VNEWS

Running

App Files Files Community

bep40 commited on Jun 14

Commit

c7da4c9

verified ·

1 Parent(s): c2a95f6

Delete static/collect_shorts.html, static/shorts_collector.html, static/test_yt.html

Browse files

Files changed (3) hide show

static/collect_shorts.html +0 -196
static/shorts_collector.html +0 -337
static/test_yt.html +0 -50

static/collect_shorts.html DELETED Viewed

@@ -1,196 +0,0 @@
-<!DOCTYPE html>
-<html lang="vi">
-<head>
-<meta charset="utf-8">
-<meta name="viewport" content="width=device-width,initial-scale=1">
-<title>Collect VTV Nam Bộ Shorts</title>
-<style>
-body{background:#111;color:#eee;font-family:monospace;padding:20px;max-width:900px;margin:0 auto}
-h1{color:#f0c040;font-size:18px;margin-bottom:10px}
-input,button,textarea{font-family:monospace;font-size:13px}
-input{width:100%;padding:8px;background:#222;border:1px solid #444;color:#eee;border-radius:6px;margin:6px 0}
-button{background:#0b6bcb;border:none;color:#fff;padding:10px 20px;border-radius:6px;cursor:pointer;margin:4px}
-button:hover{background:#1a7be0}
-textarea{width:100%;height:300px;background:#222;border:1px solid #444;color:#5cb87a;border-radius:6px;padding:8px;margin:6px 0;resize:vertical}
-.status{padding:10px;background:#1a1a1a;border-radius:6px;margin:10px 0;white-space:pre-wrap;font-size:12px}
-.short-item{padding:6px 10px;background:#1a1a1a;border-radius:4px;margin:4px 0;display:flex;gap:8px;align-items:center}
-.short-item img{width:60px;height:106px;object-fit:cover;border-radius:4px;background:#333}
-.short-item .info{flex:1}
-.short-item .title{font-size:12px;color:#ccc}
-.short-item .id{font-size:10px;color:#777}
-</style>
-</head>
-<body>
-<h1>📱 Collect YouTube Shorts from @vtvnambo</h1>
-<p style="color:#888;font-size:12px">Tool này sẽ fetch trang YouTube của VTV Nam Bộ và extract tất cả shorts. Chạy trên browser của bạn (không qua server).</p>
-<input type="text" id="proxyUrl" placeholder="CORS proxy (leave empty to try direct)" value="">
-<button onclick="collectShorts()">🚀 Collect Shorts</button>
-<button onclick="collectFromVideos()">📹 Collect from /videos (duration filter)</button>
-<button onclick="exportJSON()">📋 Export JSON</button>
-<button onclick="exportJS()">📄 Export JS Array</button>
-<div class="status" id="status">Ready...</div>
-<div id="results"></div>
-<script>
-let collectedShorts = [];
-async function fetchPage(url, useProxy = false) {
-    const el = document.getElementById('status');
-    el.textContent = `Fetching: ${url}...`;
-    let fetchUrl = url;
-    const proxy = document.getElementById('proxyUrl').value.trim();
-    if (proxy) {
-        fetchUrl = proxy + encodeURIComponent(url);
-    }
-    try {
-        const r = await fetch(fetchUrl, {
-            headers: {
-                'Accept': 'text/html,*/*'
-            },
-            signal: AbortSignal.timeout(15000)
-        });
-        if (!r.ok) throw new Error(`HTTP ${r.status}`);
-        return await r.text();
-    } catch(e) {
-        el.textContent += `\nError: ${e.message}`;
-        throw e;
-    }
-}
-function extractShortsFromHTML(html) {
-    const shorts = [];
-    const seen = new Set();
-    // Pattern 1: videoId in JSON data
-    const patterns = [
-        /"videoId":"([A-Za-z0-9_-]{11})"/g,
-        /"videoId"\s*:\s*"([A-Za-z0-9_-]{11})"/g,
-        /watch\?v=([A-Za-z0-9_-]{11})/g,
-        /\/shorts\/([A-Za-z0-9_-]{11})/g
-    ];
-    for (const pattern of patterns) {
-        let m;
-        while ((m = pattern.exec(html)) !== null) {
-            const vid = m[1];
-            if (seen.has(vid)) continue;
-            seen.add(vid);
-            // Try to find title near this video ID
-            const start = Math.max(0, m.index - 1500);
-            const end = Math.min(html.length, m.index + 1500);
-            const snippet = html.substring(start, end);
