Spaces:

vanseleem
/

subf2m

Running

subf2m / Original
Update Original
e74a3f8 verified 5 months ago
35.5 kB
'use strict';
const express = require('express');
const cheerio = require('cheerio');
const AdmZip  = require('adm-zip');
const https   = require('https');
const http    = require('http');
const zlib    = require('zlib');
const { URL } = require('url');
 
const app  = express();
const PORT = process.env.PORT || 7860;
 
app.use((req, res, next) => {
  res.header('Access-Control-Allow-Origin',  '*');
  res.header('Access-Control-Allow-Headers', '*');
  res.header('Access-Control-Allow-Methods', 'GET, OPTIONS');
  if (req.method === 'OPTIONS') return res.sendStatus(200);
  next();
});
 
// ================================================================
//  PROXY CONFIG  ← edit these three lines to swap proxies
// ================================================================
const PROXY_1 = 'https://corsproxy.io/?url=';
const PROXY_2 = 'https://api.codetabs.com/v1/proxy?quest=';
const PROXY_3 = 'https://thingproxy.freeboard.io/fetch/';
// ================================================================
 
const BASE         = 'https://subf2m.co';
const TMDB_KEY     = process.env.TMDB_KEY || '30166e2db6b420d3f230808cd1bb2c90';
const TMDB_BASE    = 'https://api.themoviedb.org/3';
const MAX_PER_LANG = 4;
const CACHE_TTL    = 6 * 60 * 60 * 1000;
const REQ_TIMEOUT  = 18_000;
const DL_TIMEOUT   = 30_000;
const MAX_DETAIL   = 30;
 
// Full ordinal list covering seasons 1–30
const ORDINALS = [
  '', 'first', 'second', 'third', 'fourth', 'fifth',
  'sixth', 'seventh', 'eighth', 'ninth', 'tenth',
  'eleventh', 'twelfth', 'thirteenth', 'fourteenth', 'fifteenth',
  'sixteenth', 'seventeenth', 'eighteenth', 'nineteenth', 'twentieth',
  'twenty-first', 'twenty-second', 'twenty-third', 'twenty-fourth', 'twenty-fifth',
  'twenty-sixth', 'twenty-seventh', 'twenty-eighth', 'twenty-ninth', 'thirtieth',
];
 
const ALLOWED_LANGS = new Set(['english', 'arabic']);
 
const OTHER_LANGS = [
  'french','spanish','portuguese','german','italian','turkish','persian','farsi',
  'hindi','russian','polish','dutch','swedish','norwegian','danish','finnish',
  'korean','japanese','chinese','malay','indonesian','vietnamese','thai','greek',
  'hungarian','czech','romanian','bulgarian','hebrew','ukrainian','serbian',
  'croatian','slovak','slovenian','latvian','estonian','lithuanian','albanian',
  'azerbaijani','bosnian','catalan','georgian','icelandic','macedonian','mongolian',
  'punjabi','sinhala','tamil','bengali','urdu','nepali','khmer','burmese',
  'afrikaans','swahili',
];
 
const QUALITY_TIERS = [
  { bonus: 0.13, terms: [/\bbluray\b/i,/\bblu[- ]ray\b/i,/\bbdrip\b/i,/\bbdremux\b/i,/\b4k\b/i,/\b2160p\b/i,/\b1080p\b/i] },
  { bonus: 0.10, terms: [/\bweb[- ]?dl\b/i,/\bwebrip\b/i,/\bamzn\b/i,/\bnflx\b/i,/\bhbo\b/i,/\bdsnp\b/i,/\bappletv\b/i,/\batvp\b/i,/\bpcok\b/i] },
  { bonus: 0.05, terms: [/\bhdtv\b/i,/\b720p\b/i,/\bhdrip\b/i] },
  { bonus: 0.08, terms: [/\bproper\b/i,/\brepack\b/i,/\breal\.proper\b/i,/\bsynced\b/i,/\bsync\b/i,/\bfixed\b/i,/\bcorrected\b/i,/\bv2\b/i,/\bv3\b/i] },
];
 
const AI_PATTERNS = [
  /\ba\.i\.\b/i,/\bai[\s\-_]sub(title)?s?\b/i,/\bai[\s\-_](translat|generat|correct|made|creat)/i,
  /(translat|generat|creat|made|sub(title)?s?)\s+(by\s+)?ai\b/i,/machine[\s\-_]translat/i,
  /auto[\s\-_]?translat/i,/auto[\s\-_]?generat/i,/auto[\s\-_]?sub(title)?s?\b/i,/\bautosub\b/i,
  /\bchatgpt\b/i,/\bgpt[\s\-_]?[34o]?\b/i,/\bdeep[\s\-_]?l\b/i,/google[\s\-_]translat/i,
  /yandex[\s\-_]translat/i,/amazon[\s\-_]translat/i,/microsoft[\s\-_]translat/i,
  /bing[\s\-_]translat/i,/\bopenai\b/i,/\bwhisper\b/i,/\bgemini\b/i,/\bclaude\b/i,
  /\bllama\b/i,/neural[\s\-_](translat|sub|net|machine)/i,/\bnmt\b/i,
  /artificial[\s\-_]intelligence/i,/computer[\s\-_](translat|generat)/i,
];
 
