prepare_patches

prepare_patches(output_folder, name: str, min_overlap: int, target_overlap: int, graph, min_patch_size: int | None = None, cluster='metis', num_clusters=10, num_iters: int | None = None, beta=0.1, levels=1, sparsify='resistance', target_patch_degree=4.0, gamma=0.0, verbose=False)[source]

initialise patch data

Parameters:

  • output_folder – experiment folder

  • name – name of data set

  • data_root – root dir for data set

  • min_overlap – minimum patch overlap

  • target_overlap – desired patch overlap

  • min_patch_size – minimum patch size

  • cluster – cluster method (one of {‘metis’, ‘louvain’, ‘distributed’, ‘fennel’}

  • num_clusters – number of clusters for metis/fennel

  • num_iters – number of iterations for fennel/distributed

  • beta – beta value for distributed

  • sparsify – sparsification method (one of {‘resistance’, ‘rmst’, ‘none’})

  • target_patch_degree – target patch degree for resistance sparsification

  • gamma – gamma value for rmst sparsification

  • verbose – print output