Skip to content
TheChatArchiveThe Chat Archive
Coherecommand-a-03-2025· CohereHugging Face LMArena Human Preference 140K

Please write a python function that uses pypdf to extract a page range from a pdf and…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Claudeclaude-opus-4-20250514’s answer.
  1. User

    Please write a python function that uses pypdf to extract a page range from a pdf and save it to a new file, the arguments of the function are the input and output file names and the from and to pages.

  2. Cohere Assistant

    Below is a Python function that uses the PyPDF2 library (now known as pypdf) to extract a specified page range from a PDF and save it to a new file. The function takes the input and output file names, as well as the from_page and to_page arguments.

    from pypdf import PdfReader, PdfWriter
    
    def extract_pages_from_pdf(input_file, output_file, from_page, to_page):
        """
        Extracts a range of pages from a PDF and saves them to a new PDF file.
    
        :param input_file: Path to the input PDF file.
        :param output_file: Path to the output PDF file.
        :param from_page: The starting page number (1-based index).
        :param to_page: The ending page number (1-based index).
        """
        try:
            # Open the input PDF file
            reader = PdfReader(input_file)
            
            # Ensure page numbers are within valid range
            if from_page < 1 or to_page > len(reader.pages) or from_page > to_page:
                raise ValueError("Invalid page range.")
            
            # Create a PdfWriter object to write the new PDF
            writer = PdfWriter()
            
            # Add the specified page range to the writer
            for page_num in range(from_page - 1, to_page):  # PyPDF uses 0-based indexing
                writer.add_page(reader.pages[page_num])
            
            # Write the new PDF to the output file
            with open(output_file, "wb") as output_pdf:
                writer.write(output_pdf)
            
            print(f"Pages {from_page}-{to_page} extracted and saved to {output_file}")
        
        except Exception as e:
            print(f"An error occurred: {e}")
    
    # Example usage:
    # extract_pages_from_pdf("input.pdf", "output.pdf", 1, 3)  # Extracts pages 1 to 3
    

    Explanation:

    1. PdfReader: Opens the input PDF file and reads its pages.
    2. Page Range Validation: Checks if the provided page range is valid.
    3. PdfWriter: Creates a new PDF file and adds the specified pages to it.
    4. Saving the Output: Writes the new PDF to the specified output file.

    Notes:

    • Page numbers are 1-based (e.g., the first page is 1), but pypdf uses 0-based indexing internally, so we subtract 1 when accessing pages.
    • Ensure the pypdf library is installed (pip install pypdf).
    • Error handling is included to catch invalid page ranges or other issues.

    Let me know if you need further assistance!

  3. User

    Is it possible to use the pdf page number? I don´t know how it is called but my pdf starts with roman numbers for the introductoin

  4. Cohere Assistant

    Yes, it's possible to handle PDFs with page numbering that includes Roman numerals or other non-standard numbering schemes. However, pypdf (or PyPDF2) does not directly interpret or expose the visual page numbers (e.g., Roman numerals) used in the PDF. Instead, it uses 0-based indexing for pages, regardless of how the pages are labeled in the document.

    To extract pages based on the visual page numbers (e.g., Roman numerals), you would need to:

    1. Identify the mapping between the visual page numbers and the internal page indices.
    2. Use that mapping to extract the correct pages.

    Here’s an updated function that allows you to specify the page range based on the visual page numbers (e.g., i, ii, iii, etc.). You'll need to manually determine the range of visual page numbers you want to extract.

    from pypdf import PdfReader, PdfWriter
    
    def extract_pages_from_pdf_by_visual_numbers(input_file, output_file, from_visual, to_visual):
        """
        Extracts a range of pages from a PDF based on visual page numbers (e.g., Roman numerals) and saves them to a new PDF file.
    
        :param input_file: Path to the input PDF file.
        :param output_file: Path to the output PDF file.
        :param from_visual: The starting visual page number (e.g., "i", "1").
        :param to_visual: The ending visual page number (e.g., "iii", "5").
        """
        try:
            # Open the input PDF file
            reader = PdfReader(input_file)
            
            # Create a mapping of visual page numbers to internal indices
            visual_to_index = {}
            for i, page in enumerate(reader.pages):
                # Extract the page number from the page (this requires parsing the page content)
                # For simplicity, assume the page number is in the first text chunk (this may not always work)
                page_text = page.extract_text()
                if page_text:
                    first_line = page_text.split('\n')[0].strip()
                    if first_line.isalpha() or first_line.isdigit():  # Check if it looks like a page number
                        visual_to_index[first_line] = i
            
            # Find the internal indices for the given visual page numbers
            from_index = visual_to_index.get(from_visual.lower())
            to_index = visual_to_index.get(to_visual.lower())
            
            if from_index is None or to_index is None:
                raise ValueError("Visual page numbers not found in the PDF.")
            
