audio_codebuddy/deploy-package/scripts/render-local-html.js

// scripts/render-local-html.js — playwright 渲染本地 HTML 文件,提取 DOM 文本
const { chromium } = require('playwright');
const fs = require('fs');
const os = require('os');
const path = require('path');

(async () => {
  const filePath = process.argv[2];
  const outPath = process.argv[3] || path.join(os.tmpdir(), path.basename(filePath).replace(/\.html$/, '-dom.txt'));

  const url = 'file:///' + filePath.replace(/\\/g, '/');
  const browser = await chromium.launch({ headless: true });
  const page = await browser.newPage();

  // 等动态内容(阿里云 SPA 的 React 渲染可能稍慢)
  await page.goto(url, { waitUntil: 'networkidle', timeout: 60000 }).catch(() => {});

  for (let i = 0; i < 6; i++) {
    await page.waitForTimeout(2000);
    const tables = await page.locator('table').count();
    console.log(`  attempt ${i + 1}: tables=${tables}`);
    if (tables >= 2) break;
  }

  const text = await page.evaluate(() => {
    document.querySelectorAll('script,style,noscript').forEach(e => e.remove());
    return document.body.innerText;
  });

  fs.writeFileSync(outPath, text, 'utf8');
  console.log(`dom text: ${text.length} bytes → ${outPath}`);
  console.log(`first 300 chars:`);
  console.log(text.slice(0, 300));

  await browser.close();
})().catch(e => { console.error(e); process.exit(1); });