提取多个html生成单个html

#!/bin/bash
# coding=utf-8
"""
    作者：gaojs
    功能：
    新增功能：
            schedule 是一个第三方轻量级的任务调度模块，可以按照秒，分，小时，日期或者自定义事件执行时间。
            schedule 允许用户使用简单、人性化的语法以预定的时间间隔定期运行 Python 函数（或其它可调用函数）
        demo如下：
            schedule.every(10).seconds.do(job)
            schedule.every(10).minutes.do(job)
            schedule.every().hour.do(job)
            schedule.every().day.at("10:30").do(job)
            schedule.every(5).to(10).minutes.do(job)
            schedule.every().monday.do(job)
            schedule.every().wednesday.at("13:15").do(job)
            schedule.every().minute.at(":17").do(job)
    日期：2022/5/29 12:45
    blog: www.gaojs.com.cn
"""
from email.mime.application import MIMEApplication
from email.mime.multipart import MIMEMultipart
from smtplib import SMTP_SSL
from email.header import Header
import schedule
from selenium import webdriver
from email.mime.text import MIMEText
from selenium.webdriver.firefox.options import Options
import re
import time
import os
def get_report_source_code():
    """
    获取test报告源码页面
    """
    chrome_options = Options()
    chrome_options.add_argument('--headless')
    chrome_options.add_argument('--disable-gpu')
    driver = webdriver.Firefox(options=chrome_options)
    release_name = ['APV', 'netIAG', 'FT', 'UAG', 'AG9.4']
    source = ''
    for release in release_name:
        time.sleep(5)
        url = f'http://192.168.120.204/builds/{release}/'
        driver.get(url)
        driver.maximize_window()
        source_code = driver.page_source
        # 删除替换网页源码部分标签
        result = re.sub('<td valign="top">(.*?)</td>', ' ', source_code)
        result = re.sub('<tr> <td><a href="/builds/">Parent(.*?)</td></tr>', ' ', result)
        result = re.sub('<address>(.*?)</address>', ' ', result)
        result = re.sub('<h1>(.*?)</h1>', f'<h1>{release} Latest Builds</h1>', result)
        result1 = re.sub('<tr><th valign="top">(.*?)</tr>', ' ', result)
        r = re.findall('<tr> <td><a href=".*?">(.*?)</a></td>', result1)
        if r:
            s1 = result1
            for i in r:
                t = f'http://192.168.120.204/builds/{release}/' + i
                s0 = re.sub(f'<tr> <td><a href=".*?">{i}</a></td>', f'<tr> <td><a href=".*?">{t}</a></td>', s1)
                s1 = s0
            source = source + s1
    driver.close()
    return source
def send_email():
    """
    发送test_report邮件
    """
    # 获取页面源码
    source_code = get_report_source_code()
    # 以126邮箱为例
    # ----------------发件相关参数----------------
    smtpserver = 'smtp.126.com'
    port = 0
    sender = 'testops_jianshuai@126.com'
    password = 'HHHAXZQYIAPTAQST'
    receicer = ['gaojs@arraynetworks.com.cn', '1315202xxxx@163.com']
    # ----------------编辑邮件内容----------------
    subject = '204 server 最新builds 一览表'
    body = f'<p>{source_code}<p>'
    msg = MIMEMultipart()
    msg['Subject'] = Header(subject, 'utf-8')  # 邮件主题
    msg['From'] = sender  # 发件人
    msg['To'] = ';'.join(receicer)
    msg.attach(MIMEText(body, 'html', 'utf-8'))
    attchment = MIMEApplication(open(r'./report.html', 'rb').read())
    attchment.add_header('Content-Disposition', 'attachment', filename="report.html")
    msg.attach(attchment)  # 添加附件到邮件
    smtp = SMTP_SSL(smtpserver)
    smtp.login('testops_jianshuai@126.com', password)
    smtp.sendmail(sender, receicer, msg.as_string())
    smtp.quit()
    print('******************* 邮件发送完成,请查收附件！ ************************')
def download():
    """
    自动下载release
    """
    chrome_options = Options()
    chrome_options.add_argument('--headless')
    chrome_options.add_argument('--disable-gpu')
    driver = webdriver.Firefox(options=chrome_options)
    print('*************************** 正在刷新新版本，请稍等 *******************************')
    url = 'http://10.3.0.20/sp2/build/rel_ag_9_4.html'
    time.sleep(5)
    driver.get(url)
    driver.maximize_window()
    print('downloading')
    start_time = time.time()
    try:
        element = driver.find_element_by_xpath(xpath='/html/body/blockquote[1]/a[1]').text
        release_url = element.split(' ')[0]
        print(release_url)
        # 分离url和文件名称
        release_name = os.path.split(release_url)[1]
        print(release_name)
    finally:
        os.chdir('/var/www/html/builds/AG9.4/')
        if not os.path.exists('/var/www/html/builds/AG9.4/' + release_name):
            os.system(r'curl -o %s  %s' % (release_name, release_url))
    end_time = time.time()
    driver.close()
    print(f'*************************** 总共用时 {end_time - start_time} s **************************')
if __name__ == '__main__':
    send_email()