2623 move action mask generation to game and fix MARL masking

src/primaite/notebooks/Action-masking.ipynb

@@ -96,7 +96,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from primaite.session.ray_envs import PrimaiteRayEnv, PrimaiteRayMARLEnv\n",
+    "from primaite.session.ray_envs import PrimaiteRayEnv\n",
     "from ray.rllib.algorithms.ppo import PPOConfig\n",
     "import yaml\n",
     "from ray import air, tune\n"
@@ -146,18 +146,59 @@
    ]
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
+   "cell_type": "markdown",
    "metadata": {},
-   "outputs": [],
-   "source": []
+   "source": [
+    "## Action masking with MARL in Ray RLLib\n",
+    "Each agent has their own action mask, this is useful if the agents have different action spaces."
+   ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "from primaite.session.ray_envs import PrimaiteRayMARLEnv\n",
+    "from primaite.config.load import data_manipulation_marl_config_path"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open(data_manipulation_marl_config_path(), 'r') as f:\n",
+    "    cfg = yaml.safe_load(f)\n",
+    "env_config = cfg\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "config = (\n",
+    "    PPOConfig()\n",
+    "    .multi_agent(\n",
+    "        policies={'defender_1','defender_2'}, # These names are the same as the agents defined in the example config.\n",
+    "        policy_mapping_fn=lambda agent_id, episode, worker, **kw: agent_id,\n",
+    "        )\n",
+    "    .environment(env=PrimaiteRayMARLEnv, env_config=cfg)\n",
+    "    .env_runners(num_env_runners=0)\n",
+    "    .training(train_batch_size=128)\n",
+    "    )\n",
+    "\n",
+    "tune.Tuner(\n",
+    "    \"PPO\",\n",
+    "    run_config=air.RunConfig(\n",
+    "        stop={\"timesteps_total\": 5 * 128},\n",
+    "    ),\n",
+    "    param_space=config\n",
+    ").fit()"
+   ]
   }
  ],
  "metadata": {