实现了自动识别形状验证码

README.md

@@ -1,3 +1,55 @@
+## 20240601 更新
+### 更新内容
+- 加入了自动验证形状码的方法
+
+### 配置步骤
+步骤较繁琐，我们需要拿到几个重要的坐标，请加油！！！
+#### 1、需要获取以下三个坐标
+```
+# 截取文字区域的左上和右下坐标
+small_img_top_left_x, small_img_top_left_y = 500, 534  # 左上角坐标
+small_img_bottom_right_x, small_img_bottom_right_y = 800, 559  # 右下角X坐标
+
+# 形状图的左上角坐标
+backend_top_left_x, backend_top_left_y = 505, 340
+```
+
+#### 2、运行脚本
+```commandline
+python locate_tool4shape.py
+```
+
+- 运行后，等待浏览器自动滑块后进入形状验证码后，进行操作脚本会捕获鼠标的点击事件。
+- 按下图1,2,3顺序点击,获取三个坐标
+- 点1位于形状图片最左上角的位置，用于计算形状的坐标
+- 点2，3用于截图文字内容，进行识别
+
+
+![PNG](./img/sharp_click.png)
+
+- 运行情况如下图
+
+![PNG](./img/run_loate4shape.png)
+
+#### 3、添加配置
+编辑config.py，填入获取到的值
+
+```commandline
+# 是否自动形状验证码识别，有时不准，就关掉吧
+auto_shape_recognition = True
+
+# 形状图的左上角坐标
+backend_top_left_x, backend_top_left_y = 505, 340
+
+# 截取文字区域的左上和右下坐标
+# 左上角坐标
+small_img_top_left_x, small_img_top_left_y = 500, 534  
+# 右下角坐标
+small_img_bottom_right_x, small_img_bottom_right_y = 800, 559  
+```
+
+
+
 # MyJdCOOKIE
 
 ### 介绍
@@ -19,6 +71,7 @@
 - 写使用文档(已实现)
 - 加一些通知如钉钉等
 - 添加获取滑块x,y坐标的工具(已实现)
+- 加入了自动验证形状码的方法(已实现)
 
 ## 使用文档
 ### 安装依赖
@@ -39,6 +92,9 @@ python locate_tool.py
 
 ![GIF](./img/get_location.gif)
 
+### 配置形状验证码的内容
+####  **详见20240601 更新**
+
 ### 添加配置
 - 复制config_example.py, 重命名为config.py, 我们基于这个config.py运行程序;
 - slide_x_position, slide_y_position用locate_tool.py拿到;
@@ -50,4 +106,7 @@ python locate_tool.py
 ### 运行脚本
 ```commandline
 python main.py
-```
+```
+
+### 特别感谢
+- 感谢 **https://github.com/sml2h3/ddddocr** 项目，牛逼项目

api/qinglong.py

@@ -47,7 +47,7 @@ class QlApi(object):
         async with aiohttp.ClientSession() as session:
             async with session.get(f"{self.url}/{QlUri.envs.value}", headers=self.headers) as response:
                 if response.status == 200:
-                    logger.info("Get Envs successful. Token obtained.")
+                    logger.info("Get Envs successful.")
                     data = await response.json()
                     return data
                 else:
@@ -57,7 +57,7 @@ class QlApi(object):
         async with aiohttp.ClientSession() as session:
             async with session.put(f"{self.url}/{QlUri.envs.value}", data=json.dumps(data), headers=self.headers) as response:
                 if response.status == 200:
-                    logger.info("Set Envs successful. Token obtained.")
+                    logger.info("Set Envs successful.")
                     data = await response.json()
                     return data
                 else:
@@ -67,7 +67,7 @@ class QlApi(object):
         async with aiohttp.ClientSession() as session:
             async with session.put(f"{self.url}/{QlUri.envs_enable.value}", data=data, headers=self.headers) as response:
                 if response.status == 200:
-                    logger.info("enable Envs successful. Token obtained.")
+                    logger.info("enable Envs successful.")
                     data = await response.json()
                     return data
                 else:
@@ -77,7 +77,7 @@ class QlApi(object):
         async with aiohttp.ClientSession() as session:
             async with session.put(f"{self.url}/{QlUri.envs_disable.value}", data=data, headers=self.headers) as response:
                 if response.status == 200:
-                    logger.info("disable Envs successful. Token obtained.")
