app.py

import streamlit as st 
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.document_loaders import PyPDFLoader
from transformers import T5Tokenizer, T5ForConditionalGeneration
from transformers import pipeline
import torch
import base64
import tempfile
#import os
#from dotenv import load_dotenv
#from huggingface_hub import HfApi

#api = HfApi()
#token = api.retrieve_token("secret_token")  # Replace with your secret name

#load_dotenv()
#token = os.environ.get("HF_TOKEN")




checkpoint = "MBZUAI/LaMini-Flan-T5-248M"
#model and tokenizer loading

tokenizer = T5Tokenizer.from_pretrained(checkpoint)
base_model = T5ForConditionalGeneration.from_pretrained(checkpoint, device_map='auto', torch_dtype=torch.float32)

#file loader and preprocessing
def file_preprocessing(file):
    loader =  PyPDFLoader(file)
    pages = loader.load_and_split()
    text_splitter = RecursiveCharacterTextSplitter(chunk_size=200, chunk_overlap=50)
    texts = text_splitter.split_documents(pages)
    final_texts = ""
    for text in texts:
        print(text)
        final_texts = final_texts + text.page_content
    return final_texts

#LLM pipeline
def llm_pipeline(filepath):
    pipe_sum = pipeline(
        'summarization',
        model = base_model,
        tokenizer = tokenizer,
        max_length = 500, 
        min_length = 50)
    input_text = file_preprocessing(filepath)
    result = pipe_sum(input_text)
    result = result[0]['summary_text']
    return result




def main():
    st.title("Document Summarization App")

    uploaded_file = st.file_uploader("Upload your PDF file", type=['pdf'])

    if uploaded_file is not None:
        if st.button("Summarize"):
            col2 = st.columns(1)
            # Use a temporary filename directly
            with tempfile.NamedTemporaryFile(delete=False) as temp_file:
                temp_file.write(uploaded_file.read())
                temp_file.flush()  # Ensure contents are written to disk
                filepath = temp_file.name
            
            with col2:
                summary = llm_pipeline(filepath)
                st.info("Summarization Complete")
                st.success(summary)
            # Clean up the temporary file
            os.remove(filepath)




if __name__ == "__main__":
    main()