Yeah, print to pdf does do that. I would assume that is normal behavior since it’s using the final markup rendered by the browser combined with the print css.
I don’t know what parsing you are using for this button (sorry I know that this is open source but I don’t know ruby 
) but I am gonna assume that it’s parsing the html source directly and that likely won’t try to read custom image attributes.
I would like to propose an option in settings that to include specified custom attributes inside image tags for the image source. (We use a few variance and it would be good to be able to support all of them)