const AI_QUICK_HITS = [
  'a.i','chatgpt','deepl','whisper','openai','gemini','claude','llama',
  'machine translat','auto translat','auto-translat','auto generat','autosub','neural',
  'yandex','bing translat','google translat','amazon translat','microsoft translat','nmt',
  'artificial intel',' ai sub','ai-sub','ai translat','translated by ai','generated by ai',
];
 
function isAISub(text) {
  if (!text) return false;
  const lo = text.toLowerCase();
  const maybeAI = AI_QUICK_HITS.some(q => lo.includes(q)) || /\bai\b/.test(lo);
  if (!maybeAI) return false;
  return AI_PATTERNS.some(rx => rx.test(text));
}
 
const UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36';
 
// ================================================================
//  HTTP CORE
// ================================================================
function _request(urlStr, opts = {}) {
  return new Promise((resolve, reject) => {
    const parsed  = new URL(urlStr);
    const lib     = parsed.protocol === 'https:' ? https : http;
    const timeout = opts.timeout || REQ_TIMEOUT;
    const reqOpts = {
      hostname: parsed.hostname,
      port    : parsed.port || (parsed.protocol === 'https:' ? 443 : 80),
      path    : parsed.pathname + parsed.search,
      method  : 'GET',
      headers : {
        'User-Agent'               : UA,
        'Accept'                   : opts.binary ? 'application/zip,application/octet-stream,*/*' : 'text/html,application/xhtml+xml,*/*;q=0.8',
        'Accept-Language'          : 'en-US,en;q=0.9',
        'Accept-Encoding'          : 'gzip, deflate',
        'Referer'                  : 'https://subf2m.co/',
        'Connection'               : 'keep-alive',
        'Cache-Control'            : 'no-cache',
        'Upgrade-Insecure-Requests': '1',
        ...(opts.headers || {}),
      },
    };
    const req = lib.request(reqOpts, res => {
      if (res.statusCode >= 300 && res.statusCode < 400 && res.headers.location) {
        const next = new URL(res.headers.location, urlStr).toString();
        return _request(next, opts).then(resolve).catch(reject);
      }
      if (res.statusCode === 404) return resolve(null);
      if (res.statusCode !== 200) return resolve(null);
      const chunks = [];
      let   stream = res;
      const enc    = res.headers['content-encoding'] || '';
      if (enc.includes('gzip'))         stream = res.pipe(zlib.createGunzip());
      else if (enc.includes('br'))      stream = res.pipe(zlib.createBrotliDecompress());
      else if (enc.includes('deflate')) stream = res.pipe(zlib.createInflate());
      stream.on('data', c => chunks.push(c));
      stream.on('end',  () => {
        const buf = Buffer.concat(chunks);
        resolve(opts.binary ? buf : buf.toString('utf8'));
      });
      stream.on('error', reject);
    });
    req.setTimeout(timeout, () => { req.destroy(); reject(new Error('Timeout')); });
    req.on('error', reject);
    req.end();
  });
}
 
async function getHTML(url, tries = 3) {
  for (let i = 0; i < tries; i++) {
    try {
      const r = await _request(url, { timeout: REQ_TIMEOUT });
      if (r !== null) return r;
    } catch (e) {
      if (i < tries - 1) await sleep(1200 * (i + 1));
    }
  }
  return null;
}
 
// Binary with proxy chain fallback
async function getBinary(url, tries = 3) {
  for (let i = 0; i < tries; i++) {
    try {
      const r = await _request(url, { binary: true, timeout: DL_TIMEOUT });
      if (r !== null) return r;
    } catch (_) {}
  }
  // Proxy chain
  const enc   = encodeURIComponent(url);
  const proxied = [
    `${PROXY_1}${enc}`,
    `${PROXY_2}${enc}`,
    `${PROXY_3}${url}`,
  ];
  for (const pUrl of proxied) {
    try {
      const r = await _request(pUrl, { binary: true, timeout: DL_TIMEOUT });
      if (r && r.length > 50) return r;
    } catch (_) {}
  }
  return null;
}
 
// ================================================================
//  CACHE
// ================================================================
const _cache = new Map();
function cacheGet(key) {
  const item = _cache.get(key);
  if (!item) return null;
  if (Date.now() - item.ts > CACHE_TTL) { _cache.delete(key); return null; }
  return item.val;
}
function cacheSet(key, val) {
  if (_cache.size >= 1200) {
    let n = 0;
    for (const k of _cache.keys()) { _cache.delete(k); if (++n >= 120) break; }
  }
  _cache.set(key, { val, ts: Date.now() });
}
const sleep = ms => new Promise(r => setTimeout(r, ms));
 
