from PyPDF2 import PdfReader, PdfWriter
# Input PDF file
input_pdf = "File.pdf"
# Output PDF files
a3_pdf = "A3_pages.pdf"
a4_pdf = "A4_pages.pdf"
# A3 and A4 standard sizes in points
MM_TO_PT = 2.83465
A4_WIDTH = 210 * MM_TO_PT
A4_HEIGHT = 297 * MM_TO_PT
A3_WIDTH = 297 * MM_TO_PT
A3_HEIGHT = 420 * MM_TO_PT
# Tolerance in points to account for scanned or resized pages
TOLERANCE = 100  # increase if needed
# Open PDF
reader = PdfReader(input_pdf)
a3_writer = PdfWriter()
a4_writer = PdfWriter()
# Iterate through pages
for i, page in enumerate(reader.pages):
    width = float(page.mediabox.width)
    height = float(page.mediabox.height)
    # Use approximate size matching with tolerance
    if ((abs(width - A4_WIDTH) < TOLERANCE and abs(height - A4_HEIGHT) < TOLERANCE) or
        (abs(width - A4_HEIGHT) < TOLERANCE and abs(height - A4_WIDTH) < TOLERANCE)):
        a4_writer.add_page(page)
        print(f"Page {i+1} -> A4")
    
    elif ((abs(width - A3_WIDTH) < TOLERANCE and abs(height - A3_HEIGHT) < TOLERANCE) or
          (abs(width - A3_HEIGHT) < TOLERANCE and abs(height - A3_WIDTH) < TOLERANCE)):
        a3_writer.add_page(page)
        print(f"Page {i+1} -> A3")
    
    # Fallback for unusual sizes
    elif width > height:
        a3_writer.add_page(page)
        print(f"Page {i+1} -> A3 (approximate)")
    else:
        a4_writer.add_page(page)
        print(f"Page {i+1} -> A4 (approximate)")
# Save new PDFs
with open(a3_pdf, "wb") as f:
    a3_writer.write(f)
with open(a4_pdf, "wb") as f:
    a4_writer.write(f)
print("PDF separation complete!")