I propose we start with the core logic for the main feature in PDF-Generator-API. This includes defining the main data structures and algorithms.