the repo has a number of custom self defined objects that allows for pretty abstracted handling
it appears the frames are drawn ad hoc using PIL, SVG and or cairo libraries with the spatial reasoning done explicitly in the source
