Update app.py

app.py

@@ -23,12 +23,24 @@ def generate_voice_response(text):
     tts.save(temp_file.name)
     return temp_file.name
 
-def restaurant_voice_assistant(input_text, state_json):
+def restaurant_voice_assistant(audio, state_json):
     global cart
     state = json.loads(state_json) if state_json else {}
     response = ""
     voice_path = None
 
+    # Convert audio input to text
+    if audio:
+        import speech_recognition as sr
+        recognizer = sr.Recognizer()
+        with sr.AudioFile(audio) as source:
+            try:
+                input_text = recognizer.recognize_google(recognizer.record(source))
+            except sr.UnknownValueError:
+                input_text = ""
+    else:
+        input_text = ""
+
     if not state.get("menu_shown", False):
         # Show menu dynamically
         response = "Welcome to our restaurant! Here is our menu:\n"
@@ -36,7 +48,7 @@ def restaurant_voice_assistant(input_text, state_json):
             response += f"{item}: ${price:.2f}\n"
         response += "\nPlease tell me the item you would like to add to your cart."
         state["menu_shown"] = True
-    elif "add_to_cart" in input_text.lower():
+    elif "add to cart" in input_text.lower():
         item = input_text.replace("add to cart", "").strip()
         if item in menu_items:
             cart.append(item)
@@ -67,13 +79,13 @@ with gr.Blocks() as demo:
     state = gr.State(value=json.dumps({}))
     
     with gr.Row():
-        user_input = gr.Textbox(label="Your Voice Input", placeholder="Type here or use voice input.")
+        user_audio = gr.Audio(source="microphone", type="filepath", label="Your Voice Input")
         output_text = gr.Textbox(label="Response Text")
     
     with gr.Row():
-        voice_output = gr.Audio(label="Response Audio")
+        voice_output = gr.Audio(label="Response Audio", autoplay=True)
 
     submit_btn = gr.Button("Submit")
-    submit_btn.click(restaurant_voice_assistant, inputs=[user_input, state], outputs=[output_text, voice_output, state])
+    submit_btn.click(restaurant_voice_assistant, inputs=[user_audio, state], outputs=[output_text, voice_output, state])
 
 demo.launch()