+                    logger.info("disable Envs successful.")
                     data = await response.json()
                     return data
                 else:

config_example.py

@@ -28,3 +28,15 @@ slide_difference = 10
 
 # 是否自动识别移动滑块，有时不准，就关掉吧
 auto_move = True
+
+# 是否自动形状验证码识别，有时不准，就关掉吧
+auto_shape_recognition = True
+
+# 形状图的左上角坐标
+backend_top_left_x, backend_top_left_y = 505, 340
+
+# 截取文字区域的左上和右下坐标
+# 左上角坐标
+small_img_top_left_x, small_img_top_left_y = 500, 534
+# 右下角坐标
+small_img_bottom_right_x, small_img_bottom_right_y = 800, 559

locate_tool4shape.py

@@ -0,0 +1,62 @@
+import asyncio
+from pynput import mouse
+from config import jd_login_url
+from main import auto_move_slide
+from playwright.async_api import async_playwright
+import random
+
+"""
+这个脚本用于形状码的坐标的
+"""
+
+
+def get_position(x_name, y_name, detail):
+    print(f"请点击{detail},将获取坐标 ...")
+    def on_click(x, y, button, pressed):
+        if pressed:
+            print(f"{x_name}, {y_name} = {x}, {y}")
+            return False  # 停止监听事件
+
+    with mouse.Listener(on_click=on_click) as listener:
+        listener.join()
+
+
+async def main():
+    async with async_playwright() as playwright:
+        try:
+            browser = await playwright.chromium.launch(headless=False)
+            context = await browser.new_context()
+            page = await context.new_page()
+            # 关闭Webdriver属性,绕过Webdriver检测
+            js = """Object.defineProperties(navigator, {webdriver:{get:()=>undefined}});"""
+            await page.add_init_script(js)
+            await page.goto(jd_login_url)
+            await page.get_by_text("账号密码登录").click()
+
+            username_input = page.get_by_placeholder("账号名/邮箱/手机号")
+            await username_input.click()
+            await username_input.type("1")
+            await asyncio.sleep(random.random() / 10)
+
+            password_input = page.get_by_placeholder("请输入密码")
+            await password_input.click()
+            await password_input.type('1')
+
+            await page.get_by_role("checkbox").check()
+            await page.get_by_text("登 录").click()
+
+            await auto_move_slide(page, retry_times=5)
+
+            await asyncio.sleep(random.random() / 10)
+            get_position("backend_top_left_x", "backend_top_left_y", "形状图的左上角")
+            get_position("small_img_top_left_x", "small_img_top_left_y", "文字区域的左上角")
+            get_position("small_img_bottom_right_x", "small_img_bottom_right_y", "文字区域的右下角")
+
+            await context.close()
+            await browser.close()
+        except Exception:
+            return
+
+
+if __name__ == '__main__':
+    asyncio.run(main())

main.py

@@ -1,13 +1,40 @@
 import asyncio
 from api.qinglong import QlApi
-from config import slide_difference, slide_x_position, slide_y_position, auto_move, qinglong_data, user_datas, jd_login_url
+from config import (
+    slide_difference,
+    slide_x_position,
+    slide_y_position,
+    auto_move,
+    qinglong_data,
+    user_datas,
+    jd_login_url,
+    auto_shape_recognition,
+    backend_top_left_x,
+    backend_top_left_y,
+    small_img_top_left_x,
+    small_img_top_left_y,
+    small_img_bottom_right_x,
+    small_img_bottom_right_y
+)
 from loguru import logger
 import time
 from playwright.async_api import Playwright, async_playwright
 import random
 import traceback
 from typing import Union
-from utils.tools import base_move, get_img_bytes, get_forbidden_users_dict, filter_forbidden_users
+from utils.consts import supported_types
+from utils.tools import (
+    base_move,
+    get_img_bytes,
+    get_forbidden_users_dict,
+    filter_forbidden_users,
+    save_img,
+    get_ocr,
+    get_word,
+    save_screenshot_img,
+    get_shape_location_by_type,
+    click_by_autogui
+)
 
 """
 基于playwright做的
@@ -42,6 +69,70 @@ async def auto_move_slide(page, retry_times: int=2):
         time.sleep(3)
 
 
+async def auto_shape(page, retry_times: int=5):
+    ocr = get_ocr(beta=True)
+    """
+    自动识别滑块验证码
+    """
+    for i in range(retry_times):
+        logger.info(f'第{i}次自动识别形状中...')
