Parsing Comment Reactions from Yahoo News

Yahoo News comments have three reaction types with counts displayed in button text.

Reaction Types

Extraction Method

const reactions: { empathy?: number; insightful?: number; disagree?: number } = {};
const reactionButtons = commentArticle.querySelectorAll('button');

reactionButtons.forEach((btn) => {
  const text = btn.textContent?.trim() || '';

  const empathyMatch = text.match(/共感した\s*(\d+)/);
  if (empathyMatch) {
    reactions.empathy = parseInt(empathyMatch[1], 10);
  }

  const insightfulMatch = text.match(/なるほど\s*(\d+)/);
  if (insightfulMatch) {
    reactions.insightful = parseInt(insightfulMatch[1], 10);
  }

  const disagreeMatch = text.match(/うーん\s*(\d+)/);
  if (disagreeMatch) {
    reactions.disagree = parseInt(disagreeMatch[1], 10);
  }
});

Output Structure

Only include reactions that have counts:

{
  empathy: 15,
  insightful: 8,
  disagree: 2
}

Or undefined if no reactions:

const reactions = Object.keys(reactions).length > 0 ? reactions : undefined;

Integration with Comment Extraction

const extractedComments = [];

commentItems.forEach((item) => {
  // ... extract username, text ...

  // Extract reactions
  const reactions: any = {};
  const reactionButtons = commentArticle.querySelectorAll('button');

  reactionButtons.forEach((btn) => {
    const text = btn.textContent?.trim() || '';
    if (text.includes('共感した')) {
      reactions.empathy = parseInt(text.match(/\d+/)?.[0] || '0', 10);
    }
    // ... other reactions
  });

  extractedComments.push({
    text: commentText,
    username: username || undefined,
    reactions: Object.keys(reactions).length > 0 ? reactions : undefined,
  });
});