Improve performance of decision tree

classes/ai.py

@@ -29,7 +29,7 @@ class AI:
         pass
 
     #co ma zrobić przy każdym ruchu <------------------------- najważniejsze
-    def updateTile(self):
+    def updateTile(self, model):
         #aktualne pola (do debugu)
         sensor = self.saper.sensor()
 
@@ -40,7 +40,7 @@ class AI:
 
 
         #podniesienie bomby jeśli jest jakaś na tym polu
-        self.saper.pick_up()
+        self.saper.pick_up(model)
 
         #poruszenie się
         if self.user_controlled:

classes/bfs.py

@@ -164,7 +164,7 @@ class BFS:
 
             # jesli tmp node to goaltest
             if tmp_node_position[:2] == goaltest:
-                print('Find')
+                print('Find\n')
 
                 while tmp_node[1].get_parent() is not None:
                     final_action_list.append(tmp_node[1].get_action())
@@ -177,7 +177,7 @@ class BFS:
             explored.append(tmp_node[1])  # add node to array of visited nodes
 
             neighbours_list_of_our_node = self.successor(tmp_node_position)  # lista możliwych akcij
-            print(neighbours_list_of_our_node)
+            # print(neighbours_list_of_our_node)
 
             for node_ in neighbours_list_of_our_node:
                 # node_ is tuple(action, [x, y, gdzie_patczy], cost)

classes/decisionTrees.py

@@ -8,7 +8,11 @@ from numpy import random
 ###############
 
 class DecisionTrees:
-    def return_predict(self):
+    def create_model(self):
+        model = chef.fit(pd.read_csv("D:\\1 Python projects\Saper\data\db.txt"), {'algorithm': 'ID3'})
+        return model
+
+    def return_predict(self, mod):
 
         # read data
         df = pd.read_csv("D:\\1 Python projects\Saper\data\db.txt")
@@ -24,7 +28,6 @@ class DecisionTrees:
         # ID3 config
         config = {'algorithm': 'ID3'}
         # create decision tree
-        model = chef.fit(df, config)
 
         # print predict
         # print(chef.predict(model, [1, 2022, 0, 0, 0, 10]))
@@ -52,4 +55,4 @@ class DecisionTrees:
             cnt += 1
 
         # return prediction
-        return chef.predict(model, mine_characteristics)
+        return chef.predict(mod, mine_characteristics)

classes/minesweeper.py

@@ -389,15 +389,15 @@ class Minesweeper:
             pygame.mixer.Channel(2).set_volume(0.5)
             pygame.mixer.Channel(2).play(pygame.mixer.Sound("assets/sounds/collision.wav"))
 
-    def pick_up(self):
+    def pick_up(self, model):
         if self.offset_x != 0 or self.offset_y != 0:
             return
         for mine in self.current_map.mines:
             if (self.position_x, self.position_y) == (mine.position_x, mine.position_y):
                 tree = decisionTrees.DecisionTrees()
 
-                decision = tree.return_predict()
-                print("Decision : ", decision)
+                decision = tree.return_predict(model)
+                print("Decision : ", decision, "\n")
 
                 self.current_map.mines.remove(mine)
                 pygame.mixer.Channel(3).set_volume(0.7)

main.py

@@ -58,6 +58,11 @@ def main():
     # główna pętla
     game_loop = True
     clock = pygame.time.Clock()
+
+    # create decision tree
+    tree = decisionTrees.DecisionTrees()
+    model = tree.create_model()
+
     while game_loop:
         # wdrożenie FPS, delta - czas od ostatniej klatki
         delta = clock.tick(FPS)
@@ -66,7 +71,7 @@ def main():
         AI.updateFPS()
 
         if saper.offset_x == 0 and saper.offset_y == 0:
-            AI.updateTile()
+            AI.updateTile(model)
 
         # narysowanie terenu i obiektów
         map.draw_tiles()