feat: /m mirror — cache HTML + assets locally, rare upstream refresh (30min)

server.js

@@ -224,6 +224,130 @@ const CHAT_OVERLAY = `
 </script>
 `;
 
+// ══════════════════════════════════════════════════════════════
+//   MIRROR  —  one-time asset mirror + in-memory HTML cache
+//   Goal: cache everything locally, refresh HTML rarely, never
+//   hammer sechenov. Steady-state: ~1 upstream request / 30 min.
+// ══════════════════════════════════════════════════════════════
+
+const MIRROR_DIR = path.join(PUBLIC_DIR, 'assets', 'mirror');
+const MIRROR_TTL_MS   = 30 * 60 * 1000;
+const MIRROR_MIN_GAP  = 60 * 1000;
+const UPSTREAM_DELAY  = 500;
+const BROWSER_UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36';
+
+let cachedMirrorHtml = null;
+let mirrorInProgress = false;
+let lastMirrorAt = 0;
+let mirrorSessionCookies = '';
+
+const sleep = ms => new Promise(r => setTimeout(r, ms));
+
+async function safeFetch(url, extraHeaders = {}) {
+    const h = {
+        'User-Agent': BROWSER_UA,
+        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8',
+        'Accept-Language': 'ru,en;q=0.9',
+        'Referer': UPSTREAM + '/',
+        ...extraHeaders,
+    };
+    if (mirrorSessionCookies) h['Cookie'] = mirrorSessionCookies;
+    return fetch(url, { dispatcher: directAgent, headers: h });
+}
+
+function extractAssetUrls(html) {
+    const urls = new Set();
+    const patterns = [
+        /(?:href|src)="(\/[^"\s]+)"/g,
+        /url\(["']?(\/[^"'\s\)]+)/g,
+    ];
+    for (const p of patterns) {
+        let m;
+        while ((m = p.exec(html)) !== null) {
+            const u = m[1];
+            if (u.startsWith('//')) continue;
+            if (u.startsWith('/api/') || u.startsWith('/assets/') || u.startsWith('/proxy/')) continue;
+            urls.add(u);
+        }
+    }
+    return [...urls];
+}
+
+async function mirrorAsset(urlPath) {
+    const pure = urlPath.split(/[?#]/)[0];
+    const local = path.join(MIRROR_DIR, pure);
+    try { await fs.promises.stat(local); return false; } catch {}
+    const r = await safeFetch(UPSTREAM + urlPath);
+    if (!r.ok) { console.warn(`[mirror] ${r.status} ${urlPath}`); return false; }
+    const ct = r.headers.get('content-type') || '';
+    let buf;
+    if (ct.includes('text/css')) {
+        let css = await r.text();
+        css = css.replace(/url\(\s*(["']?)\/(?!\/)/g, 'url($1/assets/mirror/');
+        buf = Buffer.from(css, 'utf8');
+    } else {
+        buf = Buffer.from(await r.arrayBuffer());
+    }
+    await fs.promises.mkdir(path.dirname(local), { recursive: true });
+    await fs.promises.writeFile(local, buf);
+    console.log(`[mirror] + ${pure} (${buf.length}b)`);
+    await sleep(UPSTREAM_DELAY);
+    return true;
+}
+
+async function refreshMirror() {
+    if (mirrorInProgress) return;
+    if (Date.now() - lastMirrorAt < MIRROR_MIN_GAP && cachedMirrorHtml) return;
+    mirrorInProgress = true;
+    lastMirrorAt = Date.now();
+    try {
+        console.log('[mirror] fetching HTML…');
+        const r = await safeFetch(UPSTREAM + '/auth.php');
+        if (!r.ok) throw new Error('upstream ' + r.status);
+
+        const setCookies = typeof r.headers.getSetCookie === 'function' ? r.headers.getSetCookie() : [];
+        if (setCookies.length) {
+            mirrorSessionCookies = setCookies.map(c => c.split(';')[0]).join('; ');
+        }
+
+        let html = await r.text();
+        const assetUrls = extractAssetUrls(html);
+        console.log(`[mirror] HTML ${html.length}b, ${assetUrls.length} assets referenced`);
+
+        let downloaded = 0;
+        for (const u of assetUrls) {
+            try { if (await mirrorAsset(u)) downloaded++; } catch (e) { console.warn('[mirror]', u, e.message); }
+        }
+        console.log(`[mirror] downloaded ${downloaded} new assets (cached: ${assetUrls.length - downloaded})`);
+
+        html = html
+            .replace(/(href|src)="\/(?!\/)/g, '$1="/assets/mirror/')
+            .replace(/url\(["']?\/(?!\/)/g, 'url(/assets/mirror/')
+            .replace(/https?:\/\/student\.sechenov\.ru/gi, '')
+            .replace(/<meta[^>]+http-equiv=["']?(Content-Security-Policy|X-Frame-Options)["']?[^>]*>/gi, '')
+            .replace(/\btarget=["'](_top|_parent|_blank)["']/gi, '');
+        html = html.replace(/<\/body>/i, CHAT_OVERLAY + '</body>');
+
+        cachedMirrorHtml = html;
+        console.log('[mirror] ready');
+    } catch (e) {
+        console.error('[mirror fail]', e.message);
+    } finally {
+        mirrorInProgress = false;
+    }
+}
+
+setTimeout(() => refreshMirror(), 3000);
+setInterval(() => refreshMirror(), MIRROR_TTL_MS);
+
+app.get(['/m', '/m.html', '/mirror'], (req, res) => {
+    if (!cachedMirrorHtml) return res.status(503).send('mirror initializing — try again in 60s');
+    res.setHeader('Content-Type', 'text/html; charset=utf-8');
+    res.send(cachedMirrorHtml);
+});
+
+app.use('/assets/mirror', express.static(MIRROR_DIR));
+
 // Reverse-proxy disabled: Sechenov banned us by IP. Left here for reference.
 app.use('/_disabled_reverse_proxy', async (req, res) => {
     const upstreamUrl = UPSTREAM + req.url;