fromis_9/backend/services/x-bot.js

/**
 * X 봇 서비스
 *
 * - Nitter를 통해 @realfromis_9 트윗 수집
 * - 트윗을 schedules 테이블에 저장
 * - 유튜브 링크 감지 시 별도 일정 추가
 */

import pool from "../lib/db.js";
import { addOrUpdateSchedule } from "./meilisearch.js";

// YouTube API 키
const YOUTUBE_API_KEY =
  process.env.YOUTUBE_API_KEY || "AIzaSyBmn79egY0M_z5iUkqq9Ny0zVFP6PoYCzM";

// X 카테고리 ID
const X_CATEGORY_ID = 12;

// 유튜브 카테고리 ID
const YOUTUBE_CATEGORY_ID = 7;

/**
 * UTC → KST 변환
 */
export function toKST(utcDate) {
  const date = new Date(utcDate);
  return new Date(date.getTime() + 9 * 60 * 60 * 1000);
}

/**
 * 날짜를 YYYY-MM-DD 형식으로 변환
 */
export function formatDate(date) {
  return date.toISOString().split("T")[0];
}

/**
 * 시간을 HH:MM:SS 형식으로 변환
 */
export function formatTime(date) {
  return date.toTimeString().split(" ")[0];
}

/**
 * Nitter 날짜 파싱 ("Jan 7, 2026 · 12:00 PM UTC" → Date)
 */
function parseNitterDateTime(timeStr) {
  if (!timeStr) return null;
  try {
    const cleaned = timeStr.replace(" · ", " ").replace(" UTC", "");
    const date = new Date(cleaned + " UTC");
    if (isNaN(date.getTime())) return null;
    return date;
  } catch (e) {
    return null;
  }
}

/**
 * 트윗 텍스트에서 첫 문단 추출 (title용)
 */
export function extractTitle(text) {
  if (!text) return "";

  // 빈 줄(\n\n)로 분리하여 첫 문단 추출
  const paragraphs = text.split(/\n\n+/);
  const firstParagraph = paragraphs[0]?.trim() || "";

  return firstParagraph;
}

/**
 * 텍스트에서 유튜브 videoId 추출
 */
export function extractYoutubeVideoIds(text) {
  if (!text) return [];

  const videoIds = [];

  // youtu.be/{videoId} 형식
  const shortMatches = text.match(/youtu\.be\/([a-zA-Z0-9_-]{11})/g);
  if (shortMatches) {
    shortMatches.forEach((m) => {
      const id = m.replace("youtu.be/", "");
      if (id && id.length === 11) videoIds.push(id);
    });
  }

  // youtube.com/watch?v={videoId} 형식
  const watchMatches = text.match(
    /youtube\.com\/watch\?v=([a-zA-Z0-9_-]{11})/g
  );
  if (watchMatches) {
    watchMatches.forEach((m) => {
      const id = m.match(/v=([a-zA-Z0-9_-]{11})/)?.[1];
      if (id) videoIds.push(id);
    });
  }

  // youtube.com/shorts/{videoId} 형식
  const shortsMatches = text.match(
    /youtube\.com\/shorts\/([a-zA-Z0-9_-]{11})/g
  );
  if (shortsMatches) {
    shortsMatches.forEach((m) => {
      const id = m.match(/shorts\/([a-zA-Z0-9_-]{11})/)?.[1];
      if (id) videoIds.push(id);
    });
  }

  return [...new Set(videoIds)];
}

/**
 * 관리 중인 채널 ID 목록 조회
 */
export async function getManagedChannelIds() {
  const [configs] = await pool.query(
    "SELECT channel_id FROM bot_youtube_config"
  );
  return configs.map((c) => c.channel_id);
}

/**
 * YouTube API로 영상 정보 조회
 */
async function fetchVideoInfo(videoId) {
  try {
    const url = `https://www.googleapis.com/youtube/v3/videos?part=snippet,contentDetails&id=${videoId}&key=${YOUTUBE_API_KEY}`;
    const response = await fetch(url);
    const data = await response.json();

    if (!data.items || data.items.length === 0) {
      return null;
    }

    const video = data.items[0];
    const snippet = video.snippet;
    const duration = video.contentDetails?.duration || "";