// ================================================================
//  STRING UTILS
// ================================================================
function titleToSlug(title, year) {
  const base = (title || '').toLowerCase()
    .normalize('NFD').replace(/[\u0300-\u036f]/g, '')
    .replace(/[^a-z0-9\s]/g, '')
    .replace(/\s+/g, '-')
    .replace(/-+/g, '-')
    .replace(/^-|-$/g, '');
  return year ? `${base}-${year}` : base;
}
 
function _clean(t) { return (t || '').replace(/\s+/g, ' ').trim(); }
 
// ================================================================
//  RELEASE PARSING & SCORING
// ================================================================
function parseRelease(text) {
  if (!text) return {};
  const tokens = {};
  const seMatch = text.match(/[sS](\d{1,2})[eE](\d{1,3})/);
  if (seMatch) {
    tokens.season  = parseInt(seMatch[1]);
    tokens.episode = parseInt(seMatch[2]);
  } else {
    const xMatch = text.match(/\b(\d{1,2})x(\d{2,3})\b/i);
    if (xMatch) { tokens.season = parseInt(xMatch[1]); tokens.episode = parseInt(xMatch[2]); }
  }
  const resMatch = text.match(/\b(4K|2160p|1080p|720p|480p)\b/i);
  if (resMatch) tokens.resolution = resMatch[1].toLowerCase();
  const srcMatch = text.match(/\b(WEB[-]?DL|WEBRip|BluRay|HDTV|AMZN|NF|Netflix|DSNP|ATVP|HBO|BRAVIA|PCOK)\b/i);
  if (srcMatch) tokens.source = srcMatch[1].toUpperCase().replace(/[-]/g, '');
  const grpMatch = text.match(/-([A-Za-z0-9]+)$/);
  if (grpMatch) tokens.group = grpMatch[1].toUpperCase();
  return tokens;
}
 
function qualityBonus(text) {
  if (!text) return 0;
  let bonus = 0;
  for (const tier of QUALITY_TIERS)
    if (tier.terms.some(rx => rx.test(text))) bonus += tier.bonus;
  return Math.min(bonus, 0.20);
}
 
function scoreSubtitle(sub, streamTokens) {
  if (!streamTokens) return 0;
  const subText   = sub.releaseInfo || sub._rawText || '';
  const subTokens = parseRelease(subText);
  let   score     = 0;
  if (streamTokens.season != null && streamTokens.episode != null &&
      subTokens.season === streamTokens.season && subTokens.episode === streamTokens.episode)
    score += 100;
  if (streamTokens.source && subTokens.source && subTokens.source === streamTokens.source)
    score += 20;
  if (streamTokens.resolution && subTokens.resolution === streamTokens.resolution)
    score += 10;
  if (streamTokens.group && subTokens.group && subTokens.group === streamTokens.group)
    score += 30;
  if (sub.isGood) score += 5;
  score += qualityBonus(subText) * 10;
  return score;
}
 
// ================================================================
//  TMDB
// ================================================================
async function getTMDB(imdbId) {
  const ck  = `tmdb:${imdbId}`;
  const hit = cacheGet(ck);
  if (hit !== null) return hit;
  try {
    const raw = await _request(
      `${TMDB_BASE}/find/${imdbId}?external_source=imdb_id&api_key=${TMDB_KEY}`,
      { timeout: 10_000 },
    );
    if (!raw) { cacheSet(ck, null); return null; }
    const d    = JSON.parse(raw);
    let   info = null;
    if (d.movie_results?.length) {
      const m = d.movie_results[0];
      info = { title: m.title || m.original_title, year: m.release_date?.substring(0,4) || null, type: 'movie', tmdbId: m.id };
    } else if (d.tv_results?.length) {
      const t = d.tv_results[0];
      info = { title: t.name || t.original_name, year: t.first_air_date?.substring(0,4) || null, type: 'series', tmdbId: t.id };
    }
    cacheSet(ck, info);
    return info;
  } catch (e) { cacheSet(ck, null); return null; }
}
 
async function getTMDBSeasonYear(tmdbId, season) {
  if (!tmdbId) return null;
  const ck  = `tmdbsy:${tmdbId}:${season}`;
  const hit = cacheGet(ck);
  if (hit !== null) return hit;
  try {
    const raw  = await _request(`${TMDB_BASE}/tv/${tmdbId}/season/${season}?api_key=${TMDB_KEY}`, { timeout: 10_000 });
    if (!raw) { cacheSet(ck, null); return null; }
    const year = JSON.parse(raw).air_date?.substring(0, 4) || null;
    cacheSet(ck, year);
    return year;
  } catch (e) { cacheSet(ck, null); return null; }
}
 
