Get a subplan for graph-based clustering.

Currently supported community-based detection algorithms are:

• igraph::cluster_leiden()

• igraph::cluster_louvain()

• igraph::cluster_walktrap()

See run_graph_based_clustering() for more details.

Usage

get_clustering_graph_subplan(
  sce_target_name,
  dimred,
  cluster_graph_louvain_enabled,
  cluster_graph_louvain_resolutions_,
  cluster_graph_walktrap_enabled,
  cluster_graph_leiden_enabled,
  cluster_graph_leiden_resolutions_,
  cluster_graph_snn_k_,
  cluster_graph_snn_type_,
  is_integration,
  plots_out_dir
)

Arguments

sce_target_name

A character scalar: name of target representing a SCE object that will be used for SNN graph construction.

dimred

A character scalar: which reducedDim() to use for clustering.

cluster_graph_louvain_resolutions_, cluster_graph_leiden_resolutions_

A numeric vector: resolutions to calculate for the selected clustering method.

cluster_graph_leiden_enabled, cluster_graph_louvain_enabled, cluster_graph_walktrap_enabled

A logical scalar: if FALSE, disable the selected clustering and set all corresponding targets to NULL.

cluster_graph_snn_k_, cluster_graph_snn_type_

Passed to graph_snn_fn().

is_integration

A logical scalar: if TRUE, clustering results will be named as cluster_int_graph_<method>[_r<r>]. If FALSE, they will be named as cluster_graph_<method>[_r<r>].

plots_out_dir

A character scalar: path to output directory to save plots (e.g. clustree::clustree()).

Value

drake::drake_plan()