Python 自动化办公-玩转浏览器

关注

原创最新推荐文章于 2026-06-18 08:55:36 发布 1.1k 阅读 AI 写同款· GEO 优化 ›

【Python 爬虫】Playwright 多浏览器支持（Chromium/Firefox/WebKit）手动修改navigator""")Win64;}) 阅读详情

日常工作中，我们不可能避免得使用浏览器来完成一些工作，Python 也有不少浏览器自动化的工具，我用过 selenium、splinter、playwright，最终还是选择了微软的 playwright，之所以选择它，是因为它可以自动安装浏览器，不需要手动下载浏览器的驱动程序，比如 chromedriver，这样写出来的自动化工具很容易移植到其他系统中运行。

Playwright 可通过单个 API 自动执行 Chromium，Firefox 和 WebKit浏览器，支持无头浏览器（headless），Linux、macOS、Windows 下均可以使用，Playwright提供的自动化技术是绿色的，功能强大，稳定且速度快。你可以充分发挥空间，想象它可以实现什么样的功能。

安装：
官方文档 playwright.dev/python/docs…

pip install playwright
playwright install

playwright install 将会安装 Chromium，Firefox 和 WebKit 浏览器的二进制文件，非常方便，需要 Python 3.7 及以上版本才行。

先来一段示例代码：

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    page.goto("http://playwright.dev")
    print(page.title())
    browser.close()

运行程序后，程序会自动打开浏览器，访问 playwright.dev，并打印网站的标题。

自动生成代码

Playwright 最吸引我的地方在于它可以自己记录你对浏览器的操作，并将这些操作生成可以执行的代码，这简直就是神器，大大提升了浏览器自动化的效率。生成代码只需要执行

python -m playwright codegen baidu.com

可以生成如下代码：

from playwright.sync_api import Playwright, sync_playwright
def run(playwright: Playwright) -> None:
    browser = playwright.chromium.launch(headless=False)
    context = browser.new_context()
    # Open new page
    page = context.new_page()
    # Go to https://www.baidu.com/
    page.goto("https://www.baidu.com/")
    # Click input[name="wd"]
    page.click("input[name=\"wd\"]")
    # Fill input[name="wd"]
    page.fill("input[name=\"wd\"]", "playwright ")
    # Press CapsLock
    page.press("input[name=\"wd\"]", "CapsLock")
    # Fill input[name="wd"]
    page.fill("input[name=\"wd\"]", "playwright 教程")
    # Press Enter
    # with page.expect_navigation(url="https://www.baidu.com/s?ie=utf-8&f=8&rsv_bp=1&rsv_idx=1&tn=baidu&wd=playwright%20%E6%95%99%E7%A8%8B&fenlei=256&rsv_pq=880cdb05002fe1ed&rsv_t=19abqiURFrqQT3i6%2F84nvsfVrJlI%2B1T6XbVpQkOap78JGssznOJ4%2FVasRzE&rqlang=cn&rsv_dl=tb&rsv_enter=1&rsv_sug3=23&rsv_sug1=20&rsv_sug7=100&rsv_sug2=0&rsv_btype=i&inputT=6608&rsv_sug4=11435&rsv_jmp=fail"):
    with page.expect_navigation():
        page.press("input[name=\"wd\"]", "Enter")
    # Click text=Playwright-python 教程_天下任我行-优快云博客
    # with page.expect_navigation(url="https://blog.youkuaiyun.com/lb245557472/article/details/111572119"):
    with page.expect_navigation():
        with page.expect_popup() as popup_info:
            page.click("text=Playwright-python 教程_天下任我行-优快云博客")
        page1 = popup_info.value
    # Click text=×
    page1.click("text=×")
    # ---------------------
    context.close()
    browser.close()
with sync_playwright() as playwright:
    run(playwright)

如何与浏览器的元素进行交互
先熟悉一些概念

浏览器

浏览器就是指浏览器的一个实例，可以是 Chromium, Firefox 或 WebKit，Playwright 脚本通常从打开一个浏览器开始，以关闭浏览器作为结束，可以使用无头浏览器模式，也就是说虽然打开了浏览器，但是看不到浏览器启动和操作的过程，这是隐藏的。

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch(headless=False)
    browser.close()

浏览器上下文

浏览器上下文是浏览器实例中孤立的匿名会话。浏览器上下文创建速度快且价格低廉。我们建议在自己的新浏览器上下文中运行每个测试场景，以便在测试之间隔离浏览器状态。浏览器上下文还可用于模拟涉及移动设备、权限、区域设置和配色方案的多页场景。

browser = playwright.chromium.launch()
context = browser.new_context()