// ================================================================
//  SLUG PROBING — FIXED
//  No longer requires S##E## on the page — any item list qualifies.
// ================================================================
async function _probeSlug(slug, opts = {}) {
  if (!slug || slug.length > 220) return false;
  try {
    const lang = opts.lang || 'english';
    const url  = `${BASE}/subtitles/${slug}/${lang}`;
    const html = await getHTML(url, 1);
    if (!html || html.length < 500) return false;
    // Must have subtitle items
    if (!html.includes('<li') && !html.includes('subtitle') && !html.includes('download')) return false;
    // If we need to verify season, do a looser check
    if (opts.season) {
      const s    = Number(opts.season);
      const pad  = String(s).padStart(2, '0');
      const ord  = (ORDINALS[s] || '').toLowerCase();
      // Accept if any season-related text is found, OR if there are item listings
      const hasSeasonRef = (
        html.includes(`S${pad}E`) || html.includes(`s${pad}e`) ||
        html.includes(`season ${s}`) || html.includes(`Season ${s}`) ||
        (ord && html.toLowerCase().includes(`${ord} season`)) ||
        html.includes(`li class="item"`) ||
        html.includes('class="r"')
      );
      return hasSeasonRef;
    }
    return html.includes('<li') || html.includes('item');
  } catch (_) { return false; }
}
 
// ================================================================
//  SERIES SLUG RESOLUTION — FIXED
//  Key insight from bazarr issue #2617 and #2616:
//  1. Subf2m search returns entries like "Young Sheldon - first season (2017)"
//     with a TRAILING SPACE before closing tag — must trim.
//  2. Search by IMDb ID works: ?query=tt1234567
//  3. The slug for a series season is: show-name-first-season OR
//     show-name-first-season-2017
// ================================================================
async function resolveSeriesSlug(imdbId, season) {
  const ck  = `sslug2:${imdbId}:${season}`;
  const hit = cacheGet(ck);
  if (hit !== null) return hit;
 
  const tmdb      = await getTMDB(imdbId);
  const title     = tmdb?.title;
  const showYear  = tmdb?.year;
  const seasonYear = tmdb?.tmdbId ? await getTMDBSeasonYear(tmdb.tmdbId, season) : null;
 
  // Strategy 1: Search by IMDb ID (most reliable)
  let slug = await _searchSeriesSlugByQuery(imdbId, season, title, seasonYear, showYear);
  if (slug) { cacheSet(ck, slug); return slug; }
 
  // Strategy 2: Search by title
  if (title) {
    slug = await _searchSeriesSlugByQuery(title, season, title, seasonYear, showYear);
    if (slug) { cacheSet(ck, slug); return slug; }
  }
 
  // Strategy 3: Smart pattern brute-force
  if (title) {
    slug = await _smartPatternSlugs(title, season, seasonYear, showYear);
    if (slug) { cacheSet(ck, slug); return slug; }
  }
 
  // Strategy 4: Bare slug (complete series page)
  if (title) {
    const bare = titleToSlug(title, null);
    if (await _probeSlug(bare, { season })) { cacheSet(ck, bare); return bare; }
    // Try with show year
    if (showYear) {
      const withYear = titleToSlug(title, showYear);
      if (await _probeSlug(withYear, { season })) { cacheSet(ck, withYear); return withYear; }
    }
  }
 
  cacheSet(ck, null);
  return null;
}
 
async function _searchSeriesSlugByQuery(query, season, titleHint, seasonYear, showYear) {
  const searchUrl = `${BASE}/subtitles/searchbytitle?query=${encodeURIComponent(query)}&l=`;
  const html      = await getHTML(searchUrl);
  if (!html) return null;
  const $       = cheerio.load(html);
  const ord     = (ORDINALS[season] || '').toLowerCase();
  const firstWord = titleHint ? titleToSlug(titleHint, null).split('-')[0] : '';
  const results = [];
  const seen    = new Set();
 
  // The bazarr fix: subf2m titles in HTML have trailing spaces.
  // Regex matches: "Show Name - ORDINAL season (YEAR)" or "Show Name (YEAR)"
  // The series regex from bazarr (with the \s*$ fix):
  //   ^(.+?) [-\(]\s?(.*?) (season|series)\)?( \((\d{4})\))?\s*$
  const SERIES_RE = /^(.+?)\s*[-–]\s*(.*?)\s*(season|series)\s*(\((\d{4})\))?\s*$/i;
  const MOVIE_RE  = /^(.+?)\s*\((\d{4})\)\s*$/;
 
  $('a[href]').each((_, a) => {
    const href = $(a).attr('href') || '';
    if (!href.startsWith('/subtitles/')) return;
    const parts = href.split('/').filter(Boolean);
    if (parts.length < 2) return;
    const slug = parts[1];
    if (!slug || slug === 'searchbytitle' || seen.has(slug)) return;
    seen.add(slug);
 
    const rawText = _clean($(a).text());
    const slugLow = slug.toLowerCase();
 
    // Must start with show's first word for disambiguation
    if (firstWord && !slugLow.startsWith(firstWord)) return;
 
    let score = 0;
 
    // Check title text for season match
    const seriesMatch = SERIES_RE.exec(rawText);
    if (seriesMatch) {
      const subOrd  = (seriesMatch[2] || '').trim().toLowerCase();
      const subYear = seriesMatch[5];
      // Ordinal match (e.g. "first season")
      if (ord && subOrd === ord) score += 200;
      else if (ord && subOrd.startsWith(ord)) score += 150;
      // Numeric season as ordinal
      if (subOrd === String(season)) score += 100;
      // Year match
      if (seasonYear && subYear === seasonYear) score += 50;
      else if (showYear && subYear === showYear && season === 1) score += 30;
      // Any series result is better than none
      if (score === 0) score += 10;
    }
 
