Update app.py

app.py

@@ -79,25 +79,24 @@ async def text_to_speech(text):
         await communicate.save(tmp_path)
     return tmp_path
 
-def chatbot(audio, history):
+def process_audio(audio, history):
     if audio is None:
-        return "", history
+        return history, None
     
     text = speech_to_text(audio)
     response = generate(text, history)
     history.append((text, response))
     
-    return "", history
+    return history, response
 
-def text_input(text, history):
+def process_text(text, history):
     response = generate(text, history)
     history.append((text, response))
-    return "", history
+    return history, response
 
-async def generate_audio(history):
-    if len(history) > 0:
-        last_response = history[-1][1]
-        audio_path = await text_to_speech(last_response)
+async def generate_audio(response):
+    if response:
+        audio_path = await text_to_speech(response)
         return audio_path
     return None
 
@@ -106,14 +105,31 @@ with gr.Blocks() as demo:
     gr.Markdown("# Mood-Based Music Recommender with Voice Chat")
     
     chatbot = gr.Chatbot()
-    audio_input = gr.Audio(sources="microphone", type="filepath")
+    audio_input = gr.Audio(source="microphone", type="filepath")
     text_input = gr.Textbox(placeholder="Type your message here...")
     audio_output = gr.Audio(label="AI Response")
 
-    audio_input.change(chatbot, inputs=[audio_input, chatbot], outputs=[audio_input, chatbot])
-    text_input.submit(text_input, inputs=[text_input, chatbot], outputs=[text_input, chatbot])
-    
-    chatbot.change(generate_audio, inputs=[chatbot], outputs=[audio_output])
+    state = gr.State([])
+
+    audio_input.change(
+        process_audio, 
+        inputs=[audio_input, state], 
+        outputs=[chatbot, state]
+    ).then(
+        generate_audio,
+        inputs=[state],
+        outputs=[audio_output]
+    )
+
+    text_input.submit(
+        process_text, 
+        inputs=[text_input, state], 
+        outputs=[chatbot, state]
+    ).then(
+        generate_audio,
+        inputs=[state],
+        outputs=[audio_output]
+    )
 
 if __name__ == "__main__":
     demo.launch()