Spider--实战--selenium_12306

# login12306_02
# 图像识别涉及到深度学习，这里直接将验证码识别任务发送到大佬的验证码解析地址，不过现在已经失效了，程序跑到这会报错。
# 用户名和密码存储在本地工作目录中的 username_password_12306.txt文件中。

from selenium import webdriver
from selenium.webdriver import ActionChains
from selenium.webdriver.common.by import By
from selenium.webdriver.support.wait import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
import base64
import re
import time
import requests

class Login(object):
    def __init__(self, driver):
        self.driver = driver
#         图片验证码坐标
        self.coordinate = [[-105, -20], [-35, -20], [40, -20], [110, -20], [-105, 50], [-35, 50], [40, 50], [110, 50]]

    def get_login_info(self):
        with open('username_password_12306.txt', 'r') as file_obj:
            lines = file_obj.readlines()
            username = lines[0].split()
            password = lines[1].split()
        return [username, password]

    def login(self):
        # 读取文件，获取账号和密码
        username = self.get_login_info()[0]
        password = self.get_login_info()[1]
        # 12306登陆页面
        login_url = "https://kyfw.12306.cn/otn/resources/login.html"
        # 设置浏览器长宽
        self.driver.set_window_size(1200, 900)
        # 打开登陆页面
        self.driver.get(login_url)
        # 找到账号登陆按钮  （有两种登陆方式，一种是扫码登录，一种是账号登录）。
        account = self.driver.find_element_by_class_name("login-hd-account")
        # 点击按钮
        account.click()
        # 找到用户名输入框
        userName = self.driver.find_element_by_id("J-userName")
        # 输入用户名
        userName.send_keys(username)
        # 找到密码输入框
        passWord = self.driver.find_element_by_id("J-password")
        # 输入密码
        passWord.send_keys(password)

    def getVerifyImage(self):
        try:
            # 找到图片验证码标签
            img_element = WebDriverWait(self.driver, 100).until(
                EC.presence_of_element_located((By.ID, "J-loginImg"))
            )

        except Exception as e:
            print(u"验证码未加载,请检查您的网络设置！")
        # 获取图片验证码的src属性，就是图片base64加密后的数据
        base64_str = img_element.get_attribute("src").split(",")[-1]
        # base64解码得到图片的数据
        imgdata = base64.b64decode(base64_str)
        # 存入img.jpg
        with open('img.jpg', 'wb') as file:
            file.write(imgdata)
        self.img_element = img_element

    def getVerifyResult(self):
        '''解析返回结果，将坐标存放到列表里'''
        # 12306验证码识别网址
        url = "http://littlebigluo.qicp.net:47720/"   # 目前，这个网址被封了，无法使用
        # 发送post请求把图片数据带上
        response = requests.request("POST", url, data={"type": "1"}, files={'pic_xxfile': open('img.jpg', 'rb')})
        result = []
        print(response.text)
        # 返回识别结果
        for i in re.findall("<B>(.*)</B>", response.text)[0].split(" "):
            result.append(int(i) - 1)
        self.result = result
        print(result)

    def moveAndClick(self):
        # 根据目标验证码的位置，点击相应的图片。
        try:
            # 创建鼠标对象
            Action = ActionChains(self.driver)
            for i in self.result:
                # 根据获取的结果取坐标选择图片并点击
                Action.move_to_element(self.img_element).move_by_offset(self.coordinate[i][0],
                                                                        self.coordinate[i][1]).click()
            Action.perform()
        except Exception as e:
            print(e)

    def submit(self):
        # 点击登陆按钮
        self.driver.find_element_by_id("J-login").click()

if __name__ == '__main__':
    driver=webdriver.Firefox()
    login = Login(driver)
    login.login()
    time.sleep(3)
    login.getVerifyImage()
    time.sleep(1)
    login.getVerifyResult()
    time.sleep(1)
    login.moveAndClick()
    time.sleep(1)
    login.submit()
    time.sleep(10)
#     driver.close()
#     driver.quit()

参考资料：

https://www.52pojie.cn/thread-1048861-1-1.html