VNEWS

Running

App Files Files Community

bep40 commited on Jun 14

Commit

0de9cde

verified ·

1 Parent(s): f46e9bc

Upload static/collect_shorts.html

Browse files

Files changed (1) hide show

static/collect_shorts.html +196 -0

static/collect_shorts.html ADDED Viewed

	@@ -0,0 +1,196 @@

+<!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>