Update app.py

app.py

@@ -1,7 +1,8 @@
 import gradio as gr
 from gtts import gTTS
-import tempfile
 import os
+import tempfile
+import json
 import speech_recognition as sr
 
 # Store cart in a temporary storage
@@ -23,84 +24,76 @@ def generate_voice_response(text):
     tts.save(temp_file.name)
     return temp_file.name
 
-def process_audio(audio_path):
+def calculate_total(cart):
+    return sum(menu_items[item] for item in cart)
+
+def restaurant_voice_assistant(audio, state_json):
     global cart
-    recognizer = sr.Recognizer()
+    state = json.loads(state_json) if state_json else {}
     response = ""
-
-    try:
-        with sr.AudioFile(audio_path) as source:
-            audio = recognizer.record(source)
-            input_text = recognizer.recognize_google(audio)
-            print("User said:", input_text)
-
-            if "menu" in input_text.lower():
-                response = "Here is our menu:\n"
-                for item, price in menu_items.items():
-                    response += f"{item}: ${price:.2f}\n"
-                response += "\nWhat would you like to add to your cart?"
-
-            elif any(item.lower() in input_text.lower() for item in menu_items):
-                for item in menu_items:
-                    if item.lower() in input_text.lower():
-                        cart.append(item)
-                        total = sum(menu_items[cart_item] for cart_item in cart)
-                        response = f"{item} has been added to your cart. Your current cart includes:\n"
-                        for cart_item in cart:
-                            response += f"- {cart_item}: ${menu_items[cart_item]:.2f}\n"
-                        response += f"\nTotal: ${total:.2f}. Would you like to add anything else?"
-                        break
-
-            elif "final order" in input_text.lower() or "submit order" in input_text.lower():
-                if cart:
-                    total = sum(menu_items[cart_item] for cart_item in cart)
-                    response = "Your final order includes:\n"
-                    for item in cart:
-                        response += f"- {item}: ${menu_items[item]:.2f}\n"
-                    response += f"\nTotal: ${total:.2f}. Thank you for ordering!"
-                    cart = []  # Clear cart after finalizing order
-                else:
-                    response = "Your cart is empty. Would you like to order something?"
-
-            elif "stop" in input_text.lower():
-                response = "Stopping voice assistant."
-
-            else:
-                response = "I didn’t quite catch that. Please tell me what you’d like to order."
-
-    except sr.UnknownValueError:
-        response = "Sorry, I didn’t catch that. Could you please repeat?"
-    except Exception as e:
-        response = f"An error occurred: {str(e)}"
-
-    audio_response_path = generate_voice_response(response)
-    return response, audio_response_path
-
-def run_voice_assistant(audio):
-    response, audio_path = process_audio(audio)
-    return response, (audio_path,)
-
-def javascript_autoplay():
-    return """<script>
-    document.querySelector("audio").addEventListener("loadeddata", function() {
-        this.play();
-    });
-    </script>"""
+    voice_path = None
+
+    # Convert audio input to text
+    if audio:
+        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"
+        for item, price in menu_items.items():
+            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 any(item.lower() in input_text.lower() for item in menu_items):
+        # Check if input matches a menu item
+        for item in menu_items:
+            if item.lower() in input_text.lower():
+                cart.append(item)
+                total = calculate_total(cart)
+                response = f"{item} has been added to your cart. Your current cart includes:\n"
+                for cart_item in cart:
+                    response += f"- {cart_item}: ${menu_items[cart_item]:.2f}\n"
+                response += f"\nTotal: ${total:.2f}. Would you like to add anything else?"
+                break
+    elif "menu" in input_text.lower():
+        response = "Here is our menu again:\n"
+        for item, price in menu_items.items():
+            response += f"{item}: ${price:.2f}\n"
+        response += "\nWhat would you like to add to your cart?"
+    elif "final order" in input_text.lower() or "submit order" in input_text.lower():
+        if cart:
+            total = calculate_total(cart)
+            response = "Your final order includes:\n"
+            for item in cart:
+                response += f"- {item}: ${menu_items[item]:.2f}\n"
+            response += f"\nTotal: ${total:.2f}.\nThank you for ordering!"
+            cart = []  # Clear cart after finalizing order
+        else:
+            response = "Your cart is empty. Would you like to order something?"
+    else:
+        response = "I didn’t quite catch that. Please tell me what you’d like to order."
+
+    voice_path = generate_voice_response(response)
+    return response, voice_path, json.dumps(state)
 
 with gr.Blocks() as demo:
-    gr.Markdown("### Voice-Activated Restaurant Assistant")
+    state = gr.State(value=json.dumps({}))
 
     with gr.Row():
-        voice_input = gr.Audio(type="filepath", label="Speak Now")
-        assistant_response = gr.Textbox(label="Assistant Response")
-
-    gr.Markdown("Note: The audio will play automatically.")
+        user_audio = gr.Audio(type="filepath", label="Your Voice Input")
+        output_text = gr.Textbox(label="Response Text")
 
     with gr.Row():
-        audio_output = gr.Audio(label="Audio Response")
-
-    voice_input.change(run_voice_assistant, inputs=voice_input, outputs=[assistant_response, audio_output])
+        voice_output = gr.Audio(label="Response Audio", autoplay=True)
 
-    gr.HTML(javascript_autoplay())
+    # Automatically process audio when recording stops
+    user_audio.change(restaurant_voice_assistant, inputs=[user_audio, state], outputs=[output_text, voice_output, state])
 
 demo.launch()