garden time steps param

RL_Framework/pipeline.py

@@ -21,7 +21,7 @@ class Pipeline:
     def __init__(self):
         pass
 
-    def create_config(self, rl_time_steps=3000000, garden_time_steps=40, garden_x=10, garden_y=10, num_plant_types=2, num_plants_per_type=1, step=2, action_low=0.0, action_high=1.0, obs_low=0, obs_high=1000, ent_coef=0.01, n_steps=40000, nminibatches=4, noptepochs=4, learning_rate=1e-8, cnn_args=None):
+    def create_config(self, rl_time_steps=3000000, garden_time_steps=50, garden_x=10, garden_y=10, num_plant_types=2, num_plants_per_type=1, step=1, action_low=0.0, action_high=1.0, obs_low=0, obs_high=1000, ent_coef=0.01, n_steps=40000, nminibatches=4, noptepochs=4, learning_rate=1e-8, cnn_args=None):
         config = configparser.ConfigParser()
         config.add_section('rl')
         config['rl']['time_steps'] = str(rl_time_steps)
@@ -113,19 +113,19 @@ def plot_average_reward(self, folder_path, reward, days, x_range, y_range, ticks
         pathlib.Path(folder_path + '/Graphs').mkdir(parents=True, exist_ok=True)
         plt.savefig('./' + folder_path + '/Graphs/avg_reward.png')
 
-    def plot_stddev_reward(self, folder_path, reward, reward_stddev, days, x_range, y_range, ticks):
+    def plot_stddev_reward(self, folder_path, garden_time_steps, reward, reward_stddev, days, x_range, y_range, ticks):
         fig = plt.figure(figsize=(28, 10))
         plt.xticks(np.arange(0, days, 10))
         plt.yticks(np.arange(x_range, y_range, ticks))
         plt.title('Std Dev of Reward Over ' + str(days) + ' Days', fontsize=18)
         plt.xlabel('Day', fontsize=16)
         plt.ylabel('Reward', fontsize=16)
 
-        plt.errorbar([i for i in range(40)], reward, reward_stddev, linestyle='None', marker='o', color='g')
+        plt.errorbar([i for i in range(garden_time_steps)], reward, reward_stddev, linestyle='None', marker='o', color='g')
         pathlib.Path(folder_path + '/Graphs').mkdir(parents=True, exist_ok=True)
         plt.savefig('./' + folder_path + '/Graphs/std_reward.png')
 
-    def graph_evaluations(self, folder_path, garden_x, garden_y, time_steps, step, num_evals, num_plant_types):
+    def graph_evaluations(self, folder_path, garden_time_steps, garden_x, garden_y, time_steps, step, num_evals, num_plant_types):
         obs = [0] * time_steps
         r = [0] * time_steps
         for i in range(num_evals):
@@ -152,7 +152,7 @@ def graph_evaluations(self, folder_path, garden_x, garden_y, time_steps, step, n
         min_r = min(r) - 10
         max_r = max(r) + 10
         self.plot_average_reward(folder_path, r, time_steps, min_r, max_r, abs(min_r - max_r) / 10)
-        self.plot_stddev_reward(folder_path, rewards, rewards_stddev, time_steps, min_r, max_r, abs(min_r - max_r) / 10)
+        self.plot_stddev_reward(folder_path, garden_time_steps, rewards, rewards_stddev, time_steps, min_r, max_r, abs(min_r - max_r) / 10)
 
     def evaluate_policy(self, folder_path, num_evals, env, is_baseline=False, baseline_policy=None, step=1):
         model = None
@@ -181,6 +181,8 @@ def evaluate_policy(self, folder_path, num_evals, env, is_baseline=False, baseli
                 env.render()
             done = False
 
+            # env.env_method('show_animation')
+
             pathlib.Path(folder_path + '/Returns').mkdir(parents=True, exist_ok=True)
             filename = folder_path + '/Returns' + '/predict_' + str(i) + '.json'
             f = open(filename, 'w')
@@ -205,6 +207,7 @@ def single_run(self, folder_path, num_evals, policy_kwargs=None, is_baseline=Fal
         step = config.getint('garden', 'step')
         num_plants_per_type = config.getint('garden', 'num_plants_per_type')
         num_plant_types = config.getint('garden', 'num_plant_types')
+        garden_time_steps = config.getint('garden', 'time_steps')
         garden_x = config.getint('garden', 'X')
         garden_y = config.getint('garden', 'Y')
         # Z axis contains a matrix for every plant type plus one for water levels.
@@ -234,7 +237,7 @@ def single_run(self, folder_path, num_evals, policy_kwargs=None, is_baseline=Fal
             self.evaluate_policy(folder_path=folder_path, num_evals=num_evals, env=env, is_baseline=True, baseline_policy=baseline_policy, step=1)
 
             # Graph evaluations
-            self.graph_evaluations(folder_path, garden_x, garden_y, time_steps, step, num_evals, num_plant_types)
+            self.graph_evaluations(folder_path, garden_time_steps, garden_x, garden_y, time_steps, step, num_evals, num_plant_types)
         else:
             pathlib.Path(folder_path + '/ppo_v2_tensorboard').mkdir(parents=True, exist_ok=True)
             # Instantiate the agent
@@ -252,7 +255,7 @@ def single_run(self, folder_path, num_evals, policy_kwargs=None, is_baseline=Fal
             self.evaluate_policy(folder_path=folder_path, num_evals=num_evals, env=env, is_baseline=False)
 
             # Graph evaluations
-            self.graph_evaluations(folder_path, garden_x, garden_y, time_steps, step, num_evals, num_plant_types)
+            self.graph_evaluations(folder_path, garden_time_steps, garden_x, garden_y, time_steps, step, num_evals, num_plant_types)
 
         profiler_object.disable()
 
@@ -296,7 +299,7 @@ def createBaselineSingleRunFolder(self, garden_x, garden_y, num_plant_types, num
         filename_time = str(filename_time)
 
 
-    def batch_run(self, n, rl_config, garden_x, garden_y, num_plant_types, num_plants_per_type, policy_kwargs=[], num_evals=50, is_baseline=[], baseline_policy=None):
+    def batch_run(self, n, rl_config, garden_x, garden_y, num_plant_types, num_plants_per_type, policy_kwargs=[], num_evals=1, is_baseline=[], baseline_policy=None):
         assert(len(rl_config) == n)
         assert(len(garden_x) == n)
         assert(len(garden_y) == n)
@@ -336,7 +339,7 @@ def batch_run(self, n, rl_config, garden_x, garden_y, num_plant_types, num_plant
     rl_config = [
         {
             'rl_algorithm': 'MLP', 
-            'time_steps': 7000000,
+            'time_steps': 200,
             'ent_coef': 0.0,
             'n_steps': 40000,
             'nminibatches': 4,