            # Create a PdfWriter object to write the new PDF
            writer = PdfWriter()
            
            # Add the specified page range to the writer
            for index in range(from_index, to_index + 1):
                writer.add_page(reader.pages[index])
            
            # Write the new PDF to the output file
            with open(output_file, "wb") as output_pdf:
                writer.write(output_pdf)
            
            print(f"Pages {from_visual}-{to_visual} extracted and saved to {output_file}")
        
        except Exception as e:
            print(f"An error occurred: {e}")
    
    # Example usage:
    # extract_pages_from_pdf_by_visual_numbers("input.pdf", "output.pdf", "i", "iii")  # Extracts pages i to iii
    

    Key Points:

    1. Visual Page Number Extraction: The function attempts to extract the visual page number from the first line of text on each page. This is a simplification and may not work for all PDFs, especially those with complex layouts.
    2. Mapping: It creates a mapping of visual page numbers (e.g., "i", "ii") to their corresponding internal indices.
    3. Limitations: This approach assumes the page number is present in the text and is the first line on the page. For more accurate results, you may need to use a library like pdfplumber or PyMuPDF to better parse the PDF content.

    If you need a more robust solution, consider using pdfplumber for better text extraction and page number detection:

    pip install pdfplumber
    

    Let me know if you'd like help with a pdfplumber-based solution!

  5. User

    Can you please rewrite, taking a with_label argument (boolean) and using reader.page_labels to find the page number, and exit if not found

  6. Cohere Assistant

    Here’s a revised version of the function that takes a with_label argument (boolean) to determine whether to use page labels (e.g., Roman numerals) from the PDF's metadata. If with_label is True, it uses reader.get_page_labels() to map visual page numbers to internal indices. If a label is not found, the function exits with an error.

    from pypdf import PdfReader, PdfWriter
    
    def extract_pages_from_pdf(input_file, output_file, from_page, to_page, with_label=False):
        """
        Extracts a range of pages from a PDF and saves them to a new PDF file.
        If with_label is True, uses page labels (e.g., Roman numerals) for page numbering.
    
        :param input_file: Path to the input PDF file.
        :param output_file: Path to the output PDF file.
        :param from_page: The starting page number (or label if with_label is True).
        :param to_page: The ending page number (or label if with_label is True).
        :param with_label: Boolean to use page labels instead of indices.
        """
        try:
            # Open the input PDF file
            reader = PdfReader(input_file)
            
            # Get page labels if available
            page_labels = reader.get_page_labels() or []
            
            if with_label:
                # Create a mapping of page labels to internal indices
                label_to_index = {label: i for i, label in enumerate(page_labels) if label}
                
                # Find the internal indices for the given page labels
                from_index = label_to_index.get(str(from_page))
                to_index = label_to_index.get(str(to_page))
                
                if from_index is None or to_index is None:
                    raise ValueError(f"Page labels '{from_page}' or '{to_page}' not found in the PDF.")
            else:
                # Use 0-based indices directly
                from_index = from_page - 1
                to_index = to_page - 1
                
                # Validate page numbers
                if from_index < 0 or to_index >= len(reader.pages) or from_index > to_index:
                    raise ValueError("Invalid page range.")
            
            # Create a PdfWriter object to write the new PDF
            writer = PdfWriter()
            
            # Add the specified page range to the writer
            for index in range(from_index, to_index + 1):
                writer.add_page(reader.pages[index])
            
            # Write the new PDF to the output file
            with open(output_file, "wb") as output_pdf:
                writer.write(output_pdf)
            
            if with_label:
                print(f"Pages '{from_page}'-'{to_page}' (by label) extracted and saved to {output_file}")
            else:
                print(f"Pages {from_page}-{to_page} extracted and saved to {output_file}")
        
        except Exception as e:
            print(f"An error occurred: {e}")
    
    # Example usage:
    # extract_pages_from_pdf("input.pdf", "output.pdf", 1, 3, with_label=False)  # Extracts pages 1 to 3 by index
    # extract_pages_from_pdf("input.pdf", "output.pdf", "i", "iii", with_label=True)  # Extracts pages i to iii by label
    

    Key Features:

    1. with_label Argument:
      • If True, uses page labels (e.g., Roman numerals) from reader.get_page_labels().
      • If False, uses standard 1-based page indices.
    2. Page Label Mapping:
      • Creates a mapping of page labels to internal indices for accurate extraction.
    3. Error Handling:
      • Exits with an error if the specified page labels are not found in the PDF.

    Notes:

    • Page labels are defined in the PDF's metadata. If the PDF does not have page labels, reader.get_page_labels() returns None or an empty list.
    • Ensure the pypdf library is installed (pip install pypdf).

    Let me know if you need further adjustments!

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.