--- name: oce-selector-checker description: OCE Selector Checker — 4-phase selector verification via curl, simulates CloudStream extraction from live HTML license: MIT compatibility: "*" metadata: project: oce type: selector-verification --- # OCE Selector Checker ## Purpose Verifikasi CSS selector provider terhadap HTML live website. Mensimulasikan cara CloudStream mengekstrak data dari website menggunakan selector di `config/.json`. ## When to Use - Setelah mengubah selector di provider config - Saat website berubah struktur dan provider gagal - Sebelum push perubahan selector - Saat debugging metadata extraction (poster, title, episodes) ## When NOT to Use - Untuk debug extractor video (→ `extraction`) - Untuk debug CI/build (→ `build-deploy`) ## Prerequisites ```bash curl # fetch HTML python3 # parse HTML base64 # decode base64 (opsional) ``` ## 4-Phase Verification **Rule #1: cek item ADA di HTML sebelum menulis selector.** Kalau link-nya memang tidak ada di HTML, itu masalah sisi situs — selector apa pun tidak akan menemukannya. Contoh nyata: Anichin series page hanya punya episode 1–159, sedangkan episode 160 hanya ada di episode page. **Rule #2: episode selector harus diuji di DUA halaman.** Series page (`episodeItems`) dan episode page (`linkOptions`) punya DOM berbeda. Satu selector untuk dua konteks hampir selalu salah. **Rule #3: multi-variant = satu string comma-separated.** `"h1.entry-title, h1.title"` — bukan array, bukan selector terpisah. ### Phase 1: Main Page (Search/List) **Simulates:** `getMainPage()` + `search()` **Selectors tested:** | Selector | Purpose | |----------|---------| | `searchItems` | Container per item | | `searchTitle` | Title element | | `searchHref` | Link to detail/episode | | `searchPoster` | Poster image | | `searchRating` | Rating badge | | `searchEpText` | Episode badge | **Template:** ```bash PROVIDER_URL="" curl -sL --max-time 15 \ -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36" \ "$PROVIDER_URL" | python3 -c " import sys, re html = sys.stdin.read() # searchItems items = re.findall(r']*class=\"[^\"]*bs[^\"]*\"[^>]*>(.*?)', html, re.DOTALL) print(f'searchItems: {len(items)} items') # searchTitle (dalam item pertama) if items: title = re.search(r']*>(.*?)', items[0], re.DOTALL) if title: clean = re.sub(r'<[^>]+>', '', title.group(1)).strip() print(f'searchTitle: [{clean}]') # searchPoster posters = re.findall(r']+src=\"([^\"]+)\"[^>]*>', html) print(f'searchPoster: {len(posters)} images') if posters: print(f' sample: {posters[0][:80]}') " ``` ### Phase 2: Detail Page (Load) **Simulates:** `load(url)` **Selectors tested:** | Selector | Purpose | |----------|---------| | `loadTitle` | Series title | | `loadPoster` | Series poster | | `loadDesc` | Synopsis | | `loadInfoBox` | Info box | | `loadTags` | Genre tags | | `loadStatus` | Status | | `loadRating` | Rating | | `loadRecommend` | Recommendations | **Template:** ```bash SERIES_URL="" curl -sL --max-time 15 \ -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36" \ "$SERIES_URL" | python3 -c " import sys, re html = sys.stdin.read() # loadTitle m = re.search(r']*>(.*?)', html, re.DOTALL) title = re.sub(r'<[^>]+>', '', m.group(1)).strip() if m else 'NOT FOUND' print(f'loadTitle: [{title}]') # loadPoster m = re.search(r']*class=\"[^\"]*thumb[^\"]*\"[^>]*>.*?]+src=\"([^\"]+)\"', html, re.DOTALL) poster = m.group(1)[:80] if m else 'NOT FOUND' print(f'loadPoster: [{poster}]') # loadDesc m = re.search(r']*class=\"[^\"]*entry-content[^\"]*\"[^>]*>(.*?)', html, re.DOTALL) desc = re.sub(r'<[^>]+>', '', m.group(1)).strip()[:100] if m else 'NOT FOUND' print(f'loadDesc: [{desc}]') " ``` ### Phase 3: Episode Page (loadLinks) **Simulates:** `loadLinks(data)` **Selectors tested:** | Selector | Purpose | |----------|---------| | `linkOptions` | Server options (option[value], a[data-url]) | | `episodeItems` | Episode list | | `downloadItems` | Download links | | `iframeTag` | Iframe embed | **Template:** ```bash EP_URL="" curl -sL --max-time 15 \ -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36" \ "$EP_URL" | python3 -c " import sys, re, base64 html = sys.stdin.read() # linkOptions (option[value]) options = re.findall(r']*value=\"([^\"]+)\"[^>]*>(.