fix: handle tuple audio input in transcribe function and skip empty tuples

app.py

@@ -29,6 +29,14 @@ def transcribe(audio, state=""):
     print(f"Received audio input of type: {type(audio)}")
     print(f"Audio shape: {audio.shape if isinstance(audio, np.ndarray) else 'N/A'}")
     # Append NumPy array to buffer
+    if isinstance(audio, tuple):
+        print(f"Tuple contents: {audio}")
+        # Try extracting the first element
+        audio = audio[0] if len(audio) > 0 else None
+        if not audio:
+            print("Empty tuple, skipping")
+            return state, state
+        
     if isinstance(audio, np.ndarray):
         audio_buffer.append(audio)
         # Process if buffer has enough data (e.g., 5 seconds at 16kHz)