让python自动给你报告百度新收录了多少页面，并发送到你的邮箱

import smtplib
import time
from email.mime.text import MIMEText
import parsel

import requests
def get_info(domain):

url = f'http://tool.chinaz.com/baidu/?lm=0&wd={domain}&rn=50'

headers = {

'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.138 Safari/537.36',

}

html = requests.get(url=url, headers=headers).text

html = parsel.Selector(html)

titles = html.xpath("//a[@class='col-blue02']/text()").extract()

urls = html.xpath("//a[@class='col-blue02']/@href").extract()

pages = html.xpath("//a[@class='item'][last()]/text()").extract()

try:

nums = int(html.xpath("//span[@class='col-blue02'][2]/a/text()").extract()[0])

if pages:

pages = int(pages[0]) * 10

else:

pages = 0

if pages == 0:

return {

'titles': titles,

'urls': urls,

'nums': nums

}

else:

all_title = []

all_urls = []

for i in range(0, pages, 10):

url = f'http://tool.chinaz.com/baidu/?pn={i}&wd={domain}&rn=10'

html = requests.get(url=url, headers=headers).text

html = parsel.Selector(html)

print(i)

all_title.extend(html.xpath("//a[@class='col-blue02']/text()").extract())

all_urls.extend(html.xpath("//a[@class='col-blue02']/@href").extract())

return {

'titles': all_title,

'urls': all_urls,

'nums': nums

}

except:

print("此网站被屏蔽！！！")
def send_mail(infomation, old_nums, old_titles, recever):

newnums = infomation['nums']

titles = infomation['titles']

urls = infomation['urls']

new_contents = ''

if infomation['nums'] > old_nums:

for i in range(len(titles)):

if titles[i] not in old_titles:

new_contents += f'{titles[i]}

{urls[i]}

'

mail_content = f'老大，截至{time.strftime("%Y年%m月%d日%H：%M：%S", time.localtime())}

百度新收录：{newnums - old_nums}条

总计收录：{newnums}条

新收录信息如下：'

mail_content += new_contents

mail_content += '目前收录如下：

'

# By：www.lanol.cn

# autor:Lan

for i in range(len(titles)):

mail_content += f'{titles[i]}

{urls[i]}

'

try:

content = MIMEText(mail_content, 'plain', 'utf-8')

reveivers = f"{recever}"

content['To'] = reveivers  # 设置邮件的接收者，多个接收者之间用逗号隔开

content['From'] = str("admin@lanol.cn")  # 邮件的发送者,最好写成str("这里填发送者")，不然可能会出现乱码

content['Subject'] = f"老大，你的网站又被百度收录啦！！！{time.strftime('%Y年%m月%d日%H：%M：%S', time.localtime())}"  # 邮件的主题

smtp_server = smtplib.SMTP_SSL("smtp.exmail.qq.com", 465)

smtp_server.login("{发送邮箱}", "{邮箱密码}")

smtp_server.sendmail("admin@lanol.cn", [recever], content.as_string())

smtp_server.quit()

except Exception as e:

print(str(e))
    except smtplib.SMTPException:
        print("Error: 无法发送邮件")

return {
    'newnums': newnums,
    'titles': titles,
    'urls': urls
}

if name == 'main':

domain = input("请输入你要监控的网址（如：www.lanol.cn，不需要加https这些）：")

reveiver = input("请输入收信邮箱（如：78013994@qq.com）：")

nums = 0

titles = []

urls = []

while True:

new_infomation = send_mail(get_info(domain), nums, titles, reveiver)

nums = new_infomation['newnums']

titles = new_infomation['titles']

urls = new_infomation['urls']

print(f'{time.strftime("%Y年%m月%d日%H：%M：%S", time.localtime())}检查成功')

time.sleep(3600)