nn update 3

.gitignore

@@ -0,0 +1,3 @@
+.idea
+.DS_Store
+__pycache__

main.py

@@ -4,90 +4,10 @@ import random
 import land
 import tractor
 import blocks
+import nn
 import astar_search
-import neural_network.inference
 from pygame.locals import *
-
-examples = [
-    ['piasek', 'sucha', 'jalowa', 'żółty'],
-    ['czarnoziem', 'wilgotna', 'bogata', 'brazowa'],
-    ['rzedzina', 'wilgotna', 'bogata', 'zielona'],
-    ['gleby murszowe', 'wilgotna', 'bogata', 'szara'],
-    ['pustynne gleby', 'sucha', 'jalowa', 'pomarańczowa'],
-    ['torfowiska', 'sucha', 'jalowa', 'czerwona']
-]
-
-attributes = ['typ_gleby', 'wilgotność', 'zawartość_składników', 'kolor']
-
-# Tworzenie obiektu TreeLearn i nauka drzewa decyzyjnego
-# tree_learner = TreeLearn()
-# default_class = 'nieznane'
-
-# tree_learner.train(examples, attributes, default_class)
-
-class TreeLearn:
-    def __init__(self):
-        self.tree = None
-    
-    def train(self, examples, attributes, default_class):
-        self.tree = self.build_tree(examples, attributes, default_class)
-    
-    def build_tree(self, examples, attributes, default_class):
-        if not examples:
-            return Node(default_class)
-        
-        if self.all_same_class(examples):
-            return Node(examples[0][-1])
-        
-        if not attributes:
-            class_counts = self.get_class_counts(examples)
-            default_class = max(class_counts, key=class_counts.get)
-            return Node(default_class)
-        
-        best_attribute = self.choose_attribute(examples, attributes)
-        root = Node(best_attribute)
-        
-        attribute_values = self.get_attribute_values(examples, best_attribute)
-        
-        for value in attribute_values:
-            new_examples = self.filter_examples(examples, best_attribute, value)
-            new_attributes = attributes[:]
-            new_attributes.remove(best_attribute)
-            new_default_class = max(self.get_class_counts(new_examples), key=lambda k: class_counts.get(k, 0))
-            
-            subtree = self.build_tree(new_examples, new_attributes, new_default_class)
-            root.add_child(value, subtree)
-        
-        return root
-    
-    def all_same_class(self, examples):
-        return len(set([example[-1] for example in examples])) == 1
-    
-    def get_class_counts(self, examples):
-        class_counts = {}
-        for example in examples:
-            class_label = example[-1]
-            class_counts[class_label] = class_counts.get(class_label, 0) + 1
-        return class_counts
-    
-    def choose_attribute(self, examples, attributes):
-        # Placeholder for attribute selection logic
-        return attributes[0]
-    
-    def get_attribute_values(self, examples, attribute):
-        return list(set([example[attribute] for example in examples]))
-    
-    def filter_examples(self, examples, attribute, value):
-        return [example for example in examples if example[attribute] == value]
-
-
-class Node:
-    def __init__(self, label):
-        self.label = label
-        self.children = {}
-    
-    def add_child(self, value, child):
-        self.children[value] = child
+import numpy as np
 
 
 class Game:
@@ -105,6 +25,8 @@ class Game:
         self.grass_body = []
         self.red_block = [] #aim block
 
+        #self.one_body = []
+
         self.fawn_seed_body = []
         self.fawn_wheat_body = []
 
@@ -141,6 +63,15 @@ class Game:
         # self.potato = blocks.Blocks(self.surface, self.cell_size)
         # self.potato.locate_soil('black earth', 6, 1, [])
 
+        #class_names = ['Pumpkin', 'Tomato', 'Carrot']
+
+        self.neural_network = nn.NNModel("neural_network/save/second_model.pth")
+
+        # self.pumpkin_batch = self.neural_network.input_image("resources/pampkin.png")
+        # self.tomato_batch = self.neural_network.input_image("resources/tomato.png")
+        # self.carrot_batch = self.neural_network.input_image("resources/carrot.png")
+        
+
         self.tractor = tractor.Tractor(self.surface, self.cell_size)
         self.tractor.draw()
 
@@ -152,13 +83,10 @@ class Game:
         clock = pygame.time.Clock()
 
         move_tractor_event = pygame.USEREVENT + 1
-        pygame.time.set_timer(move_tractor_event, 500)  # tractor moves every 1000 ms
+        pygame.time.set_timer(move_tractor_event, 100)  # tractor moves every 1000 ms
         tractor_next_moves = []
         astar_search_object = astar_search.Search(self.cell_size, self.cell_number)
 
-        veggies = dict()
-        veggies_debug = dict()
-
         while running:
             clock.tick(60)  # manual fps control not to overwork the computer
             for event in pygame.event.get():
@@ -186,28 +114,21 @@ class Game:
                         random_x = random.randrange(0, self.cell_number * self.cell_size, 50)
                         random_y = random.randrange(0, self.cell_number * self.cell_size, 50)
                         print("Generated target: ",random_x, random_y)
+                        #aim-blue block
                         if self.red_block:
                             self.red_block.pop()
                         self.red_block.append([random_x/50, random_y/50])
+
+                        self.path_image = "resources/2.png"
+                        self.aim_batch = self.neural_network.input_image(self.path_image)
+                        self.predicate = self.neural_network.predicte(self.aim_batch)
+
+
                         # below line should be later moved into tractor.py
                         angles = {0: 'UP', 90: 'RIGHT', 270: 'LEFT', 180: 'DOWN'}
                         #bandaid to know about stones
                         tractor_next_moves = astar_search_object.astarsearch(
                             [self.tractor.x, self.tractor.y, angles[self.tractor.angle]], [random_x, random_y], self.stone_body, self.flower_body)
-                        current_veggie = next(os.walk('./neural_network/images/test'))[1][random.randint(0, len(next(os.walk('./neural_network/images/test'))[1])-1)]
-                        if(current_veggie in veggies_debug):
-                            veggies_debug[current_veggie]+=1
-                        else:
-                            veggies_debug[current_veggie] = 1
-
-                        current_veggie_example = next(os.walk(f'./neural_network/images/test/{current_veggie}'))[2][random.randint(0, len(next(os.walk(f'./neural_network/images/test/{current_veggie}'))[2])-1)]
-                        predicted_veggie = neural_network.inference.main(f"./neural_network/images/test/{current_veggie}/{current_veggie_example}")
-                        if predicted_veggie in veggies:
-                            veggies[predicted_veggie]+=1
-                        else:
-                            veggies[predicted_veggie] = 1
-                        print("Debug veggies: ", veggies_debug, "Predicted veggies: ", veggies)
-
                     else:
                         self.tractor.move(tractor_next_moves.pop(0)[0], self.cell_size, self.cell_number)
                 elif event.type == QUIT: