Of course they would. Not only would they get their hands on data users fully voluntarily give them by using their platform, but they’d get their hands on verified IDs and quite reliable family tie information. The potential loss of users is definately worth it for them (from their perspective).

Isn’t that illegal?

The content (i.e. text, tables, images, etc. ) of the book is under copyleft, while the only thing the publisher can argue that’s theirs is the design (cover, font, copyright claim text, etc.) There are things like page layout and stuff that may’ve been created by the author or the publishers so it’s in a grey area.

All in all, I think scanning the book and OCRing it, removing stuff like page numbers and those first few pages of junk would remove all “infringing” elements.

Or, as always, you can email tye author and they’re 99% sure to give you their manuscript directly if they didn’t publish it somewhere else already.