No images detected for PDF with a clear image

[mikethea1]

Here is the file of interest:
repro_p1.pdf

Clearly this file has an image yet pdfpg does not find it:

using var doc = PdfDocument.Open(path);
var page = doc.GetPage(1);
Console.WriteLine(page.NumberOfImages); // 0

I dug into this a bit, and I believe the reason is that the image is being referenced via a "cs" operator (SetNonStrokeColorSpace) which references /Resources/Pattern and ultimately the image.

It would be cool if PdfPig could detect images referenced in this way!

[BobLd]

@mikethea1 thx for sharing the sample document, I'll look into that shortly

[BobLd]

@mikethea1 I had a look and you are correct, the image is inside the Pattern color. I think this is a bit of a corner case and it might not be beneficial to the library to include those.

You can still extract them relatively easily though by implementing your own ContentStreamProcessor. I'll try to create a sample to show you how to achieve that. If there's not real performance cost, we might include that in the cor library

[BobLd]

@mikethea1 I finally manage to find the time to do a proof of concept for what you are looking for. Please have a look at this branch https://github.com/BobLd/PdfPig/tree/issues/940

The logic might not be very straightforward. I'll give it a second look soon.

For reference, the logic is based on my Skia pdf renderer. See here https://github.com/BobLd/PdfPig.Rendering.Skia/blob/f2456c0e2a1ea08e0befe1ac6695c23f39f8227e/UglyToad.PdfPig.Rendering.Skia/SkiaStreamProcessor.Shading.cs#L393

[mikethea1]

Nice!