Add KADABRA betweenness centrality approximation algorithm

CHANGELOG.md

@@ -3,6 +3,7 @@
 We follow SemVer as most of the Julia ecosystem. Below you might see the "breaking" label even for minor version bumps -- we use it a bit more loosely to denote things that are not breaking by SemVer's definition but might cause breakage to people using internal or experimental APIs or undocumented implementation details. 
 
 ## unreleased
+- KADABRA betweenness centrality approximation algorithm
 - `is_articulation(g, v)` for checking whether a single vertex is an articulation point
 - The iFUB algorithm is used for faster diameter calculation and now supports weighted graph diameter calculation
 - ECG community detection algorithm

docs/src/algorithms/centrality.md

@@ -14,6 +14,7 @@ Pages   = ["centrality.md"]
 Modules = [Graphs]
 Pages   = [
     "centrality/betweenness.jl",
+    "centrality/kadabra.jl",
     "centrality/closeness.jl",
     "centrality/degree.jl",
     "centrality/eigenvector.jl",

src/Graphs.jl

@@ -24,6 +24,8 @@ using Random:
     AbstractRNG,
     GLOBAL_RNG,
     MersenneTwister,
+    Xoshiro,
+    default_rng,
     randperm,
     randsubseq!,
     seed!,
@@ -277,6 +279,8 @@ export
     eigenvector_centrality,
     stress_centrality,
     radiality_centrality,
+    kadabra_centrality,
+    kadabra_top_k,
 
     # spectral
     adjacency_matrix,
@@ -548,6 +552,7 @@ include("centrality/betweenness.jl")
 include("centrality/closeness.jl")
 include("centrality/stress.jl")
 include("centrality/degree.jl")
+include("centrality/kadabra.jl")
 include("centrality/katz.jl")
 include("centrality/pagerank.jl")
 include("centrality/eigenvector.jl")