how to remove the invisible contents after show_pdf_page

[xulifan]

I need to copy a middle section of a pdf page into the top of a new page. However, in the new page, the original contents are still there, they are just invisible. I tried to use redaction. If I keep the middle section in the middle of the new page, the redaction can remove the top and bottom section. However, if I move the middle section to top, the redaction does not work. Any one can help?

[JorjMcKie]

It depends on which of the pages you use redactions.
You should use in on the source page. Then only use the remaining part (after applying redactions) using the "clip" parameter.

[xulifan]

Hi I attached my simple code here. The redaction to the bottom page

bottom_page.add_redact_annot(first_split_rect)
bottom_page.add_redact_annot(second_split_rect)

is not working, the second page will contain the original text, they are just invisible:

def split_page_into_two(pdf_file_path, page_num, split_text1, split_text2, output_file_path):
    # First segment the page into three parts, then put them into two pages.
    # The first segments stays on the first page.
    # The second and third segments are put into the second page.
    # However, in the second page, move the second segment to the top of the page.  

    # Open the PDF file
    pdf_document = fitz.open(pdf_file_path)
    new_pdf = fitz.open()  # Create a new PDF document

    page = pdf_document[page_num]
    
    text_instances_1 = page.search_for(split_text1)
    if not text_instances_1:
        print(f"No text instance found for {split_text1}")
        return
    text_instances_2 = page.search_for(split_text2)
    if not text_instances_2:
        print(f"No text instance found for {split_text2}")
        return
    text_instances_1 = text_instances_1[0]
    text_instances_2 = text_instances_2[0]
    
        # Get the dimensions of the page
    rect = page.rect
    print(f"Splitting Page {page_num + 1}...")

    # Create the first page with content before split_text1
    first_split_y = text_instances_1[1] - 1
    first_split_rect = fitz.Rect(rect[0], rect[1], rect[2], first_split_y)

    second_split_rect = fitz.Rect(rect[0], text_instances_1[1], rect[2], text_instances_2[3])
    second_split_rect_height = text_instances_2[3] - text_instances_1[1]
    # move the second segment to the top of the page with 40 points margin
    second_split_location_rect = fitz.Rect(rect[0], 40+rect[1], rect[2], 40+second_split_rect_height)

    third_split_rect = fitz.Rect(rect[0], text_instances_2[3], rect[2], rect[3])

    top_page = new_pdf.new_page(width=rect.width, height=rect.height)
    top_page.show_pdf_page(first_split_rect, pdf_document, page_num, clip=first_split_rect)
    # add redact annotations to the second and third segments so these segments are not included in the OCR
    # otherwise, get_text("") will include these segments even they are invisible
    # check the notes for https://pymupdf.readthedocs.io/en/latest/page.html#Page.show_pdf_page
    top_page.add_redact_annot(second_split_rect)
    top_page.add_redact_annot(third_split_rect)
    top_page.apply_redactions()

    # Create the second page with content starting from split_text1
    bottom_page = new_pdf.new_page(width=rect.width, height=rect.height)
    bottom_page.show_pdf_page(third_split_rect, pdf_document, page_num, clip=third_split_rect)

    # the redaction here does not work because it will be added back by the show_pdf_page
    bottom_page.add_redact_annot(first_split_rect)
    bottom_page.add_redact_annot(second_split_rect)
    bottom_page.apply_redactions()
    bottom_page.show_pdf_page(second_split_location_rect, pdf_document, page_num, clip=second_split_rect)


    # Replace the original page with the new PDF
    pdf_document.delete_page(page_num)
    pdf_document.insert_pdf(new_pdf, start_at=page_num)
    pdf_document.save(output_file_path)
    new_pdf.close()
    pdf_document.close()

    print(f"Pages split and saved to {output_file_path}")

[JorjMcKie]

