corpora.js

function getVisibleText(element) {
  if (element.nodeType === Node.TEXT_NODE) {
    return element.textContent.trim();
  }

  if (window.getComputedStyle(element).display === 'none') {
    return '';
  }

  return Array.from(element.childNodes)
    .map(child => getVisibleText(child))
    .join('');
}

function words() {
  // 获取目标表格
  const frame = document.querySelector('frame[name="x1"]');
  if (!frame) throw new Error('未找到目标 frame');
  const table = frame.contentDocument.querySelector('#zabba > p > table');
  if (!table) throw new Error('未找到目标表格');

  // 收集结果（跳过第一行表头）
  const results = [...table.querySelectorAll('tr:not([style*="display: none"]')]
    .slice(1)
    .map(tr => {
      const visibleTds = [...tr.querySelectorAll('td:not([style*="display: none"])')];
      const rank = getVisibleText(visibleTds[1]);
      const freq = getVisibleText(visibleTds[2]);
      const word = getVisibleText(visibleTds[3]);
      const pos = getVisibleText(visibleTds[4]).toLowerCase();
      return `${rank}\t${word}\t${pos}\t${freq}`;
    });

  // 用换行符连接结果
  const finalText = results.join('\n');

  // 复制到剪贴板
  if (finalText.length === 0) throw new Error('未找到有效的数据');
  window.prompt("请按Ctrl+C复制文本", finalText);
}

words();