Can we get an exact description of the algorithm this library uses? Or source code?

It works great and I'd much prefer using this over struggling with diffusers.
But for research publication purposes I need to know how it's working.
The "image processing tool using the library" contains a brief description but not much detail.

