Multisample GraphPCA examples

The ordinary hierarchical API remains unchanged:

from GraphPCA import Run_Hierarchical_Multi_GPCA

Z_list, W0, Ws_list = Run_Hierarchical_Multi_GPCA(
    adatas=[adata_1, adata_2],
    locations=[adata_1.obsm["spatial"], adata_2.obsm["spatial"]],
    n_components=30,
    lambdas=0.5,
    rhos=1.0,
    n_neighbors=6,
)

For datasets with many sections, see out_of_core_hierarchical.py. It is a complete, runnable synthetic example of the optional disk-backed workflow. That workflow requires adatas=None and a section store created by create_hierarchical_section_store; it does not change the default tutorial workflow.