import gradio as gr
from huggingface_hub import InferenceClient
from transformers import pipeline
from datasets import load_dataset
import soundfile as sf
import torch

# Initialize the chat model
chat_client = InferenceClient("Futuresony/future_ai_12_10_2024.gguf")

# Initialize the TTS pipeline
tts_synthesizer = pipeline("text-to-speech", model="Futuresony/Output")

# Load the speaker embeddings dataset
embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
speaker_embedding = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)

def chat_with_tts(message, history, system_message, max_tokens, temperature, top_p):
    # Step 1: Generate response using the chat model
    messages = [{"role": "system", "content": system_message}]
    
    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})
    
    messages.append({"role": "user", "content": message})
    
    response = ""
    for msg in chat_client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = msg.choices[0].delta.content
        response += token

    # Step 2: Generate speech using TTS
    speech = tts_synthesizer(response, forward_params={"speaker_embeddings": speaker_embedding})
    output_file = "generated_speech.wav"
    sf.write(output_file, speech["audio"], samplerate=speech["sampling_rate"])
    
    # Update the chat history
    history.append((message, response))
    
    # Return both text response, audio file, and updated history
    return response, output_file, history

# Create the Gradio interface
demo = gr.Interface(
    fn=chat_with_tts,
    inputs=[
        gr.Textbox(label="User Input", placeholder="Type your message..."),
        gr.State([]),  # Initialize history as an empty list
        gr.Textbox(value="You are a friendly chatbot.", label="System Message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max New Tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
    ],
    outputs=[
        gr.Textbox(label="Generated Response"),
        gr.Audio(label="Generated Speech"),
        gr.State(),  # Add State as an output to update the history
    ],
    title="Chat with TTS",
    description="Enter text to chat with an AI chatbot. The chatbot will generate a response, which will also be converted to speech using TTS."
)

if __name__ == "__main__":
    demo.launch()