fix: korrekte Track-Nummerierung, Scanner-Rekursion, M3U-Reihenfolge

scanner: nicht in Unterordner wenn Root Audio-Dateien enthält (verhindert
  Doppel-Scan bei versehentlichen Unterordner-Kopien); nur Disc-Ordner
  (CD1, Disc 2…) werden bei Multi-CD-Alben rekursiert.

hint_extractor: M3U/Playlist-Dateien als Track-Reihenfolge-Quelle; BOM-
  Bereinigung; Tracklist-Matching auch per Titel (nicht nur per Nummer);
  tracknumber=0 wird als 'keine Nummer' gewertet.

metadata_resolver: sequenzielle Fallback-Nummerierung (1,2,3…) für Tracks
  ohne Tracknummer — verhindert '00'-Präfix beim --rename; dir_artist hat
  Vorrang vor 'Various Artists'-Heuristik; LLM darf bei Konfidenz <0.3
  auch bestehende Werte korrigieren (Tippfehler im Verzeichnisnamen).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

hint_extractor.py

@@ -51,6 +51,8 @@ _DISC_SECTION_RE = re.compile(r"(?i)(?:cd|disc|disk|side)[_ \-]*(\d{1,2})")
 def _clean(s: Optional[str]) -> str:
     if not s:
         return ""
+    # BOM (U+FEFF), Zero-Width-Space (U+200B), Soft-Hyphen (U+00AD) entfernen
+    s = re.sub(r"[​­]", "", s)
     return re.sub(r"\s+", " ", s.replace("_", " ")).strip(" -._")
 
 
@@ -143,6 +145,34 @@ def _parse_tracklist(text: str) -> List[Dict[str, str]]:
     return tracks
 
 
+def _parse_m3u(text: str) -> List[Dict[str, str]]:
+    """M3U/M3U8 → geordnete Liste: [{filename, title, position}].
+    Reihenfolge der Einträge = gewünschte Trackreihenfolge.
+    """
+    tracks: List[Dict[str, str]] = []
+    pending_title: Optional[str] = None
+    position = 0
+    for line in text.splitlines():
+        line = line.strip()
+        if not line:
+            continue
+        if line.upper().startswith("#EXTINF:"):
+            parts = line.split(",", 1)
+            pending_title = parts[1].strip() if len(parts) > 1 else None
+        elif not line.startswith("#"):
+            filename = Path(line.replace("\\", "/")).name
+            if not filename:
+                continue
+            position += 1
+            tracks.append({
+                "position": str(position),
+                "filename": filename,
+                "title": pending_title or "",
+            })
+            pending_title = None
+    return tracks
+
+
 def _read_tracklist_file(path: Path) -> Optional[str]:
     try:
         if path.suffix.lower() in (".htm", ".html"):
@@ -203,6 +233,29 @@ def extract_hints(scan: AlbumScan) -> AlbumHints:
 
     parsed_tracklist = _parse_tracklist(hints.tracklist_text) if hints.tracklist_text else []
 
+    # M3U/Playlist-Reihenfolge: filename (stem, normalisiert) → Tracknummer
+    m3u_order: Dict[str, int] = {}
+    m3u_titles: Dict[str, str] = {}
+    for pf in scan.playlist_files:
+        try:
+            text = pf.read_text(encoding="utf-8", errors="replace")
+            for entry in _parse_m3u(text):
+                stem = _clean(Path(entry["filename"]).stem).casefold()
+                pos = int(entry["position"])
+                if stem and stem not in m3u_order:
+                    m3u_order[stem] = pos
+                    if entry.get("title"):
+                        m3u_titles[stem] = entry["title"]
+        except Exception as e:
+            print(f"  ⚠️ Playlist-Lesefehler {pf.name}: {e}", file=sys.stderr)
+
+    # Tracklist-Lookup: normalisierter Titel → Eintrag (für titelbasiertes Matching)
+    tl_by_title: Dict[str, Dict[str, str]] = {}
+    for entry in parsed_tracklist:
+        key = _clean(entry.get("title", "")).casefold()
+        if key:
+            tl_by_title[key] = entry
+
     # Build TrackHints per audio file
     for audio_path in sorted(scan.audio_files):
         tags, duration = _read_tags(audio_path)
@@ -215,10 +268,18 @@ def extract_hints(scan: AlbumScan) -> AlbumHints:
         raw_tn = tags.get("tracknumber") or fn_hints.get("track")
         if raw_tn:
             try:
-                track_num = int(str(raw_tn).split("/")[0])
+                tn_int = int(str(raw_tn).split("/")[0])
+                if tn_int > 0:   # 0 gilt als "keine Nummer"
+                    track_num = tn_int
             except ValueError:
                 pass
 
+        # Track number aus M3U-Reihenfolge (Vorrang vor Dateiname, aber nicht vor Tag)
+        if track_num is None:
+            stem_key = _clean(audio_path.stem).casefold()
+            if stem_key in m3u_order:
+                track_num = m3u_order[stem_key]
+
         # Disc number: tag > filename > path segment
         raw_dn = tags.get("discnumber") or fn_hints.get("disc")
         if raw_dn:
@@ -236,16 +297,27 @@ def extract_hints(scan: AlbumScan) -> AlbumHints:
         title = tags.get("title") or fn_hints.get("title")
         artist = tags.get("artist") or fn_hints.get("artist")
 
-        # Enrich from parsed tracklist if track_num matches
-        if parsed_tracklist and track_num:
-            for tl_entry in parsed_tracklist:
-                tl_track = tl_entry.get("track")
-                tl_disc = tl_entry.get("disc", "1")
-                if (tl_track and int(tl_track) == track_num
-                        and int(tl_disc) == (disc_num or 1)):
-                    if not _is_good(title) and _is_good(tl_entry.get("title")):
-                        title = tl_entry["title"]
-                    break
+        # Tracklist: erst nach Nummer, dann nach Titel
+        if parsed_tracklist:
+            matched_tl: Optional[Dict[str, str]] = None
+            if track_num:
+                for tl_entry in parsed_tracklist:
+                    tl_track = tl_entry.get("track")
+                    tl_disc = tl_entry.get("disc", "1")
+                    if (tl_track and int(tl_track) == track_num
+                            and int(tl_disc) == (disc_num or 1)):
+                        matched_tl = tl_entry
+                        break
+            if matched_tl is None and title:
+                matched_tl = tl_by_title.get(_clean(title).casefold())
+            if matched_tl and not _is_good(title) and _is_good(matched_tl.get("title")):
+                title = matched_tl["title"]
+
+        # M3U-Titel als Fallback (enthält "Composer - Title" — nur nutzen wenn kein besserer Titel)
+        if not _is_good(title):
+            stem_key = _clean(audio_path.stem).casefold()
+            if stem_key in m3u_titles:
+                title = m3u_titles[stem_key]
 
         hints.tracks.append(TrackHints(
             path=audio_path,