-            let title = "VTV Nam Bộ Short";
-            const titleMatch = snippet.match(/"title"\s*:\s*\{"runs"\s*:\s*\[\{"text"\s*:\s*"([^"]+)"/);
-            if (titleMatch) {
-                title = titleMatch[1].replace(/\\n/g, ' ').replace(/\\"/g, '"').trim();
-            } else {
-                const titleMatch2 = snippet.match(/"title"\s*:\s*"([^"]+)"/);
-                if (titleMatch2) {
-                    title = titleMatch2[1].replace(/\\n/g, ' ').trim();
-                } else {
-                    const titleMatch3 = snippet.match(/"accessibilityText"\s*:\s*"([^"]+)"/);
-                    if (titleMatch3) {
-                        title = titleMatch3[1].replace(/\\n/g, ' ').trim();
-                    }
-                }
-            }
-            // Check if page is shorts page
-            const isShortsPage = html.includes('/shorts/') || html.includes('"shorts"');
-            const titleLower = title.toLowerCase();
-            const isShort = isShortsPage || titleLower.includes('#shorts') || titleLower.includes('#short');
-            shorts.push({
-                id: vid,
-                title: title,
-                channel: 'vtvnambo',
-                img: `https://i.ytimg.com/vi/${vid}/hqdefault.jpg`,
-                isShort: isShort
-            });
-        }
-    }
-    return shorts;
-}
-async function collectShorts() {
-    const el = document.getElementById('status');
-    collectedShorts = [];
-    try {
-        // Try shorts page
-        const html = await fetchPage('https://www.youtube.com/@vtvnambo/shorts');
-        const shorts = extractShortsFromHTML(html);
-        collectedShorts = shorts;
-        el.textContent = `Found ${shorts.length} videos from /shorts page`;
-        displayResults(shorts);
-    } catch(e) {
-        el.textContent += `\nFailed. Try using a CORS proxy like https://corsproxy.io/?`;
-    }
-}
-async function collectFromVideos() {
-    const el = document.getElementById('status');
-    collectedShorts = [];
-    try {
-        const html = await fetchPage('https://www.youtube.com/@vtvnambo/videos');
-        const all = extractShortsFromHTML(html);
-        // First page might have limited results, try to find continuation
-        collectedShorts = all;
-        el.textContent = `Found ${all.length} videos from /videos page`;
-        displayResults(all);
-    } catch(e) {
-        el.textContent += `\nFailed. Try using a CORS proxy.`;
-    }
-}
-function displayResults(shorts) {
-    const el = document.getElementById('results');
-    let h = `<h3 style="color:#5cb87a;margin:10px 0">${shorts.length} shorts found:</h3>`;
-    shorts.forEach(s => {
-        h += `<div class="short-item">
-            <img src="${s.img}" onerror="this.style.background='#333'">
-            <div class="info">
-                <div class="title">${s.isShort ? '📱' : '🎬'} ${s.title}</div>
-                <div class="id">${s.id}</div>
-            </div>
-        </div>`;
-    });
-    el.innerHTML = h;
-}
-function exportJSON() {
-    if (!collectedShorts.length) return alert('Chưa collect được shorts nào!');
-    const json = JSON.stringify(collectedShorts, null, 2);
-    document.getElementById('status').textContent = json;
-    navigator.clipboard.writeText(json).then(() => alert('Đã copy JSON!'));
-}
-function exportJS() {
-    if (!collectedShorts.length) return alert('Chưa collect được shorts nào!');
-    let js = 'const VTV_SHORTS = [\n';
-    collectedShorts.forEach(s => {
-        js += `    {"id":"${s.id}","title":${JSON.stringify(s.title)},"channel":"vtvnambo"},\n`;
-    });
-    js += '];';
-    document.getElementById('status').textContent = js;
-    navigator.clipboard.writeText(js).then(() => alert('Đã copy JS!'));
-}
-// Auto-collect on load
-setTimeout(collectShorts, 500);
-</script>
-</body>
-</html>

