Source-Web-23070-DataAnalysis/v2/scrapper/scrapper.js

import puppeteer from 'puppeteer';

/**
 * 極限深度查找 (Ultimate Meta Relay Lookup)
 * 專門針對 ScheduledServerJS -> __bbox -> RelayPrefetchedStreamCache 的參數結構
 */
function nestedLookup(obj, keys = ["thread_items", "replies", "child_posts", "threaded_replies", "post", "posts", "edges", "node", "reply_threads"]) {
    let results = [];
    if (typeof obj !== "object" || obj === null) return results;

    if (Array.isArray(obj)) {
        // 關鍵修正：檢查這是否為 RelayPrefetchedStreamCache 的呼叫陣列
        // 格式通常是 ["RelayPrefetchedStreamCache", "next", null, ["key", {數據物件}]]
        if (obj[0] === "RelayPrefetchedStreamCache" && Array.isArray(obj[3])) {
            const relayData = obj[3][1]; // 取得包含數據的第二個參數
            if (relayData) results = results.concat(nestedLookup(relayData, keys));
        }
        
        for (const item of obj) {
            results = results.concat(nestedLookup(item, keys));
        }
    } else {
        // 處理 __bbox 封裝結構
        if (obj.__bbox?.require) {
            results = results.concat(nestedLookup(obj.__bbox.require, keys));
        }
        if (obj.__bbox?.define) {
            results = results.concat(nestedLookup(obj.__bbox.define, keys));
        }

        for (const k in obj) {
            // 匹配目標鍵值
            if (keys.includes(k) && obj[k] !== null) {
                // 如果是列表結構，進一步遞迴
                if ((k === "edges" || k === "thread_items" || k === "reply_threads") && Array.isArray(obj[k])) {
                    results = results.concat(nestedLookup(obj[k], keys));
                } else {
                    results.push(obj[k]);
                }
            }

            // 繼續深度遍歷
            if (typeof obj[k] === "object" && obj[k] !== null) {
                results = results.concat(nestedLookup(obj[k], keys));
            }
        }
    }
    return results;
}

/**
 * 解析單一貼文資料 (支援更多 GraphQL 變體路徑)
 */
function parseThread(data) {
    if (!data) return null;
    
    // 遍歷所有可能的 Post 容器路徑
    let post = data.post || data.node?.post || data.thread_items?.[0]?.post || (data.posts && data.posts[0]) || (data.caption ? data : null);
    
    // 處理 GraphQL result.data 結構
    if (!post && data.result?.data) {
        const d = data.result.data;
        post = d.text_post_app_info || d.post || d;
    }

    if (!post || (!post.id && !post.pk)) return null;

    const mediaType = post.media_type;
    let images = [];
    let videos = [];
    let videoThumbnail = null;

    if (mediaType === 2 && post.video_versions?.length > 0) {
        const highestResVideo = post.video_versions.reduce((max, cur) => (cur.width > max.width ? cur : max), post.video_versions[0]);
        if (highestResVideo?.url) videos.push(highestResVideo.url);
        if (post.image_versions2?.candidates?.length > 0) videoThumbnail = post.image_versions2.candidates[0].url;
    } 
    else if (mediaType === 1 && post.image_versions2?.candidates?.length > 0) {
        const highestResImage = post.image_versions2.candidates.reduce((max, cur) => (cur.width > max.width ? cur : max), post.image_versions2.candidates[0]);
        if (highestResImage?.url) images.push(highestResImage.url);
    } 
    else if (mediaType === 8 && post.carousel_media?.length > 0) {
        post.carousel_media.forEach(m => {
            if (m.media_type === 1 && m.image_versions2?.candidates?.length > 0) images.push(m.image_versions2.candidates[0].url);
            if (m.media_type === 2 && m.video_versions?.length > 0) videos.push(m.video_versions[0].url);
        });
    }

    const result = {
        text: post.caption?.text || post.text_post_app_info?.share_info?.quoted_post?.caption?.text || "",
        published_on: post.taken_at,
        id: post.id || post.pk,
        code: post.code,
        username: post.user?.username,
        // user_pic: post.user?.profile_pic_url,
        like_count: post.like_count || 0,
        reply_count: post.direct_reply_count || post.reply_count || 0,
        // images,
        // videos,
        // video_thumbnail: videoThumbnail,
        // url: post.user?.username && post.code ? `https://www.threads.net/@${post.user.username}/post/${post.code}` : null
    };

    return result.id ? result : null;
}

export async function getThread(postUrl) {
    if (!postUrl?.includes("threads.")) throw new Error("無效的 Threads 網址");

    const postCodeFromUrl = postUrl.split("/post/")[1]?.split("/")[0];
    console.log(`[目標鎖定] 貼文代碼: ${postCodeFromUrl}`);

    const browser = await puppeteer.launch({
        headless: "new",
        args: ['--disable-blink-features=AutomationControlled', '--no-sandbox', '--disable-setuid-sandbox']
    });

    try {
        const page = await browser.newPage();
        await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36');
        
        // 增加等待時間確保網路請求啟動
        await page.goto(postUrl, { waitUntil: 'networkidle2', timeout: 30000 });

        // 滾動是必須的，因為 Relay 串流需要滾動來觸發 JS 解析
        await page.evaluate(async () => {
            await new Promise((resolve) => {
                let totalHeight = 0;
                let distance = 500;
                let timer = setInterval(() => {
                    let scrollHeight = document.body.scrollHeight;
                    window.scrollBy(0, distance);
                    totalHeight += distance;
                    if (totalHeight >= scrollHeight || totalHeight > 5000) {
                        clearInterval(timer);
                        resolve();
                    }
                }, 100);
            });
        });

        // 增加等待時間，讓 Streaming 區塊載入完畢
        await new Promise(r => setTimeout(r, 5000));

        const allScripts = await page.$$eval(
            'script[type="application/json"]',
            (scripts) => scripts.map((s) => s.textContent)
        );

        let allParsedItems = new Map();

        allScripts.forEach((content) => {
            if (!content || !content.includes("ScheduledServerJS")) return;
            try {
                const data = JSON.parse(content);
                const rawItems = nestedLookup(data).flat();

                rawItems.forEach(item => {
                    if (!item) return;
                    const parsed = parseThread(item);
                    if (parsed && parsed.id) {
                        allParsedItems.set(parsed.id, parsed);
                    }
                });
            } catch (e) {}
        });

        const itemsArray = Array.from(allParsedItems.values());
        const mainThread = itemsArray.find(t => t.code === postCodeFromUrl);

        if (!mainThread) {
            throw new Error("無法定位主貼文。這通常是因為 Auth 阻擋或頁面未完全渲染。");
        }

        const authorName = mainThread.username;
        const replies = itemsArray.filter(t => 
            t.code !== postCodeFromUrl
        ).sort((a, b) => a.published_on - b.published_on);

        console.log(`[解析成功] 找到主貼文，作者: ${authorName}，回覆數量: ${replies.length}`);

        return { thread: mainThread, replies };

    } finally {
        await browser.close();
    }
}