页面和框架

浏览器上下文可以有多个页面。页面是指浏览器上下文中的单个选项卡或弹出窗口。它应该用于导航到 URL 并与页面内容交互。

page = context.new_page()

# 显式导航，类似于在浏览器中输入URL。
page.goto('http://example.com')
# 填写输入。
page.fill('#search', 'query')

# 点击链接隐式导航。
page.click('#submit')
# 期待一个新的网址。
print（page.url）

# 页面可以从脚本导航 - 剧作家将接取。
# window.location.href = 'https://example.com'

一个页面可以附加一个或多个 Frame 对象。每个页面都有一个主框架，假设页面级交互（如点击）在主框架中运行。

一个页面可以有附加的附加框架和 iframe HTML 标签。可以访问这些 iframe：

# Get frame using the frame's name attribute
frame = page.frame('frame-login')

# Get frame using frame's URL
frame = page.frame(url=r'.*domain.*')

# Get frame using any other selector
frame_element_handle = page.query_selector('.frame-class')
frame = frame_element_handle.content_frame()

# Interact with the frame
frame.fill('#username-input', 'John')

选择器

选择器就是选取 html 页面内元素的工具。

Playwright 可以使用 CSS 选择器、XPath 选择器、id 等 HTML 属性、data-test-id 甚至文本内容搜索元素。

您可以显式指定您正在使用的选择器引擎，或让 Playwright 检测到它。

Playwright 的选择器是非常直观好用的，在这里了解更多关于选择器和选择器引擎的信息。

实现视频网站的自动播放
以下是简单的打开视频网站，并通过刷新浏览器达到感知视频播放完毕的代码实现。

from playwright.sync_api import sync_playwright
import re, sys
import progressbar
from log import logger
from urllib.parse import urlparse

import time
from config import chromium, browser_path


current_milli_time = lambda: int(round(time.time() * 1000))


class AutoLearning(object):
    @staticmethod
    def get_total_seconds(time_str):
        hour, minute, seconds = 0, 0, 0
        time = [int(i) for i in time_str.split(":")]
        if len(time) == 2:
            minute, seconds = time
        elif len(time) == 1:
            seconds = time[0]
        elif len(time) == 3:
            hour, minute, seconds = time
        else:
            pass
        return hour * 60 * 60 + minute * 60 + seconds

    def __init__(self, username, passwd, base_url, key=None):
        self.username = username
        self.passwd = passwd
        urlparseObj = urlparse(base_url)
        self.base_url = f"{urlparseObj.scheme}://{urlparseObj.hostname}"
        self.hostname = urlparseObj.hostname
        self.sync_playwright = sync_playwright()
        self.playwright = self.sync_playwright.start()
        if chromium:
            self.browser = self.playwright.chromium.launch(executable_path=browser_path, headless=False)
        else:
            self.browser = self.playwright.firefox.launch(executable_path=browser_path,headless=False)
        self.context = self.browser.new_context()
        self.current_page = self.context.new_page()
        self.cookies = {}
        self.corp_code = "default"
        self.map_url = f"{self.base_url}/els/html/index.parser.do?id=0007"

        self.headers = {
            "Host": self.hostname,
            "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 11_2_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4389.114 Safari/537.36",
            "Origin": self.base_url,
        }

        self.eln_session_id = ""

    def __del__(self):
        self.context.close()
        self.browser.close()
        self.sync_playwright.__exit__()

    def login(self):
        logger.info(self.base_url)
        self.current_page.goto(url=self.base_url)
        page = self.current_page
        # self.context.set_default_timeout(6000)
        try:
            # Click [placeholder="请输入用户名"]
            page.click('[name="loginName"]')

            # Fill [placeholder="请输入用户名"]
            page.fill('[name="loginName"]', self.username)

            # Click [placeholder="请输入密码"]
            page.click('[name="password"]')

            # Fill [placeholder="请输入密码"]
            page.fill('[name="password"]', self.passwd)

            # Click text=登录
            page.click("input.login_Btn")
            print("如有验证码，请在浏览器上手动登陆")
            # Click text=继续登录
            # if page.is_visible("text=继续登录", timeout=15000):
            page.click("text=继续登录")

        except Exception:
            print("请在浏览器上手动登陆")
            time.sleep(3)

        while True:
            try:
                if (
                        page.is_visible("text='课程中心'", timeout=3000)
                        or page.is_visible("text='个人中心'", timeout=3000)
                        or page.is_visible("text='学习中心'", timeout=3000)
                ):
                    logger.info("登陆成功")
                    self.current_page = page
                    break
            except Exception:
                print("请在浏览器上手动登陆")
            time.sleep(5)