static/shorts_collector.html DELETED Viewed

@@ -1,337 +0,0 @@
-<!DOCTYPE html>
-<html lang="vi">
-<head>
-<meta charset="utf-8">
-<meta name="viewport" content="width=device-width,initial-scale=1">
-<title>VTV Nam Bộ Shorts Collector</title>
-<style>
-*{box-sizing:border-box;margin:0;padding:0}
-body{background:#0a0a0a;color:#eee;font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Roboto,sans-serif;padding:16px;max-width:800px;margin:0 auto}
-h1{color:#f0c040;font-size:20px;margin-bottom:8px}
-h2{color:#5cb87a;font-size:14px;margin:16px 0 8px}
-p{color:#888;font-size:12px;margin-bottom:12px}
-button{background:#0b6bcb;border:none;color:#fff;padding:10px 20px;border-radius:8px;font-size:13px;cursor:pointer;margin:4px}
-button:hover{background:#1a7be0}
-button:disabled{background:#333;color:#666;cursor:not-allowed}
-button.danger{background:#c0392b}
-textarea{width:100%;height:200px;background:#1a1a1a;border:1px solid #333;color:#5cb87a;border-radius:8px;padding:10px;font-size:11px;font-family:monospace;resize:vertical}
-.status{background:#1a1a1a;border:1px solid #333;border-radius:8px;padding:12px;margin:10px 0;font-size:12px;white-space:pre-wrap;max-height:300px;overflow-y:auto}
-.short-item{display:flex;gap:10px;padding:8px;background:#1a1a1a;border-radius:6px;margin:4px 0;align-items:center}
-.short-item img{width:80px;height:142px;object-fit:cover;border-radius:4px;background:#333;flex-shrink:0}
-.short-item .info{flex:1;min-width:0}
-.short-item .title{font-size:12px;color:#ccc;word-break:break-word}
-.short-item .id{font-size:10px;color:#666;margin-top:4px}
-.short-item .copy-btn{flex:0 0 auto;background:#333;border:none;color:#ccc;padding:6px 10px;border-radius:4px;font-size:10px;cursor:pointer}
-.short-item .copy-btn:hover{background:#444}
-.grid{display:grid;grid-template-columns:repeat(auto-fill,minmax(180px,1fr));gap:8px;margin:10px 0}
-.grid .card{background:#1a1a1a;border-radius:8px;overflow:hidden;cursor:pointer;border:2px solid transparent;transition:border-color .2s}
-.grid .card:hover{border-color:#0b6bcb}
-.grid .card.selected{border-color:#5cb87a}
-.grid .card img{width:100%;aspect-ratio:9/16;object-fit:cover;background:#333;display:block}
-.grid .card .title{font-size:10px;color:#ccc;padding:4px 6px;line-height:1.3;display:-webkit-box;-webkit-line-clamp:2;-webkit-box-orient:vertical;overflow:hidden}
-.grid .card .id{font-size:9px;color:#666;padding:0 6px 4px}
-.actions{display:flex;gap:8px;flex-wrap:wrap;margin:10px 0}
-</style>
-</head>
-<body>
-<h1>📱 VTV Nam Bộ Shorts Collector</h1>
-<p>Tool này chạy hoàn toàn trên browser của bạn. Nó sẽ fetch trang YouTube của @vtvnambo và extract tất cả shorts IDs.</p>
-<h2>Bước 1: Chọn phương pháp</h2>
-<div class="actions">
-<button onclick="methodDirect()">🌐 Method 1: Fetch trực tiếp</button>
-<button onclick="methodProxy()">🔄 Method 2: Qua CORS proxy</button>
-<button onclick="methodManual()">✏️ Method 3: Dán HTML thủ công</button>
-</div>
-<h2>Bước 2: Kết quả</h2>
-<div class="status" id="status">Nhấn một nút ở trên để bắt đầu...</div>
-<div class="actions" id="resultActions" style="display:none">
-<button onclick="exportJSON()">📋 Copy JSON</button>
-<button onclick="exportJSArray()">📄 Copy JS Array</button>
-<button onclick="exportPythonList()">🐍 Copy Python List</button>
-<button onclick="testAllVideos()">🧪 Test tất cả videos</button>
-<button class="danger" onclick="clearAll()">🗑 Xóa tất cả</button>
-</div>
-<div id="results"></div>
-<h2>Output (copy cái này vào code)</h2>
-<textarea id="output" placeholder="Kết quả sẽ hiện ở đây..."></textarea>
-<script>
-let allShorts = [];
-function setStatus(msg) {
-    document.getElementById('status').textContent = msg;
-    console.log(msg);
-}
-function addShorts(newShorts) {
-    const seen = new Set(allShorts.map(s => s.id));
-    let added = 0;
-    for (const s of newShorts) {
-        if (!seen.has(s.id)) {
-            seen.add(s.id);
-            allShorts.push(s);
-            added++;
-        }
-    }
-    if (added > 0) {
-        setStatus(`✅ Thêm ${added} shorts mới. Tổng: ${allShorts.length} shorts.`);
-        displayResults();
-        updateOutput();
-    }
-    return added;
-}
-function extractFromHTML(html, source) {
-    const shorts = [];
-    const seen = new Set();
