Migrate babysnake to retro_gamer's observation_function

BabySnake's tabular Q-learning exercise needs a small discrete state,
not the flattened board encoding retro_gamer normally produces. Define
get_state(game) in babysnake_env.py, returning a plain
(agent_x, agent_y, food_x, food_y) tuple, and point babysnake's
observation_function at it instead of declaring a character_set.
Disable the on-screen state overlay so it doesn't clutter the
terminal watch view.

babysnake/__init__.py

@@ -101,6 +101,7 @@ def create_game():
         },
         board_size=(bw, bh),
         framerate=6,
+        show_state=False,
     )
     forager.position = (randint(0, bw - 1), randint(0, bh - 1))
     food.relocate(game)

babysnake/pyproject.toml

@@ -1,4 +1,4 @@
 [tool.retro-gamer]
 actions = ["KEY_RIGHT", "KEY_UP", "KEY_LEFT", "KEY_DOWN"]
 reward = "reward"
-character_set = ["@", "*"]
+observation_function = "babysnake_env:get_state"

babysnake_env.py

@@ -0,0 +1,13 @@
+"""BabySnake's observation_function: maps a game to its tabular Q-learning state.
+
+Referenced from babysnake/pyproject.toml's [tool.retro-gamer] section, and
+used directly by train_babysnake.py via GameEnvironment.
+"""
+
+ACTIONS = ["KEY_RIGHT", "KEY_DOWN", "KEY_LEFT", "KEY_UP"]
+
+
+def get_state(game):
+    """Return BabySnake's state as a hashable (agent_x, agent_y, food_x, food_y) tuple."""
+    s = game.state
+    return (int(s['agent_x']), int(s['agent_y']), int(s['food_x']), int(s['food_y']))