refactor: Improve some imports, remove whitespace

.github/workflows/pre-commit.yml

@@ -1,4 +1,4 @@
-name: pre-commit hooks
+name: CI
 
 on:
   push:

README.md

@@ -5,7 +5,7 @@
 ## **Generals.io RL**
 
 [![CodeQL](https://github.com/strakam/Generals-RL/actions/workflows/codeql.yml/badge.svg)](https://github.com/strakam/Generals-RL/actions/workflows/codeql.yml)
-[![CI](https://github.com/strakam/Generals-RL/actions/workflows/tests.yml/badge.svg)](https://github.com/strakam/Generals-RL/actions/workflows/tests.yml)
+[![CI](https://github.com/strakam/Generals-RL/actions/workflows/pre-commit.yml/badge.svg)](https://github.com/strakam/Generals-RL/actions/workflows/pre-commit.yml)
 
 
 
@@ -25,11 +25,8 @@ Highlights:
 * 🔬 **analysis tools**: leverage features like replays for deeper insights
 
 > [!Note]
-> This repository is based on the generals.io game. Its one and only goal is to provide a bot development
-> platform, particularly for Machine-Learning agents.
-
-> [!Tip]
-> Try out the [original game](https://generals.io), its a lot of fun!
+> This repository is based on the [generals.io](https://generals.io) game (check it out, it's a lot of fun!).
+The one and only goal of this project is to provide a bot development platform, especially for Machine Learning based agents.
 
 ## 📦 Installation
 You can install the latest stable version via `pip` for reliable performance
@@ -173,7 +170,7 @@ The `observation` is a `Dict`. Values are either `numpy` matrices with shape `(N
 | `is_winner`          |     —     | Indicates whether the agent won                                              |
 | `timestep`           |     —     | Current timestep of the game                                                 |
 
-The `action_mask` is a 3D array with shape `(N, M, 4)`, where each element corresponds to whether a move is valid from cell 
+The `action_mask` is a 3D array with shape `(N, M, 4)`, where each element corresponds to whether a move is valid from cell
 `[i, j]` in one of four directions: `0 (up)`, `1 (down)`, `2 (left)`, or `3 (right)`.
 
 ### ⚡ Action
@@ -191,7 +188,8 @@ Actions are in a `dict` format with the following `key - value` format:
 > ```
 
 ### 🎁 Reward
-It is possible to implement custom reward function. The default is `1` for winner and `-1` for loser, otherwise `0`.
+It is possible to implement custom reward function. The default reward is awarded only at the end of a game
+and gives `1` for winner and `-1` for loser, otherwise `0`.
 ```python
 def custom_reward_fn(observation, action, done, info):
     # Give agent a reward based on the number of cells they own

examples/gymnasium_example.py

@@ -2,15 +2,15 @@
 
 from generals import AgentFactory
 
-# Initialize opponent agent ("random" or "expander")
-npc = AgentFactory.make_agent("random")
+# Initialize opponent agent
+npc = AgentFactory.make_agent("expander")
 
 # Create environment
 env = gym.make("gym-generals-v0", npc=npc, render_mode="human")
 
 observation, info = env.reset()
 terminated = truncated = False
 while not (terminated or truncated):
-    action = env.action_space.sample()  # Here you put your agent's action
+    action = env.action_space.sample()  # Here you put an action of your agent
     observation, reward, terminated, truncated, info = env.step(action)
     env.render()

examples/pettingzoo_example.py

@@ -1,15 +1,16 @@
-from generals.agents import AgentFactory
+from generals.agents import RandomAgent, ExpanderAgent
 from generals.envs import PettingZooGenerals
 
 # Initialize agents
-random = AgentFactory.make_agent("random")
-expander = AgentFactory.make_agent("expander")
+random = RandomAgent()
+expander = ExpanderAgent()
 
+# Store agents in a dictionary
 agents = {
     random.id: random,
     expander.id: expander,
 }
-agent_ids = list(agents.keys()) # Environment calls agents by name
+agent_ids = list(agents.keys())
 
 # Create environment
 env = PettingZooGenerals(agents=agent_ids, render_mode="human")

generals/__init__.py

@@ -1,6 +1,6 @@
 from gymnasium.envs.registration import register
 
-from generals.agents.agent_factory import AgentFactory
+from generals.agents import AgentFactory
 from generals.core.grid import Grid, GridFactory
 from generals.core.replay import Replay
 from generals.envs.pettingzoo_generals import PettingZooGenerals

generals/agents/__init__.py

@@ -2,6 +2,14 @@
 
 from .agent import Agent
 from .agent_factory import AgentFactory
+from .expander_agent import ExpanderAgent
+from .random_agent import RandomAgent
 
 # You can also define an __all__ list if you want to restrict what gets imported with *
-__all__ = ["Agent", "AgentFactory"]
+__all__ = [
+    "Agent",
+    "AgentFactory",
+    "RandomAgent",
+    "ExpanderAgent",
+    "AgentFactory",
+]

generals/core/config.py

@@ -5,6 +5,7 @@
 import gymnasium as gym
 import numpy as np
 
+# Type aliases
 Observation: TypeAlias = dict[str, np.ndarray | dict[str, gym.Space]]
 Action: TypeAlias = dict[str, int | np.ndarray]
 Info: TypeAlias = dict[str, Any]
@@ -13,9 +14,7 @@
 RewardFn: TypeAlias = Callable[[Observation, Action, bool, Info], Reward]
 AgentID: TypeAlias = str
 
-#################
-# Game Literals #
-#################
+# Game Literals
 PASSABLE: Literal["."] = "."
 MOUNTAIN: Literal["#"] = "#"