fix: filter media extras from standalone scan + fix Wanted list layout (#91)

* fix: setup auth redirect and plugins nav label

- SetupPage: handle "Already configured" error by navigating to home
  instead of showing an error toast; fixes "Continue without password"
  button doing nothing when AUTH_ENABLED=false
- i18n: add missing nav.plugins key to EN and DE common locale files
  so Sidebar shows "Plugins" instead of the raw key

* fix: standalone app context, library view, scan UI indicator

- standalone/__init__.py: push Flask app context in _initial_scan and
  _on_new_file background threads; store app ref via start(app=app)
- app.py: pass app to standalone_mgr.start()
- routes/library.py: include standalone series/movies in /library when
  no Sonarr/Radarr configured; enrich with live wanted_count; filter
  misidentified trailer files (tvshow/movie/trailer titles)
- useApi.ts: useStandaloneStatus polls every 2s while scanner_scanning
- AdvancedTab.tsx: Scan All button shows Scanning... spinner during active scan
- Setup.tsx: navigate home on Already configured error (fixes Continue
  without password button when AUTH_ENABLED=false)
- i18n: add missing nav.plugins key to EN and DE locales

* docs: V9 Sublarr integration plan

- Switch to /api/chat with explicit system/user turns
- Add series_context parameter to translate_batch interface
- Fix quality evaluator prompt + score parser (was returning 50 for everything)
- Ollama Modelfile for V9 (no SYSTEM block — injected dynamically)
- Backward-compatible: use_chat_api=False preserves V6 behavior

* feat: NFO metadata integration for standalone scanner

- Add nfo_parser.py: reads tvshow.nfo / movie.nfo (Kodi/Emby/Jellyfin
  format), extracts title, year, TVDB/TMDB/IMDB IDs, is_anime flag,
  and local poster path; uses stdlib xml.etree only
- Integrate into scanner._process_series_group() and _process_movie():
  NFO is tried first (offline, accurate); MetadataResolver only called
  when NFO lacks IDs; local poster.jpg used instead of API-fetched URL
- Add GET /api/v1/standalone/series/<id>/poster and
  GET /api/v1/standalone/movies/<id>/poster endpoints to serve local
  poster images with is_safe_path() path-traversal protection
- ruff format routes/library.py

* fix: NFO/poster lookup in Season subfolder layout

parse_series_nfo() and find_local_poster() now also check the parent
directory when the given folder_path is a Season subfolder — handles
the common case where episodes are in Season 1/ but tvshow.nfo and
poster.jpg live in the series root.

* fix: poster endpoint path security and library poster URL

- poster endpoints now allow poster_url to reside in series_folder OR
  its parent directory (handles Season-subfolder layout where poster.jpg
  is in the series root but folder_path points to Season 1/)
- library endpoint returns /api/v1/standalone/series/<id>/poster URLs
  instead of raw file paths (browsers block file:// local resources)

* fix: standalone series detail fallback for Sonarr-free mode

* fix: standalone fallback for series subtitles sidecar endpoint

* docs: fansub scoring integration design spec

* docs: revise fansub scoring spec — fix layer architecture, process_wanted_item gap, settings tab placement

* fix: filter media extras from standalone scan + fix Wanted list layout

- Skip trailers, featurettes, samples etc. during standalone filesystem scan.
  Matches Jellyfin/Kodi extras naming convention (-trailer, -featurette,
  -behindthescenes, -deleted, -interview, -scene, -short, -sample, -theme)
  and standalone stems (tvshow, movie, trailer, sample).
- Cleanup step also removes existing wanted entries for matched extras.
- New config field standalone_skip_extras (default: true), exposed as
  toggle in Settings → Library Sources (advanced).
- Fix Wanted page scroll list not filling available viewport: replaced
  hardcoded calc(100vh-300px) with proper flex-1/min-h-0 chain so the
  virtualizer always gets the correct container height.

---------

Co-authored-by: Abrechen2 <179815050+DennisSico@users.noreply.github.com>

Author: Abrechen2

backend/config.py

@@ -248,6 +248,7 @@ class Settings(BaseSettings):
     standalone_enabled: bool = False
     standalone_scan_interval_hours: int = 6  # 0 = disabled
     standalone_debounce_seconds: int = 10
+    standalone_skip_extras: bool = True  # Skip trailers/featurettes/samples during scan
     tmdb_api_key: str = ""  # TMDB API v3 Bearer token
     tvdb_api_key: str = ""  # TVDB API v4 key (optional)
     tvdb_pin: str = ""  # TVDB PIN (optional)

backend/standalone/scanner.py

@@ -14,6 +14,28 @@
 
 logger = logging.getLogger(__name__)
 
