fromis_9/backend/scripts/refetch-retweets.js

/**
 * 리트윗 데이터 재수집 스크립트
 * 잘못 저장된 리트윗 일정을 Nitter에서 다시 가져와 수정합니다.
 *
 * 사용법: node scripts/refetch-retweets.js [scheduleId1,scheduleId2,...]
 */
import mysql from 'mysql2/promise';
import { fetchSingleTweet, extractTitle } from '../src/services/x/scraper.js';

const NITTER_URL = process.env.NITTER_URL || 'http://nitter:8080';

const pool = mysql.createPool({
  host: process.env.DB_HOST || 'mariadb',
  port: parseInt(process.env.DB_PORT || '3306'),
  user: process.env.DB_USER || 'fromis9',
  password: process.env.DB_PASSWORD || 'fromis9',
  database: process.env.DB_NAME || 'fromis9',
});

async function main() {
  // CLI에서 특정 ID 지정 가능
  const argIds = process.argv[2]?.split(',').map(Number).filter(Boolean);

  let rows;
  if (argIds && argIds.length > 0) {
    [rows] = await pool.query(
      `SELECT sx.schedule_id, sx.post_id, sx.username, sx.content
       FROM schedule_x sx WHERE sx.schedule_id IN (?)`,
      [argIds]
    );
  } else {
    [rows] = await pool.query(
      `SELECT sx.schedule_id, sx.post_id, sx.username, sx.content
       FROM schedule_x sx
       WHERE sx.content LIKE 'RT @%' OR sx.content LIKE '%nitter%t.co%'`
    );
  }

  console.log(`대상: ${rows.length}건`);
  if (rows.length === 0) {
    await pool.end();
    return;
  }

  let updated = 0;
  let failed = 0;

  for (const row of rows) {
    try {
      // RT @username: 에서 원본 작성자 추출
      const rtMatch = row.content?.match(/^RT @(\w+):/);
      const fetchUsername = rtMatch ? rtMatch[1] : (row.username || 'realfromis_9');

      console.log(`[${row.schedule_id}] post_id=${row.post_id}, from=@${fetchUsername}`);

      const tweet = await fetchSingleTweet(NITTER_URL, fetchUsername, row.post_id);

      // RT @ 프리픽스 제거
      let newContent = tweet.text;
      const rtPrefixMatch = newContent.match(/^RT @\w+:\s*/);
      if (rtPrefixMatch) {
        newContent = newContent.slice(rtPrefixMatch[0].length);
      }
      // 끝의 … 제거
      newContent = newContent.replace(/…$/, '').trim();

      const newTitle = extractTitle(newContent);
      const newImageUrls = tweet.imageUrls.length > 0 ? JSON.stringify(tweet.imageUrls) : null;

      // DB 업데이트
      await pool.query('UPDATE schedules SET title = ? WHERE id = ?', [newTitle, row.schedule_id]);
      await pool.query(
        'UPDATE schedule_x SET username = ?, content = ?, image_urls = ? WHERE schedule_id = ?',
        [fetchUsername, newContent, newImageUrls, row.schedule_id]
      );

      console.log(`  -> title: ${newTitle.substring(0, 60)} | images: ${tweet.imageUrls.length}`);
      updated++;

      // Nitter 부하 방지
      await new Promise(r => setTimeout(r, 500));
    } catch (err) {
      console.error(`  -> 실패: ${err.message}`);
      failed++;
    }
  }

  console.log(`\n완료: ${updated}건 수정, ${failed}건 실패`);
  await pool.end();
}

main().catch(err => {
  console.error(err);
  process.exit(1);
});