python 爬虫 晋江 python爬取晋江小说

直接上成品代码吧 简单易懂

import requests
from bs4 import BeautifulSoup
import re
import datetime
import time
import pyautogui as pag
import pyperclip

def check_recent():
    headers = {
        # 假装自己是浏览器，
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/105.0.0.0 Safari/537.36',
        # 把你刚刚拿到的Cookie塞进来
        'Cookie':'你的cookie',
    }
    session = requests.Session()
#这里填写你自己文章的地址
    response = session.get('https://my.jjwxc.net/backend/managenovel.php?novelid=00000000&jsid=3000000_4000000', headers=headers)
    response.encoding = response.apparent_encoding  # 自动转码

    html = response.text
    soup = BeautifulSoup(html, features="html.parser")  # 创建一个BeautifulSoup对象
    table_node = soup.find_all('table')
    res = ''
    for table in table_node:
        res = res + table.get_text()
    #把空格去一去
    res = res.replace(' ', '')
    res = res.replace('\n', '')
    res = res.replace('\r', '')
    res = res.replace('\t', '')
    
    #用正则匹配一下
    match1 = re.findall("总点击：\d+\xa0总书评：\d+\xa0精华评：\d+\xa0收藏数：\d+", res)
    message = match1[0].split('\xa0')
    # print(message)

    new_info = []
    for i in [0,1,3]:
        new_info.append(int(message[i].split('：')[1]))
    return new_info


def compare_with_before(recent_info, new_info) :
    difference = [new_info[i] - recent_info[i] for i in range(3)]
    if difference[0] > 0:
        text = "总点击增加了" + str(difference[0])
        print(text)
        send_message(text)
    if difference[1] > 0:
        text = "总书评增加了" + str(difference[1])
        print(text)
        send_message(text)
    if difference[2] > 0:
        text = "收藏数增加了" + str(difference[2])
        print(text)
        send_message(text)


def send_message(text):

    #填入你自己的微信输入框坐标
    pag.click(-937, 95)  # 鼠标点击并定位到聊天窗口
    pyperclip.copy(text)  # 复制该行
    pag.hotkey("ctrl", "v")  # 粘贴，mac电脑则把ctrl换成command
    pag.typewrite("\n")  # 发送
    time.sleep(5)  # 每次发完间隔5s


def time_printer():
    now = datetime.datetime.now()
    ts = now.strftime('%Y-%m-%d %H:%M:%S')
    print('do func time :', ts)


if __name__ == '__main__':
    #初始化默认为0，也可以改成你的数据
    recent_info = [0,0,0]
    while True:
        time_printer()
        new_info = check_recent()
        compare_with_before(recent_info,new_info)
        recent_info = new_info
        time.sleep(1200)