+        try:
+            # 查找小图
+            await page.wait_for_selector('#cpc_img', state='visible', timeout=3000)
+        except Exception as e:
+            # 未找到元素，认为成功，退出循环
+            logger.info('未找到形状图,退出识别状态')
+            break
+        # 获取 图片的src 属性和button按键
+        background_src = await page.locator('#cpc_img').get_attribute('src')
+        button = page.locator('div.captcha_footer button.sure_btn')
+
+        # 找到刷新按钮
+        refresh_button = page.locator('div.captcha_header img.jcap_refresh')
+
+        # 截文字小图,返回小图path
+        small_img_path = save_screenshot_img(small_img_top_left_x, small_img_top_left_y, small_img_bottom_right_x, small_img_bottom_right_y, 'small_img')
+
+        # 获取大图并保存
+        background_img_bytes = get_img_bytes(background_src)
+        background_img_path = save_img('background_img', background_img_bytes)
+
+        # 获取问题的文字
+        word = get_word(ocr, small_img_path)
+
+        if word.find('色') > 0:
+            logger.info(f'不支持颜色,刷新中......')
+            # 刷新
+            await refresh_button.click()
+            await asyncio.sleep(random.random() * 2)
+            continue
+        else:
+            type = word.split('请选出图中的')[1]
+            if type in supported_types:
+                logger.info(f'已找到图形,点击中......')
+                # 获取点的中心点
+                center_x, center_y = get_shape_location_by_type(background_img_path, type)
+                if center_x is None and center_y is None:
+                    logger.info(f'识别失败,刷新中......')
+                    await refresh_button.click()
+                    continue
+                # 得到网页上的中心点
+                x, y = backend_top_left_x + center_x ,backend_top_left_y + center_y
+                # 点击图片
+                click_by_autogui(x, y)
+                await asyncio.sleep(random.random() * 3)
+                # 点击确定
+                await button.click()
+                await asyncio.sleep(3)
+                continue
+            else:
+                logger.info(f'不支持该类型形状,刷新中......')
+                # 刷新
+                await refresh_button.click()
+                await asyncio.sleep(random.random() * 3)
+                continue
+
+
 async def get_jd_pt_key(playwright: Playwright, user) -> Union[dict, None]:
     """
     获取jd的pt_key
@@ -77,6 +168,11 @@ async def get_jd_pt_key(playwright: Playwright, user) -> Union[dict, None]:
         time.sleep(2)
         await auto_move_slide(page, retry_times=5)
 
+    # 自动验证形状验证码
+    if auto_shape_recognition:
+        time.sleep(2)
+        await auto_shape(page, retry_times=30)
+
     # 等待滑块验证码通过
     await page.wait_for_selector('#msShortcutMenu', state='visible', timeout=120000)
 

requirements.txt

@@ -3,4 +3,5 @@ requests
 aiohttp
 playwright
 loguru
-pyautogui
+pyautogui
+pynput

utils/consts.py

@@ -0,0 +1,9 @@
+supported_types = [
+    "三角形",
+    "正方形",
+    "长方形",
+    "五角星",
+    "六边形",
+    "圆形",
+    "梯形"
+]

utils/tools.py

@@ -1,8 +1,19 @@
 import base64
+import cv2
 import ddddocr
+import io
 import pyautogui
 import random
+import os
+from PIL import Image, ImageGrab
 import re
+import time
+
+def get_tmp_dir(tmp_dir:str = './tmp'):
+    # 检查并创建 tmp 目录（如果不存在）
+    if not os.path.exists(tmp_dir):
+        os.makedirs(tmp_dir)
+    return tmp_dir
 
 
 def ddddocr_find_files_pic(target_file, background_file) -> int:
@@ -41,6 +52,41 @@ def get_img_bytes(img_src: str) -> bytes:
         raise "image is empty"
 
 
+def get_ocr(**kwargs):
+    return ddddocr.DdddOcr(show_ad=False, **kwargs)
+
+
+def save_img(img_name, img_bytes):
