability to environments using seeds and enforcing the difference between optimal and
suboptimal arms
- Added "standard" names for environments. Not everyone will get my naming scheme.
- Environments use info to return necessary information to derive the solution to the environment.