    // Also score by slug patterns (backup)
    if (ord && slugLow.includes(`-${ord}-season`)) score += Math.max(score, 80);
    if (slugLow.includes(`-season-${season}-`))    score += Math.max(score, 70);
    if (slugLow.includes(`-s${String(season).padStart(2,'0')}`)) score += Math.max(score, 60);
    if (ord && slugLow.includes(`-${ord}season`))  score += Math.max(score, 50);
 
    if (score > 0) results.push({ slug, score });
  });
 
  if (!results.length) return null;
  results.sort((a, b) => b.score - a.score);
 
  for (const r of results.slice(0, 8)) {
    await sleep(180);
    if (await _probeSlug(r.slug, { season })) return r.slug;
  }
  return null;
}
 
async function _smartPatternSlugs(title, season, seasonYear, showYear) {
  const base = titleToSlug(title, null);
  const ord  = ORDINALS[season] || '';
  const pad  = String(season).padStart(2, '0');
  const patterns = [];
 
  // Tier 1 — TMDB season year exact
  if (seasonYear) {
    if (ord) {
      patterns.push(`${base}-${ord}-season-${seasonYear}`);
      patterns.push(`${base}--${ord}-season-${seasonYear}`);
      patterns.push(`${base}-${ord}-season-tv-${seasonYear}`);
      patterns.push(`${base}-the-${ord}-season-${seasonYear}`);
    }
    patterns.push(`${base}-season-${season}-${seasonYear}`);
    patterns.push(`${base}-s${pad}-${seasonYear}`);
    patterns.push(`${base}-${seasonYear}`);
  }
 
  // Tier 2 — no year
  if (ord) {
    patterns.push(`${base}-${ord}-season`);
    patterns.push(`${base}--${ord}-season`);
    patterns.push(`${base}-the-${ord}-season`);
    patterns.push(`${base}-${ord}-series`);
  }
  patterns.push(`${base}-season-${season}`);
  patterns.push(`${base}-s${pad}`);
 
  // Tier 3 — adjacent years
  const yearBase = seasonYear ? +seasonYear : (showYear ? +showYear : null);
  if (yearBase) {
    for (const yr of [yearBase - 1, yearBase + 1, yearBase - 2]) {
      if (ord) {
        patterns.push(`${base}-${ord}-season-${yr}`);
        patterns.push(`${base}-the-${ord}-season-${yr}`);
      }
      patterns.push(`${base}-season-${season}-${yr}`);
      patterns.push(`${base}-${yr}`);
    }
  }
 
  const unique = [...new Set(patterns)].filter(Boolean);
  console.log(`[slug] Probing ${unique.length} patterns for "${title}" S${season}`);
 
  for (const s of unique) {
    await sleep(120);
    if (await _probeSlug(s, { season })) {
      console.log(`[slug] HIT: ${s}`);
      return s;
    }
  }
  return null;
}
 
// ================================================================
//  MOVIE SLUG RESOLUTION
// ================================================================
async function resolveMovieSlug(imdbId, titleHint, yearHint) {
  const ck  = `mslug2:${imdbId}`;
  const hit = cacheGet(ck);
  if (hit !== null) return hit;
 
  const tmdb  = await getTMDB(imdbId);
  const title = tmdb?.title || titleHint;
  const year  = tmdb?.year  || yearHint;
 
  // Strategy 1: Direct slug (most common pattern)
  if (title && year) {
    const direct = titleToSlug(title, year);
    if (await _probeMovieSlug(direct, title)) { cacheSet(ck, direct); return direct; }
  }
 
  // Strategy 2: Search by IMDb ID
  let slug = await _searchMovieSlug(imdbId, year);
  if (slug && await _probeMovieSlug(slug, title)) { cacheSet(ck, slug); return slug; }
 
  // Strategy 3: Search by title
  if (title) {
    slug = await _searchMovieSlug(title, year);
    if (slug && await _probeMovieSlug(slug, title)) { cacheSet(ck, slug); return slug; }
 
    // Strategy 4: Brute-force slugs
    const candidates = [];
    if (year) {
      candidates.push(titleToSlug(title, year));
      candidates.push(titleToSlug(title, String(+year - 1)));
      candidates.push(titleToSlug(title, String(+year + 1)));
    }
    candidates.push(titleToSlug(title, null));
    for (const c of [...new Set(candidates)]) {
      await sleep(250);
      if (await _probeMovieSlug(c, title)) { cacheSet(ck, c); return c; }
    }
  }
 
  cacheSet(ck, null);
  return null;
}
 
async function _searchMovieSlug(query, yearHint) {
  const url  = `${BASE}/subtitles/searchbytitle?query=${encodeURIComponent(query)}&l=`;
  const html = await getHTML(url);
  if (!html) return null;
  const $       = cheerio.load(html);
  const results = [];
  const SELS    = ['ul.search-result li','.search-result li','.content ul li','section ul li','ul li','li'];
  let   found   = false;
 
