Image with Filter "[/FlateDecode/JPXDecode]" not extracted

[ghost]

Describe the bug (mandatory)

I'm trying to iterate over all images in the document, but for one filter combination it fails.
PyMuPDF works correctly for all filter combinations except Filter [/FlateDecode/JPXDecode]. Pdfs with such image filters are correctly read by pdf readers and other python pdf libs, but PyMuPDF fails to extract image and get correct filters by xref_get_key(xref, "/Filters").

these:
<</ColorSpace/DeviceRGB/BitsPerComponent 8/Width 1672/Length 1964389/Height 1124/Name/im1/Subtype/Image/Type/XObject/Filter/JPXDecode>>

and

<</ID 26 0 R/Type/XObject/Length 476/Filter[/FlateDecode/DCTDecode]/Subtype/Image/BitsPerComponent 8/Width 126/Height 81/ColorSpace/DeviceRGB>>

are ok

this
<</ColorSpace/DeviceGray/BitsPerComponent 8/Width 67/Length 1958/Height 68/Name/im2/Subtype/Image/Type/XObject/Filter[/FlateDecode/JPXDecode]>>

fails

However, document.xref_stream(xref) correctly decompresses the stream and output is valid jpeg2000 stream.

To Reproduce (mandatory)

    document = fitz.Document(srcFileName)
    allXrefsLength = document.xref_length()
    for xref in range(1, allXrefsLength):
        if document.xref_get_key(xref, "Subtype")[1] != "/Image":
            continue

        imgDict = document.extract_image(xref)
        if not imgDict:
            tmpFilters = document.xref_get_key(xref, "/Filters")
            print("subtype of xref {0} is /Image, but pymupdf can not extract it as image. filters: {1}".format(xref, tmpFilters))

output for images with such filters:
subtype of xref 76 is /Image, but pymupdf can not extract it as image. filters: ('null', 'null')

Your configuration (mandatory)

Windows 10 x64, python 3.10, pymupdf 1.21, installed by pip install pymupdf

[JorjMcKie]

Duplicate of #1995: the maximum supported values for image width and height currently are 2**16 = 65,536.
See that issue for details.

For the other part of the issue, extraction of the /Filter value, please allow me an example.

[ghost]

Duplicate of #1995: the maximum supported values for image width and height currently are 2**16 = 65,536. See that issue for details.

It happens with all image sizes, in my example it is 67 x 68 px. Only the filter combination matters.

[JorjMcKie]

Does doc.xref_get_key(xref, "Filter") work correctly?
Only image extraction is the problem?

[JorjMcKie]

You have a typo in extracting the filter names: /Filters does not exist, only /Filter.

[JorjMcKie]

Please provide a reproducer file.

[ghost]

You have a typo in extracting the filter names: /Filters does not exist, only /Filter.

My bad :( "Filter" works correctly, so image extraction is the only problem,
Attached sample file.

massage.pdf

[JorjMcKie]

Thanks for the file.
Fixed it.
Will be published with next version.

[ghost]

Thank you!

[julian-smith-artifex-com]

Fixed in PyMuPDF-1.21.1.