Puppeteer不会在无头模式下返回HTML标记,但在非无头模式下会返回-为什么?

时间:2019-03-25 00:58:49

标签: javascript node.js web-scraping puppeteer

我最近开始在JavaScript中尝试使用Puppeteer进行网络抓取,并注意到当我使用无头Puppeteer实例运行代码时,它不会返回我正在寻找的标签,而是返回[]。但是,当关闭无头模式时,代码可以很好地返回标签。它返回的标签是

[ ElementHandle {
    _context:
     ExecutionContext { _client: [CDPSession], _world: [DOMWorld], _contextId: 3 },
    _client:
     CDPSession {
       _events: [Object],
       _eventsCount: 27,
       _maxListeners: undefined,
       _callbacks: Map {},
       _connection: [Connection],
       _targetType: 'page',
       _sessionId: '6950CA7711045DA500204C2EFDAFC5B3' },
    _remoteObject:
     { type: 'object',
       subtype: 'node',
       className: 'HTMLHeadingElement',
       description: 'h1',
       objectId: '{"injectedScriptId":3,"id":3}' },
    _disposed: false,
    _page:
     Page {
       _events: {},
       _eventsCount: 0,
       _maxListeners: undefined,
       _closed: false,
       _client: [CDPSession],
       _target: [Target],
       _keyboard: [Keyboard],
       _mouse: [Mouse],
       _timeoutSettings: [TimeoutSettings],
       _touchscreen: [Touchscreen],
       _accessibility: [Accessibility],
       _networkManager: [NetworkManager],
       _frameManager: [FrameManager],
       _emulationManager: [EmulationManager],
       _tracing: [Tracing],
       _pageBindings: Map {},
       _ignoreHTTPSErrors: false,
       _coverage: [Coverage],
       _javascriptEnabled: true,
       _viewport: [Object],
       _screenshotTaskQueue: [TaskQueue],
       _workers: Map {} },
    _frameManager:
     FrameManager {
       _events: [Object],
       _eventsCount: 3,
       _maxListeners: undefined,
       _client: [CDPSession],
       _page: [Page],
       _networkManager: [NetworkManager],
       _timeoutSettings: [TimeoutSettings],
       _frames: [Map],
       _contextIdToContext: [Map],
       _isolatedWorlds: [Set],
       _mainFrame: [Frame] } } ]

我尝试使用await page.goto(url, {"waitUntil" : "networkidle0"});来使程序等到页面完全加载后再返回标记,但还是没有运气。这是我的功能异常的代码:

const initializeBrowserTask = async (proxy) => {
  console.log("Initializing browser task...");
  if (proxy == '') {
    console.log('No proxy.')
    const browser = await puppeteer.launch({ headless: false });
    return browser;
  } else {
    const browser = await puppeteer.launch({
      'args': ['--proxy-server=`${proxy}`']
    });
    console.log("Proxy: `${proxy}`")
    return browser;
  };

}

const browserTask = async (proxy, url) => {
  const browser = await initializeBrowserTask(proxy);
  let check = 0
  const page = await browser.newPage();
  console.log('Running...')
  await page.goto(url);
  console.log("Arrived at URL!")
  let status = await page.$x('//*[@id="app"]/div[1]/div/div[1]/h1');
  console.log(status)


};

browserTask('', 'http://adidas.com/yeezy')

我确实删除了{"waitUntil" : "networkidle0"},因为它不能解决问题。由于某种原因,此代码在浏览器不是无头的情况下将起作用,但在浏览器不是无头的情况下将返回[]。为什么会出现此问题,我该如何解决?

1 个答案:

答案 0 :(得分:0)

只需弄清楚我的代码出了什么问题。显然,我尝试抓取的网站(adidas.com)会根据标头过滤掉请求。我没有意识到的是,在无头模式下,我实际上到达了Adidas的403 Access Denied页面,这就是为什么我要查找的标签不存在的原因。我猜无头木偶只是发送HTTP请求的另一种方式,而无头无人驾驶更像是实际的网络驱动程序,这就是为什么无头无人驾驶模式没有被阿迪达斯阻止的原因。

要修复无头模式并添加适当的标题,我在“ page.goto(url);”之前使用了这一行代码行:

await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/72.0.3626.109 Safari/537.36');

事实证明,我所缺少的只是使Adidas网站无法阻止我的标题。