app.py

import gradio as gr
import speech_recognition as sr
import torch
import os
from transformers import pipeline
from gtts import gTTS
import time

# Load ASR Model (Whisper)
device = "cuda" if torch.cuda.is_available() else "cpu"
speech_to_text = pipeline("automatic-speech-recognition", model="openai/whisper-base", device=0 if device == "cuda" else -1)

# Initialize Speech Recognition
recognizer = sr.Recognizer()

# Function to Play Audio Prompt
def play_audio(text):
    tts = gTTS(text=text, lang='en')
    filename = "prompt.mp3"
    tts.save(filename)
    os.system(f"mpg321 {filename}" if os.name != "nt" else f"start {filename}")  # Works on Linux & Windows
    time.sleep(2)  # Give some time for the speech to play

# Function to Capture Name
def capture_name(audio):
    play_audio("Tell me your name")
    try:
        text = speech_to_text(audio)["text"]
        return f"👤 Name Captured: {text}", "Please provide your email address."
    except Exception as e:
        return f"❌ Error: {str(e)}", ""

# Function to Capture Email
def capture_email(audio):
    play_audio("Please provide your email address")
    try:
        text = speech_to_text(audio)["text"]
        return f"📧 Email Captured: {text}"
    except Exception as e:
        return f"❌ Error: {str(e)}"

# Gradio Interface
def gradio_interface():
    with gr.Blocks() as demo:
        gr.Markdown("<h1 style='text-align: center;'>🍽️ AI Dining Assistant</h1>")
        
        with gr.Column():
            gr.Image("/mnt/data/image.png", elem_id="header_image", show_label=False)  # Upload the image you provided
            gr.Markdown("<p style='text-align: center;'>Press the mic button to start...</p>")

            gr.Markdown("#### 🎤 Step 1: Tell me your name")
            mic_button = gr.Button("🎙️ Tap to Speak Your Name")
            audio_input_name = gr.Audio(type="filepath", visible=False)
            name_output = gr.Textbox(label="Your Name:")
            email_prompt_output = gr.Textbox(label="Next Step:", interactive=False)

            mic_button.click(capture_name, inputs=audio_input_name, outputs=[name_output, email_prompt_output])

            gr.Markdown("#### 🎤 Step 2: Provide your email")
            mic_button_email = gr.Button("🎙️ Tap to Speak Your Email")
            audio_input_email = gr.Audio(type="filepath", visible=False)
            email_output = gr.Textbox(label="Your Email:")

            mic_button_email.click(capture_email, inputs=audio_input_email, outputs=email_output)

    return demo

# Launch the Gradio Interface
demo = gradio_interface()
demo.launch(debug=True)