Puppeteer远程服务器始终加载移动端脚本问题排查
解决无头Chromium爬取时加载移动端脚本的问题
根据你排查出的window.matchMedia判断触发移动端脚本的问题,给出以下几个直接可行的解决思路:
1. 设置桌面级视口尺寸
无头模式下的Chromium默认视口可能偏小,直接触发了移动端媒体查询。在创建页面后添加视口设置:
await page.setViewport({ width: 1920, height: 1080 });
2. 指定桌面端User-Agent
部分网站会结合User-Agent判断设备类型,配合视口设置一起添加桌面UA:
await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/118.0.0.0 Safari/537.36');
3. 强制覆盖目标媒体查询的返回值
如果前两种方法不生效,直接在页面加载前注入脚本,修改matchMedia行为,让移动端查询返回false:
await page.evaluateOnNewDocument(() => { const originalMatchMedia = window.matchMedia; window.matchMedia = (query: string) => { if (query.includes('min-device-width : 320px') && query.includes('max-device-width : 480px')) { return { matches: false, media: query, addListener: () => {}, removeListener: () => {}, addEventListener: () => {}, removeEventListener: () => {}, dispatchEvent: () => false }; } return originalMatchMedia(query); }; });
4. 检查远程服务器环境配置
- 确认远程服务器上的Chromium版本与本地一致,旧版本无头模式可能存在默认行为差异;
- 如果是Docker等容器化部署,检查是否有配置限制了浏览器的设备模拟参数。
修改后的完整代码示例
public async fetchScripts(url: string, waitFor = 'cdn-pipes.js') { const page = await this.browser.newPage(); try { // 添加桌面视口与UA设置 await page.setViewport({ width: 1920, height: 1080 }); await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/118.0.0.0 Safari/537.36'); // 可选:覆盖matchMedia行为 await page.evaluateOnNewDocument(() => { const originalMatchMedia = window.matchMedia; window.matchMedia = (query: string) => { if (query.includes('min-device-width : 320px') && query.includes('max-device-width : 480px')) { return { matches: false, media: query, addListener: () => {}, removeListener: () => {}, addEventListener: () => {}, removeEventListener: () => {}, dispatchEvent: () => false }; } return originalMatchMedia(query); }; }); await page.goto(url, {timeout: 10000, waitUntil: 'domcontentloaded'}); const func = waitFor ? `document.documentElement.innerHTML.indexOf("${waitFor}") !== -1 || document.documentElement.innerHTML.indexOf("spa-detector") !== -1` : 'document.readyState === "complete"'; await page.waitForFunction(func, {polling: 500, timeout: 8000}).catch(reason => { console.error('page.waitForFunction', {error: reason, url}); }); const pageUrls = await page.evaluate(() => { const urlArray = Array.from(document.scripts).map((link) => link.src).filter(value => value.includes('taboola.com')); return [...new Set(urlArray)]; }); console.log('fetchMinimal - urlsArray ', {pageUrls}); return pageUrls; } catch (e) { console.error('fetchMinimal - error ', e); } finally { await page.close(); } }
内容的提问来源于stack exchange,提问作者daniel gi
相关产品推荐
相关产品推荐

