app.py

import streamlit as st
from transformers import pipeline
import pandas as pd

# Load the text summarization pipeline
summarizer = pipeline("summarization", model="astro21/bart-cls_n")

chunk_counter = 0


def summarize_text(input_text):
    global chunk_counter
    chunk_counter = 0

    max_chunk_size = 1024
    chunks = [input_text[i:i + max_chunk_size] for i in range(0, len(input_text), max_chunk_size)]

    summarized_chunks = []
    chunk_lengths = []
    summarized_chunks_only = []

    for chunk in chunks:
        chunk_counter += 1
        summarized_chunk = summarizer(chunk, max_length=128, min_length=64, do_sample=False)[0]['summary_text']
        summarized_chunks.append(f"Chunk {chunk_counter}:\n{summarized_chunk}")
        summarized_chunks_only.append(summarized_chunk)

        chunk_lengths.append(len(chunk))

    summarized_text = "\n".join(summarized_chunks)
    summarized_text_only = "\n".join(summarized_chunks_only)

    # Save the merged summary to a file
    with open("summarized.txt", "w") as output_file:
        output_file.write(summarized_text_only)

    chunk_df = pd.DataFrame({'Chunk Number': range(1, chunk_counter + 1), 'Chunk Length': chunk_lengths})

    return summarized_text_only, chunk_df, "summarized.txt"


def summarize_text_file(file):
    if file is not None:
        content = str(file.read(), 'utf-8')
        return summarize_text(content)


st.title("Text Summarization")
st.write("Summarize text using BART")

uploaded_file = st.file_uploader("Upload a text file", type=["txt"])

if uploaded_file is not None:
    result = summarize_text_file(uploaded_file)
    st.subheader("Summarized Text")
    st.write(result[0])

    st.subheader("Chunk Information")
    st.write(result[1])