app.py

# Import necessary libraries
import whisper
from gtts import gTTS
import os
import tempfile
import gradio as gr
from groq import Groq
from dotenv import load_dotenv

# Load environment variables from .env file (if available)
load_dotenv()

# Set up Groq API key (fallback method if not set as an environment variable)
if "GROQ_API_KEY" not in os.environ:
    os.environ["GROQ_API_KEY"] = "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"  # Replace with your Groq API key

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

# Load the Whisper model
whisper_model = whisper.load_model("base")  # "base" is lightweight; use "large" for higher accuracy

# Define chatbot function
def voice_to_voice_chatbot(input_audio):
    """
    Transcribes the audio input using Whisper, queries Groq's LLM API, 
    converts the response to audio, and returns the audio file.
    """
    # Step 1: Transcribe audio to text using Whisper
    try:
        transcription_result = whisper_model.transcribe(input_audio)
        user_query = transcription_result["text"]
        print("User Query:", user_query)
    except Exception as e:
        return f"Error in transcription: {str(e)}"

    # Step 2: Query the Groq API with the transcribed text
    try:
        chat_completion = client.chat.completions.create(
            messages=[{"role": "user", "content": user_query}],
            model="llama3-8b-8192",
            stream=False
        )
        bot_response = chat_completion.choices[0].message.content
        print("Bot Response:", bot_response)
    except Exception as e:
        return f"Error in LLM response: {str(e)}"

    # Step 3: Convert the LLM response to speech using gTTS
    try:
        tts = gTTS(text=bot_response, lang="en")
        output_audio_path = tempfile.NamedTemporaryFile(suffix=".mp3").name
        tts.save(output_audio_path)
    except Exception as e:
        return f"Error in text-to-speech conversion: {str(e)}"

    # Return the audio file path
    return output_audio_path

# Create Gradio interface for the chatbot
interface = gr.Interface(
    fn=voice_to_voice_chatbot,
    inputs=gr.Audio(type="filepath"),  # Input: File path of recorded audio
    outputs=gr.Audio(type="filepath"), # Output: File path of the generated audio
    live=True,                         # Enable real-time interaction
    description="This Voice to Voice Chatbot is created by Rizwan Sajad using OpenAI Whisper, Groq API, and gTTS."
)

# Launch Gradio interface
interface.launch()