使用 selenium 处理计算验证码实现168网课登录效果

whz1998 发表于 2024-4-12 03:50

~~~ python
from selenium import webdriver
import time
from selenium.webdriver.common.by import By
from PIL import Image
import ddddocr
import time
driver = webdriver.Chrome()
option = webdriver.ChromeOptions()
option.add_experimental_option("detach", True)
# 忽略证书错误
option.add_argument('--ignore-certificate-errors')
# 忽略 Bluetooth: bluetooth_adapter_winrt.cc:1075 Getting Default Adapter failed. 错误
option.add_experimental_option('excludeSwitches', ['enable-automation'])
# 忽略 DevTools listening on ws://127.0.0.1... 提示
option.add_experimental_option('excludeSwitches', ['enable-logging'])
# 获取driver对象, 并将配置好的option传入进去运行结束不关闭窗口
driver = webdriver.Chrome(options=option)
driver.get('https://xatu.168wangxiao.com/web/login?redirect=%2Fhome')
driver.maximize_window()
time.sleep(3)
driver.find_element(By.XPATH,"/html/body/div/div/div/div/div/div/div/div/div/form/div/div/div/input").send_keys("username")
driver.find_element(By.XPATH,"/html/body/div/div/div/div/div/div/div/div/div/form/div/div/div/input").send_keys("password")
# 防止图片识别错误 try except 处理
while True:

try:
   svgHeml=driver.find_element(By.XPATH,'/html/body/div/div/div/div/div/div/div/div/div/form/div/div/div/div')

   # 获取元素的位置和尺寸
   location = svgHeml.location
   size = svgHeml.size

   # 计算四个角的坐标
   x1 = location['x']
   y1 = location['y']
   x2 = location['x'] + size['width']
   y2 = location['y'] + size['height']

   # 输出四个角的坐标
   print("左上角坐标: (x1, y1) = ({0}, {1})".format(x1, y1))
   print("右上角坐标: (x2, y1) = ({0}, {1})".format(x2, y1))
   print("左下角坐标: (x1, y2) = ({0}, {1})".format(x1, y2))
   print("右下角坐标: (x2, y2) = ({0}, {1})".format(x2, y2))

   # 截全图
   screenshot_path = 'screenshot.png'
   driver.save_screenshot(screenshot_path)

   # # time.sleep(3)

   # # # 截取指定区域
   im = Image.open(screenshot_path)
   region = im.crop((x1, y1, x2, y2))

   # # 保存截图
   cropped_path = 'result_screenshot.png'
   region.save(cropped_path)

   start = time.time() # 开始时间

   # 1. 创建DdddOcr对象
   ocr = ddddocr.DdddOcr(show_ad=False)

   # 2. 读取图片
   with open(cropped_path, 'rb') as f:
         img = f.read()

   # 3. 识别图片内验证码并返回字符串
   result = ocr.classification(img)
   print("识别结果：",result)
   num = eval(result)
   # eval()函数的安全性问题，因为它会执行字符串中的任何Python代码，可能会导致安全漏洞，所以仅在可信的字符串上使用它
   # 方法二
   # import ast

   # expression = "3 + 4 * 2"
   # result = ast.literal_eval(expression)
   # print(result)

   print(result)
   print(num)
   end = time.time()
   print("耗时：%s 秒" % str(start-end))
   driver.find_element(By.XPATH,"/html/body/div/div/div/div/div/div/div/div/div/form/div/div/div/input").send_keys(5)
   driver.find_element(By.XPATH,"/html/body/div/div/div/div/div/div/div/div/div/form/div/div/button").click()
except:
   # 以防万一图片识别错误重新点击图片重新获取验证码执行try
   driver.find_element(By.XPATH,'/html/body/div/div/div/div/div/div/div/div/div/form/div/div/div/div').click()
   time.sleep(3)


~~~

爱飞的猫 发表于 2024-4-12 07:14

`while True:` 加上 `try/except` 死循环了哦。

改写方案：

```
while True:
try:
pass # 处理代码
except:
pass # 遇到错误时
else:
break # 没有错误时，跳出 while True 的循环
```

78zhanghao87 发表于 2024-4-12 08:00

ocr识别？

baliao 发表于 2024-4-12 08:09

78zhanghao87 发表于 2024-4-12 08:00
ocr识别？

ddddocr（带带弟弟 OCR ）

1e3e 发表于 2024-4-12 08:40

太强了，给楼主点个赞赞啊

Pwaerm 发表于 2024-4-12 08:53

本帖最后由 Pwaerm 于 2024-4-12 08:54 编辑

168哇哇叫

168平台很多bug
里面源码一看就是菜鸟写的
不用登陆都可以可以直接灌入数据

流泪的小白 发表于 2024-4-12 09:18

感谢分享

chengmao2 发表于 2024-4-12 09:19

新技术，感谢分享！学习💪

tyq2003 发表于 2024-4-12 10:32

太厉害了！给你点赞

blindcat 发表于 2024-4-12 11:14

ocr识别率怎么样？

页: [1] 2

吾爱破解 - 52pojie.cn's Archiver

使用 selenium 处理计算验证码实现168网课登录效果