Python Playwright库全面详解

Playwright 是 Microsoft 开发的一个现代化的端到端测试和浏览器自动化库，支持 Chromium、WebKit 和 Firefox 浏览器。它提供了跨浏览器、跨平台的自动化能力，且具有高性能和可靠性。

一、核心特性

1. 多浏览器支持：

   • Chromium (Chrome, Edge)
   • WebKit (Safari)
   • Firefox

2. 跨平台支持：

   • Windows
   • macOS
   • Linux

3. 多语言支持：

   • JavaScript/TypeScript (原生)
   • Python
   • Java
   • .NET

4. 自动等待机制：

   • 自动等待元素出现
   • 自动等待网络请求完成

5. 网络拦截：

   • 可以模拟和修改网络请求

6. 设备模拟：

   • 支持移动设备模拟

二、安装与设置

安装 Playwright Python 包

pip install playwright

安装浏览器二进制文件

playwright install

三、核心 API 结构

1. 主要模块

• playwright.sync_api: 同步 API
• playwright.async_api: 异步 API
• playwright.chromium, playwright.firefox, playwright.webkit: 浏览器特定模块

2. 核心类

• Browser: 浏览器实例
• BrowserContext: 浏览器上下文(类似隐身会话)
• Page: 标签页
• Frame: 页面中的框架
• ElementHandle: DOM 元素
• Locator: 元素定位器

四、常用方法与属性

Browser 类

browser = playwright.chromium.launch(headless=False)  # 启动浏览器
browser.close()  # 关闭浏览器
browser.version  # 浏览器版本

Page 类

page.goto(url)  # 导航到URL
page.title()  # 获取页面标题
page.url  # 当前URL
page.content()  # 获取页面HTML内容
page.screenshot()  # 截图
page.pdf()  # 生成PDF
page.click(selector)  # 点击元素
page.fill(selector, text)  # 填充文本
page.wait_for_selector(selector)  # 等待元素出现
page.evaluate(script)  # 执行JavaScript

Locator 类

locator = page.locator(selector)
locator.click()
locator.fill(text)
locator.text_content()
locator.count()
locator.all_inner_texts()

五、实战案例

案例1：基本页面操作

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch(headless=False)
    page = browser.new_page()
    page.goto("https://example.com")
    print(page.title())
    page.screenshot(path="example.png")
    browser.close()

案例2：表单填写与提交

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    page.goto("https://demoqa.com/automation-practice-form")
    
    # 填写表单
    page.fill("#firstName", "John")
    page.fill("#lastName", "Doe")
    page.fill("#userEmail", "[email protected]")
    page.click("#gender-radio-1", force=True)  # 选择男性
    page.fill("#userNumber", "1234567890")
    
    # 提交表单
    page.click("#submit")
    
    # 验证提交结果
    page.wait_for_selector("#example-modal-sizes-title-lg")
    browser.close()

案例3：处理动态内容与等待

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    
    # 监听控制台日志
    page.on("console", lambda msg: print(f"Console: {msg.text}"))
    
    page.goto("https://dynamic-load.vercel.app/")
    
    # 等待并点击加载按钮
    page.wait_for_selector("button:has-text('Load')")
    page.click("button:has-text('Load')")
    
    # 等待动态内容加载
    page.wait_for_selector(".dynamic-content")
    print("Dynamic content loaded:", page.text_content(".dynamic-content"))
    
    browser.close()

案例4：网络请求拦截

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    
    # 拦截并修改请求
    def handle_request(route, request):
        if "api.example.com" in request.url:
            route.fulfill(
                status=200,
                content_type="application/json",
                body='{"mock": "data"}'
            )
        else:
            route.continue_()
    
    page.route("**/*", handle_request)
    
    page.goto("https://example.com")
    browser.close()

案例5：多页面与上下文

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    # 创建两个独立的浏览器上下文
    browser = p.chromium.launch()
    context1 = browser.new_context()
    context2 = browser.new_context()
    
    # 在每个上下文中创建页面
    page1 = context1.new_page()
    page2 = context2.new_page()
    
    # 独立操作
    page1.goto("https://example.com")
    page2.goto("https://google.com")
    
    print("Page1 title:", page1.title())
    print("Page2 title:", page2.title())
    
    context1.close()
    context2.close()
    browser.close()

六、高级功能

1. 设备模拟

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    # 使用iPhone 11模拟
    iphone_11 = p.devices["iPhone 11"]
    browser = p.chromium.launch()
    context = browser.new_context(**iphone_11)
    page = context.new_page()
    page.goto("https://example.com")
    page.screenshot(path="mobile.png")
    browser.close()

2. 录制测试

# 使用Playwright CLI录制测试
# 命令: playwright codegen https://example.com

3. 视频录制

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    context = browser.new_context(record_video_dir="videos/")
    page = context.new_page()
    page.goto("https://example.com")
    # ...执行操作...
    context.close()
    browser.close()
    # 视频会自动保存

4. 性能分析

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    
    # 开始跟踪
    page.context.tracing.start(screenshots=True, snapshots=True, sources=True)
    
    page.goto("https://example.com")
    # ...执行操作...
    