-    // Pattern 1: /shorts/VIDEO_ID
-    let m;
-    const shortsPattern = /\/shorts\/([A-Za-z0-9_-]{11})/g;
-    while ((m = shortsPattern.exec(html)) !== null) {
-        const vid = m[1];
-        if (seen.has(vid)) continue;
-        seen.add(vid);
-        // Find title near this match
-        const start = Math.max(0, m.index - 2000);
-        const end = Math.min(html.length, m.index + 2000);
-        const snippet = html.substring(start, end);
-        let title = "VTV Nam Bộ Short";
-        const tm = snippet.match(/"title"\s*:\s*\{"runs"\s*:\s*\[\{"text"\s*:\s*"([^"]+)"/);
-        if (tm) title = tm[1].replace(/\\n/g,' ').replace(/\\"/g,'"').trim();
-        shorts.push({ id: vid, title, channel: 'vtvnambo', source });
-    }
-    // Pattern 2: "videoId":"XXXX" near shorts context
-    const vidPattern = /"videoId"\s*:\s*"([A-Za-z0-9_-]{11})"/g;
-    while ((m = vidPattern.exec(html)) !== null) {
-        const vid = m[1];
-        if (seen.has(vid)) continue;
-        // Check if this is in a shorts context
-        const start = Math.max(0, m.index - 500);
-        const end = Math.min(html.length, m.index + 500);
-        const ctx = html.substring(start, end);
-        if (ctx.includes('shorts') || ctx.includes('#shorts') || ctx.includes('#short')) {
-            seen.add(vid);
-            let title = "VTV Nam Bộ Short";
-            const tm = ctx.match(/"title"\s*:\s*\{"runs"\s*:\s*\[\{"text"\s*:\s*"([^"]+)"/);
-            if (tm) title = tm[1].replace(/\\n/g,' ').trim();
-            shorts.push({ id: vid, title, channel: 'vtvnambo', source });
-        }
-    }
-    return shorts;
-}
-// ===== Method 1: Direct fetch =====
-async function methodDirect() {
-    setStatus('🔄 Đang fetch trực tiếp từ YouTube...');
-    const urls = [
-        'https://www.youtube.com/@vtvnambo/shorts',
-        'https://www.youtube.com/@vtvnambo/videos',
-    ];
-    for (const url of urls) {
-        try {
-            setStatus(`🔄 Fetching: ${url}...`);
-            const r = await fetch(url, {
-                headers: { 'Accept': 'text/html' },
-                signal: AbortSignal.timeout(15000)
-            });
-            if (r.ok) {
-                const html = await r.text();
-                const shorts = extractFromHTML(html, url);
-                setStatus(`✅ Từ ${url}: tìm được ${shorts.length} shorts`);
-                addShorts(shorts);
-            } else {
-                setStatus(`❌ HTTP ${r.status} từ ${url}`);
-            }
-        } catch(e) {
-            setStatus(`❌ Lỗi fetch ${url}: ${e.message}`);
-        }
-    }
-    if (allShorts.length === 0) {
-        setStatus('❌ Fetch trực tiếp thất bại (CORS blocked). Thử Method 2 hoặc 3.');
-    }
-}
-// ===== Method 2: CORS proxy =====
-async function methodProxy() {
-    setStatus('🔄 Đang fetch qua CORS proxy...');
-    const proxies = [
-        'https://corsproxy.io/?',
-        'https://api.allorigins.win/raw?url=',
-        'https://cors-anywhere.herokuapp.com/',
-    ];
-    const targetUrl = 'https://www.youtube.com/@vtvnambo/shorts';
-    for (const proxy of proxies) {
-        try {
-            setStatus(`🔄 Thử proxy: ${proxy.substring(0, 40)}...`);
-            const fetchUrl = proxy + encodeURIComponent(targetUrl);
-            const r = await fetch(fetchUrl, { signal: AbortSignal.timeout(20000) });
-            if (r.ok) {
-                const html = await r.text();
-                const shorts = extractFromHTML(html, 'proxy');
-                setStatus(`✅ Qua proxy: tìm được ${shorts.length} shorts`);
-                addShorts(shorts);
-                if (shorts.length > 0) return; // Success!
-            } else {
-                setStatus(`❌ Proxy trả về HTTP ${r.status}`);
-            }
-        } catch(e) {
-            setStatus(`❌ Proxy lỗi: ${e.message.substring(0, 80)}`);
-        }
-    }
-    // Also try /videos page
-    try {
-        const proxy = proxies[0];
-        const targetUrl2 = 'https://www.youtube.com/@vtvnambo/videos';
-        setStatus(`🔄 Thử fetch /videos page...`);
-        const r = await fetch(proxy + encodeURIComponent(targetUrl2), { signal: AbortSignal.timeout(20000) });
-        if (r.ok) {
-            const html = await r.text();