+    tmp_dir = get_tmp_dir()
+    img_path = os.path.join(tmp_dir, f'{img_name}.png')
+    # with open(img_path, 'wb') as file:
+    #     file.write(img_bytes)
+    # 使用 Pillow 打开图像
+    with Image.open(io.BytesIO(img_bytes)) as img:
+        # 保存图像到文件
+        img.save(img_path)
+    return img_path
+
+
+def get_word(ocr, img_path):
+    image_bytes = open(img_path, "rb").read()
+    result = ocr.classification(image_bytes, png_fix=True)
+    return result
+
+
+def save_screenshot_img(left, top, right, bottom, img_name):
+    tmp_dir = get_tmp_dir()
+    img_path = os.path.join(tmp_dir, f'{img_name}.png')
+    # 等待片刻以确保截图区域的准备
+    time.sleep(2)
+
+    # 获取屏幕截图
+    screenshot = ImageGrab.grab(bbox=(left, top, right, bottom))
+
+    # 保存截图
+    screenshot.save(img_path)
+    return img_path
+
 def slide_by_autogui(x, y, offset):
     """
     使用pyautogui实现滑块并自定义轨迹方程
@@ -92,3 +138,85 @@ def base_move(slide_x_position, slide_y_position, small_img_bytes, background_im
     x = ddddocr_find_bytes_pic(small_img_bytes, background_img_bytes) + slide_difference
 
     slide_by_autogui(slide_x_position, slide_y_position, x)
+
+
+def sort_rectangle_vertices(vertices):
+    """
+    获取左上、右上、右下、左下顺序的坐标
+    """
+    # 根据 y 坐标对顶点排序
+    vertices = sorted(vertices, key=lambda x: x[1])
+
+    # 根据 x 坐标对前两个和后两个顶点分别排序
+    top_left, top_right = sorted(vertices[:2], key=lambda x: x[0])
+    bottom_left, bottom_right = sorted(vertices[2:], key=lambda x: x[0])
+
+    return [top_left, top_right, bottom_right, bottom_left]
+
+
+def is_trapezoid(vertices):
+    """
+    判断四边形是否为梯形。
+    vertices: 四个顶点按顺序排列的列表。
+    返回值: 如果是梯形返回 True，否则返回 False。
+    """
+    top_width = abs(vertices[1][0] - vertices[0][0])
+    bottom_width = abs(vertices[2][0] - vertices[3][0])
+    return top_width < bottom_width
+
+
+def get_shape_location_by_type(img_path, type: str):
+    """
+    获取指定形状在图片中的坐标
+    """
+    img = cv2.imread(img_path)
+    imgGray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)  # 转灰度图
+    imgBlur = cv2.GaussianBlur(imgGray, (5, 5), 1)  # 高斯模糊
+    imgCanny = cv2.Canny(imgBlur, 60, 60)  # Canny算子边缘检测
+    contours, hierarchy = cv2.findContours(imgCanny, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)  # 寻找轮廓点
+    for obj in contours:
+        perimeter = cv2.arcLength(obj, True)  # 计算轮廓周长
+        approx = cv2.approxPolyDP(obj, 0.02 * perimeter, True)  # 获取轮廓角点坐标
+        CornerNum = len(approx)  # 轮廓角点的数量
+        x, y, w, h = cv2.boundingRect(approx)  # 获取坐标值和宽度、高度
+
+        # 轮廓对象分类
+        if CornerNum == 3:
+            obj_type = "三角形"
+        elif CornerNum == 4:
+            if w == h:
+                obj_type = "正方形"
+            else:
+                approx = sort_rectangle_vertices([vertex[0] for vertex in approx])
+                if is_trapezoid(approx):
+                    obj_type = "梯形"
+                else:
+                    obj_type = "长方形"
+        elif CornerNum == 6:
+            obj_type = "六边形"
+        elif CornerNum == 20:
+            obj_type = "五角星"
+        elif 4 < CornerNum < 6 or 6 < CornerNum < 20 or CornerNum > 20:
+            # 圆形
+            obj_type = "圆形"
+        else:
+            obj_type = "未知"
+
+        if obj_type == type:
+            # 获取中心点
+            center_x, center_y = x + w // 2, y + h // 2
+            return center_x, center_y
+
+    # 如果获取不到,则返回空
+    return None, None
+
+
+def click_by_autogui(x, y):
+    """
+    点击指定坐标的元素
+    """
+    # 移动鼠标到指定坐标
+    pyautogui.moveTo(x, y)
+
+    # 点击鼠标左键
+    pyautogui.click()