import streamlit as st
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

# Load model and tokenizer
model_name = "lora_adapter"  # Update this to your LoRA model path

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")

# Chat function with prompt-based filtering
def chat(instruction):
    prompt = f"""You are a helpful and expert Python programming tutor.
If the question is about Python, explain clearly with examples.
If the question is unrelated to Python, respond with "Sorry, I can only answer Python-related questions."

### Instruction:
{instruction}

### Response:
"""
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_new_tokens=150,
            temperature=0.7,
            top_p=0.95,
            do_sample=True,
            pad_token_id=tokenizer.eos_token_id
        )
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response.split("### Response:")[-1].strip()

# Streamlit UI
st.set_page_config(page_title="Python Tutor Chatbot", page_icon="🐍")
st.title("🐍 Python Tutor Chatbot")
st.write("Ask me Python programming questions!")

user_input = st.text_input("Your question:")

if user_input:
    with st.spinner("Generating response..."):
        response = chat(user_input)
        st.markdown("**Answer:**")
        st.markdown(response)