app.py

import os
import gradio as gr
from openai import OpenAI
import tempfile
import torch
import whisper

# Load API key from environment
GROQ_API_KEY = os.environ.get("GROQ_API_KEY")

# Initialize client for Groq-compatible OpenAI API
client = OpenAI(
    api_key=GROQ_API_KEY,
    base_url="https://api.groq.com/openai/v1"
)

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

# Chat history storage
chat_history = []

def chat_with_bot(message, history):
    global chat_history
    chat_history = history or []

    # Append user message
    chat_history.append({"role": "user", "content": message})

    # Get model response
    response = client.chat.completions.create(
        model="llama3-8b-8192",
        messages=chat_history
    )
    reply = response.choices[0].message.content

    # Append assistant reply
    chat_history.append({"role": "assistant", "content": reply})

    return chat_history, chat_history

def transcribe_audio(audio_file):
    if audio_file is None:
        return ""
    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=torch.cuda.is_available())
    result = whisper.decode(whisper_model, mel, options)
    return result.text

with gr.Blocks() as demo:
    gr.Markdown("# 🤖 Neobot - Chat with Voice, File & Text")

    chatbot = gr.Chatbot(type="messages", label="Chat")
    state = gr.State([])

    with gr.Row():
        txt = gr.Textbox(placeholder="Type a message or upload audio/file...", show_label=False)
        send_btn = gr.Button("Send")

    with gr.Row():
        audio_upload = gr.Audio(type="filepath", label="Upload Audio")  # Removed `source=`
        transcribe_btn = gr.Button("Transcribe Audio")

    send_btn.click(chat_with_bot, inputs=[txt, state], outputs=[chatbot, state])
    transcribe_btn.click(transcribe_audio, inputs=audio_upload, outputs=txt)

demo.launch()