import base64 import random import numpy as np import speech_recognition as sr import edge_tts import tempfile import json # Load the food menu from menu.json def load_food_menu(): with open("menu.json") as f: return json.load(f) food_menu = load_food_menu() class AudioProcessor: @staticmethod def encode_audio(data, sample_rate): encoded = base64.b64encode(data.tobytes()).decode("UTF-8") return { "realtimeInput": { "mediaChunks": [ { "mimeType": f"audio/pcm;rate={sample_rate}", "data": encoded, } ], }, } @staticmethod def process_audio_response(data): audio_data = base64.b64decode(data) return np.frombuffer(audio_data, dtype=np.int16) class OrderAssistant: def __init__(self, food_menu): self.food_menu = food_menu self.audio_processor = AudioProcessor() def get_food_suggestion(self, food_type, dietary_preference): if food_type in self.food_menu: return random.choice(self.food_menu[food_type].get(dietary_preference, [])) return "Sorry, we don't have that option." def recognize_speech_from_mic(self): recognizer = sr.Recognizer() with sr.Microphone() as source: print("Say something!") audio = recognizer.listen(source) try: return recognizer.recognize_google(audio) # Google Speech Recognition except sr.UnknownValueError: return "Sorry I didn't catch that" except sr.RequestError: return "Sorry, I'm having trouble reaching the service" async def text_to_speech(self, text): # Change pitch from '0Hz' to a valid value, e.g., '+0Hz' communicate = edge_tts.Communicate(text, "en-US-AriaNeural", rate="+0%", pitch="+0Hz") # Changed '0Hz' to '+0Hz' with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file: tmp_path = tmp_file.name await communicate.save(tmp_path) return tmp_path async def food_order_assistant(self, audio_input, food_type): greeting_text = "Welcome to the restaurant! What would you like to order today? Are you looking for Vegan, Halal, or Guilt-Free options?" audio_path = await self.text_to_speech(greeting_text) dietary_preference = self.recognize_speech_from_mic().lower() if dietary_preference not in ["vegan", "halal", "guilt-free"]: dietary_preference = "vegan" # Default to Vegan if unrecognized suggestion = self.get_food_suggestion(food_type, dietary_preference) suggestion_text = f"I suggest you try {suggestion}. Does that sound good?" suggestion_audio = await self.text_to_speech(suggestion_text) confirmation = self.recognize_speech_from_mic().lower() if "yes" in confirmation: confirmation_text = f"Your order for {suggestion} has been confirmed. Sending the order to the kitchen." confirmation_audio = await self.text_to_speech(confirmation_text) return confirmation_audio, f"Confirmed order: {suggestion}" else: cancellation_text = "Okay, let's try again." cancellation_audio = await self.text_to_speech(cancellation_text) return cancellation_audio, "Order not confirmed"