import gradio as gr
import subprocess
import os
from huggingface_hub import InferenceClient

# Initialize Chatbot Model (Futuresony.gguf)
chat_client = InferenceClient("Futuresony/future_ai_12_10_2024.gguf")  # Change if needed

def asr_chat_tts(audio):
    """
    1. Convert Speech to Text using asr.py
    2. Process text through Chat Model (Futuresony.gguf)
    3. Convert response to Speech using tts.py
    """
    # Step 1: Run ASR (Speech-to-Text)
    asr_output = subprocess.run(["python3", "asr.py", audio], capture_output=True, text=True)
    transcription = asr_output.stdout.strip()

    # Step 2: Process text through the chat model
    messages = [{"role": "system", "content": "You are a helpful AI assistant."}]
    messages.append({"role": "user", "content": transcription})

    response = ""
    for msg in chat_client.chat_completion(messages, max_tokens=512, stream=True):
        token = msg.choices[0].delta.content
        response += token

    # Step 3: Run TTS (Text-to-Speech)
    tts_output_file = "generated_speech.wav"
    subprocess.run(["python3", "tts.py", response, tts_output_file])

    return transcription, response, tts_output_file

# Gradio Interface
with gr.Blocks() as demo:
    gr.Markdown("<h2 style='text-align: center;'>ASR → Chatbot → TTS</h2>")
    
    with gr.Row():
        audio_input = gr.Audio(source="microphone", type="filepath", label="🎤 Speak Here")
        text_transcription = gr.Textbox(label="📝 Transcription", interactive=False)
        text_response = gr.Textbox(label="🤖 Chatbot Response", interactive=False)
        audio_output = gr.Audio(label="🔊 Generated Speech")

    submit_button = gr.Button("Process Speech 🔄")

    submit_button.click(fn=asr_chat_tts, inputs=[audio_input], outputs=[text_transcription, text_response, audio_output])

# Run the App
if __name__ == "__main__":
    demo.launch()