Initial code finished

README.md

@@ -1,2 +1,15 @@
-# gartic-dalle
-
+# Gartic DALL·E
+Uma implementação para gerar e desenhar imagens usando inteligência artificial no Gartic.com.br
+
+## Instalação e Uso
+- Abra o arquivo ``main.py`` e defina sua chave de API da OpenAI [(pegue-a aqui)](https://platform.openai.com/account/api-keys) na primeira linha
+- Execute o arquivo ``main.py``
+
+## Avisos
+- O Gartic implementou uma proteção contra desenhos automatizados, então se a imagem possuir muitos pixeis, você será desconectado pelo sistema
+- O script roda baseado no Google Chrome (pois o Firefox não possui as APIs necessárias), portanto ele também é um requisito para funcionamento.
+
+## TODO
+- Uma forma de "bypassear" a proteção contra desenhos automatizados
+- Suportar também o [gartic.io](https://gartic.io)
+- Um overlay para melhor visualização do stdout

main.py

@@ -1,31 +1,53 @@
+OPENAI_API_KEY = "API KEY here"
+
+import importlib.util
+import sys
+import subprocess
+
+
+# Check if all dependencies are installed
+if selenium := importlib.util.find_spec("selenium") is None:
+    subprocess.check_call([sys.executable, "-m", "pip", "install", 'selenium'])
+
+if openai := importlib.util.find_spec("openai") is None:
+    subprocess.check_call([sys.executable, "-m", "pip", "install", 'openai'])
+
+if PIL := importlib.util.find_spec("PIL") is None:
+    subprocess.check_call([sys.executable, "-m", "pip", "install", 'Pillow'])
+
+if requests := importlib.util.find_spec("requests") is None:
+    subprocess.check_call([sys.executable, "-m", "pip", "install", 'requests'])
+
+
+
 from selenium import webdriver
 from selenium.webdriver.common.by import By
 from time import sleep
-from env import OPENAI_API_KEY
 import openai
 from PIL import Image
 from io import BytesIO
 import requests
-from datetime import datetime
 
-openai.api_key = "API KEY here" #OPENAI_API_KEY
+openai.api_key = OPENAI_API_KEY
 
 browser = webdriver.Chrome()
 browser.get("https://gartic.com.br/")
 
-alreadyDrawed: list[dict[str, str]] = []
+# this list is used to fetch the word only once, {word: "word", url: "url"}
+alreadyDrawed: list[dict[str, str]] = [] 
 
 while True:
+    # If user is in a room
     if browser.find_elements(By.CSS_SELECTOR, ".btPular").__len__() == 0:
-        print("Not in a Room")
-        sleep(1)
+        sleep(5)
         continue
 
+    # If user is drawing
     if not browser.find_element(By.CSS_SELECTOR, ".btPular").is_displayed():
-        print("Not drawing")
         sleep(1)
         continue
 
+    # Get the word to draw
     element = browser.find_elements(By.CSS_SELECTOR, ".contentSpan")
     if element.__len__() == 0:
         sleep(5)
@@ -39,6 +61,7 @@
 
     print("Drawing: " + value)
 
+    # Check if the word was already drawed, prevent to fetch the same word twice
     if not any(d["word"] == value for d in alreadyDrawed):
         print("Fetching image...")
 
@@ -51,44 +74,51 @@
         alreadyDrawed.append(
             {
                 "word": value,
-                "url": "https://i.imgur.com/FTgo0DQ.png",
+                "url": response['images'][0]['url'],
             }
         )
-        print("Fetching done...")
-
-    print(
-        f"Image link: {next((item for item in alreadyDrawed if item['word'] == value), None)['url']}"
-    )
 
     img = Image.open(BytesIO(requests.get(next((item for item in alreadyDrawed if item['word'] == value), None)['url']).content))
-
-    browser.execute_script("arguments[0].value = 1; arguments[0].dispatchEvent(new Event('change'))", browser.find_element(By.CSS_SELECTOR, "#tamanho"))        
+
+    #Change brush size to 1px    
+    browser.execute_script("arguments[0].value = 1; arguments[0].dispatchEvent(new Event('change'))", browser.find_element(By.CSS_SELECTOR, "#tamanho"))
 
     drawElement = browser.find_element(By.CSS_SELECTOR ,"#telaCanvas > canvas:nth-child(2)")    
     action = webdriver.ActionChains(driver=browser, duration=0)
 
+
+    #Block gartic drawing backend url to prevent disconnects while drawing
+    print("Blocking Gartic servers...")
     browser.execute_cdp_cmd('Network.setBlockedURLs', {"urls": ["gartic.com.br/room/atualizar.php"]})
     browser.execute_cdp_cmd('Network.enable', {})
 
 
+    print("Start drawing...")
     for x in range(img.width):
         for y in range(img.height):
             pixel = img.load()[x, y]
 
-            if (pixel[0] > 250 and pixel[1] > 250 and pixel[2] > 250):
+            # If pixel is white then skip (background already is white, best performance)
+            if (pixel[0] > 250 and pixel[1] > 250 and pixel[2] > 250): 
                 continue
 
+            # Change color to pixel color
             browser.execute_script(f"arguments[0].setAttribute('codigo', '{'%02x%02x%02x' % (pixel[0], pixel[1], pixel[2])}'); arguments[0].dispatchEvent(new Event('click'))", browser.find_element(By.CSS_SELECTOR, "#cores > div:nth-child(1)"))
             action.move_to_element(drawElement)
+
+            # Selenium moves to center of the element, so we need to move to the top left corner and add relative position
             action.move_by_offset((-drawElement.size['width']/2 + 10) + x , (-drawElement.size['height']/2 + 10) + y)
             action.click()
             action.perform()
-            print(f"{datetime.now().time()}")
 
+    print("Drawing finished...")
+
+    print("Unblocking Gartic servers...")
     browser.execute_cdp_cmd('Network.setBlockedURLs', {"urls": [""]})
     browser.execute_cdp_cmd('Network.disable', {})
 
+    # Wait for round to finish
     while browser.find_element(By.CSS_SELECTOR, ".btPular").is_displayed():
         sleep(1)
 
-    print("Ended drawing...")
+    print("Drawing round finished...")

requirements.txt

@@ -0,0 +1,4 @@
+openai==0.27.8
+Pillow==10.0.0
+Requests==2.31.0
+selenium==4.10.0