selenium实现登录百度(自动识别简单验证码)

Posted andy9468

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了selenium实现登录百度(自动识别简单验证码)相关的知识,希望对你有一定的参考价值。

 

# 导入webdriver
import os
import time

import re
import os
from lxml import etree

from selenium import webdriver
from PIL import Image, ImageEnhance
import pytesseract
from func import base642str, str2base64

# 当前目录设置为根路径
ROOT_PATH = os.getcwd()
# print(os.path.join(ROOT_PATH,‘1.txt‘))

# 将mm隐匿化
# import sys
# # print(str2base64("123456"))
# # str1_base64="MTIzNDU2"
# # print(base642str(str1_base64))
# sys.exit()

print(‘000-正在启用selenium...‘)
# 调用环境变量指定的PhantomJS浏览器创建浏览器对象
chromedriver_exe_path = os.path.join(ROOT_PATH, "chromedriver.exe")
driver = webdriver.Chrome(chromedriver_exe_path)
print(‘000-OK‘)
#
url = ‘https://passport.baidu.com/v2/?login‘
print(‘111-selenium正在请求页面:%s‘ % url)
driver.get(url)  # get方法请求页面,获取响应
print(‘111-请求OK‘)

print("打印标题")
print(driver.title)

# 点击账号和密码登录
input_box0 = driver.find_element_by_xpath("//p[@id=‘TANGRAM__PSP_3__footerULoginBtn‘]")
input_box0.click()

print(‘222-selenium正在填写表单...‘)
time.sleep(1)
account = "abcdef"
pwd = "MTIzNDU2"
verify_code = "1234"

input_box1 = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_3__userName‘]")
input_box1.send_keys(account)
time.sleep(0.5)
input_box2 = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_3__password‘]")
input_box2.send_keys(base642str(pwd))
time.sleep(1.5)

# 截图或验证码图片保存地址
screenImg = os.path.join(ROOT_PATH, "screenImg.png")

# 判断是否需要验证码
imgsrc = driver.find_element_by_id("TANGRAM__PSP_3__verifyCodeImg").get_attribute(‘src‘)
if re.match(r‘https://passport.baidu.com/cgi-bin/genimage.*‘, imgsrc):
    print("需要验证码")
    # 自动填写验证码
    # input_box3 = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_3__verifyCode‘]")
    # input_box3.send_keys(verify_code)
    # 浏览器页面截屏
    driver.get_screenshot_as_file(screenImg)

    # 定位验证码位置及大小
    location = driver.find_element_by_id(‘TANGRAM__PSP_3__verifyCodeImg‘).location
    size = driver.find_element_by_id(‘TANGRAM__PSP_3__verifyCodeImg‘).size
    left = location[‘x‘]
    top = location[‘y‘]
    right = location[‘x‘] + size[‘width‘]
    bottom = location[‘y‘] + size[‘height‘]
    # 从文件读取截图,截取验证码位置再次保存
    img = Image.open(screenImg).crop((left, top, right, bottom))
    img = img.convert(‘L‘)  # 转换模式:L | RGB
    img = ImageEnhance.Contrast(img)  # 增强对比度
    img = img.enhance(2.0)  # 增加饱和度
    img.save(screenImg)
    print("图片验证码以保存:%s" % screenImg)
    # time.sleep(100)
    # 再次读取识别验证码
    print("开始读取识别图片验证码")
    # screenImg=os.path.join(ROOT_PATH, "1.jpg")
    print(screenImg)
    img = Image.open(screenImg)
    code = pytesseract.image_to_string(img)
    print("识别结果:%s" % code)
    # time.sleep(100)
    # code= pytesser.image_file_to_string(screenImg)
    driver.find_element_by_id("TANGRAM__PSP_3__verifyCode").send_keys(code.strip())
    print(code.strip())
    time.sleep(2)

    # 手动填写验证码
    # man_verify_code = input("请填写验证码:")
    # input_box3 = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_3__verifyCode‘]")
    # input_box3.send_keys(man_verify_code)
    # print("验证码填写完成请等待")
else:
    print("不需要验证码")

print(‘222-填写表单OK‘)
time.sleep(1)

print(‘333-selenium提交表单...‘)
one_click = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_3__submit‘]")
one_click.click()
time.sleep(2)

# 处理弹窗(手机验证码):点击关闭按钮
i = 0
while True:  # 处理手机验证码认证反复弹窗
    try:
        one3_click = driver.find_element_by_id("TANGRAM__%s__header_a" % (22 + i))  # 22+i在应对弹窗的关闭按钮id名称发生变化。
    except Exception as e:
        print("无需手机验证码")
        break
    else:
        print("第 %s 次弹出安全验证,要求获取手机验证码" % (i + 1))
        time.sleep(0.5)
        print("1s后自动选择无需手机验证码")
        time.sleep(1)
        one3_click.click()
        print("1s后自动点击登陆")
        time.sleep(1)
        one_click.click()
        time.sleep(2)
        # 判断是否成功登陆
        current_title = driver.title.strip()
        if current_title.startswith("登录"):
            print(‘333-登陆失败...‘)
            time.sleep(0.5)
            print(‘333-2秒后自动重试...‘)
            i = i + 1
            time.sleep(2)
            continue
        else:
            print("打印标题")
            print(driver.title)
            print(‘333-登录成功...‘)

            import sys

            sys.exit()

# 处理弹窗(绑定手机号确认):点击不需要修改
# 判断是否需要手机号绑定确认
try:
    # 绑定手机号确认
    one2_click = driver.find_element_by_xpath("//input[@id=‘TANGRAM__PSP_27__rebindGuideCancel‘]")
except Exception as e:
    print("无需绑定手机号确认")
else:
    print("弹出了绑定手机号确认,1s后自动选择不需要")
    time.sleep(1)
    one2_click.click()

# 打印标题
print("打印标题")
print(driver.title)
time.sleep(2)

# 判断是否成功登陆
current_title = driver.title.strip()
if current_title.startswith("登录"):
    print(‘333-登陆失败...‘)
else:
    print(‘333-登录成功...‘)

  

以上是关于selenium实现登录百度(自动识别简单验证码)的主要内容,如果未能解决你的问题,请参考以下文章

selenium识别登录验证码---基于python实现

#私藏项目实操分享#Python模拟登录,selenium模块,Python识别图形验证码实现自动登录

Selenium识别验证码

python selenium-webdriver 登录验证码的处理

最新selenium+验证码识别模拟登陆B站

selenium 自动化测试登录,登录页面有一个随机图片,应该怎么处理?