fixed an issue with inital state after reset

Agent/agent.py

@@ -90,7 +90,6 @@ class QLearning:
         print(dirs)
 
     # returns all possible actions
-    # TODO: maby possible != not kill
     def possible_directions(self):
         poss_actions = []
         # if not state.kill_l:
@@ -140,6 +139,7 @@ def learning(al, ga, ep, games):
     print("Starting learning process!")
     for i in range(max_games):
         playerwin = play_game_learning(qagent, q_0, poss_dirs_0)
+        q_0, poss_dirs_0 = reset()
         print("Game ", i+1, "/", max_games, " finished!")
         if playerwin == 1:
             score[0] += 1

Agent/snake_logic.py

@@ -523,7 +523,6 @@ def step(action):
                 reward = 0
 
             print("Apples eaten: ", snake1.getscore())
-            reset()
             agent_apples = []
             for a, id in apples:
                 for apple in a:
@@ -577,8 +576,8 @@ def step(action):
             playerwin = 2
 
         if check_if_crossing(snake1.pos[0] / block_size, snake1.pos[1] / block_size):
-            # print("Action needed!")
-            reward -= 1
+            if not init:
+                reward -= 1
             agent_apples = []
             for a, id in apples:
                 for apple in a:
@@ -662,6 +661,8 @@ def reset():
     snake2.set_enemy_norm_snake(snake1)
     gameExit = False
     playerwin = 0
+    state, _, _, poss_dirs = step("init")
+    return state, poss_dirs
 
 
 # exit game