+# Filename stems and suffixes that mark non-episode extras (Jellyfin/Kodi convention).
+# Files matching these are excluded from subtitle discovery.
+_EXTRA_STEMS = frozenset({"tvshow", "movie", "trailer", "sample"})
+_EXTRA_SUFFIXES = (
+    "-trailer",
+    "-featurette",
+    "-behindthescenes",
+    "-deleted",
+    "-interview",
+    "-scene",
+    "-short",
+    "-sample",
+    "-theme",
+)
+
+
+def _is_extra_file(path: str) -> bool:
+    """Return True if *path* is a media extra (trailer, featurette, …) that should
+    be excluded from subtitle discovery."""
+    stem = os.path.splitext(os.path.basename(path))[0].lower()
+    return stem in _EXTRA_STEMS or any(stem.endswith(s) for s in _EXTRA_SUFFIXES)
+
 
 class StandaloneScanner:
     """Scans watched folders for video files, resolves metadata, and populates wanted_items.
@@ -152,19 +174,22 @@ def _scan_folder(self, folder: dict) -> tuple:
         Returns:
             Tuple of (series_count, movie_count, wanted_count).
         """
+        from config import get_settings
         from standalone.parser import group_files_by_series, is_video_file, parse_media_file
 
         folder_path = folder["path"]
         if not os.path.isdir(folder_path):
             logger.warning("Watched folder does not exist: %s", folder_path)
             return (0, 0, 0)
 
+        skip_extras = getattr(get_settings(), "standalone_skip_extras", True)
+
         # Collect all video files
         video_files = []
         for root, _dirs, files in os.walk(folder_path, followlinks=True):
             for filename in files:
                 full_path = os.path.join(root, filename)
-                if is_video_file(full_path):
+                if is_video_file(full_path) and (not skip_extras or not _is_extra_file(full_path)):
                     video_files.append(full_path)
 
         if not video_files:
@@ -613,8 +638,10 @@ def _cleanup_stale_wanted(self) -> int:
             Number of items removed.
         """
         try:
+            from config import get_settings
             from db import _db_lock, get_db
 
+            skip_extras = getattr(get_settings(), "standalone_skip_extras", True)
             db = get_db()
             with _db_lock:
                 rows = db.execute(
@@ -623,7 +650,7 @@ def _cleanup_stale_wanted(self) -> int:
 
             to_remove = []
             for row in rows:
-                if not os.path.exists(row[1]):
+                if not os.path.exists(row[1]) or (skip_extras and _is_extra_file(row[1])):
                     to_remove.append(row[0])
 
             if to_remove:

frontend/src/pages/Settings/index.tsx

@@ -234,6 +234,9 @@ const FIELDS: FieldConfig[] = [
   { key: 'standalone_debounce_seconds', label: 'File Detection Debounce (seconds)', type: 'number', placeholder: '10', tab: 'Library Sources',
     description: 'Wartezeit nach letzter Dateiänderung bevor Scan ausgelöst wird.',
     advanced: true },
+  { key: 'standalone_skip_extras', label: 'Skip Extra Files (Trailer, Featurettes …)', type: 'toggle', tab: 'Library Sources',
+    description: 'Trailer, Featurettes, Samples und ähnliche Dateien beim Scan ignorieren.',
+    advanced: true },
   // Automation — Sidecar Cleanup
   { key: 'auto_cleanup_after_extract', label: 'Nach Extraktion bereinigen', type: 'toggle', tab: 'Automation',
     description: 'Nach batch-extract automatisch nicht-benötigte Sprachen löschen. Benötigt keep_languages.' },

frontend/src/pages/Settings/settingsHelpText.ts

@@ -58,6 +58,7 @@ export const HELP_TEXT: Record<string, string> = {
   tvdb_pin: 'Optional TVDB subscriber PIN.',
   standalone_scan_interval_hours: 'How often to scan watched folders for new files. 0 = disabled.',
   standalone_debounce_seconds: 'Wait this many seconds after file detection before processing.',
+  standalone_skip_extras: 'When enabled, files matching common extras naming conventions are ignored (e.g. movie-trailer.mp4, tvshow.mp4, movie-featurette.mkv). Follows the Jellyfin/Kodi extras standard.',
 
   // Translation Backends
   ollama_url: 'Full URL to your Ollama instance (e.g. http://localhost:11434).',

frontend/src/pages/Wanted.tsx

@@ -454,7 +454,7 @@ export function WantedPage() {
   }
 
   return (
-    <div className="space-y-5">
+    <div className="flex flex-col gap-5" style={{ height: 'calc(100vh - 40px)' }}>
       {/* Batch Probe Progress Banner */}
       {probeStatus?.running && (
         <div
@@ -834,10 +834,10 @@ export function WantedPage() {
       {/* Table */}
       <div
         data-testid="wanted-list"
-        className="rounded-lg overflow-hidden"
+        className="rounded-lg overflow-hidden flex-1 min-h-0 flex flex-col"
         style={{ backgroundColor: 'var(--bg-surface)', border: '1px solid var(--border)' }}
       >
-        <div ref={wantedParentRef} style={{ height: 'calc(100vh - 300px)', overflowY: 'auto' }}>
+        <div ref={wantedParentRef} className="flex-1 min-h-0" style={{ overflowY: 'auto' }}>
           <table className="w-full min-w-[800px]">
             <thead style={{ position: 'sticky', top: 0, zIndex: 1, backgroundColor: 'var(--bg-elevated)' }}>
               <tr style={{ borderBottom: '1px solid var(--border)' }}>