-            const shorts = extractFromHTML(html, 'proxy-videos');
-            setStatus(`✅ Từ /videos: tìm được ${shorts.length} shorts`);
-            addShorts(shorts);
-        }
-    } catch(e) {
-        setStatus(`❌ /videos cũng lỗi: ${e.message.substring(0, 80)}`);
-    }
-    if (allShorts.length === 0) {
-        setStatus('❌ Tất cả proxy đều thất bại. Dùng Method 3: Dán HTML thủ công.');
-    }
-}
-// ===== Method 3: Manual paste =====
-async function methodManual() {
-    setStatus('📋 Hướng dẫn:\n1. Mở https://www.youtube.com/@vtvnambo/shorts trên trình duyệt\n2. Chuột phải → "View Page Source" (hoặc Ctrl+U)\n3. Copy toàn bộ HTML\n4. Dán vào ô bên dưới\n5. Nhấn "Extract"');
-    const html = prompt('Dán HTML source của trang YouTube @vtvnambo/shorts vào đây:');
-    if (!html) return;
-    setStatus('🔄 Đang extract từ HTML...');
-    const shorts = extractFromHTML(html, 'manual');
-    setStatus(`✅ Extract được ${shorts.length} shorts từ HTML`);
-    addShorts(shorts);
-}
-// ===== Display results =====
-function displayResults() {
-    const el = document.getElementById('results');
-    document.getElementById('resultActions').style.display = 'flex';
-    let h = `<h2>Tìm được ${allShorts.length} shorts:</h2><div class="grid">`;
-    allShorts.forEach((s, i) => {
-        h += `<div class="card" onclick="toggleSelect(${i})" id="card-${i}">
-            <img src="https://i.ytimg.com/vi/${s.id}/hqdefault.jpg" onerror="this.src='data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 width=%2280%22 height=%22142%22><rect fill=%22%23333%22 width=%2280%22 height=%22142%22/><text fill=%22%23666%22 x=%2250%22 y=%2276%22 text-anchor=%22middle%22 font-size=%2210%22>📱</text></svg>'">
-            <div class="title">${s.title}</div>
-            <div class="id">${s.id}</div>
-        </div>`;
-    });
-    h += '</div>';
-    el.innerHTML = h;
-}
-function toggleSelect(i) {
-    const card = document.getElementById('card-' + i);
-    card.classList.toggle('selected');
-}
-// ===== Export =====
-function updateOutput() {
-    document.getElementById('output').value = exportJSONString();
-}
-function exportJSONString() {
-    return JSON.stringify(allShorts.map(s => ({
-        id: s.id,
-        title: s.title,
-        channel: 'vtvnambo'
-    })), null, 2);
-}
-function exportJSON() {
-    const json = exportJSONString();
-    navigator.clipboard.writeText(json).then(() => {
-        setStatus('✅ Đã copy JSON vào clipboard!\n\n' + json.substring(0, 200) + '...');
-    });
-}
-function exportJSArray() {
-    let js = 'const VTV_SHORTS = [\n';
-    allShorts.forEach(s => {
-        js += `    {"id":"${s.id}","title":${JSON.stringify(s.title)},"channel":"vtvnambo"},\n`;
-    });
-    js += '];';
-    navigator.clipboard.writeText(js).then(() => setStatus('✅ Đã copy JS Array!'));
-}
-function exportPythonList() {
-    let py = 'VTV_SHORTS = [\n';
-    allShorts.forEach(s => {
-        py += `    {"id": "${s.id}", "title": """${s.title.replace(/"""/g, '\\"\\"\\"')}""", "channel": "vtvnambo"},\n`;
-    });
-    py += ']';
-    navigator.clipboard.writeText(py).then(() => setStatus('✅ Đã copy Python list!'));
-}
-// ===== Test videos =====
-async function testAllVideos() {
-    setStatus('🧪 Đang test tất cả videos...');
-    let working = 0, broken = 0;
-    for (const s of allShorts) {
-        try {
-            const r = await fetch(`https://i.ytimg.com/vi/${s.id}/hqdefault.jpg`, {
-                method: 'HEAD',
-                signal: AbortSignal.timeout(5000)
-            });
-            if (r.ok) {
-                working++;
-            } else {
-                broken++;
-                setStatus(`❌ Broken: ${s.id}`);
-            }
-        } catch(e) {
-            broken++;
-        }
-    }
-    setStatus(`✅ Test xong: ${working} working, ${broken} broken`);
-}
-function clearAll() {
-    allShorts = [];
-    document.getElementById('results').innerHTML = '';
-    document.getElementById('resultActions').style.display = 'none';
-    document.getElementById('output').value = '';
-    setStatus('Đã xóa tất cả. Nhấn nút để collect lại.');
-}
-// Auto-try on load
-setTimeout(() => {
-    setStatus('🚀 Tự động thử fetch...\nNhấn "Method 1" hoặc "Method 2" để bắt đầu.');
-}, 500);
-</script>
-</body>
-</html>

