I wouldn't feel good about publishing it as a book as the project stands right now. The quality of the articles, the explanation style, the notation, the code style, ... varies a lot between articles. Some of the articles are known to have mistakes in them (although I hope only a handful at most), for some the explanation is completely lacking (like the suffix tree), and some very basic topics like binary search or DP are missing. And on the other side there are some obscure topics like the MPM algorithm, the Lyndon factorization, Montgomery multiplication, ..., that probably nobody would ever need (well, at least not many).

It would be a lot of effort going over all the content and bringing it up to a form, fix everything, bring it into good shape, filter out the important bits, add images, and extend the missing parts.

And I'm not even sure if the project licenses allows it.