AIVoice15 / app.py
geethareddy's picture
Update app.py
e8c5011 verified
import os
import speech_recognition as sr
import logging
from flask import Flask, render_template_string, request, jsonify
from tempfile import NamedTemporaryFile
import ffmpeg
from fuzzywuzzy import process, fuzz
# Initialize Flask app
app = Flask(__name__)
logging.basicConfig(level=logging.INFO)
# Global variables
cart = [] # Stores items as [item_name, price, quantity] in the cart
menu_preferences = None # Tracks the current menu preference
section_preferences = None # Tracks the current section preference
default_menu_preferences = "all" # To reset menu preferences
default_sections = {
"biryanis": ["veg biryani", "paneer biryani", "chicken biryani", "mutton biryani"],
"starters": ["samosa", "onion pakoda", "chilli gobi", "chicken manchurian", "veg manchurian"],
"curries": ["paneer butter", "chicken curry", "fish curry", "chilli chicken"],
"desserts": ["gulab jamun", "ice cream"],
"soft drinks": ["cola", "lemon soda"]
}
prices = {
"samosa": 9,
"onion pakoda": 10,
"chilli gobi": 12,
"chicken biryani": 14,
"mutton biryani": 16,
"veg biryani": 12,
"paneer butter": 10,
"fish curry": 12,
"chicken manchurian": 14,
"veg manchurian": 12,
"chilli chicken": 14,
"paneer biryani": 13,
"chicken curry": 14,
"gulab jamun": 8,
"ice cream": 6,
"cola": 5,
"lemon soda": 6
}
menus = {
"all": list(prices.keys()),
"vegetarian": [
"samosa", "onion pakoda", "chilli gobi", "veg biryani", "paneer butter", "veg manchurian", "paneer biryani", "gulab jamun", "ice cream", "cola", "lemon soda"
],
"non-vegetarian": [
"chicken biryani", "mutton biryani", "fish curry", "chicken manchurian", "chilli chicken", "chicken curry", "gulab jamun", "ice cream", "cola", "lemon soda"
]
}
@app.route("/")
def index():
return render_template_string(html_code)
@app.route("/reset-cart", methods=["GET"])
def reset_cart():
global cart, menu_preferences, section_preferences
cart = []
menu_preferences = None
section_preferences = None
return "Cart reset successfully."
@app.route("/process-audio", methods=["POST"])
def process_audio():
try:
# Handle audio input
audio_file = request.files.get("audio")
if not audio_file:
return jsonify({"response": "Oops! I didn't catch any audio. Please try again."}), 400
# Save and convert audio to WAV format
temp_file = NamedTemporaryFile(delete=False, suffix=".webm")
audio_file.save(temp_file.name)
converted_file = NamedTemporaryFile(delete=False, suffix=".wav")
ffmpeg.input(temp_file.name).output(
converted_file.name, acodec="pcm_s16le", ac=1, ar="16000"
).run(overwrite_output=True)
# Recognize speech
recognizer = sr.Recognizer()
recognizer.dynamic_energy_threshold = True
recognizer.energy_threshold = 4000 # Increased sensitivity
with sr.AudioFile(converted_file.name) as source:
recognizer.adjust_for_ambient_noise(source, duration=1)
audio_data = recognizer.record(source)
# Use multiple recognition services with fallbacks
try:
raw_command = recognizer.recognize_google(audio_data).lower()
except sr.UnknownValueError:
raw_command = "Sorry, I couldn't understand that."
except sr.RequestError as e:
raw_command = f"Request error from the service: {e}"
logging.info(f"User said: {raw_command}") # Print user speech in the console
# Display the transcribed text and AI voice response
response = process_command(raw_command)
except Exception as e:
response = f"An error occurred: {str(e)}"
finally:
os.unlink(temp_file.name)
os.unlink(converted_file.name)
return jsonify({"response": response})
def preprocess_command(command):
"""
Normalize the user command to improve matching.
"""
command = command.strip().lower()
return command
def process_command(command):
global cart, menu_preferences, section_preferences
# Finalize order
if "final order" in command or "complete order" in command:
if not cart:
return "Your cart is empty. Please add items before finalizing the order."
order_summary = "\n".join([f"{item[2]} x {item[0]} for {item[1] * item[2]} INR" for item in cart])
total_price = sum(item[1] * item[2] for item in cart)
cart.clear() # Clear the cart after finalizing
menu_preferences = None
section_preferences = None
return f"Your order has been placed successfully:\n{order_summary}\nTotal: {total_price} INR.\nThank you for ordering!"
# Greet the user and ask for preferences when first started
if menu_preferences is None:
if "hello" in command or "hi" in command:
return "Hello, welcome to Biryani Hub! Please choose your preference: All, Vegetarian, or Non-Vegetarian."
preferences = ["non-vegetarian", "vegetarian", "all"]
if command in preferences:
menu_preferences = command
return f"You've selected the {command} menu! Which section would you like to browse next? (e.g., biryanis, starters, curries, desserts, soft drinks)"
# Use fuzzy matching to help recognize similar inputs
closest_match = process.extractOne(command, preferences, scorer=fuzz.partial_ratio)
if closest_match and closest_match[1] > 75:
menu_preferences = closest_match[0]
return f"Great choice! You've chosen the {menu_preferences} menu. Which section would you like to browse next?"
return "I couldn't recognize your choice. Please say either 'Non-Vegetarian', 'Vegetarian', or 'All'."
if section_preferences is None:
sections = list(default_sections.keys())
for section in sections:
if section in command:
section_preferences = section
return f"Here are the items in the {section_preferences} section: {', '.join(default_sections[section_preferences])}. Please choose an item."
closest_match = process.extractOne(command, sections, scorer=fuzz.partial_ratio)
if closest_match and closest_match[1] > 75:
section_preferences = closest_match[0]
return f"Here are the items in the {section_preferences} section: {', '.join(default_sections[section_preferences])}. What would you like to add?"
return "I didn't catch that. Please say a section like 'biryanis', 'starters', 'curries', 'desserts', or 'soft drinks'."