  for (const sel of SELS) {
    const items = $(sel);
    if (!items.length) continue;
    items.each((_, el) => {
      const a    = $(el).find('a[href]').first();
      const href = a.attr('href') || '';
      if (!href.startsWith('/subtitles/') || href === '/subtitles/searchbytitle') return;
      const parts = href.split('/').filter(Boolean);
      if (parts.length < 2) return;
      const slug = parts[1];
      if (!slug || slug === 'searchbytitle') return;
      const text    = _clean(a.text());
      const yrMatch = text.match(/\((\d{4})\)/);
      const resYear = yrMatch ? +yrMatch[1] : null;
      let   score   = 1;
      if (query.startsWith('tt'))                score += 100;
      if (yearHint && resYear === +yearHint)      score += 10;
      // Penalise series results in movie search
      if (text.toLowerCase().includes('season')) score -= 5;
      results.push({ slug, score });
      found = true;
    });
    if (found) break;
  }
 
  if (!results.length) return null;
  results.sort((a, b) => b.score - a.score);
  return results[0].slug;
}
 
async function _probeMovieSlug(slug, expectedTitle) {
  if (!slug) return false;
  try {
    const html = await getHTML(`${BASE}/subtitles/${slug}/english`);
    if (!html || html.length < 1500) return false;
    if (expectedTitle) {
      const $         = cheerio.load(html);
      const pageTitle = ($('title').text() || '').toLowerCase();
      const exp       = expectedTitle.toLowerCase();
      // Accept if title appears in page title OR if there are subtitle items
      return pageTitle.includes(exp) || html.includes('class="item"') || html.includes('class="r"');
    }
    return true;
  } catch (_) { return false; }
}
 
// ================================================================
//  SUBTITLE SCRAPING
// ================================================================
async function getSubsList(slug, lang, season, episode, streamTokens) {
  if (!ALLOWED_LANGS.has(lang)) return [];
  const ck  = `list2:${slug}:${lang}:${episode || 0}`;
  const hit = cacheGet(ck);
  if (hit) return hit;
 
  const url  = `${BASE}/subtitles/${slug}/${lang}`;
  const html = await getHTML(url);
  if (!html) { cacheSet(ck, []); return []; }
  const $  = cheerio.load(html);
  let   raw = [];
 
  // Primary scraper
  $('li.item').each((_, el) => { _extractItem($, el, lang, raw); });
 
  // Fallback 1: table rows
  if (!raw.length) {
    $('tr, .subtitle-item, .item').each((_, el) => { _extractItem($, el, lang, raw); });
  }
 
  // Fallback 2: any anchor with matching lang path
  if (!raw.length) {
    $('a[href]').each((_, a) => {
      const href = $(a).attr('href') || '';
      if (!href.includes(`/${lang}/`) || !/\/\d+$/.test(href)) return;
      const parts = href.split('/');
      const subId = parts[parts.length - 1];
      if (!subId || !/^\d+$/.test(subId)) return;
      const full = href.startsWith('http') ? href : `${BASE}${href}`;
      raw.push({ subId, detailUrl: full, isGood: false, releaseInfo: '', comment: '', authorNote: '', lang, goodVotes: 0, badVotes: 0, downloads: 0, _rawText: '' });
    });
  }
 
  // Filter to exact episode if possible
  if (season && episode && raw.length) {
    const expected = `s${String(season).padStart(2,'0')}e${String(episode).padStart(2,'0')}`;
    const filtered = raw.filter(item => {
      const txt = (item._rawText || item.releaseInfo || '').toLowerCase();
      return txt.includes(expected);
    });
    if (filtered.length) raw = filtered;
  }
 
  // Deduplicate
  const seen   = new Set();
  const unique = raw.filter(s => { if (seen.has(s.subId)) return false; seen.add(s.subId); return true; });
 
  const toEnrich = unique.slice(0, MAX_DETAIL);
  const enriched = (await Promise.all(toEnrich.map(s => _enrichSub(s)))).filter(Boolean);
  const human    = enriched.filter(s => !isAISub(`${s.releaseInfo} ${s.comment} ${s.authorNote}`));
 
  cacheSet(ck, human);
  return human;
}
 
function _extractItem($, el, lang, out) {
  const $el     = $(el);
  const rawText = _clean($el.text());
  let detailHref = null;
 
