extract: images using Predictor and BitsPerComponent != 8

[ogimenezb]

I'm trying to extract the QR image from the Covid PDF and I have an error.

Fatal: pdfcpu: filter FlateDecode: "BitsPerComponent" must be 8

Attaching crash log.
crash.log

[hhrutter]

duplicate of #380

Thanks for reporting this!

[hhrutter]

Image decoding for images using a Predictor and bitsPerPixel != 8 is untested and therefore disabled.
If anybody out there can share a sample file with such an image I am happy to implement this.

What I would need in order to reproduce and fix the error reported originally on the top is an image object using a Predictor.
This does not show up in the output of pdfcpu images list but you can first use this command to identify the object number and then check the concrete image dict in the output of pdfcpu val -v test.pdf.

<<
	<BitsPerComponent, 4>
	<ColorSpace, [Indexed [CalRGB
			<<
				<Gamma, [2.20 2.20 2.20]>
				<Matrix, [0.41 0.21 0.02 0.36 0.72 0.12 0.18 0.07 0.95]>
				<WhitePoint, [0.95 1 1.09]>
			>>
			] 6 (üÝ\tél�Ú��Þ'�åU�á?�ÿÿÿ)]>
	<DecodeParms, <<
		<BitsPerComponent, 4>
		<Colors, 1>
		<Columns, 640>
		<Predictor, 15>
	>>>
	<Filter, FlateDecode>
	<Height, 427>
	<Length, 389>
	<Subtype, Image>
	<Type, XObject>
	<Width, 640>
>>

[petervwyatt]

<</Type /XObject /Subtype /Image /Width 172 /Height 119 /ColorSpace [/Indexed /DeviceRGB 15 30 0 R] /BitsPerComponent 4 /Filter /FlateDecode /DecodeParms << /Predictor 15 /Colors 1 /BitsPerComponent 4 /Columns 172 >> /Length 2333 >>

Predictor15-BPC4.pdf

[hhrutter]

This is fixed with the latest commit!
Thanks everybody for their input 👍

[ogimenezb]

Working perfectly.
Awesome, thanks!!!