# Filter items based on the menu preference (vegetarian/non-vegetarian)
available_items = []
if menu_preferences == "vegetarian":
available_items = [item for item in default_sections[section_preferences] if item in menus["vegetarian"]]
elif menu_preferences == "non-vegetarian":
available_items = [item for item in default_sections[section_preferences] if item in menus["non-vegetarian"]]
elif menu_preferences == "all":
available_items = [item for item in default_sections[section_preferences]]
for item in available_items:
if item in command:
quantity = extract_quantity(command)
if quantity:
cart.append([item, prices[item], quantity])
return f"Added {quantity} x {item} to your cart. Your current cart: {', '.join([f'{i[0]} x{i[2]}' for i in cart])}. Would you like to add more items?"
return "I didn't recognize the item you mentioned. Please say the item name clearly, or choose from the available items."
def extract_quantity(command):
"""
Extract quantity from the command (e.g., 'two', '3', '5').
"""
number_words = {
"one": 1, "two": 2, "three": 3, "four": 4, "five": 5,
"six": 6, "seven": 7, "eight": 8, "nine": 9, "ten": 10,
"1": 1, "2": 2, "3": 3, "4": 4, "5": 5, "6": 6, "7": 7, "8": 8, "9": 9, "10": 10
}
command_words = command.split()
for word in command_words:
if word in number_words:
return number_words[word]
return None
html_code = """
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>AI Dining Assistant</title>
<style>
body {
display: flex;
flex-direction: column;
align-items: center;
justify-content: center;
min-height: 100vh;
margin: 0;
font-family: Arial, sans-serif;
background-color: #f4f4f9;
}
h1 {
color: #333;
}
.mic-button {
font-size: 2rem;
padding: 1rem 2rem;
color: white;
background-color: #007bff;
border: none;
border-radius: 50px;
cursor: pointer;
transition: background-color 0.3s;
}
.mic-button:hover {
background-color: #0056b3;
}
.status, .response {
margin-top: 1rem;
text-align: center;
color: #555;
font-size: 1.2rem;
}
.response {
background-color: #e8e8ff;
padding: 1rem;
border-radius: 10px;
box-shadow: 0 4px 8px rgba(0, 0, 0, 0.2);
display: none;
}
</style>
</head>
<body>
<h1>AI Dining Assistant</h1>
<button class="mic-button" id="mic-button">🎤</button>
<div class="status" id="status">Press the mic button to start...</div>
<div class="response" id="response">Response will appear here...</div>
<script>
const micButton = document.getElementById('mic-button');
const status = document.getElementById('status');
const response = document.getElementById('response');
let mediaRecorder;
let audioChunks = [];
let isConversationActive = false;
micButton.addEventListener('click', () => {
if (!isConversationActive) {
isConversationActive = true;
startConversation();
}
});
function startConversation() {
const utterance = new SpeechSynthesisUtterance('Hello, welcome to Biryani Hub! Please choose your preference: All, Vegetarian, or Non-Vegetarian.');
speechSynthesis.speak(utterance);
utterance.onend = () => {
status.textContent = 'Listening...';
startListening();
};
}
function startListening() {
navigator.mediaDevices.getUserMedia({ audio: true }).then(stream => {
mediaRecorder = new MediaRecorder(stream, { mimeType: 'audio/webm;codecs=opus' });
mediaRecorder.start();
audioChunks = [];
mediaRecorder.ondataavailable = event => audioChunks.push(event.data);
mediaRecorder.onstop = async () => {
const audioBlob = new Blob(audioChunks, { type: 'audio/webm' });
const formData = new FormData();
formData.append('audio', audioBlob);
status.textContent = 'Processing...';
try {
const result = await fetch('/process-audio', { method: 'POST', body: formData });
const data = await result.json();
response.textContent = 'You said: ' + data.response; // Display user text
response.style.display = 'block';
const utterance = new SpeechSynthesisUtterance(data.response);
speechSynthesis.speak(utterance);
utterance.onend = () => {
console.log("Speech synthesis completed.");
if (data.response.includes("final order") || data.response.includes("Thank you for ordering")) {
status.textContent = 'Order completed. Press the mic button to start again.';
isConversationActive = false;
} else {
status.textContent = 'Listening...';
setTimeout(() => {
startListening();
}, 100);
}
};
utterance.onerror = (e) => {
console.error("Speech synthesis error:", e.error);
status.textContent = 'Error with speech output.';
isConversationActive = false;
};
} catch (error) {
response.textContent = 'Sorry, I could not understand. Please try again.';
response.style.display = 'block';
status.textContent = 'Press the mic button to restart the conversation.';
isConversationActive = false;
}
};
setTimeout(() => mediaRecorder.stop(), 5000);
}).catch(() => {
status.textContent = 'Microphone access denied.';
isConversationActive = false;
});
}
</script>
</body>
</html>
"""
if __name__ == "__main__":
app.run(host="0.0.0.0", port=7860)