    # 停止跟踪并保存
    page.context.tracing.stop(path="trace.zip")
    browser.close()

七、注意事项

1. 元素定位策略：

   • 优先使用 page.locator() 而不是 page.$() 或 page.$$()
   • 使用语义化的选择器，如 page.get_by_role("button", name="Submit")
   • 避免使用XPath，除非必要

2. 等待机制：

   • Playwright 有自动等待机制，通常不需要手动添加等待
   • 需要时使用 page.wait_for_* 方法

3. 并行执行：

   • 每个 Playwright 实例应该是独立的
   • 使用多个 BrowserContext 来实现并行

4. 资源管理：

   • 确保关闭浏览器和页面
   • 使用 with 语句自动管理资源

5. 调试技巧：

   • 使用 headless=False 查看浏览器操作
   • 添加 slow_mo 参数减慢执行速度便于观察

   browser = p.chromium.launch(headless=False, slow_mo=100)
   

八、常见问题与解决方案

Q1: 元素点击不生效

• 原因：元素可能被遮挡或不可见
• 解决：

  page.click(selector, force=True)  # 强制点击
  # 或
  page.locator(selector).click(timeout=5000)  # 增加超时
  

Q2: 页面加载超时

• 原因：网络慢或页面资源过多
• 解决：

  page.goto(url, timeout=60000)  # 增加超时时间
  # 或
  page.goto(url, wait_until="domcontentloaded")  # 只等待DOM加载
  

Q3: iframe 操作问题

• 解决：

  frame = page.frame_locator("iframe").locator("button")
  frame.click()
  

Q4: 文件下载

• 解决：

  with page.expect_download() as download_info:
      page.click("a#download-link")
  download = download_info.value
  path = download.path()
  download.save_as("/path/to/save")
  

Q5: 处理弹窗

• 解决：

  page.on("dialog", lambda dialog: dialog.accept())
  

九、扩展知识

1. 与 Pytest 集成

# conftest.py
import pytest
from playwright.sync_api import sync_playwright

@pytest.fixture(scope="module")
def browser():
    with sync_playwright() as p:
        browser = p.chromium.launch()
        yield browser
        browser.close()

@pytest.fixture
def page(browser):
    page = browser.new_page()
    yield page
    page.close()

2. 自定义选择器引擎

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    # 注册自定义选择器
    def text_selector(selector, page):
        return page.evaluate_handle("""selector => {
            return Array.from(document.querySelectorAll('*'))
                .find(el => el.textContent === selector);
        }""", selector)
    
    p.selectors.register("text", text_selector)
    
    browser = p.chromium.launch()
    page = browser.new_page()
    page.goto("https://example.com")
    element = page.locator("text=Example Domain")
    print(element.text_content())
    browser.close()

3. 性能监控

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    
    # 监听性能指标
    page.on("metrics", lambda metrics: print(metrics))
    
    page.goto("https://example.com")
    # 获取特定指标
    metrics = page.evaluate("""() => {
        return {
            memory: window.performance.memory,
            timing: window.performance.timing
        }
    }""")
    print(metrics)
    browser.close()

十、最佳实践

1. 使用 Page Object 模式：

   class LoginPage:
       def __init__(self, page):
           self.page = page
           self.username = page.locator("#username")
           self.password = page.locator("#password")
           self.submit = page.locator("#submit")
       
       def login(self, username, password):
           self.username.fill(username)
           self.password.fill(password)
           self.submit.click()
   

2. 配置重用：

   # 创建可重用的浏览器上下文配置
   def create_context(browser, **kwargs):
       return browser.new_context(
           viewport={"width": 1920, "height": 1080},
           locale="en-US",
           timezone_id="America/New_York",
           **kwargs
       )
   

3. 错误处理：

   try:
       page.click(selector, timeout=5000)
   except TimeoutError:
       print(f"元素 {selector} 未找到")
       page.screenshot(path="error.png")
       raise
   

4. 日志记录：

   import logging
   from playwright.sync_api import sync_playwright
   
   logging.basicConfig(level=logging.INFO)
   logger = logging.getLogger(__name__)
   
   with sync_playwright() as p:
       browser = p.chromium.launch()
       page = browser.new_page()
       
       def log_request(request):
           logger.info(f"Request: {request.method} {request.url}")
       
       page.on("request", log_request)
       page.goto("https://example.com")
       browser.close()
   

Playwright 是一个功能强大且灵活的浏览器自动化工具，适用于测试、爬虫和各种自动化场景。通过合理利用其丰富的 API 和功能，可以构建高效可靠的自动化解决方案。