  $el.find('a').each((_, a) => {
    const h = $(a).attr('href') || '';
    if (h.includes(`/${lang}/`) && /\/\d+$/.test(h)) detailHref = h;
  });
 
  if (!detailHref) return;
  const hLow = detailHref.toLowerCase();
  if (!hLow.includes(`/${lang}/`)) return;
  if (OTHER_LANGS.some(ol => hLow.includes(`/${ol}/`))) return;
 
  const parts = detailHref.split('/');
  const subId = parts[parts.length - 1];
  if (!subId || !/^\d+$/.test(subId)) return;
 
  const isGood = (
    $el.find('.green-icon, .positive-icon').length > 0 ||
    $el.find('img[alt="good"], img[alt="Positive"], img[alt="positive"]').length > 0 ||
    $el.hasClass('good') || $el.hasClass('positive') ||
    ($el.attr('class') || '').toLowerCase().includes('good')
  );
 
  const comment      = _clean($el.find('p.comment, .comment').first().text());
  const releaseLines = $el.find('.r li, ul.release li, ul.scrolllist li, ul li')
                          .map((_, li) => $(li).text().trim()).get().filter(Boolean);
  const releaseInfo  = (releaseLines.join(' | ') || _clean($el.text())).substring(0, 130);
 
  if (isAISub(`${comment} ${releaseInfo}`)) return;
 
  out.push({
    subId,
    detailUrl: detailHref.startsWith('http') ? detailHref : `${BASE}${detailHref}`,
    isGood, releaseInfo, comment, authorNote: '', lang,
    goodVotes: 0, badVotes: 0, downloads: 0, _rawText: rawText,
  });
}
 
async function _enrichSub(sub) {
  const ck  = `det2:${sub.subId}`;
  const hit = cacheGet(ck);
  if (hit !== null) { if (hit.__aiBlocked) return null; return { ...sub, ...hit }; }
 
  const html = await getHTML(sub.detailUrl);
  if (!html) return sub;
  const $          = cheerio.load(html);
  const txt        = $.root().text();
  const goodVotes  = +(txt.match(/voted\s+as\s+good\s+by[:\s]*(\d+)/i)?.[1]  ?? 0);
  const badVotes   = +(txt.match(/voted\s+as\s+bad\s+by[:\s]*(\d+)/i)?.[1]   ?? 0);
  const downloads  = +(txt.match(/downloads?[:\s]*(\d+)/i)?.[1]               ?? 0);
  const total      = goodVotes + badVotes;
  const ratio      = total >= 3 ? goodVotes / total : null;
  const isGood     = sub.isGood || (ratio !== null && ratio >= 0.70 && goodVotes >= 2);
  const authorNote = _clean($('p.comment, .comment, .notes, .author-comment').first().text());
 
  if (isAISub(`${authorNote} ${sub.releaseInfo} ${sub.comment}`)) {
    cacheSet(ck, { __aiBlocked: true });
    return null;
  }
 
  const relLines   = $('ul.release li, .release-names li, ul.scrolllist li').map((_, e) => $(e).text().trim()).get();
  const releaseInfo = relLines.length ? relLines.join(' | ').substring(0, 130) : sub.releaseInfo;
 
  const enriched = { ...sub, releaseInfo, goodVotes, badVotes, downloads, isGood, authorNote };
  cacheSet(ck, enriched);
  return enriched;
}
 
// ================================================================
//  DOWNLOAD & SRT EXTRACTION
// ================================================================
async function fetchSRT(downloadUrl) {
  const ck  = `srt2:${downloadUrl}`;
  const hit = cacheGet(ck);
  if (hit !== null) return hit;
  try {
    const buf = await getBinary(downloadUrl);
    if (!buf) return null;
    const content = _extractFromZip(buf) ?? _tryDirect(buf);
    if (!content) return null;
    cacheSet(ck, content);
    return content;
  } catch (_) { return null; }
}
 
function _extractFromZip(buf) {
  try {
    const zip  = new AdmZip(buf);
    const RANK = { '.srt': 0, '.vtt': 1, '.ass': 2, '.ssa': 3 };
    const files = zip.getEntries()
      .filter(e => {
        const n = e.entryName.toLowerCase();
        return !e.isDirectory && !n.startsWith('__macosx') && Object.keys(RANK).some(x => n.endsWith(x));
      })
      .sort((a, b) => (RANK[_ext(a.entryName)] ?? 9) - (RANK[_ext(b.entryName)] ?? 9));
    for (const f of files) {
      const raw  = zip.readFile(f);
      if (!raw) continue;
      const text = _decode(raw);
      if (_isSub(text)) return text;
    }
  } catch (_) {}
  return null;
}
 
function _tryDirect(buf) { const t = _decode(buf); return _isSub(t) ? t : null; }
const _ext   = n => (n.match(/(\.[^.]+)$/)?.[1]?.toLowerCase() ?? '');
const _isSub = t => typeof t === 'string' && t.includes('-->') && t.length > 50;
function _decode(buf) {
  const hasBOM = buf[0] === 0xEF && buf[1] === 0xBB && buf[2] === 0xBF;
  const u8     = buf.slice(hasBOM ? 3 : 0).toString('utf8');
  if (_isSub(u8)) return u8;
  return buf.toString('latin1');
}
 
