Fix issues

Author: ArzhelaR

environment/gymnasium_envs/quadmesh_env/envs/quadmesh.py

@@ -51,13 +51,13 @@ class QuadMeshEnv(gym.Env):
     def __init__(
             self,
             mesh=None,
-            max_episode_steps: int =50,
-            n_darts_selected: int =20,
-            deep: int =6,
+            max_episode_steps: int = 50,
+            n_darts_selected: int = 20,
+            deep: int = 6,
             render_mode:  Optional[str] = None,
-            with_degree_obs: bool =True,
-            action_restriction: bool =False,
-            obs_count: bool=False,
+            with_degree_obs: bool = True,
+            action_restriction: bool = False,
+            obs_count: bool = False,
     ) -> None:
 
         assert render_mode is None or render_mode in self.metadata["render_modes"]

training/exploit_PPO_perso.py

@@ -5,7 +5,7 @@
 import torch
 from torch.distributions import Categorical
 from model_RL.PPO_model_pers import Actor
-from stable_baselines3 import PPO
+
 from mesh_model.mesh_analysis.global_mesh_analysis import global_score
 from mesh_model.mesh_struct.mesh import Mesh
 from mesh_model.reader import read_gmsh
@@ -57,7 +57,7 @@ def testPolicy(
             ep_mesh_rewards: int = 0
             ep_length: int = 0
             observation, info = env.reset(options={"mesh": copy.deepcopy(mesh)})
-            while terminated == False and truncated == False:
+            while terminated is False and truncated is False:
                 obs = torch.tensor(observation.flatten(), dtype=torch.float32)
                 pmf = actor.forward(obs)
                 dist = Categorical(pmf)
@@ -96,6 +96,7 @@ def isBetterMesh(best_mesh, actual_mesh):
     else:
         return False
 
+
 if __name__ == '__main__':
 
 

training/exploit_trimesh.py

@@ -32,7 +32,9 @@ def exploit():
         plot_test_results(avg_rewards, avg_wins, avg_steps)
     plot_dataset(final_meshes)
 
+
 if __name__ == '__main__':
+
     mesh = read_gmsh("../mesh_files/t1_quad.msh")
 
     #Create a dataset of 9 meshes
@@ -56,7 +58,7 @@ def exploit():
     actor.load_state_dict(torch.load('policy_saved/actor_network.pth'))
     avg_steps, avg_wins, avg_rewards, final_meshes = testPolicy(actor, 15, dataset, 20)
 
-    plot_test_results(avg_rewards, avg_wins, avg_steps)
+    plot_test_results(avg_rewards, avg_wins, avg_steps, avg_wins)
     plot_dataset(final_meshes)
     for m in final_meshes:
         smoothing_mean(m)