Refactor CLI/interface: init command, factory field, remove extras_size

- Rename `retro-gamer create` to `retro-gamer init` with positional
  args (GAME OUTPUT) instead of --game/--output flags
- Add [tool.retro-gamer].factory = "module:attr" to declare the game
  factory function from pyproject.toml instead of relying on a
  hard-coded create_game attribute
- Remove extras_size from user-facing config; it is now measured
  automatically from a sample observation and never declared
- Update docs throughout: create→init, runs/→training/, add factory
  field documentation, clarify [model] vs [training] hyperparameter
  sections, remove stale version-history explanations
- Bump version to 0.3.0; require retro-games>=2.5.0

docs/integration.rst

@@ -19,14 +19,14 @@ Both approaches start by creating a :class:`retro_gamer.TrainedPolicy`:
 
    from retro_gamer import TrainedPolicy
 
-   ai = TrainedPolicy("runs/snake/")
+   ai = TrainedPolicy("training/snake/")
 
 This reads ``config.toml``, rebuilds the network, and loads the latest
 checkpoint. To load a specific checkpoint instead:
 
 .. code-block:: python
 
-   ai = TrainedPolicy("runs/snake/", checkpoint="ep_0500")
+   ai = TrainedPolicy("training/snake/", checkpoint="ep_0500")
 
 PolicyInput: model as player
 ----------------------------
@@ -40,7 +40,7 @@ it to ``game.play()`` and everything else works exactly as usual:
    from retro.examples.snake import create_game
    from retro_gamer import TrainedPolicy, PolicyInput
 
-   ai = TrainedPolicy("runs/snake/")
+   ai = TrainedPolicy("training/snake/")
    game = create_game()
    game.play(input_source=PolicyInput(ai, game))
 
@@ -62,7 +62,7 @@ loaded from disk once — not once per episode.
    from retro.examples.snake import Apple, SnakeHead
    from retro_gamer import TrainedPolicy
 
-   _ai = TrainedPolicy("runs/snake/")
+   _ai = TrainedPolicy("training/snake/")
 
    class AISnake(SnakeHead):
        def handle_keystroke(self, k, game): pass  # ignore keyboard
@@ -112,9 +112,8 @@ from the game state. To train an enemy:
 
    .. code-block:: console
 
-      % retro-gamer create --game my_game:create_enemy_training_game \
-                           --output runs/enemy/
-      % retro-gamer train runs/enemy/
+      % retro-gamer init games/my_game training/enemy
+      % retro-gamer train training/enemy/
 
 3. **Embed the trained model in your main game** using ``get_action``, exactly
    as shown above.
@@ -145,7 +144,7 @@ once per episode:
    # enemy_training_game.py
    from retro_gamer import TrainedPolicy
 
-   _player = TrainedPolicy("runs/player/")   # loaded once when the module is imported
+   _player = TrainedPolicy("training/player/")   # loaded once when the module is imported
 
    def create_game():
        enemy = EnemyAgent()
@@ -156,9 +155,9 @@ You then alternate training runs:
 
 .. code-block:: console
 
-   % retro-gamer train runs/player/   # train player against current enemy
-   % retro-gamer train runs/enemy/    # train enemy against updated player
-   % retro-gamer train runs/player/   # train player again
+   % retro-gamer train training/player/   # train player against current enemy
+   % retro-gamer train training/enemy/    # train enemy against updated player
+   % retro-gamer train training/player/   # train player again
    # ...
 
 How many episodes to run before switching is itself a design decision: too