static/test_yt.html DELETED Viewed

@@ -1,50 +0,0 @@
-<!DOCTYPE html>
-<html lang="vi">
-<head>
-<meta charset="utf-8">
-<title>Test yt-dlp</title>
-<style>
-body{background:#111;color:#eee;font-family:monospace;padding:20px;max-width:800px;margin:0 auto}
-button{background:#0b6bcb;border:none;color:#fff;padding:10px 20px;border-radius:6px;cursor:pointer;margin:4px}
-pre{background:#1a1a1a;padding:10px;border-radius:6px;overflow:auto;max-height:400px;font-size:11px}
-</style>
-</head>
-<body>
-<h2>🧪 Test yt-dlp Scraper</h2>
-<button onclick="testAPI()">Test /api/shorts/vtvnamo</button>
-<button onclick="testWC()">Test /api/shorts/wc</button>
-<pre id="output">Click a button to test...</pre>
-<script>
-async function testAPI() {
-    const el = document.getElementById('output');
-    el.textContent = 'Loading /api/shorts/vtvnamo...';
-    try {
-        const r = await fetch('/api/shorts/vtvnamo?count=50');
-        const data = await r.json();
-        el.textContent = `Found ${data.length} shorts:\n\n`;
-        data.forEach(s => {
-            el.textContent += `${s.id} | ${s.title}\n`;
-        });
-    } catch(e) {
-        el.textContent = 'Error: ' + e.message;
-    }
-}
-async function testWC() {
-    const el = document.getElementById('output');
-    el.textContent = 'Loading /api/shorts/wc...';
-    try {
-        const r = await fetch('/api/shorts/wc?count=50');
-        const data = await r.json();
-        el.textContent = `Found ${data.length} WC shorts:\n\n`;
-        data.forEach(s => {
-            el.textContent += `${s.id} | ${s.title}\n`;
-        });
-    } catch(e) {
-        el.textContent = 'Error: ' + e.message;
-    }
-}
-</script>
-</body>
-</html>