DCT Decoding Error

[admercs]

I'm getting the following error:

$ document.pdf
ERROR:root:Partially decoded. Filters applied: []
Traceback (most recent call last):
  File "/HOME/quicksand/lib/python3.6/site-packages/pdfreader/types/native.py", line 55, in apply_filter_multi
    binary = apply_filter(fname, binary, params)
  File "/HOME/quicksand/lib/python3.6/site-packages/pdfreader/filters/__init__.py", line 14, in apply_filter
    return decoder.decode(binary, params or {})
  File "/HOME/quicksand/lib/python3.6/site-packages/pdfreader/filters/dct.py", line 5, in decode
    raise NotImplementedError('DCTDecode')
NotImplementedError: DCTDecode

Any idea how to resolve it?

[maxpmaxp]

DCT decoder is not supported at this point. Feel free to contribute.

[maxpmaxp]

@admercs can you share the file please? I can try to add the decoder.

[admercs]

I cannot, sorry.

…

On Sat, Nov 5, 2022 at 15:25 Maksym Polshcha ***@***.***> wrote: @admercs <https://github.com/admercs> can you share the file please? I can try to add the decoder. — Reply to this email directly, view it on GitHub <#101 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABL4FIFI7M2WKOAORFAYGJTWG2YEPANCNFSM6AAAAAAQONW3GQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[canbolukbas]

What would you think about not raising an error, but returning the same bytes back to the caller?

pypdf library does this : https://github.com/py-pdf/pypdf/blob/e92b20e0b35e4feb5a2a7f347de7a4c3f713011a/pypdf/filters.py#L510

LMK if you want me to create the MR, I'd be happy to contribute.

[maxpmaxp]

@canbolukbas

Raw stream data can be accessed directly for any Stream object, use obj.stream instead of obj.filtered . See

pdfreader/pdfreader/types/native.py

Lines 88 to 89 in fb8189a

	self.dictionary = info_dict
	self.stream = binary_stream

This should work for any Image object, as technically it's a descendant of Stream.

As for the suggestion to return raw data with unimplemented filters - I see pros and cons. Ideally we need to have this decoder implemented. Feel free to create a PR and contribute.

[maxpmaxp]

@canbolukbas can you also attach your file please? I don't have PDFs with DCT streams. Thanks!

[maxpmaxp]

Just realized that it's a very trivial patch. It's on master. The support added on #132

[mara004]

can you also attach your file please? I don't have PDFs with DCT streams.

For what it's worth, DCT corresponds to JPEG, so should be trivial to create a sample. Just run img2pdf on an arbitrary JPEG image from the web, or drag one into Libreoffice and export to PDF.
If you have some PDFs on your disk, it's quite likely there will be one with DCT, as it's basically the most common PDF image encoding.