1.前言

实验网站：网易云（https://music.163.com/）
本次实验主要用于练习cookie的使用

2.环境

python 3.8.7
selenium 3.141.0

3.代码

导入库

import json
import os
import re
import time
import pymysql
from selenium.webdriver import Chrome, ChromeOptions

输入账号密码

print("******欢迎使用tzh的程序，输入网易云账号密码自动获取近一周常听歌单******")
user = input("输入账号:")
passw = input("输入密码:")

设置等待时间

driver.implicitly_wait(20)

初始化谷歌驱动

def init_chrome():
    # 读取驱动
    driver_path = r"chromedriver.exe"
    # 浏览器exe路径（如果使用的是chrome内核的浏览器而不是chrome，则这行代码一定不能忽略）
    broswer_path = r"C:\Program Files\Google\Chrome\Application\chrome.exe"
    Options = ChromeOptions()
    Options.binary_location = broswer_path
    # 更换头部，防止被网站识别
    Options.add_argument(
        'user-agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) ""AppleWebKit/537.36 '
        '(KHTML, like Gecko) Chrome/93.0.4577.63 Safari/537.36"')
    # 后台运行selenium
    Options.add_argument('headless')
    # 启动chromedriver，根据浏览器路径启动浏览器
    return Chrome(options=Options, executable_path=driver_path)

登录操作

def login(username, password):
    # 登录获取cookie
    driver = init_chrome()
    driver.implicitly_wait(30)
    driver.get("https://music.163.com/")
    driver.delete_all_cookies()
    driver.maximize_window()
    driver.find_element_by_xpath('/html/body/div[1]/div[1]/div/div[1]/a').click()
    driver.find_element_by_xpath('//a[@class="u-btn2 other"]').click()
    driver.find_element_by_xpath('//*[@id="j-official-terms"]').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[1]/div[1]/div[2]/a').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[3]/div[3]/a').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[1]/div/div/input').send_keys(username)
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[2]/div[1]/input').send_keys(password)
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[4]/a').click()
    time.sleep(1)
    cookie = driver.get_cookies()
    with open("./user_cookies/" + str(user) + ".txt", "w+", encoding="utf-8") as f:
        f.write(json.dumps(cookie))
    driver.close()
    print("获取cookie成功")

进入榜单页面

def get_hobby(username):
    # 此函数用于验证cookie是否可用
    driver = init_chrome()
    driver.implicitly_wait(20)
    driver.get("https://music.163.com/#")
    driver.delete_all_cookies()
    with open("./user_cookies/" + username + ".txt", "r")as f:
        for line in json.loads(f.read()):
            driver.add_cookie(line)
    driver.get("https://music.163.com/#")
    driver.maximize_window()
    with open("./user_html/" + str(user) + "_html.txt", "w+", encoding="utf-8") as f:
        f.write(driver.page_source)
    if len(re.findall("动态", driver.page_source)) == 1:
        print("fail")
    else:
        print("success")
        get_top10(driver)

获取本周歌单播放量前十并写入数据库（数据库操作可自行删除省略）

def get_top10(driver):
    # 获取自己听歌排行前十
    driver.switch_to.window(driver.window_handles[-1])
    # /html/body/div[4]/div[2]/div[1]/div/ul/li[1]/a
    id = re.findall(r'GUser={userId:(.*?),nickname:"KeepppGoing",avatarUrl', driver.page_source)
    if len(id) == 1:
        id = id[0]
    driver.get('https://music.163.com/#/user/home?id=' + id)
    driver.switch_to.window(driver.window_handles[-1])
    driver.switch_to.frame('g_iframe')
    time.sleep(1)
    song_name = re.findall('<b title="(.*?)">', driver.page_source)
    singer_name = re.findall('<span title="(.*?)">', driver.page_source)
    listen_time = re.findall('<span class="times f-ff2">(.*?)次</span>', driver.page_source)
    if len(song_name) == len(singer_name) == len(listen_time):
        for i in range(10):
            connect = pymysql.connect(user="", password="",
                                      database="", host="",
                                      port=)
            cursor = connect.cursor()

            cursor.execute("""
            CREATE TABLE IF NOT EXISTS `wyymusic` (
              `username` varchar(50) NOT NULL,
              `password` varchar(50) NOT NULL,
              `song_name` varchar(50) NOT NULL,
              `singer_name` varchar(50) NOT NULL,
              `listen_time` varchar(50) NOT NULL,
              `select_time` varchar(50) NOT NULL
            ) """)
            connect.commit()
            cursor.execute('insert into wyymusic values (%s,%s,%s,%s,%s,%s)',
                           (user, passw, song_name[i], singer_name[i], listen_time[i],
                            str(time.asctime(time.localtime(time.time())))))
            print("歌名:" + song_name[i] + "歌手名::" + singer_name[i] + "听歌次数:" + listen_time[i])
            connect.commit()
            cursor.close()
            connect.close()
    driver.close()

定义一个运行自检的function

def judge_exist(username, password):
    # 存在用户cookie则直接执行 否则自动登录获取
    if username + '.txt' in list(os.walk(r"./user_cookies"))[0][2]:
        print("用户cookie存在")
        try:
            print(get_hobby(username))
        except Exception as e:
            print(e)
            print("cookie失效")
            login(username, password)
            get_hobby(username)
        finally:
            print("运行完毕")
    else:
        print("无用户cookie")
        try:
            login(username, password)
            get_hobby(username)
        except Exception as e:
            print(e)
            print("获取cookie失败")
        finally:
            print("运行完毕")

