Roadmap to Version 1.0

The following is a tentative roadmap for the project to reach version 1.0.

v0.1

~~Sprint game mode is supported~~
~~Items in cells can be parsed by the game state~~
API for vector state representations is complete
- Finish monster vector in monster.py get_monster_vector.py
  - Next step - return monster details like health, ac, ev, etc.
- Obtain item properties, see list in state/itemproperty.py
API for PDDL state representations is complete
Update models/fastdownward_simple.pddl to be consistent with PDDL state representation API
Create simple RL agent that uses vector-based API
Create simple planning agent that uses PDDL-based API
~~New agent class that acts as a passthrough for a human player to play the game~~

v0.2

Simple planning agent working on tutorials
Simple reinforcement learning agent working on tutorials

v0.3

Add support for autobahn for terminal version of the game
Refactor game connection code to be less redundant between browser and terminal versions of the game
Ensure all game modes are supported in the terminal version running on linux

v0.4

Add metrics for measuring API performance (speed of actions being sent to game, etc.)
Add metrics for measuring agent performance (number of tiles visited, experience level, number of monsters killed, etc.)

v0.5

Add player skills into the gamestate data
Add actions to change experience point allocation for skills

v0.6

Enable agent to receive spectator commands through spectator chat, including 'pause' and 'resume'

v0.7

~~Refactor code into a pip installable library~~

v0.8

Code refactoring to adhere to better design principles

v 0.9

Replace most print statements with logging statements
Log files separated into agent logs and network logs

v 1.0

Documentation of all functions
Documentation hosted on pypi packages page