spatial neighbors and datasets integration

[wangjiawen2013]

Hi, I don't quite understand about the "library_key" parameter in sq.gr.spatial_neighbors. I have integrated several datasets in an anndata object, the datasets are from different 10xvisium samples. In this case, how the neighbors are computed ? The spatial coordinates of the spots will be the same among different samples, so I wanna the ways that squidpy process this kind of object. In fact, all the downstream analysis, such moranI, ripleyL, co occurrence are related to this kind of problems and this is a question about spatial transcriptome data integration.