app.py

import streamlit as st
from transformers import pipeline
import textwrap

st.title('Hugging Face BERT Summarizer')

# List of models
models = ["sshleifer/distilbart-cnn-12-6", "facebook/bart-large-cnn", "t5-base", "t5-large", "google/pegasus-newsroom"]

# Dropdown model selector
model = st.sidebar.selectbox("Choose a model", models)

uploaded_file = st.file_uploader("Choose a .txt file", type="txt")

if uploaded_file is not None:
    user_input = uploaded_file.read().decode('utf-8')
    total_length = len(user_input.split())

    # Add sliders to the sidebar
    min_length_percentage = st.sidebar.slider('Minimum Length %', min_value=10, max_value=100, value=50)
    max_length_percentage = min_length_percentage + 10
    st.sidebar.text(f'Maximum Length %: {max_length_percentage}')

    if st.button('Summarize'):
        summarizer = pipeline('summarization', model=model)
        summarized_text = ""

        # Split the text into chunks of approximately 500 words each
        chunks = textwrap.wrap(user_input, 500)
        
        # Summarize each chunk
        for chunk in chunks:
            min_length = max(int(total_length * min_length_percentage / 100), 1)  # Calculate min_length based on the percentage of the total length
            max_length = int(total_length * max_length_percentage / 100)  # Calculate max_length based on the percentage of the total length
            summarized = summarizer(chunk, max_length=max_length, min_length=min_length, do_sample=False)
            summarized_text += summarized[0]['summary_text'] + " "

        st.text_area('Summarized Text', summarized_text, height=200)