*?)', html, re.DOTALL) print(f'linkOptions: {len(options)} options') for val, label in options[:5]: if val: try: decoded = base64.b64decode(val).decode('utf-8', errors='replace') src = re.search(r'src=\"([^\"]+)\"', decoded) if src: print(f' [{label.strip()}]: {src.group(1)[:60]}') except: print(f' [{label.strip()}]: (decode failed)') # episodeItems eps = re.findall(r']*>.*?]+href=\"([^\"]+)\"[^>]*>(.*?)', html, re.DOTALL) print(f'episodeItems: {len(eps)} episodes') " ``` ### Phase 4: Base64 Decode (Anichin/Animasu specific) ```bash # Decode base64 option value echo "PGlmcmFtZSB..." | base64 -d 2>/dev/null | grep -oP 'src="([^"]+)"' # Atau dengan python python3 -c " import base64, re val = 'PGlmcmFtZSB...' decoded = base64.b64decode(val).decode() src = re.search(r'src=\"([^\"]+)\"', decoded) if src: print('SRC:', src.group(1)) " ``` ## Provider URLs **SEBELUM verifikasi, ambil URL dari config provider:** ```bash # Baca mainUrl dari config python3 -c " import json with open('BaseProvider/src/main/kotlin/com/baseprovider/config/.json') as f: print(json.load(f).get('mainUrl', 'UNKNOWN')) " ``` **JANGAN hardcode URL.** URL bisa berubah. Selalu baca dari config. ## Output Format ### Per-Item (Phase 1) ``` ── Item 1 ── SEARCH_TITLE : Example Title SEARCH_HREF : https://example.com/series/slug/ SEARCH_POSTER : https://example.com/poster.webp ✅ SEARCH_EP_TEXT : Ep 140 ✅ SEARCH_RATING : 8.5 ✅ ``` ### Per-Selector (Phase 2-3) ``` LOAD_TITLE : Example Title ✅ LOAD_POSTER : https://... ✅ LOAD_DESC : Synopsis text... ✅ LINK_OPTIONS : 7 options ✅ EPISODE_ITEMS : 12 episodes ✅ ``` ### Summary ``` PHASE 1: 46 items — title/href/poster OK ✅ PHASE 2: title/poster/desc/tags/status OK ✅ PHASE 3: 7 server options, 1 iframe ✅ PHASE 4: Base64 decode OK ✅ ``` ### Status Icons - ✅ = working - ⚠️ = attention needed (not critical) - ❌ = broken / not found ## Troubleshooting ### 0 items found 1. Cek apakah website berubah tema/struktur 2. Inspect HTML manual: `curl -sL "$URL" | grep -c "class-name"` 3. Cari class baru dengan grep 4. **Konfirmasi item memang ada di HTML** sebelum ubah selector — kalau tidak ada, itu sisi situs 5. Update selector di `config/.json` — multi-variant comma-separated untuk dua layout ### 403 / Cloudflare saat curl UA biasa bisa kena CF challenge sementara app berhasil (WebView solver). Selector tetap bisa diverifikasi, tapi note hasilnya sebagai keterbatasan, bukan bukti selector salah. Bypass: `curl_cffi` dengan `impersonate='chrome'`. ### Poster not found (❌) 1. Cek apakah src pakai lazy loading (`data-src`, `data-original`) 2. Cek apakah path root-relative (`/wp-content/...`) 3. Cek `attrImage` priority di config 4. Cek `SelectorValidator.isValidPoster` — butuh `http://` atau `//` ### Title has extra words 1. Cek `bloatRegex` — bisa strip bloat dari title 2. Cek selector mengambil element yang benar (bukan `a[title]`) ### Base64 decode failed 1. Pastikan value tidak kosong 2. Coba padding: `val + '=='` 3. Mungkin plain text, bukan base64 ## Verification Checklist Per provider, pastikan: - [ ] Phase 1: `searchItems` match, items > 0 - [ ] Phase 1: `searchTitle`, `searchHref`, `searchPoster` valid - [ ] Phase 2: `loadTitle`, `loadPoster`, `loadDesc` match - [ ] Phase 3: `episodeItems` match di **series page** - [ ] Phase 3: `linkOptions` match di **episode page** - [ ] Phase 4: base64 decode works (if applicable) - [ ] Semua URL accessible (no 404, no CF block) - [ ] Jumlah item sesuai ekspektasi (mis. 160 episode, bukan 159) — kalau kurang, cek apakah link-nya ada di HTML ## Related Skills - `provider` — edit config, tambah field - `architecture` — `SelectorResolver` multi-variant, data flow - `logging` — `SELECTOR_FAILURE` investigation - `extraction` — Phase 3 lanjutan (link → extractor)