博客

基于selenium模块自动登录boss直聘

超级鹰识别验证码操作详细

.

import time

from PIL import Image
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver import ChromeOptions, ActionChains
from selenium.webdriver.common.by import By

from practiceproject.spiders.chaojiying import get_coord

# 定义函数，从selenium截取的整个网页图片里截取验证码区域的图片，location是验证码区域坐标，size是验证码区域大小
def get_bossyzm(location, size):
    try:
        im = Image.open('boss_login.png')
        rangle = (int(location['x']),
                  int(location['y']),
                  int(location['x']) + int(size['width']),
                  int(location['y']) + int(size['height']),
                  )
        print(rangle)
        frame = im.crop(rangle)
        # 保存到本地
        frame.save('boss_yzm.png')
        print('截取成功')
    except Exception as e:
        print('本地截取失败', e.__class__, e)

# 设置chrom浏览器无头模式的参数
chrom_options = Options()
chrom_options.add_argument('--headless')
chrom_options.add_argument('--disable-gpu')
# 设置chrom浏览器无法检测到selenium的参数
option = ChromeOptions()
option.add_experimental_option('excludeSwitches', ['enable-automation'])

# boss登录网址
url = 'https://login.zhipin.com/?ka=header-login'

if __name__ == '__main__':
    # 不设置无头模式，便于观察
    driver = webdriver.Chrome(options=option)
    # 发起请求
    driver.get(url)
    # 窗口最大化
    driver.maximize_window()
    # 隐式等待10s
    driver.implicitly_wait(5)
    # 输入用户名和密码
    driver.find_element(By.XPATH,
                        '//form[@action="/wapi/zppassport/login/accountV2"]//div/span/input[@name="account"]'
                        ).click()
    # 这里输入账号
    driver.find_element(By.XPATH,
                        '//form[@action="/wapi/zppassport/login/accountV2"]//div/span/input[@name="account"]'
                        ).send_keys('username')
    driver.implicitly_wait(2)
    driver.find_element(By.XPATH,
                        '//form[@action="/wapi/zppassport/login/accountV2"]//div/span/input[@name="password"]'
                        ).click()
    # 这里输入密码
    driver.find_element(By.XPATH,
                        '//form[@action="/wapi/zppassport/login/accountV2"]//div/span/input[@name="password"]'
                        ).send_keys('password')
    driver.implicitly_wait(2)
    # 点击生成验证码按钮
    driver.find_element(By.XPATH, '//div[@class="verify-init-btn"]').click()
    # 获取验证码图片区域坐标及大小
    time.sleep(3)
    yzm = driver.find_element(By.XPATH, '//div[@class="geetest_fullpage_click_box"]')
    yzm_location = driver.find_element(By.XPATH, '//div[@class="geetest_fullpage_click_box"]').location
    yzm_size = driver.find_element(By.XPATH, '//div[@class="geetest_fullpage_click_box"]').size
    # 保存当前浏览器界面
    driver.save_screenshot('boss_login.png')
    # 使用PIL模块根据坐标和大小截取验证码图片
    print(yzm_location, yzm_size, '坐标，大小')
    get_bossyzm(yzm_location, yzm_size)
    # 将验证码图片发送到超级鹰，获取返回坐标
    result = get_coord('boss_yzm.png')
	# 遍历结果集分别点击对应坐标
    for i in result:
        action = ActionChains(driver)
        action.move_to_element_with_offset(yzm, i[0], i[1]).click().perform()
        
    # 这里使用find_element()方法点击确认按钮的话会报错无法找到元素，暂时还不知道什么原因，所以就直接根据坐标来使用动作链操作了
    action = ActionChains(driver)
    action.move_to_element_with_offset(yzm, 290, 420).click().perform()
    time.sleep(5)
    # 点击登录按钮
    driver.find_element(By.XPATH,
                        '//form[@action="/wapi/zppassport/login/accountV2"]/div[@class="form-btn"]/button').click()
    # driver.quit()