Feature request: avoid altering images passed to FPDF.image() if there is no need to

[eroux]

(orignal title: "not parsing images when possible")

A feature I really love in img2pdf is the fact that it doesn't load and save the image if it doesn't need to. For our use case this means a huge performance gain and also in the tests I made, a gain in size of the PDF since re-compression doesn't always produce smaller pictures. See this function for instance to get the CCITT payload from a tiff file, so that it can then be copied directly without even having to look at it

[Lucas-C]

Hi @eroux

I am sorry but I am not sure to fully understand your intent here:
is this a feature request? You would like fpdf2 to not alter the image passed to FPDF.image() if it does not need to?

(if you confirm that this your correct intent, we could rename this issue title accordingly)

This would be possible for JPEG images, and CCITT/TIFF images I guess, but some images like PNG cannot be embedded in their native format in a PDF document.

In terms of code, this would mean a refactoring in fpdf/image_parsing.py and fpdf/fpdf.py to not systematically convert images to PIL.Image instances. This may not play well with existing features such as oversized images detection & downscaling, but is in theory perfectly doable.

I don't personnally plan to work on this task, but you or other contributors are welcome to do so 😊

[eroux]

That's exactly the intent indeed

[Lucas-C]

@allcontributors please add @eroux for ideas

[allcontributors]

@Lucas-C

I've put up a pull request to add @eroux! 🎉