This is where I'm having a problem converting from. I'm not sure how I handle the unique id per node, or how to recalculate or anything else along those lines
The way you might want to handle it in the DB is to have each table have a parent_id column. This parent_id would refer to the unique id of a row entry. So a 'Chapter 1' row entry in the chapter table might have a parent id of '32'. Where '32' is the book_id of the book 'Moby Dick'. Or something along those lines. I assume you would want this for your chapter and sentance tables. As previously stated, to do this you would have to have multiple entries of 'Chapter X' dependant on each book.