Upload taobao_auto.py with huggingface_hub

taobao_auto.py

@@ -0,0 +1,216 @@
+#!/usr/bin/env python3
+"""
+Taobao Auto Scraper - Sistema Completo com Visão
+Combina RPA (mouse/teclado) + Visão Computacional para extrair produtos do Taobao
+"""
+
+import sys
+import json
+import time
+import subprocess
+import re
+from pathlib import Path
+from ctypes import *
+from urllib.parse import quote
+from datetime import datetime
+
+# ==================== CONFIGURAÇÃO ====================
+SCREEN_WIDTH = 1360
+SCREEN_HEIGHT = 768
+OUTPUT_DIR = Path("/tmp/taobao_results")
+OUTPUT_DIR.mkdir(exist_ok=True)
+
+# ==================== X11 MOUSE & KEYBOARD ====================
+X11 = CDLL("libX11.so.6")
+Xtst = CDLL("libXtst.so.6")
+
+XOpenDisplay = X11.XOpenDisplay
+XOpenDisplay.restype = c_void_p
+XFlush = X11.XFlush
+XFlush.argtypes = [c_void_p]
+
+XTestFakeKeyEvent = Xtst.XTestFakeKeyEvent
+XTestFakeKeyEvent.argtypes = [c_void_p, c_uint, c_int, c_ulong]
+
+XTestFakeMotionEvent = Xtst.XTestFakeMotionEvent
+XTestFakeMotionEvent.argtypes = [c_void_p, c_int, c_int, c_ulong]
+
+XTestFakeButtonEvent = Xtst.XTestFakeButtonEvent
+XTestFakeButtonEvent.argtypes = [c_void_p, c_uint, c_int, c_ulong]
+
+XKeysymToKeycode = X11.XKeysymToKeycode
+XKeysymToKeycode.argtypes = [c_void_p, c_uint]
+XKeysymToKeycode.restype = c_uint
+
+XSync = X11.XSync
+XSync.argtypes = [c_void_p, c_int]
+
+display = XOpenDisplay(None)
+
+# Keycodes
+XK_Return = 0xFF0D
+XK_Tab = 0xFF09
+XK_Escape = 0xFF1B
+XK_Down = 0xFF54
+XK_Up = 0xFF52
+XK_Left = 0xFF51
+XK_Right = 0xFF53
+XK_Home = 0xFF50
+XK_End = 0xFF57
+XK_BackSpace = 0xFF08
+XK_Control_L = 0xFFE3
+XK_Shift_L = 0xFFE1
+
+# ==================== FUNÇÕES X11 ====================
+
+def keycode(keysym):
+    return XKeysymToKeycode(display, keysym)
+
+def press(keycode, delay=0.05):
+    XTestFakeKeyEvent(display, keycode, 1, 0)
+    XFlush(display)
+    time.sleep(delay)
+    XTestFakeKeyEvent(display, keycode, 0, 0)
+    XFlush(display)
+    time.sleep(delay/2)
+
+def move(x, y):
+    XTestFakeMotionEvent(display, 0, x, y, 0)
+    XFlush(display)
+    time.sleep(0.15)
+
+def click(button=1):
+    XTestFakeButtonEvent(display, button, 1, 0)
+    XFlush(display)
+    time.sleep(0.15)
+    XTestFakeButtonEvent(display, button, 0, 0)
+    XFlush(display)
+
+def scroll(times=3):
+    for _ in range(times):
+        press(keycode(XK_Down), 0.03)
+        time.sleep(0.05)
+
+def type_text(text, delay=0.03):
+    for char in text:
+        if char == '\n':
+            press(keycode(XK_Return))
+        elif char == ' ':
+            press(0x20)
+        elif char.isupper():
+            press(keycode(XK_Shift_L), 0)
+            kc = keycode(ord(char.lower()))
+            if kc:
+                press(kc, delay)
+            press(keycode(XK_Shift_L), 0)
+        else:
+            kc = keycode(ord(char))
+            if kc:
+                press(kc, delay)
+        time.sleep(0.02)
+
+def ctrl_a():
+    press(keycode(XK_Control_L), 0)
+    press(keycode(0x61), 0.05)
+    press(keycode(0x61), 0)
+    press(keycode(XK_Control_L), 0)
+
+# ==================== SCREENSHOT ====================
+
+def screenshot(name=None):
+    if name is None:
+        name = f"screen_{datetime.now().strftime('%H%M%S')}.png"
+    path = OUTPUT_DIR / name
+
+    subprocess.run([
+        "ffmpeg", "-f", "x11grab",
+        "-video_size", f"{SCREEN_WIDTH}x{SCREEN_HEIGHT}",
+        "-i", ":0",
+        "-frames:v", "1",
+        "-y", str(path)
+    ], capture_output=True, timeout=10)
+
+    return path
+
+# ==================== TAOBAO RPA ====================
+
+def open_browser(url="https://taobao.com"):
+    print(f"🌐 Abrindo: {url}")
+    subprocess.Popen(['firefox', '--new-window', url])
+    time.sleep(5)
+    return screenshot("01_opened.png")
+
+def search(query):
+    print(f"🔍 Buscando: {query}")
+
+    # Clica barra de busca
+    move(SCREEN_WIDTH//2, 180)
+    time.sleep(0.3)
+    click()
+    time.sleep(0.5)
+
+    # Limpa e digita
+    ctrl_a()
+    time.sleep(0.2)
+    type_text(query)
+    time.sleep(0.5)
+
+    # Enter
+    press(keycode(XK_Return))
+    time.sleep(6)
+
+    return screenshot("02_search.png")
+
+def scrape_page(page_num):
+    print(f"📦 Página {page_num}")
+
+    path = screenshot(f"03_page_{page_num}.png")
+    print(f"   📸 {path}")
+
+    # Scroll para mais produtos
+    scroll(5)
+    time.sleep(2)
+
+    return path
+
+# ==================== MAIN ====================
+
+def main():
+    query = sys.argv[1] if len(sys.argv) > 1 else "DDR4 16GB"
+
+    print("=" * 60)
+    print("  TAOBAO AUTO SCRAPER")
+    print("=" * 60)
+    print(f"Query: {query}")
+    print(f"Output: {OUTPUT_DIR}")
+    print("=" * 60)
+
+    # Fluxo
+    print("\n1️⃣ Abrindo navegador...")
+    open_browser()
+
+    print("2️⃣ Buscando produto...")
+    search(query)
+
+    print("3️⃣ Extraindo produtos...")
+    for i in range(3):
+        scrape_page(i+1)
+
+    # Salva manifest
+    manifest = {
+        'query': query,
+        'timestamp': datetime.now().isoformat(),
+        'screenshots': [str(p) for p in sorted(OUTPUT_DIR.glob("*.png"))],
+        'output_dir': str(OUTPUT_DIR)
+    }
+
+    with open(OUTPUT_DIR / "manifest.json", 'w') as f:
+        json.dump(manifest, f, indent=2, ensure_ascii=False)
+
+    print("\n✅ Concluído!")
+    print(f"📁 Screenshots salvos em: {OUTPUT_DIR}")
+    print(f"📋 Manifest: {OUTPUT_DIR / 'manifest.json'}")
+    print("\n💡 Use a ferramenta de visão para analisar as imagens")
+
+if __name__ == "__main__":
+    main()