Initial commit: workspace setup with skills, memory, config

skills/local-whisper-stt/scripts/telegram_voice_handler.py

@@ -0,0 +1,96 @@
+#!/usr/bin/env python3
+"""
+Handle Telegram voice messages - download and transcribe
+Usage: telegram_voice_handler.py <bot_token> <file_id> [--model MODEL]
+"""
+
+import argparse
+import os
+import sys
+import json
+import urllib.request
+import tempfile
+
+def download_voice_file(bot_token, file_id, output_path):
+    """Download voice file from Telegram"""
+    
+    # Step 1: Get file path from Telegram
+    file_info_url = f"https://api.telegram.org/bot{bot_token}/getFile?file_id={file_id}"
+    
+    try:
+        with urllib.request.urlopen(file_info_url) as response:
+            data = json.loads(response.read().decode())
+            if not data.get("ok"):
+                print(f"Error getting file info: {data}", file=sys.stderr)
+                sys.exit(1)
+            
+            file_path = data["result"]["file_path"]
+    except Exception as e:
+        print(f"Error fetching file info: {e}", file=sys.stderr)
+        sys.exit(1)
+    
+    # Step 2: Download the actual file
+    download_url = f"https://api.telegram.org/file/bot{bot_token}/{file_path}"
+    
+    try:
+        urllib.request.urlretrieve(download_url, output_path)
+        return output_path
+    except Exception as e:
+        print(f"Error downloading file: {e}", file=sys.stderr)
+        sys.exit(1)
+
+def transcribe_with_whisper(audio_path, model_size="base"):
+    """Transcribe using local Faster-Whisper"""
+    
+    from faster_whisper import WhisperModel
+    
+    # Load model (cached after first use)
+    model = WhisperModel(model_size, device="cpu", compute_type="int8")
+    
+    # Transcribe
+    segments, info = model.transcribe(audio_path, beam_size=5)
+    
+    # Collect text
+    full_text = []
+    for segment in segments:
+        full_text.append(segment.text.strip())
+    
+    return {
+        "text": " ".join(full_text),
+        "language": info.language,
+        "language_probability": info.language_probability
+    }
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Download and transcribe Telegram voice message")
+    parser.add_argument("bot_token", help="Telegram bot token")
+    parser.add_argument("file_id", help="Telegram voice file_id")
+    parser.add_argument("--model", default="base",
+                       choices=["tiny", "base", "small", "medium", "large"],
+                       help="Whisper model size (default: base)")
+    
+    args = parser.parse_args()
+    
+    # Allow override from environment
+    model = os.environ.get("WHISPER_MODEL", args.model)
+    
+    # Create temp file for download
+    with tempfile.NamedTemporaryFile(suffix=".ogg", delete=False) as tmp:
+        temp_path = tmp.name
+    
+    try:
+        # Download
+        print(f"Downloading voice file...", file=sys.stderr)
+        download_voice_file(args.bot_token, args.file_id, temp_path)
+        
+        # Transcribe
+        print(f"Transcribing with {model} model...", file=sys.stderr)
+        result = transcribe_with_whisper(temp_path, model)
+        
+        # Output result
+        print(json.dumps(result))
+        
+    finally:
+        # Cleanup
+        if os.path.exists(temp_path):
+            os.remove(temp_path)