Kapitel 13/Tutorial.md aktualisiert

Kapitel 13/Tutorial.md

@@ -1352,55 +1352,91 @@ source "$ENV_FILE"
 VOD_ID="${1:?VOD-ID muss übergeben werden}"
 MODEL="${WHISPER_MODEL:-small}"
 
+LOG_FILE="/srv/clipper/logs/$VOD_ID/subtitle.log"
+mkdir -p "$(dirname "$LOG_FILE")"
+exec >> "$LOG_FILE" 2>&1
+
 echo "[INFO] Starte Untertitelung für VOD: $VOD_ID mit Modell: $MODEL"
 
-# === Pfade setzen ===
+# Pfade
 VOD_OUT="/srv/clipper/out/$VOD_ID"
 CLIPS_ROOT="$VOD_OUT/clips"
 WAV_DIR="/srv/clipper/temp/$VOD_ID/whisper_wav"
-LOG_FILE="/srv/clipper/logs/$VOD_ID/subtitle.log"
-
 mkdir -p "$WAV_DIR"
-exec >> "$LOG_FILE" 2>&1
 
-echo "[INFO] WAV-Verzeichnis: $WAV_DIR"
+# Vorprüfung: Whisper installiert?
-echo "[INFO] Clipverzeichnis: $CLIPS_ROOT"
+if ! /srv/clipper/.venv/bin/python3 -c "import whisper" 2>/dev/null; then
+  echo "[FATAL] Python-Modul 'whisper' fehlt im venv." >&2
+  exit 1
+fi
 
-# === Whisper auf alle Clips anwenden ===
+for clip_folder in "$CLIPS_ROOT"/*; do
-find "$CLIPS_ROOT" -mindepth 1 -maxdepth 1 -type d | sort | while read -r CLIP_FOLDER; do
+  [ -d "$clip_folder" ] || continue
-  CLIP_PATH=$(find "$CLIP_FOLDER" -maxdepth 1 -type f -name "*.mp4" | head -n1)
+  clip_file="$(find "$clip_folder" -maxdepth 1 -name '*.mp4' | head -n1)"
+  [ -f "$clip_file" ] || { echo "[WARN] Kein Clip in $clip_folder – überspringe"; continue; }
 
-  if [[ ! -f "$CLIP_PATH" ]]; then
+  basename="$(basename "$clip_file" .mp4)"
-    echo "[WARN] Kein Clip gefunden in $CLIP_FOLDER – überspringe"
+  wav_file="$WAV_DIR/$basename.wav"
-    continue
+  srt_file="$clip_folder/$basename.srt"
-  fi
+  json_file="$clip_folder/$basename.json"
+  txt_file="$clip_folder/$basename.txt"
 
-  BASENAME="$(basename "$CLIP_PATH" .mp4)"
+  echo "[INFO] → Clip: $basename"
-  WAV_FILE="$WAV_DIR/$BASENAME.wav"
-  SRT_FILE="$CLIP_FOLDER/$BASENAME.srt"
-  JSON_FILE="$CLIP_FOLDER/$BASENAME.json"
-  TXT_FILE="$CLIP_FOLDER/$BASENAME.txt"
 
-  echo "[INFO] → Clip: $BASENAME"
+  if [ ! -f "$wav_file" ]; then
-
-  if [[ ! -f "$WAV_FILE" ]]; then
     echo "[INFO] Extrahiere WAV..."
-    ffmpeg -hide_banner -loglevel error -y -i "$CLIP_PATH" -vn -ac 1 -ar 16000 -f wav "$WAV_FILE"
+    ffmpeg -hide_banner -loglevel error -y \
+      -i "$clip_file" -vn -ac 1 -ar 16000 "$wav_file"
   fi
 
-  echo "[INFO] Starte Whisper..."
+  echo "[INFO] Starte Whisper-Transkription für $basename..."
-  whisper "$WAV_FILE" \
+
-    --model "$MODEL" \
+  # Python-Block korrekt im venv ausführen
-    --output_format srt \
+  /srv/clipper/.venv/bin/python3 -u <<EOF >> "$LOG_FILE" 2>&1
-    --output_format json \
+import os
-    --output_format txt \
+import whisper
-    --output_dir "$CLIP_FOLDER" \
+import json
-    --fp16 False
+from pathlib import Path
+import logging
+
+print = lambda *a, **k: __builtins__.print(*a, flush=True, **k)
+
+logger = logging.getLogger("subtitle_logger")
+logger.setLevel(logging.INFO)
+handler = logging.StreamHandler()
+handler.setFormatter(logging.Formatter("[%(levelname)s] %(message)s"))
+logger.addHandler(handler)
+
+MODEL = os.environ.get("MODEL", "$MODEL")
+wav_path = Path("$wav_file")
+srt_path = Path("$srt_file")
+json_path = Path("$json_file")
+txt_path = Path("$txt_file")
+
+logger.info(f"Modell laden: {MODEL}")
+model = whisper.load_model(MODEL)
+
+result = model.transcribe(str(wav_path), fp16=False)
+
+logger.info("Schreibe SRT...")
+with srt_path.open("w", encoding="utf-8") as f:
+    for seg in result["segments"]:
+        f.write(f"{seg['id']+1}\\n{seg['start']:.3f} --> {seg['end']:.3f}\\n{seg['text'].strip()}\\n\\n")
+
+logger.info("Schreibe JSON...")
+with json_path.open("w", encoding="utf-8") as f:
+    json.dump(result, f, ensure_ascii=False, indent=2)
+
+logger.info("Schreibe TXT...")
+with txt_path.open("w", encoding="utf-8") as f:
+    f.write(result["text"].strip())
+
+logger.info(f"[DONE] $basename abgeschlossen.")
+EOF
 
-  echo "[DONE] Untertitel erstellt: $SRT_FILE, $JSON_FILE, $TXT_FILE"
 done
 
-echo "[DONE] Untertitel abgeschlossen."
+echo "[DONE] Untertitelung abgeschlossen für VOD $VOD_ID."
 ```
 Mit 
 ```bash