Strongly Connected Components Tree
HierarchicalHotNet.SCCTree
— TypeThe tree of strongly connected components. Organizes the strongly connected components of the weighted directed graph into a tree.
Cutting the tree at specific threshold with cut
gives the corresponding strongly connected components. The root of the tree corresponds to the weakest threshold.
HierarchicalHotNet.SCCTreeNode
— TypeNode of the SCCTree
.
HierarchicalHotNet.scctree
— Functionscctree(g::Union{AbstractSimpleWeightedGraph, AbstractMatrix};
method=:bisect, skipval=0, rev=false) -> SCCTree
Computes the hierarchical decomposition of the weighted directed graph g
into strongly connected components.
Supports weighted graph objects as well as their adjacency matrix representations.
Keyword arguments
method::Symbol
(defaults to:bisect
): the method for partitioning. The supported methods are:bisect
(the fastest) and:bottomup
(slow, but simpler).skipval::Number
(defaults to zero): what value of the adjacency matrix should be treated as "no edge".rev::Bool
(defaults tofalse
): iftrue
, bigger edges weights are considered weaker than smaller ones
HierarchicalHotNet.cut
— Functioncut(tree::SCCTree, threshold; minsize=1) -> IndicesPartition
Cuts the tree at the given edge weight threshold
to get the corresponding strongly connected components of the original graph.
Keyword arguments
minsize
: the minimal number of vertices in the component. Smaller connected components are skipped. By default returns components of any size.