Merge Multiple PDF's into PDF's under 500 MB each (JFK File Research)

merge_500.py

import os
import PyPDF2

def get_pdf_size(pdf_writer):
    """Estimate the size of the PDF by writing it to a temporary file."""
    import io
    temp_stream = io.BytesIO()
    pdf_writer.write(temp_stream)
    return temp_stream.tell() / (1024 * 1024)  # Convert bytes to MB

def merge_pdfs(input_directory, output_prefix, size_limit_mb=500):
    pdf_files = sorted([f for f in os.listdir(input_directory) if f.lower().endswith('.pdf')])
    
    if not pdf_files:
        print("No PDF files found in the directory.")
        return
    
    print(f"Merging {len(pdf_files)} PDF files...")
    
    part_number = 1
    pdf_writer = PyPDF2.PdfWriter()
    output_file = f"{output_prefix}_part{part_number}.pdf"
    
    for pdf in pdf_files:
        pdf_path = os.path.join(input_directory, pdf)
        try:
            pdf_reader = PyPDF2.PdfReader(pdf_path)
            for page in pdf_reader.pages:
                pdf_writer.add_page(page)
                
                if get_pdf_size(pdf_writer) >= size_limit_mb:
                    # Save the current part and start a new one
                    with open(output_file, "wb") as output_pdf:
                        pdf_writer.write(output_pdf)
                    print(f"Created: {output_file}")
                    
                    # Start new PDF part
                    part_number += 1
                    output_file = f"{output_prefix}_part{part_number}.pdf"
                    pdf_writer = PyPDF2.PdfWriter()
        except Exception as e:
            print(f"Error adding {pdf}: {e}")
    
    # Save the last part if it has any content
    if pdf_writer.pages:
        with open(output_file, "wb") as output_pdf:
            pdf_writer.write(output_pdf)
        print(f"Created: {output_file}")
    
if __name__ == "__main__":
    input_directory = "./jfk_records_2025"  # Change to your directory
    output_prefix = "merged_output"  # Prefix for output files
    merge_pdfs(input_directory, output_prefix)