import streamlit as st
from transformers import pipeline
import torch 

MODEL_PATH = "HuggingFaceH4/zephyr-7b-beta"

def main():
    st.title("Chatbot with Hugging Face Model")

    # Check if the model is already saved locally
    try:
        pipe = pipeline("text-generation", model=MODEL_PATH, torch_dtype=torch.bfloat16, device_map="auto")
    except:
        # If not saved, load the model and save it
        st.warning("Model not found locally. Downloading and saving the model. Please wait...")
        pipe = pipeline("text-generation", model="HuggingFaceH4/zephyr-7b-beta", torch_dtype=torch.bfloat16, device_map="auto")
        pipe.save_pretrained(MODEL_PATH)

    # Define chat messages
    messages = [
        {"role": "system", "content": "You are a friendly chatbot who always responds in the style of a pirate"},
        {"role": "user", "content": st.text_input("User Input", "How many helicopters can a human eat in one sitting?")},
    ]

    # Generate response
    prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)

    # Display generated text
    st.text(outputs[0]["generated_text"])

if __name__ == "__main__":
    main()