Update app.py

app.py

@@ -12,6 +12,7 @@ import requests
 import mimetypes
 import urllib.parse
 import subprocess
+import json
 
 # Configure logging
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
@@ -94,6 +95,15 @@ def transcribe_audio_chunks(chunks):
             os.unlink(temp_audio_file.name)
     return ' '.join(transcriptions)
 
+def get_file_info(file_path):
+    try:
+        result = subprocess.run(['ffprobe', '-v', 'quiet', '-print_format', 'json', '-show_format', '-show_streams', file_path], 
+                                capture_output=True, text=True, check=True)
+        return json.loads(result.stdout)
+    except subprocess.CalledProcessError as e:
+        logger.error(f"Error getting file info: {str(e)}")
+        return None
+
 def download_file(url):
     local_filename = url.split('/')[-1]
     with requests.get(url, stream=True) as r:
@@ -126,10 +136,27 @@ def process_media(file_path, is_url=False):
             temp_file = temp_file.name
             logger.info(f"Uploaded file saved: {temp_file}")
 
+        # Get file info
+        file_info = get_file_info(temp_file)
+        if not file_info:
+            return "Unable to process file: Could not determine file type", False
+
+        # Determine if it's audio or video
+        is_audio = any(stream['codec_type'] == 'audio' for stream in file_info['streams'])
+        is_video = any(stream['codec_type'] == 'video' for stream in file_info['streams'])
+
         # Convert to WAV using ffmpeg
         wav_path = tempfile.NamedTemporaryFile(delete=False, suffix='.wav').name
         try:
-            subprocess.run(['ffmpeg', '-i', temp_file, '-acodec', 'pcm_s16le', '-ar', '44100', wav_path], check=True)
+            if is_video:
+                # Extract audio from video
+                subprocess.run(['ffmpeg', '-i', temp_file, '-vn', '-acodec', 'pcm_s16le', '-ar', '44100', '-ac', '2', wav_path], check=True)
+            elif is_audio:
+                # Convert audio to WAV
+                subprocess.run(['ffmpeg', '-i', temp_file, '-acodec', 'pcm_s16le', '-ar', '44100', '-ac', '2', wav_path], check=True)
+            else:
+                return "Unsupported file type: Neither audio nor video detected", False
+            
             logger.info(f"Audio extracted to WAV: {wav_path}")
         except subprocess.CalledProcessError as e:
             logger.error(f"FFmpeg conversion failed: {str(e)}")
@@ -157,7 +184,7 @@ def process_media(file_path, is_url=False):
             os.unlink(temp_file)
         if wav_path and os.path.exists(wav_path):
             os.unlink(wav_path)
-
+            
 @app.callback(
     [Output('output-media-upload', 'children'),
      Output('transcription-status', 'children'),