import streamlit as st
# From here down is all the StreamLit UI.
st.set_page_config(page_title="LangChain Demo", page_icon=":robot:")  # Call it first
st.header("Hey, I'm your Chat GPT")

import torch

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Define SystemMessage class if needed
class SystemMessage:
    def __init__(self, content):
        self.content = content


# Choose your desired free model from the Hugging Face Hub
model_name = "t5-small"  # Replace with your choice (e.g.,t5-small or facebook/bart-base or EleutherAI/gpt-neo-125M)

# Load the model and tokenizer

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

system_message = SystemMessage("Welcome to my Chatbot! Ask me anything.")
st.write(system_message)


if "sessionMessages" not in st.session_state:
     st.session_state.sessionMessages = [
        SystemMessage(content="You are a helpful assistant.")
    ]

class HumanMessage:
    def __init__(self, content):
        self.content = content

class AIMessage:
    def __init__(self, content):
        self.content = content


def load_answer(question):
    st.session_state.sessionMessages.append(HumanMessage(content=question))

    inputs = tokenizer(question, return_tensors="pt")
    outputs = model.generate(**inputs)
    response_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

    # Create AIMessage object after defining response_text
    assistant_answer = AIMessage(content=response_text)
    st.session_state.sessionMessages.append(assistant_answer)

    return response_text

def get_text():
    input_text = st.text_input("You: ", key= input)
    return input_text





user_input=get_text()
submit = st.button('Generate')  

if submit:
    
    response = load_answer(user_input)
    st.subheader("Answer:")

    st.write(response,key= 1)