    // duration 파싱
    const durationMatch = duration.match(/PT(?:(\d+)H)?(?:(\d+)M)?(?:(\d+)S)?/);
    let seconds = 0;
    if (durationMatch) {
      seconds =
        parseInt(durationMatch[1] || 0) * 3600 +
        parseInt(durationMatch[2] || 0) * 60 +
        parseInt(durationMatch[3] || 0);
    }

    const isShorts = seconds > 0 && seconds <= 60;

    return {
      videoId,
      title: snippet.title,
      description: snippet.description || "",
      channelId: snippet.channelId,
      channelTitle: snippet.channelTitle,
      publishedAt: new Date(snippet.publishedAt),
      isShorts,
      videoUrl: isShorts
        ? `https://www.youtube.com/shorts/${videoId}`
        : `https://www.youtube.com/watch?v=${videoId}`,
    };
  } catch (error) {
    console.error(`영상 정보 조회 오류 (${videoId}):`, error.message);
    return null;
  }
}

/**
 * Nitter에서 트윗 수집 (첫 페이지만)
 */
async function fetchTweetsFromNitter(nitterUrl, username) {
  const url = `${nitterUrl}/${username}`;

  const response = await fetch(url);
  const html = await response.text();

  const tweets = [];
  const tweetContainers = html.split('class="timeline-item ');

  for (let i = 1; i < tweetContainers.length; i++) {
    const container = tweetContainers[i];
    const tweet = {};

    // 고정 트윗 체크
    tweet.isPinned =
      tweetContainers[i - 1].includes("pinned") || container.includes("Pinned");

    // 리트윗 체크
    tweet.isRetweet = container.includes('class="retweet-header"');

    // 트윗 ID 추출
    const linkMatch = container.match(/href="\/[^\/]+\/status\/(\d+)/);
    tweet.id = linkMatch ? linkMatch[1] : null;

    // 시간 추출
    const timeMatch = container.match(
      /<span class="tweet-date"[^>]*><a[^>]*title="([^"]+)"/
    );
    tweet.time = timeMatch ? parseNitterDateTime(timeMatch[1]) : null;

    // 텍스트 내용 추출
    const contentMatch = container.match(
      /<div class="tweet-content[^"]*"[^>]*>([\s\S]*?)<\/div>/
    );
    if (contentMatch) {
      tweet.text = contentMatch[1]
        .replace(/<br\s*\/?>/g, "\n")
        .replace(/<a[^>]*>([^<]*)<\/a>/g, "$1")
        .replace(/<[^>]+>/g, "")
        .trim();
    }

    // URL 생성
    tweet.url = tweet.id
      ? `https://x.com/${username}/status/${tweet.id}`
      : null;

    if (tweet.id && !tweet.isRetweet && !tweet.isPinned) {
      tweets.push(tweet);
    }
  }

  return tweets;
}

/**
 * Nitter에서 전체 트윗 수집 (페이지네이션)
 */
async function fetchAllTweetsFromNitter(nitterUrl, username) {
  const allTweets = [];
  let cursor = null;
  let pageNum = 1;
  let consecutiveEmpty = 0;
  const DELAY_MS = 1000;

  while (true) {
    const url = cursor
      ? `${nitterUrl}/${username}?cursor=${cursor}`
      : `${nitterUrl}/${username}`;

    console.log(`[페이지 ${pageNum}] 스크래핑 중...`);

    try {
      const response = await fetch(url);
      const html = await response.text();

      const tweets = [];
      const tweetContainers = html.split('class="timeline-item ');

      for (let i = 1; i < tweetContainers.length; i++) {
        const container = tweetContainers[i];
        const tweet = {};

        tweet.isPinned =
          tweetContainers[i - 1].includes("pinned") ||
          container.includes("Pinned");
        tweet.isRetweet = container.includes('class="retweet-header"');

        const linkMatch = container.match(/href="\/[^\/]+\/status\/(\d+)/);
        tweet.id = linkMatch ? linkMatch[1] : null;

        const timeMatch = container.match(
          /<span class="tweet-date"[^>]*><a[^>]*title="([^"]+)"/
        );
        tweet.time = timeMatch ? parseNitterDateTime(timeMatch[1]) : null;

        const contentMatch = container.match(
          /<div class="tweet-content[^"]*"[^>]*>([\s\S]*?)<\/div>/
        );
        if (contentMatch) {
          tweet.text = contentMatch[1]
            .replace(/<br\s*\/?>/g, "\n")
            .replace(/<a[^>]*>([^<]*)<\/a>/g, "$1")
            .replace(/<[^>]+>/g, "")
            .trim();
        }

        tweet.url = tweet.id
          ? `https://x.com/${username}/status/${tweet.id}`
          : null;

        if (tweet.id && !tweet.isRetweet && !tweet.isPinned) {
          tweets.push(tweet);
        }
      }

      if (tweets.length === 0) {
        consecutiveEmpty++;
        console.log(`  -> 트윗 없음 (연속 ${consecutiveEmpty}회)`);
        if (consecutiveEmpty >= 3) break;
      } else {
        consecutiveEmpty = 0;
        allTweets.push(...tweets);
        console.log(`  -> ${tweets.length}개 추출 (누적: ${allTweets.length})`);
      }

