From 1960989eefb1b50972634d0d9405f700632f586a Mon Sep 17 00:00:00 2001
From: dschlueter <dschlueter@kitux.de>
Date: Wed, 29 Apr 2026 05:57:27 +0200
Subject: [PATCH] Fix YouTube ID detection: use last _-token instead of broken
 lookbehind regex

The previous regex lookbehind (?<![A-Za-z0-9_-]) excluded _ as valid preceding
character, so IDs after underscores were never matched. New approach: split stem
by _ and check if the last token is an 11-char YouTube ID (mixed case + digit).
Also strips the ID token from the stem before _parse_filename() to prevent it
from leaking into the track title or being misread as an artist-title separator.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 hint_extractor.py | 38 ++++++++++++++++++++++++--------------
 1 file changed, 24 insertions(+), 14 deletions(-)

diff --git a/hint_extractor.py b/hint_extractor.py
index 433c71d..a5dcced 100755
--- a/hint_extractor.py
+++ b/hint_extractor.py
@@ -321,18 +321,21 @@ def _check_cover_images(paths: List[Path]) -> List[Path]:
     return good
 
 
-# YouTube-Video-ID: 11 Zeichen aus [A-Za-z0-9_-], eingebettet im Dateinamen
-_YT_ID_RE = re.compile(r"(?<![A-Za-z0-9_-])([A-Za-z0-9_-]{11})(?![A-Za-z0-9_-])")
+# YouTube-Video-ID: exakt 11 Zeichen aus [A-Za-z0-9_-], typischerweise letztes _-Token
+_YT_ID_CHARS = re.compile(r"^[A-Za-z0-9_-]{11}$")
 
 
 def _extract_youtube_id(path: Path) -> Optional[str]:
-    """Sucht eine YouTube-Video-ID im Dateinamen (Stem oder Suffix)."""
-    name = path.stem + path.suffix
-    for m in _YT_ID_RE.finditer(name):
-        candidate = m.group(1)
-        # Einfache Plausibilitätsprüfung: muss gemischte Zeichen haben
-        if re.search(r"[A-Z]", candidate) and re.search(r"[0-9a-z]", candidate):
-            return candidate
+    """
+    Erkennt YouTube-Video-ID als letztes '_'-getrenntes Token im Dateinamen.
+    Plausibilitätsprüfung: mind. ein Großbuchstabe UND mind. ein Kleinbuchstabe/Ziffer.
+    """
+    candidate = path.stem.split("_")[-1]   # letztes Token nach Unterstrich
+    if (len(candidate) == 11
+            and _YT_ID_CHARS.match(candidate)
+            and re.search(r"[A-Z]", candidate)
+            and re.search(r"[0-9a-z]", candidate)):
+        return candidate
     return None
 
 
@@ -470,8 +473,18 @@ def extract_hints(scan: AlbumScan, use_ocr: bool = True) -> AlbumHints:
 
     # Build TrackHints per audio file
     for audio_path in sorted(scan.audio_files):
+        stem_key = _clean(audio_path.stem).casefold()
+        yt_id_for_file = yt_ids_by_stem.get(stem_key)
+
+        # Stem ohne YouTube-ID für Dateiname-Parsing
+        parse_stem = audio_path.stem
+        if yt_id_for_file:
+            tokens = parse_stem.rsplit("_", 1)
+            if len(tokens) == 2 and tokens[1] == yt_id_for_file:
+                parse_stem = tokens[0]
+
         tags, duration = _read_tags(audio_path)
-        fn_hints = _parse_filename(audio_path.stem)
+        fn_hints = _parse_filename(parse_stem)
 
         track_num: Optional[int] = None
         disc_num: Optional[int] = None
@@ -549,20 +562,17 @@ def extract_hints(scan: AlbumScan, use_ocr: bool = True) -> AlbumHints:
 
         # M3U-Reihenfolge nur als letzter Fallback (wenn Tracklist kein Match liefert)
         if track_num is None:
-            stem_key = _clean(audio_path.stem).casefold()
             if stem_key in m3u_order:
                 track_num = m3u_order[stem_key]
 
         # 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]
 
         # YouTube-Titel als letzter Fallback (bei einzelner Datei = das ganze Video)
         if not _is_good(title):
-            stem_key = _clean(audio_path.stem).casefold()
-            yt_id = yt_ids_by_stem.get(stem_key)
+            yt_id = yt_id_for_file
             if yt_id:
                 meta = yt_meta_by_id.get(yt_id)
                 if meta: