final working i think

nim.py

@@ -62,7 +62,7 @@ class NimAI():
         float: The Q-value associated with the (state, action) pair. 
                Returns 0 if the pair is not yet in the Q-table.
     """
-        print(self.q, state, action)
+        # print(self.q, state, action)
         try:
             return self.q[(tuple(state), action)]
         except:
@@ -123,10 +123,13 @@ class NimAI():
             # keys = [key[1] for key in self.q.key if key[0] == state]
             # for key in keys:
         else:
-            try:
+            # state = tuple(state)
-                return max([key[1] for key in self.q.keys() if key[0] == state])
+            # max(self.q[state, action]) for action in [key[1] for key in self.q if key[0] == state]
-            except:
+            # for q in self.q:
-                return (0,0)
+            #     if q[0] == state:
+
+            return list(self.q.keys())[list(self.q.values()).index(self.best_future_reward(state))][1]
+
 
 def train(n):
     player = NimAI()

play.py

@@ -5,6 +5,7 @@ if __name__ == "__main__":
     # Train the AI with 1000 games
     print("START TRAINING \n")
     ai = train(1000)
+    print(ai.q)
 
     # Start the game and play against the trained AI
     print("STARTING THE GAME \n")

test.py

@@ -25,6 +25,8 @@ def test_best_future_reward(ai):
 
 def test_choose_action(ai):
     print("\n--- Testing choose_action ---")
+    print(ai.choose_action([1,1,1,0]))
+    print(ai.choose_action([1,1,1,0], epsilon=False))
 
 
 if __name__ == "__main__":