app.py

import os
import gradio as gr
import whisper
from gtts import gTTS
import io
from groq import Groq

# Initialize the Groq client
client = Groq(api_key=os.environ.get("GROQ_API_KEY"))

# Load the Whisper model
model = whisper.load_model("base")

def process_audio(file_path):
    try:
        # Load the audio file
        audio = whisper.load_audio(file_path)

        # Transcribe the audio using Whisper
        result = model.transcribe(audio)
        text = result["text"]

        # Generate a response using Groq
        chat_completion = client.chat.completions.create(
            messages=[{"role": "user", "content": text}],
            model="llama3-8b-8192",  # Replace with the correct model if necessary
        )

        # Access the response using dot notation
        response_message = chat_completion.choices[0].message.content.strip()

        # Convert the response text to speech
        tts = gTTS(response_message)
        response_audio_io = io.BytesIO()
        tts.write_to_fp(response_audio_io)  # Save the audio to the BytesIO object
        response_audio_io.seek(0)

        # Save audio to a file to ensure it's generated correctly
        response_audio_path = "response.mp3"
        with open(response_audio_path, "wb") as audio_file:
            audio_file.write(response_audio_io.getvalue())

        # Return the response text and the path to the saved audio file
        return response_message, response_audio_path

    except Exception as e:
        return f"An error occurred: {e}", None

# Create the Gradio interface with customized UI
with gr.Blocks() as demo:
    gr.Markdown(
        """
        <style>
        .gradio-container {
            font-family: Arial, sans-serif;
            background-color: #f0f4c3;  /* Light green background color */
            border-radius: 10px;
            padding: 20px;
            box-shadow: 0 4px 12px rgba(0,0,0,0.2);
            text-align: center;
        }
        .gradio-input, .gradio-output {
            border-radius: 6px;
            border: 1px solid #ddd;
            padding: 10px;
        }
        .gradio-button {
            background-color: #ff7043;
            color: white;
            border-radius: 6px;
            border: none;
            padding: 10px 20px;  /* Adjusted padding */
            font-size: 16px;  /* Adjusted font size */
            cursor: pointer;
        }
        .gradio-button:hover {
            background-color: #e64a19;
        }
        .gradio-title {
            font-size: 28px;
            font-weight: bold;
            margin-bottom: 20px;
            color: #37474f;
        }
        .gradio-description {
            font-size: 16px;
            margin-bottom: 20px;
            color: #616161;
        }
        </style>
        """
    )

    gr.Markdown("# Voice-to-Voice Chatbot\nDeveloped by Salman Maqbool ❤️")
    gr.Markdown("Upload an audio file to interact with the voice-to-voice chatbot. The chatbot will transcribe the audio, generate a response, and provide a spoken reply.")

    with gr.Row():
        with gr.Column():
            audio_input = gr.Audio(type="filepath", label="Upload Audio File")
            submit_button = gr.Button("Submit")
        
        with gr.Column():
            response_text = gr.Textbox(label="Response Text", placeholder="The AI-generated response will appear here", lines=5)
            response_audio = gr.Audio(label="Response Audio", type="filepath")

    # Link the submit button to the process_audio function
    submit_button.click(fn=process_audio, inputs=audio_input, outputs=[response_text, response_audio])

# Launch the Gradio app
demo.launch()