      // 다음 페이지 cursor 추출
      const cursorMatch = html.match(
        /class="show-more"[^>]*>\s*<a href="\?cursor=([^"]+)"/
      );
      if (!cursorMatch) {
        console.log("\n다음 페이지 없음. 스크래핑 완료.");
        break;
      }

      cursor = cursorMatch[1];
      pageNum++;

      await new Promise((r) => setTimeout(r, DELAY_MS));
    } catch (error) {
      console.error(`  -> 오류: ${error.message}`);
      consecutiveEmpty++;
      if (consecutiveEmpty >= 5) break;
      await new Promise((r) => setTimeout(r, DELAY_MS * 3));
    }
  }

  return allTweets;
}

/**
 * 트윗을 일정으로 저장
 */
async function createScheduleFromTweet(tweet) {
  // source_url로 중복 체크
  const [existing] = await pool.query(
    "SELECT id FROM schedules WHERE source_url = ?",
    [tweet.url]
  );

  if (existing.length > 0) {
    return null; // 이미 존재
  }

  const kstDate = toKST(tweet.time);
  const date = formatDate(kstDate);
  const time = formatTime(kstDate);
  const title = extractTitle(tweet.text);
  const description = tweet.text;

  // 일정 생성
  const [result] = await pool.query(
    `INSERT INTO schedules (title, description, date, time, category_id, source_url, source_name)
     VALUES (?, ?, ?, ?, ?, ?, NULL)`,
    [title, description, date, time, X_CATEGORY_ID, tweet.url]
  );

  const scheduleId = result.insertId;

  // Meilisearch 동기화
  try {
    const [categoryInfo] = await pool.query(
      "SELECT name, color FROM schedule_categories WHERE id = ?",
      [X_CATEGORY_ID]
    );
    await addOrUpdateSchedule({
      id: scheduleId,
      title,
      description,
      date,
      time,
      category_id: X_CATEGORY_ID,
      category_name: categoryInfo[0]?.name || "",
      category_color: categoryInfo[0]?.color || "",
      source_name: null,
      source_url: tweet.url,
      members: [],
    });
  } catch (searchError) {
    console.error("Meilisearch 동기화 오류:", searchError.message);
  }

  return scheduleId;
}

/**
 * 유튜브 영상을 일정으로 저장
 */
async function createScheduleFromYoutube(video) {
  // source_url로 중복 체크
  const [existing] = await pool.query(
    "SELECT id FROM schedules WHERE source_url = ?",
    [video.videoUrl]
  );

  if (existing.length > 0) {
    return null; // 이미 존재
  }

  const kstDate = toKST(video.publishedAt);
  const date = formatDate(kstDate);
  const time = formatTime(kstDate);

  // 일정 생성 (source_name에 채널명 저장)
  const [result] = await pool.query(
    `INSERT INTO schedules (title, date, time, category_id, source_url, source_name)
     VALUES (?, ?, ?, ?, ?, ?)`,
    [
      video.title,
      date,
      time,
      YOUTUBE_CATEGORY_ID,
      video.videoUrl,
      video.channelTitle || null,
    ]
  );

  const scheduleId = result.insertId;

  // Meilisearch 동기화
  try {
    const [categoryInfo] = await pool.query(
      "SELECT name, color FROM schedule_categories WHERE id = ?",
      [YOUTUBE_CATEGORY_ID]
    );
    await addOrUpdateSchedule({
      id: scheduleId,
      title: video.title,
      description: "",
      date,
      time,
      category_id: YOUTUBE_CATEGORY_ID,
      category_name: categoryInfo[0]?.name || "",
      category_color: categoryInfo[0]?.color || "",
      source_name: video.channelTitle || null,
      source_url: video.videoUrl,
      members: [],
    });
  } catch (searchError) {
    console.error("Meilisearch 동기화 오류:", searchError.message);
  }

  return scheduleId;
}

/**
 * 새 트윗 동기화 (첫 페이지만 - 1분 간격 실행용)
 */
export async function syncNewTweets(botId) {
  try {
    // 봇 정보 조회
    const [bots] = await pool.query(
      `SELECT b.*, c.username, c.nitter_url 
       FROM bots b 
       LEFT JOIN bot_x_config c ON b.id = c.bot_id 
       WHERE b.id = ?`,
      [botId]
    );

    if (bots.length === 0) {
      throw new Error("봇을 찾을 수 없습니다.");
    }

    const bot = bots[0];

    if (!bot.username) {
      throw new Error("Username이 설정되지 않았습니다.");
    }

    const nitterUrl = bot.nitter_url || "http://nitter:8080";

