app.py

# app.py
import os
import gradio as gr
from groq import Groq
from gtts import gTTS
import tempfile
import whisper



# Initialize Groq client
GROQ_API_KEY = "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"
client = Groq(api_key = GROQ_API_KEY)


# Load the Whisper model
whisper_model = whisper.load_model("base")  # You can use "small", "medium", or "large" depending on your preference

# Function to convert audio to text using OpenAI Whisper
def audio_to_text(audio_file):
    audio = whisper.load_audio(audio_file)
    audio = whisper.pad_or_trim(audio)
    mel = whisper.log_mel_spectrogram(audio).to(whisper_model.device)
    options = whisper.DecodingOptions(fp16=False)
    result = whisper.decode(whisper_model, mel, options)
    return result.text

# Function to interact with Groq API and generate a response
def interact_with_groq(user_input):
    try:
        chat_completion = client.chat.completions.create(
            messages=[{"role": "user", "content": user_input}],
            model="llama3-8b-8192",  # Use the appropriate model
            stream=False,
        )
        return chat_completion.choices[0].message.content
    except Exception as e:
        return f"Error interacting with Groq API: {e}"

# Function to convert text to speech using gTTS
def text_to_audio(response_text):
    tts = gTTS(response_text)
    output_path = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3").name
    tts.save(output_path)
    return output_path

# Main function for the chatbot
def voice_to_voice(audio_file):
    try:
        # Step 1: Convert voice input to text
        print("Transcribing audio...")
        transcribed_text = audio_to_text(audio_file)
        print(f"Transcribed Text: {transcribed_text}")

        # Step 2: Interact with LLM via Groq API
        print("Getting LLM response...")
        response_text = interact_with_groq(transcribed_text)
        print(f"LLM Response: {response_text}")

        # Step 3: Convert LLM response to audio
        print("Generating audio response...")
        audio_response = text_to_audio(response_text)
        return transcribed_text, audio_response
    except Exception as e:
        return f"Error processing request: {e}", None

# Gradio Interface
interface = gr.Interface(
    fn=voice_to_voice,
    inputs=gr.Audio(type="filepath"),
    outputs=[gr.Textbox(label="Transcribed Text"), gr.Audio(label="Response Audio")],
    title="Real-Time Voice-to-Voice Chatbot",
    description="A real-time voice-to-voice chatbot using Whisper for transcription, Groq API for LLM, and gTTS for audio response.",
)

# Launch the interface
if __name__ == "__main__":
    interface.launch()