import streamlit as st
from transformers import pipeline
from huggingface_hub import login
from dotenv import load_dotenv
import os

# Load the environment variables from the .env file
load_dotenv()

# Retrieve the token from the .env file
huggingface_token = os.getenv("HUGGINGFACE_TOKEN")

# Log in using the retrieved token
login(token=huggingface_token)

# Available models for summarization
models = {
    "T5": "Sandaruth/T5_Full_Fine_Tuned_FINDSUM",
    "BERT": "bert-base-uncased",  # Note: BERT isn't designed for summarization; you can change this
    "LongT5": "google/long-t5-local-base",
    "Pegasus": "google/pegasus-xsum"
}

# Streamlit app layout
st.title("Summarization with Multiple Models")

# Dropdown to select the model
model_choice = st.selectbox("Select a model for summarization", models.keys())

# Text area for input
input_text = st.text_area("Enter the long text you want to summarize", height=300)

# Button to generate the summary
if st.button("Generate Summary"):
    # Show a spinner while generating the summary
    with st.spinner("Generating summary, please wait..."):
        # Load the selected model and summarizer pipeline
        summarizer = pipeline("summarization", model=models[model_choice])
        
        # Log the model choice
        st.write(f"Using model: **{model_choice}** for summarization.")
        
        if input_text:
            # Generate the summary
            summary = summarizer(input_text, max_length=350, min_length=30, do_sample=False)
            
            # Log the success message
            st.success("Summary generated successfully!")
            
            # Display the summary
            st.subheader("Generated Summary")
            st.write(summary[0]['summary_text'])
        else:
            st.warning("Please enter text to summarize!")

# Optionally, you can add a footer or additional instructions
st.markdown("---")
st.write("Provide a long text and select a model to see the summarization in action!")