Code Editor for app.py

import streamlit as st
from PyPDF2 import PdfReader
from pdf2image import convert_from_bytes
import tempfile

st.title("PDF Comparison Tool")

# Create two columns for PDF uploads
col1, col2 = st.columns(2)

with col1:
    st.subheader("First PDF")
    pdf1 = st.file_uploader("Upload first PDF", type=['pdf'], key='pdf1')
    
    if pdf1 is not None:
        # Read PDF content
        pdf_reader1 = PdfReader(pdf1)
        text1 = ""
        for page in pdf_reader1.pages:
            text1 += page.extract_text()
        
        # Display PDF preview
        with tempfile.NamedTemporaryFile(suffix='.pdf') as tmp_file:
            tmp_file.write(pdf1.getvalue())
            images = convert_from_bytes(pdf1.getvalue())
            if images:
                st.image(images[0], caption='First page preview', use_column_width=True)
        
        # Display extracted text
        with st.expander("Show extracted text"):
            st.text(text1)

with col2:
    st.subheader("Second PDF")
    pdf2 = st.file_uploader("Upload second PDF", type=['pdf'], key='pdf2')
    
    if pdf2 is not None:
        # Read PDF content
        pdf_reader2 = PdfReader(pdf2)
        text2 = ""
        for page in pdf_reader2.pages:
            text2 += page.extract_text()
        
        # Display PDF preview
        with tempfile.NamedTemporaryFile(suffix='.pdf') as tmp_file:
            tmp_file.write(pdf2.getvalue())
            images = convert_from_bytes(pdf2.getvalue())
            if images:
                st.image(images[0], caption='First page preview', use_column_width=True)
        
        # Display extracted text
        with st.expander("Show extracted text"):
            st.text(text2)

# Comparison section
st.divider()
st.subheader("PDF Comparison")

if pdf1 is not None and pdf2 is not None:
    st.info("""
    This is where the LLM comparison would take place. 
    The comparison could analyze:
    - Content differences
    - Structure similarities
    - Key points comparison
    - Summary of changes
    
    To implement the actual comparison, you would need to:
    1. Set up an LLM API connection (e.g., OpenAI)
    2. Format the extracted text appropriately
    3. Send both texts to the LLM with specific comparison instructions
    4. Display the results
    """)
else:
    st.warning("Please upload both PDFs to enable comparison")