调用函数

if __name__ == '__main__':
    judge_exist(user, passw)

完整代码

import json
import os
import re
import time
import pymysql
from selenium.webdriver import Chrome, ChromeOptions

print("******欢迎使用tzh的程序，输入网易云账号密码自动获取近一周常听歌单******")
user = input("输入账号:")
passw = input("输入密码:")


def init_chrome():
    # 读取驱动
    driver_path = r"chromedriver.exe"
    # 浏览器exe路径（如果使用的是chrome内核的浏览器而不是chrome，则这行代码一定不能忽略）
    broswer_path = r"C:\Program Files\Google\Chrome\Application\chrome.exe"
    Options = ChromeOptions()
    Options.binary_location = broswer_path
    # 更换头部，防止被网站识别
    Options.add_argument(
        'user-agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) ""AppleWebKit/537.36 '
        '(KHTML, like Gecko) Chrome/93.0.4577.63 Safari/537.36"')
    # 后台运行selenium
    Options.add_argument('headless')
    # 启动chromedriver，根据浏览器路径启动浏览器
    return Chrome(options=Options, executable_path=driver_path)


def login(username, password):
    # 登录获取cookie
    driver = init_chrome()
    driver.implicitly_wait(30)
    driver.get("https://music.163.com/")
    driver.delete_all_cookies()
    driver.maximize_window()
    driver.find_element_by_xpath('/html/body/div[1]/div[1]/div/div[1]/a').click()
    driver.find_element_by_xpath('//a[@class="u-btn2 other"]').click()
    driver.find_element_by_xpath('//*[@id="j-official-terms"]').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[1]/div[1]/div[2]/a').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[3]/div[3]/a').click()
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[1]/div/div/input').send_keys(username)
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[2]/div[1]/input').send_keys(password)
    driver.find_element_by_xpath('/html/body/div[6]/div[2]/div/div[1]/div[4]/a').click()
    time.sleep(1)
    cookie = driver.get_cookies()
    with open("./user_cookies/" + str(user) + ".txt", "w+", encoding="utf-8") as f:
        f.write(json.dumps(cookie))
    driver.close()
    print("获取cookie成功")


def get_hobby(username):
    # 此函数用于验证cookie是否可用
    driver = init_chrome()
    driver.implicitly_wait(20)
    driver.get("https://music.163.com/#")
    driver.delete_all_cookies()
    with open("./user_cookies/" + username + ".txt", "r")as f:
        for line in json.loads(f.read()):
            driver.add_cookie(line)
    driver.get("https://music.163.com/#")
    driver.maximize_window()
    with open("./user_html/" + str(user) + "_html.txt", "w+", encoding="utf-8") as f:
        f.write(driver.page_source)
    if len(re.findall("动态", driver.page_source)) == 1:
        print("fail")
    else:
        print("success")
        get_top10(driver)


def get_top10(driver):
    # 获取自己听歌排行前十
    driver.switch_to.window(driver.window_handles[-1])
    # /html/body/div[4]/div[2]/div[1]/div/ul/li[1]/a
    id = re.findall(r'GUser={userId:(.*?),nickname:"KeepppGoing",avatarUrl', driver.page_source)
    if len(id) == 1:
        id = id[0]
    driver.get('https://music.163.com/#/user/home?id=' + id)
    driver.switch_to.window(driver.window_handles[-1])
    driver.switch_to.frame('g_iframe')
    time.sleep(1)
    song_name = re.findall('<b title="(.*?)">', driver.page_source)
    singer_name = re.findall('<span title="(.*?)">', driver.page_source)
    listen_time = re.findall('<span class="times f-ff2">(.*?)次</span>', driver.page_source)
    if len(song_name) == len(singer_name) == len(listen_time):
        for i in range(10):
            connect = pymysql.connect(user="", password="",
                                      database="", host="",
                                      port=)
            cursor = connect.cursor()

            cursor.execute("""
            CREATE TABLE IF NOT EXISTS `wyymusic` (
              `username` varchar(50) NOT NULL,
              `password` varchar(50) NOT NULL,
              `song_name` varchar(50) NOT NULL,
              `singer_name` varchar(50) NOT NULL,
              `listen_time` varchar(50) NOT NULL,
              `select_time` varchar(50) NOT NULL
            ) """)
            connect.commit()
            cursor.execute('insert into wyymusic values (%s,%s,%s,%s,%s,%s)',
                           (user, passw, song_name[i], singer_name[i], listen_time[i],
                            str(time.asctime(time.localtime(time.time())))))
            print("歌名:" + song_name[i] + "歌手名::" + singer_name[i] + "听歌次数:" + listen_time[i])
            connect.commit()
            cursor.close()
            connect.close()
    driver.close()


def judge_exist(username, password):
    # 存在用户cookie则直接执行 否则自动登录获取
    if username + '.txt' in list(os.walk(r"./user_cookies"))[0][2]:
        print("用户cookie存在")
        try:
            print(get_hobby(username))
        except Exception as e:
            print(e)
            print("cookie失效")
            login(username, password)
            get_hobby(username)
        finally:
            print("运行完毕")
    else:
        print("无用户cookie")
        try:
            login(username, password)
            get_hobby(username)
        except Exception as e:
            print(e)
            print("获取cookie失败")
        finally:
            print("运行完毕")


if __name__ == '__main__':
    judge_exist(user, passw)