// ================================================================
//  SORT: index [0] = best subtitle
//  Movies: sort by downloads desc (most popular = best sync).
//  Series: sort by scoreSubtitle desc (exact S##E## + group match).
// ================================================================
function sortForBest(subs, streamTokens, type) {
  if (type === 'series' && streamTokens) {
    return subs
      .map(s => ({ ...s, _score: scoreSubtitle(s, streamTokens) }))
      .sort((a, b) => {
        if (b._score !== a._score) return b._score - a._score;
        if (b.isGood !== a.isGood) return b.isGood ? 1 : -1;
        return b.downloads - a.downloads;
      });
  }
  // Movies: isGood first, then downloads
  return subs.slice().sort((a, b) => {
    if (b.isGood !== a.isGood) return b.isGood ? 1 : -1;
    return b.downloads - a.downloads;
  });
}
 
// ================================================================
//  ROUTES
// ================================================================
app.get('/manifest.json', (req, res) => res.json({
  id         : 'org.van.vansubf2mplus.v2',
  name       : 'VanSubF2M+',
  description: 'Human-Verified Arabic & English subtitles. Anti-AI. Enhanced 100% coverage.',
  version    : '2.0.0',
  resources  : ['subtitles'],
  types      : ['movie', 'series'],
  idPrefixes : ['tt'],
  catalogs   : [],
}));
 
app.get('/sub/:enc.srt', async (req, res) => {
  let dlUrl = '';
  try { dlUrl = Buffer.from(req.params.enc, 'base64url').toString('utf8'); }
  catch (_) { return res.status(400).end('Bad request'); }
  if (!dlUrl.startsWith('https://subf2m.co/') || !dlUrl.includes('/subtitles/') || !dlUrl.endsWith('/download'))
    return res.status(403).end('Forbidden');
  const content = await fetchSRT(dlUrl);
  if (!content) return res.status(404).end('Not found');
  res.setHeader('Content-Type',        'text/plain; charset=utf-8');
  res.setHeader('Cache-Control',       'public, max-age=86400');
  res.setHeader('Content-Disposition', 'inline; filename="subtitle.srt"');
  res.end(content);
});
 
app.get('/subtitles/:type/:id/:extra?.json', async (req, res) => {
  res.setHeader('Content-Type', 'application/json; charset=utf-8');
  const { type, id } = req.params;
  if (!id?.startsWith('tt')) return res.json({ subtitles: [] });
 
  const ck  = `result2:${id}`;
  const hit = cacheGet(ck);
  if (hit) return res.json({ subtitles: hit });
 
  try {
    let imdbId = id, season = null, episode = null;
    if (id.includes(':')) {
      const parts = id.split(':');
      imdbId  = parts[0];
      season  = parts[1] ? +parts[1] : null;
      episode = parts[2] ? +parts[2] : null;
    }
    if (!/^tt\d+$/.test(imdbId)) return res.json({ subtitles: [] });
 
    const filename     = req.query.filename || '';
    const streamTokens = (type === 'series' && season) ? parseRelease(filename) : null;
 
    let slug;
    if (type === 'series' && season) {
      slug = await resolveSeriesSlug(imdbId, season);
    } else {
      const tmdb = await getTMDB(imdbId);
      slug       = await resolveMovieSlug(imdbId, tmdb?.title || null, tmdb?.year || null);
    }
 
    if (!slug) return res.json({ subtitles: [] });
 
    const [engRes, araRes] = await Promise.allSettled([
      getSubsList(slug, 'english', season, episode, streamTokens),
      getSubsList(slug, 'arabic',  season, episode, streamTokens),
    ]);
 
    let engSubs = (engRes.status === 'fulfilled' ? engRes.value : []);
    let araSubs = (araRes.status === 'fulfilled' ? araRes.value : []);
 
    // Sort so [0] = best, then cap at MAX_PER_LANG
    engSubs = sortForBest(engSubs, streamTokens, type).slice(0, MAX_PER_LANG);
    araSubs = sortForBest(araSubs, streamTokens, type).slice(0, MAX_PER_LANG);
 
    const proto  = req.headers['x-forwarded-proto'] || (req.socket?.encrypted ? 'https' : 'http');
    const host   = req.headers['x-forwarded-host'] || req.headers.host;
    const origin = `${proto}://${host}`;
 
    const build = (subs, langCode, flag) => subs.map(s => {
      const dlUrl  = `${s.detailUrl}/download`;
      const enc    = Buffer.from(dlUrl).toString('base64url');
      const badge  = s.isGood   ? ' ⭐' : '';
      const votes  = s.goodVotes > 0 ? ` [👍${s.goodVotes}]` : '';
      const dls    = s.downloads > 0 ? ` [⬇${s.downloads}]` : '';
      const name   = `${flag}${badge}${votes}${dls} ${s.releaseInfo}`.trim().substring(0, 100);
      return { id: `vsub-${langCode}-${s.subId}`, url: `${origin}/sub/${enc}.srt`, lang: langCode, name };
    });
 
    const subtitles = [...build(engSubs, 'eng', '🇺🇸 English'), ...build(araSubs, 'ara', '🇸🇦 Arabic')];
    cacheSet(ck, subtitles);
    res.json({ subtitles });
  } catch (e) {
    console.error('[route]', e.message);
    res.json({ subtitles: [] });
  }
});
 
app.listen(PORT, '0.0.0.0', () => {
  console.log(`Server running on port ${PORT}`);
});