    // 관리 중인 채널 목록 조회
    const managedChannelIds = await getManagedChannelIds();

    // Nitter에서 트윗 수집 (첫 페이지만)
    const tweets = await fetchTweetsFromNitter(nitterUrl, bot.username);

    let addedCount = 0;
    let ytAddedCount = 0;

    for (const tweet of tweets) {
      // 트윗 저장
      const scheduleId = await createScheduleFromTweet(tweet);
      if (scheduleId) addedCount++;

      // 유튜브 링크 처리
      const videoIds = extractYoutubeVideoIds(tweet.text);
      for (const videoId of videoIds) {
        const video = await fetchVideoInfo(videoId);
        if (!video) continue;

        // 관리 중인 채널이면 스킵
        if (managedChannelIds.includes(video.channelId)) continue;

        // 유튜브 일정 저장
        const ytScheduleId = await createScheduleFromYoutube(video);
        if (ytScheduleId) ytAddedCount++;
      }
    }

    // 봇 상태 업데이트
    // 추가된 항목이 있을 때만 last_added_count 업데이트 (0이면 이전 값 유지)
    const totalAdded = addedCount + ytAddedCount;
    if (totalAdded > 0) {
      await pool.query(
        `UPDATE bots SET 
          last_check_at = NOW(),
          schedules_added = schedules_added + ?,
          last_added_count = ?,
          error_message = NULL
         WHERE id = ?`,
        [totalAdded, totalAdded, botId]
      );
    } else {
      await pool.query(
        `UPDATE bots SET 
          last_check_at = NOW(),
          error_message = NULL
         WHERE id = ?`,
        [botId]
      );
    }

    return { addedCount, ytAddedCount, total: tweets.length };
  } catch (error) {
    // 오류 상태 업데이트
    await pool.query(
      `UPDATE bots SET 
        last_check_at = NOW(),
        status = 'error',
        error_message = ?
       WHERE id = ?`,
      [error.message, botId]
    );
    throw error;
  }
}

/**
 * 전체 트윗 동기화 (전체 페이지 - 초기화용)
 */
export async function syncAllTweets(botId) {
  try {
    // 봇 정보 조회
    const [bots] = await pool.query(
      `SELECT b.*, c.username, c.nitter_url 
       FROM bots b 
       LEFT JOIN bot_x_config c ON b.id = c.bot_id 
       WHERE b.id = ?`,
      [botId]
    );

    if (bots.length === 0) {
      throw new Error("봇을 찾을 수 없습니다.");
    }

    const bot = bots[0];

    if (!bot.username) {
      throw new Error("Username이 설정되지 않았습니다.");
    }

    const nitterUrl = bot.nitter_url || "http://nitter:8080";

    // 관리 중인 채널 목록 조회
    const managedChannelIds = await getManagedChannelIds();

    // Nitter에서 전체 트윗 수집
    const tweets = await fetchAllTweetsFromNitter(nitterUrl, bot.username);

    let addedCount = 0;
    let ytAddedCount = 0;

    for (const tweet of tweets) {
      // 트윗 저장
      const scheduleId = await createScheduleFromTweet(tweet);
      if (scheduleId) addedCount++;

      // 유튜브 링크 처리
      const videoIds = extractYoutubeVideoIds(tweet.text);
      for (const videoId of videoIds) {
        const video = await fetchVideoInfo(videoId);
        if (!video) continue;

        // 관리 중인 채널이면 스킵
        if (managedChannelIds.includes(video.channelId)) continue;

        // 유튜브 일정 저장
        const ytScheduleId = await createScheduleFromYoutube(video);
        if (ytScheduleId) ytAddedCount++;
      }
    }

    // 봇 상태 업데이트
    // 추가된 항목이 있을 때만 last_added_count 업데이트 (0이면 이전 값 유지)
    const totalAdded = addedCount + ytAddedCount;
    if (totalAdded > 0) {
      await pool.query(
        `UPDATE bots SET 
          last_check_at = NOW(),
          schedules_added = schedules_added + ?,
          last_added_count = ?,
          error_message = NULL
         WHERE id = ?`,
        [totalAdded, totalAdded, botId]
      );
    } else {
      await pool.query(
        `UPDATE bots SET 
          last_check_at = NOW(),
          error_message = NULL
         WHERE id = ?`,
        [botId]
      );
    }

    return { addedCount, ytAddedCount, total: tweets.length };
  } catch (error) {
    await pool.query(
      `UPDATE bots SET 
        status = 'error',
        error_message = ?
       WHERE id = ?`,
      [error.message, botId]
    );
    throw error;
  }
}

export default {
  syncNewTweets,
  syncAllTweets,
  extractTitle,
  extractYoutubeVideoIds,
  toKST,
};