Exception in insert_image with mask specified

[cmyers009]

Description of the bug

This is the code block that produces the error:

create new page

width, height = page.mediabox_size
new_page = new_pdf.new_page(width=width, height=height)

get bounding box info and extract the position of the images

page_bounding_box_info = doc[index].get_bboxlog()

image_bounding_box_info = []
for bbx in page_bounding_box_info:
if(bbx[0] == 'fill-image'):
image_bounding_box_info.append(bbx[1])

insert these images into the new page

for index, bbx in enumerate(image_bounding_box_info):
positional_info = bbx

# Ensure positional_info has four elements
if len(positional_info) == 4:
    xref = img_list[index][0]
    base_image = doc.extract_image(xref)
    image_bytes = base_image["image"]
    smask_xref = base_image.get('smask')

    if smask_xref:
        # place smask in pdf

        # make image reference

        # Handling images with SMask
        smask_img = doc.extract_image(smask_xref)["image"]
        smask_pil = io.BytesIO(smask_img)
        print(smask_xref)
        print(smask_pil)
        new_image_xref = new_page.insert_image(positional_info, stream=image_bytes, mask=smask_pil)
    else:
        new_image_xref = new_page.insert_image(positional_info, stream=image_bytes)

Error happens on this line:

new_page.insert_image(positional_info, stream=image_bytes, mask=smask_pil)

Error happens in PyMuPDF/src/init.py
zimg = mupdf.fz_new_image_from_compressed_buffer(
w, h,
bpc, colorspace, xres, yres, 1, 0, None,
None, cbuf1, mask
)

How to reproduce the bug

Traceback (most recent call last): File "C:\Users\{user}\Desktop\{my_file}.py", line 375, in extract_image_from_pdf new_image_xref = new_page.insert_image(positional_info, stream=image_bytes, mask=smask_pil) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\{user}\AppData\Local\Programs\Python\Python311\Lib\site-packages\fitz\utils.py", line 364, in insert_image xref, digests = page._insert_image( ^^^^^^^^^^^^^^^^^^^ File "C:\Users\{user}\AppData\Local\Programs\Python\Python311\Lib\site-packages\fitz\__init__.py", line 7823, in _insert_image zimg = mupdf.fz_new_image_from_compressed_buffer( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ TypeError: fz_new_image_from_compressed_buffer_outparams_fn() takes 10 positional arguments but 12 were given Failed to extract image(s) from page, ran into error fz_new_image_from_compressed_buffer_outparams_fn() takes 10 positional arguments but 12 were given

PyMuPDF version

1.23.15

Operating system

Windows

Python version

3.11

[JorjMcKie]

I have changed the unusable title of this post to something more meaningful.
Otherwise, let me check.
You are aware, that you do not need to use the mask parameter? You can pass a pixmap with alpha channel without problems.

[JorjMcKie]

Confirming the problem.
We will check it out now.

[cmyers009]

@JorjMcKie I am trying to not use pixmap if possible, because it seems to increase file size, and reduce image quality for raster images.

Am I incorrect about this?

[JorjMcKie]

Larger file sizes can be coped with by using proper save() arguments: deflate=True,garbage=3 or, preferably, doc.ez_save().
Quality reduction going over the pixmap may happen because internal conversion to PNG is done - inevitably. However, depending on the colorspace, this may have an overall beneficial effect when b/w images are stake: then the generated mask uses the best-in-class compression method CCITTFaxDecode.

[julian-smith-artifex-com]

Fixed in 1.23.19.