    def learn_course_from_learn_map(self, which_one_to_learn=1, skip_num=0):

        logger.info("learn_course_from_learn_map begin.")
        self.current_page.goto(self.map_url)
        self.current_page.wait_for_selector(
            f":nth-match(div.track-list-tit,{which_one_to_learn})"
        )
        item = self.current_page.query_selector(
            f":nth-match(div.track-list-tit,{which_one_to_learn})"
        )
        link = self.current_page.query_selector(
            f":nth-match(a.track-list-linktoName,{which_one_to_learn})"
        )
        item_title = item.inner_text()
        link_title = link.inner_text()
        if "学习进度：100%" in item_title:
            logger.info(f"{link_title} 已经学习完成，退出")
            return

        logger.info(f"开始学习 {item_title}")
        link.click()

        self.current_page.wait_for_selector("a.innercan.goCourseByStudy")
        courses = self.current_page.query_selector_all("a.innercan.goCourseByStudy")

        for course in courses[skip_num:]:
            time.sleep(2)
            with self.current_page.expect_popup() as popup_info:
                course.click()
            new_page = popup_info.value
            new_page.wait_for_load_state(timeout=60000)
            time.sleep(1)
            h3 = new_page.query_selector("h3.cs-test-title")
            if h3:
                logger.info("本课程视频已播放完毕，无需播放")
                if h3.inner_text() == "课程评估":
                    self.evaluation(new_page)
                new_page.close()
                continue

            course_item = {
                "courseId": course.get_attribute("id"),
                "courseName": course.get_attribute("title"),
            }

            logger.info(
                f"正在播放 {course_item['courseName']}，courseId = {course_item['courseId']}"
            )

            if new_page.is_visible("iframe.url-course-content"):
                self.play_single_course2(new_page)

            if new_page.is_visible("text='确定'",timeout = 3000):
                new_page.click("text='确定'")

            if new_page.is_visible("a:has-text('下一步')"):
                new_page.click("a:has-text('下一步')")
                self.evaluation(new_page)
            new_page.close()
        logger.info("学习地图任务已完成")







    def play_single_course2(self, page):
        """
        一分屏、二分屏播放
        """

        page.wait_for_selector("time.cl-time")
        page.wait_for_selector("id=studiedTime")
        time.sleep(5)
        total_time_ele = page.query_selector("time.cl-time")
        total_minutes = int(0 if total_time_ele.inner_text() == '' else total_time_ele.inner_text())
        alread_time_ele = page.query_selector("id=studiedTime")
        alread_minutes = int(0 if alread_time_ele.inner_text() == '' else alread_time_ele.inner_text())

        chapters = page.query_selector_all("a.scormItem-no.cl-catalog-link.cl-catalog-link-sub.item-no")
        if len(chapters) > 0:
            logger.info(f"本次需要播放 {len(chapters)} 节")
            chapters[0].click()
            logger.info(f"正在播放 {chapters[0].get_attribute('title')} 总共需要时间 {total_minutes} 分钟")

        bar = None
        if sys.platform == "win32":
            bar = progressbar.bar.ProgressBar(max_value=total_minutes)
        else:
            bar = progressbar.ProgressBar(max_value=total_minutes)

        bar.update(alread_minutes)
        wait_count = 0
        while True:
            time.sleep(60)
            wait_count += 1
            if wait_count >= 7:
                page.reload()
                wait_count = 0
            if wait_count % 3 == 0:
                chapters = page.query_selector_all("a.scormItem-no.cl-catalog-link.cl-catalog-link-sub.item-no")
                if len(chapters) > 0:
                    logger.info(f"本次需要播放 {len(chapters)} 节")
                    chapters[0].click()
                    logger.info(f"正在播放 {chapters[0].get_attribute('title')} 总共需要时间 {total_minutes} 分钟")

            page.wait_for_selector("id=studiedTime")
            alread_time_ele = page.query_selector("id=studiedTime")
            alread_minutes = int(0 if alread_time_ele.inner_text() == '' else alread_time_ele.inner_text())
            if page.is_visible("a:has-text('下一步')"):
                break
            bar.update(alread_minutes)
        time.sleep(1)
        bar.update(total_minutes)
        logger.info(f"本页视频已播放完成")



if __name__ == "__main__":
    auto = AutoLearning(username='****', passwd='*', base_url='http://*****.net')
    auto.login()
    auto.learn_course_from_learn_map(which_one_to_learn=1, skip_num=0)
    time.sleep(100)