Methods: Regular neighboring function to technology adaptability

[Jieran-S]

Methods such as BayesSpace, GraphST, and SC.MEB only applies to Visium and ST due to their assumption that spots and cells in the ST datasets are regularly spatially distributed. Well, this is not the case for imaging-based technologies, meaning that those methods are definitely not applicable for those datasets, they can have the potential to be applied to more recent technologies that fulfill such assumptions such as Visium HD, StereoSeq, and SlideSeq.

Maybe we can check the neighborhood arrangement for those technologies and see if we're able to find some correspondence. (Say, for instance, Visium HD is using the same grid-neighboring as ST). We can make a list of those correspondences, putting them into the implementation of those 3 methods, to expand their applicability range.