When I have books in form of webpages I normally write a small crawler in python, extract the text div with beautifulsoup, add <hn> tags for chapter names and throw them all together in html form. Add a cover image and combine everything with pandoc.

Nothing fancy but works reliable in an automated fashion