feat: Add width and height attributes on non-responsive images

[patrickfournier]

This pull request enhances the image_process plugin by updating image processing logic for non-responsive images to automatically add width and height attributes to <img> tags based on the processed image actual dimensions. The tests have been updated to reflect this new behavior, ensuring that images without pre-existing size attributes receive them, while existing attributes remain unchanged.

Having width and height attributes helps the browsers to compute the page layout more quickly, since it does not have to wait for the images to be loaded to know their size.

Image processing logic improvements:

• The process_image function now returns the processed image's width and height, and process_img_tag uses these values to set width and height attributes on <img> tags if they are not already present.

Test suite updates:

• Test cases in test_image_process.py have been updated to expect width and height attributes in processed <img> tags, and mocks for process_image now return fake image dimensions.
• Test test_path_normalization now use BeautifulSoup to parse and assert on individual tag attributes, making them more robust to attribute order and formatting.

[justinmayer]

Seems sensible to me. Any thoughts, @minchinweb?

[minchinweb]

I'm just able to do a "desktop" (ok, phone) review at the moment, but it looks good. Thanks for the PR, and thanks for including tests!

I was vaguely aware that Exif dimensions could mis-match the "physical" dimensions, as it came up on Immich's list of "Cursed Knowledge", bit hadn't drawn the connection it could affect us here.

(For future reference,) Is there a reason not to add these dimensions to images that don't already have them?

Also, should add a note to our list of "known issues" that Exif and "physical" dimensions will drift of ExifTool isn't installed?

Edit: Oops, comment got added to the wrong issue!

[minchinweb]

This is adding a width attribute to the HTML <img> tag, right? (i.e. nothing to do with the Exif tags)

[patrickfournier]

Right. Here, img is the HTML image tag parsed from the HTML generated from the user content.

[minchinweb]

Look good to me!

I had to look at this on my computer to make sense of it, but I was mostly over-complicating things.

I added a doc string for the function, so it that looks good on your side, please merge.

@justinmayer Will you merge this one? Also, with these two, it's probably time to cut a release; are you up for that?

[justinmayer]

Yes, I will issue a release tomorrow. Many thanks to @patrickfournier for the enhancement and to @minchinweb for the review and doc-string addition. Much appreciated! 💫

[justinmayer]

Release issued 🎉 https://github.com/pelican-plugins/image-process/releases/tag/3.2.0