fix youtube video url

src/handlers/transcription_handler.py

@@ -29,52 +29,121 @@ def transcribe_file(file_path):
     logging.info(f"File transcription result: {transcript[:100]}...")
     return transcript
 
-def transcribe_video_url(url):
-    logging.info(f"transcribe_video_url called with: {url}")
-    if not url:
-        return "No URL provided"
 
-    # Define a clean output path
-    # Using a fixed name in a temp directory prevents 'ffprobe' access errors
-    output_dir = "temp_downloads"
-    os.makedirs(output_dir, exist_ok=True)
-    temp_audio_path = os.path.join(output_dir, "youtube_audio.wav")
 
-    ffmpeg_bin = get_ffmpeg_path()
+import os
+import logging
+import yt_dlp
+import tempfile
+
+def transcribe_audio(source=None, is_url=False):
+    """
+    Transcribe audio either from a file upload or from a YouTube URL.
+    
+    Parameters:
+        source (str): File path if `is_url=False` or YouTube URL if `is_url=True`.
+        is_url (bool): True if source is a YouTube URL, False if it's a local file.
+        
+    Returns:
+        str: Transcription text or error message.
+    """
+    logging.info(f"transcribe_audio called with source: {source}, is_url={is_url}")
     
-    ydl_opts = {
-        'format': 'bestaudio/best',
-        # FIX: yt-dlp needs the template to NOT include the extension if using ExtractAudio
-        'outtmpl': os.path.join(output_dir, 'youtube_audio'), 
-        'ffmpeg_location': ffmpeg_bin, # Explicitly tell yt-dlp where ffmpeg is
-        'postprocessors': [{
-            'key': 'FFmpegExtractAudio',
-            'preferredcodec': 'wav',
-            'preferredquality': '192',
-        }],
-        'quiet': False, # Set to False temporarily to debug if it fails again
-    }
+    if not source:
+        return "No source provided"
     
     try:
-        logging.info("Downloading and extracting audio")
-        with yt_dlp.YoutubeDL(ydl_opts) as ydl:
-            ydl.download([url])
+        if is_url:
+            # YouTube download mode
+            ffmpeg_bin = get_ffmpeg_path()
+            
+            with tempfile.TemporaryDirectory() as temp_dir:
+                temp_audio_path = os.path.join(temp_dir, "youtube_audio.wav")
+                
+                ydl_opts = {
+                    'format': 'bestaudio/best',
+                    'outtmpl': os.path.join(temp_dir, 'youtube_audio'),
+                    'ffmpeg_location': ffmpeg_bin,
+                    'postprocessors': [{
+                        'key': 'FFmpegExtractAudio',
+                        'preferredcodec': 'wav',
+                        'preferredquality': '192',
+                    }],
+                    'quiet': False,
+                }
+                
+                logging.info("Downloading and extracting audio from YouTube")
+                with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+                    ydl.download([source])
+                
+                if not os.path.exists(temp_audio_path):
+                    raise FileNotFoundError(f"FFmpeg failed to create {temp_audio_path}")
+                
+                logging.info("Audio downloaded, transcribing")
+                transcript = transcribe_file(temp_audio_path)
+                return transcript
+
+        else:
+            # Local file mode (works on Spaces and locally)
+            if not os.path.exists(source):
+                return f"File does not exist: {source}"
+            
+            logging.info("Transcribing local audio file")
+            transcript = transcribe_file(source)
+            return transcript
+    
+    except Exception as e:
+        logging.error(f"Error processing audio: {str(e)}")
+        return f"Error processing audio: {str(e)}"
+
+
+
+# def transcribe_video_url(url):
+#     logging.info(f"transcribe_video_url called with: {url}")
+#     if not url:
+#         return "No URL provided"
+
+#     # Define a clean output path
+#     # Using a fixed name in a temp directory prevents 'ffprobe' access errors
+#     output_dir = "temp_downloads"
+#     os.makedirs(output_dir, exist_ok=True)
+#     temp_audio_path = os.path.join(output_dir, "youtube_audio.wav")
+
+#     ffmpeg_bin = get_ffmpeg_path()
+    
+#     ydl_opts = {
+#         'format': 'bestaudio/best',
+#         # FIX: yt-dlp needs the template to NOT include the extension if using ExtractAudio
+#         'outtmpl': os.path.join(output_dir, 'youtube_audio'), 
+#         'ffmpeg_location': ffmpeg_bin, # Explicitly tell yt-dlp where ffmpeg is
+#         'postprocessors': [{
+#             'key': 'FFmpegExtractAudio',
+#             'preferredcodec': 'wav',
+#             'preferredquality': '192',
+#         }],
+#         'quiet': False, # Set to False temporarily to debug if it fails again
+#     }
+    
+#     try:
+#         logging.info("Downloading and extracting audio")
+#         with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+#             ydl.download([url])
         
-        # After post-processing, the file will be 'youtube_audio.wav'
-        final_wav_path = temp_audio_path
+#         # After post-processing, the file will be 'youtube_audio.wav'
+#         final_wav_path = temp_audio_path
         
-        if not os.path.exists(final_wav_path):
-            raise FileNotFoundError(f"FFmpeg failed to create {final_wav_path}")
+#         if not os.path.exists(final_wav_path):
+#             raise FileNotFoundError(f"FFmpeg failed to create {final_wav_path}")
 
-        logging.info("Audio downloaded, transcribing")
-        transcript = transcribe_file(final_wav_path)
+#         logging.info("Audio downloaded, transcribing")
+#         transcript = transcribe_file(final_wav_path)
         
-        # Clean up
-        if os.path.exists(final_wav_path):
-            os.remove(final_wav_path)
+#         # Clean up
+#         if os.path.exists(final_wav_path):
+#             os.remove(final_wav_path)
             
-        return transcript
+#         return transcript
 
-    except Exception as e:
-        logging.error(f"Error processing video: {str(e)}")
-        return f"Error processing video: {str(e)}"
+#     except Exception as e:
+#         logging.error(f"Error processing video: {str(e)}")
+#         return f"Error processing video: {str(e)}"