modify the logic of response_audio and output binding to make sure current lyrics to be correctly set up

app.py

@@ -680,7 +680,7 @@ def build_interface():
         )
 
         respond_audio = input_audio.stop_recording(
-            response_audio, [state, input_audio, genre, mood, theme], [state, chatbot]
+            response_audio, [state, input_audio, genre, mood, theme], [state, chatbot, lyrics_display]
         )
 
         restart = respond_audio.then(start_recording_user, [state], [input_audio]).then(
@@ -859,7 +859,7 @@ def response_audio(state: AppState, audio: tuple, genre_value, mood_value, theme
     for msg in state.conversation:
         conversation_display.append({"role": msg["role"], "content": msg["content"]})
     
-    return state, conversation_display
+    return state, conversation_display, state.lyrics
 
 def extract_lyrics_from_conversation(conversation):
     """