思路來源於此篇文章,歸納的非常全面,但在具體繞過方案上有些部分講的並不夠明晰,本文着重講具體如何配置。
因為puppeteer的page.evaluateOnNewDocument在selenium中並無對應功能,所以selenium無法使用相同方案。
const browser = await puppeteer.launch({ignoreDefaultArgs: ["--enable-automation"], headless: true}); //去除自動化測試的提醒 const page = await browser.newPage(); await page.evaluateOnNewDocument(() => { //在每個新頁面打開前執行以下腳本 const newProto = navigator.__proto__; delete newProto.webdriver; //刪除navigator.webdriver字段 navigator.__proto__ = newProto; window.chrome = {}; //添加window.chrome字段,為增加真實性還需向內部填充一些值 window.chrome.app = {"InstallState":"hehe", "RunningState":"haha", "getDetails":"xixi", "getIsInstalled":"ohno"}; window.chrome.csi = function(){}; window.chrome.loadTimes = function(){}; window.chrome.runtime = function(){}; Object.defineProperty(navigator, 'userAgent', { //userAgent在無頭模式下有headless字樣,所以需覆寫 get: () => "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.113 Safari/537.36", }); Object.defineProperty(navigator, 'plugins', { //偽裝真實的插件信息 get: () => [{"description": "Portable Document Format", "filename": "internal-pdf-viewer", "length": 1, "name": "Chrome PDF Plugin"}] }); Object.defineProperty(navigator, 'languages', { //添加語言 get: () => ["zh-CN", "zh", "en"], }); const originalQuery = window.navigator.permissions.query; //notification偽裝 window.navigator.permissions.query = (parameters) => ( parameters.name === 'notifications' ? Promise.resolve({ state: Notification.permission }) : originalQuery(parameters) ); await page.goto("https://www.aabbccc.com"); //... await browser.close();
該配置足以應付絕大部分針對無頭瀏覽器的檢測。