This script will remove all content above and below the middle third rectangle of a source page and insert the result into the top third of a target page.
Nothing from outside the inserted middle stripe of the source will land in the target page:

import pymupdf

src = pymupdf.open("input.pdf")
spage = src[0]
w, h = spage.rect.br  # width / height of source page
r1 = spage.rect  # upper third rectangle
r1.y1 = h / 3
r2 = spage.rect  # lower third rectangle
r2.y0 = h / 3 * 2
spage.add_redact_annot(r1)
spage.add_redact_annot(r2)
spage.apply_redactions()  # empty upper / lower rectangles
clip = spage.rect  # remaining middle third
clip.y0 = r1.y1
clip.y1 = r2.y0
doc = pymupdf.open()  # target doc
page = doc.new_page()
target = page.rect  # receiving rect on target page
target.y1 = page.rect.height / 3
page.show_pdf_page(target, src, 0, clip=clip)
doc.ez_save("x.pdf")

[xulifan]

Thank you for your reply. I found out the trick is you need to do multiple passes, in each pass, just deal with one section. Here is my full working code for future reference:

def split_page_into_two(pdf_file_path, page_num, split_text1, split_text2, output_file_path): 
    # First segment the page into three parts, then put them into two pages.
    # The first segments stays on the first page.
    # The second and third segments are put into the second page.
    # However, in the second page, move the second segment to the top of the page.

    src_document = fitz.open(pdf_file_path)
    src_page = src_document[page_num]
    width = src_page.rect.width
    height = src_page.rect.height

    text_instances_1 = src_page.search_for(split_text1)
    if not text_instances_1:
        print(f"No text instance found for {split_text1}")
        return
    text_instances_2 = src_page.search_for(split_text2)
    if not text_instances_2:
        print(f"No text instance found for {split_text2}")
        return
    text_instances_1 = text_instances_1[0]
    text_instances_2 = text_instances_2[0]

    first_rect = src_page.rect  # upper rectangle
    first_rect.y1 = text_instances_1[1]

    second_rect = src_page.rect  # middle rectangle
    second_rect.y0 = text_instances_1[1]
    second_rect.y1 = text_instances_2[3]

    third_rect = src_page.rect  # bottom rectangle
    third_rect.y0 = text_instances_2[3]

    target_doc = fitz.open()  # target doc
    target_page = target_doc.new_page(width=width, height=height)
    target_rect = target_page.rect  # receiving rect on target page

    # In the first pass, move the middle segment to the top of the new page with 40 points margin
    margin = 40
    target_rect.y0 = margin
    target_rect.y1 = second_rect.y1 - second_rect.y0 + margin
    # Add redaction annotations to the source page to empty the upper and bottom rectangles
    src_page.add_redact_annot(first_rect)
    src_page.add_redact_annot(third_rect)
    src_page.apply_redactions()
    # Show the middle segment on the new page
    target_page.show_pdf_page(target_rect, src_document, page_num, clip=second_rect)
    first_pass_output_file_path = output_file_path+"_1.pdf"
    target_doc.ez_save(first_pass_output_file_path)
    src_document.close()
    target_doc.close()

    # In the second pass, copy the bottom segment to the new page
    src_document = fitz.open(pdf_file_path)
    src_page = src_document[page_num]
    target_doc = fitz.open(first_pass_output_file_path)  # target doc
    target_page = target_doc[0]
    # Add redaction annotations to the source page to empty the upper and middle rectangles
    src_page.add_redact_annot(first_rect)
    src_page.add_redact_annot(second_rect)
    src_page.apply_redactions()
    # Show the bottom segment on the new page
    target_page.show_pdf_page(third_rect, src_document, page_num, clip=third_rect)
    second_pass_output_file_path = output_file_path+"_2.pdf"
    target_doc.ez_save(second_pass_output_file_path)
    src_document.close()
    target_doc.close()

    # In the third pass, remove the middle and bottom segment from the old page and insert the new page after the old page
    src_document = fitz.open(pdf_file_path)
    src_page = src_document[page_num]
    # Add redaction annotations to the source page to empty the middle and bottom rectangles
    src_page.add_redact_annot(second_rect)
    src_page.add_redact_annot(third_rect)
    src_page.apply_redactions()
    target_doc = fitz.open(second_pass_output_file_path)
    # Insert the new page with middle and bottom segments after the old page
    src_document.insert_pdf(target_doc, start_at=page_num+1)
    src_document.save(output_file_path)
    src_document.close()
    target_doc.close()

    if os.path.exists(first_pass_output_file_path):
        os.remove(first_pass_output_file_path)
    if os.path.exists(second_pass_output_file_path):
        os.remove(second_pass_output_file_path)
    print(f"Pages split and saved to {output_file_path}")