import gradio as gr
import speech_recognition as sr
import torch
from transformers import pipeline

# 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 Capture Name
def capture_name(audio):
    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):
    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("### 🎙️ Welcome to Biryani Hub")

        with gr.Column():
            gr.Markdown("#### Step 1: Tell me your name")
            audio_input_name = gr.Audio(type="filepath", label="Record your Name")
            name_output = gr.Textbox(label="Your Name:")
            email_prompt_output = gr.Textbox(label="Next Step:", interactive=False)
            audio_input_name.change(capture_name, inputs=audio_input_name, outputs=[name_output, email_prompt_output])

            gr.Markdown("#### Step 2: Provide your email")
            audio_input_email = gr.Audio(type="filepath", label="Record your Email")
            email_output = gr.Textbox(label="Your Email:")
            audio_input_email.change(capture_email, inputs=audio_input_email, outputs=email_output)

    return demo

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