all_simple_paths: update PR JuliaGraphs#20

thchr · thchr · commit e9fa195f75bf · 2024-03-18T16:45:52.000+01:00
- this updates the port of sbromberger/LightGraphs.jl#1540 from JuliaGraphs#20 - has a number of simplifications relative to original implementation - original implementation by @i_aki_y - cutoff now defaults to `nv(g)` Co-authored-by: i_aki_y Co-authored-by: etiennedeg
diff --git a/src/Graphs.jl b/src/Graphs.jl
@@ -23,7 +23,8 @@ using DataStructures:
     union!,
     find_root!,
     BinaryMaxHeap,
-    BinaryMinHeap
+    BinaryMinHeap,
+    Stack
 using LinearAlgebra: I, Symmetric, diagm, eigen, eigvals, norm, rmul!, tril, triu
 import LinearAlgebra: Diagonal, issymmetric, mul!
 using Random:
@@ -196,6 +197,9 @@ export
 
     # eulerian
     eulerian,
+    
+    # all simple paths
+    all_simple_paths,
 
     # coloring
     greedy_color,
@@ -496,6 +500,7 @@ include("traversals/maxadjvisit.jl")
 include("traversals/randomwalks.jl")
 include("traversals/diffusion.jl")
 include("traversals/eulerian.jl")
+include("traversals/all_simple_paths.jl")
 include("connectivity.jl")
 include("distance.jl")
 include("editdist.jl")
diff --git a/src/traversals/all_simple_paths.jl b/src/traversals/all_simple_paths.jl
@@ -0,0 +1,156 @@
+"""
+    all_simple_paths(g, u, v; cutoff=nv(g)) --> Graphs.SimplePathIterator
+    
+Returns an iterator that generates all simple paths in the graph `g` from a source vertex
+`u` to a target vertex `v` or iterable of target vertices `vs`.
+
+The iterator's elements (i.e., the paths) can be materialized via `collect` or `iterate`.
+Paths are iterated in the order of a depth-first search.
+
+## Keyword arguments
+The maximum path length (i.e., number of edges) is limited by the keyword argument `cutoff`
+(default, `nv(g)`). If a path's path length is greater than or equal to `cutoff`, it is
+omitted.
+
+## Examples
+```jldoctest
+julia> using Graphs
+julia> g = complete_graph(4)
+julia> spi = all_simple_paths(g, 1, 4)
+ Graphs.SimplePathIterator(1 → 4)
+julia> collect(spi)
+5-element Vector{Vector{Int64}}:
+ [1, 4]
+ [1, 3, 4]
+ [1, 3, 2, 4]
+ [1, 2, 4]
+ [1, 2, 3, 4]
+```
+We can restrict the search to paths of length less than a specified cut-off (here, 2 edges):
+```jldoctest
+julia> collect(all_simple_paths(g, 1, 4; cutoff=2))
+ [1, 2, 4]
+ [1, 3, 4]
+ [1, 4]
+```
+"""
+function all_simple_paths(
+            g::AbstractGraph{T}, 
+            u::T, 
+            vs;
+            cutoff::T=nv(g)
+            ) where T <: Integer
+
+    vs = vs isa Set{T} ? vs : Set{T}(vs)
+    return SimplePathIterator(g, u, vs, cutoff)
+end
+
+"""
+    SimplePathIterator{T <: Integer}
+
+Iterator that generates all simple paths in `g` from `u` to `vs` of a length at most
+`cutoff`.
+"""
+struct SimplePathIterator{T <: Integer, G <: AbstractGraph{T}}
+    g::G
+    u::T       # start vertex
+    vs::Set{T} # target vertices
+    cutoff::T  # max length of resulting paths
+end
+
+function Base.show(io::IO, spi::SimplePathIterator)
+    print(io, "SimplePathIterator{", typeof(spi.g), "}(", spi.u, " → ")
+    if length(spi.vs) == 1
+        print(io, only(spi.vs))
+    else
+        print(io, '[')
+        join(io, spi.vs, ", ")
+        print(io, ']')
+    end
+    print(io, ')')
+end
+Base.IteratorSize(::Type{<:SimplePathIterator}) = Base.SizeUnknown()
+Base.eltype(::SimplePathIterator{T}) where T = Vector{T}
+
+mutable struct SimplePathIteratorState{T <: Integer}
+    stack::Stack{Vector{T}} # used to restore iteration of child vertices; each vector has
+                            # two elements: a parent vertex and an index of children
+    visited::Stack{T}       # current path candidate
+    queued::Vector{T}       # remaining targets if path length reached cutoff
+end
+function SimplePathIteratorState(spi::SimplePathIterator{T}) where T <: Integer
+    stack = Stack{Vector{T}}()
+    visited = Stack{T}()
+    queued = Vector{T}()
+    push!(visited, spi.u)    # add a starting vertex to the path candidate
+    push!(stack, [spi.u, 1]) # add a child node with index 1
+    SimplePathIteratorState{T}(stack, visited, queued)
+end
+
+function _stepback!(state::SimplePathIteratorState) # updates iterator state.
+    pop!(state.stack)
+    pop!(state.visited)
+end
+
+
+"""
+    Base.iterate(spi::SimplePathIterator{T}, state=nothing)
+
+Returns the next simple path in `spi`, according to a depth-first search.
+"""
+function Base.iterate(
+            spi::SimplePathIterator{T}, 
+            state::SimplePathIteratorState=SimplePathIteratorState(spi)
+            ) where T <: Integer
+
+    while !isempty(state.stack)
+        if !isempty(state.queued) # consume queued targets
+            target = pop!(state.queued)
+            result = vcat(reverse(collect(state.visited)), target)
+            if isempty(state.queued)
+                _stepback!(state)
+            end
+            return result, state
+        end
+
+        parent_node, next_childe_index = first(state.stack)
+        children = outneighbors(spi.g, parent_node)
+        if length(children) < next_childe_index
+            # all children have been checked, step back.
+            _stepback!(state)
+            continue
+        end
+
+        child = children[next_childe_index]    
+        first(state.stack)[2] += 1 # move child index forward
+        child in state.visited && continue
+
+        if length(state.visited) == spi.cutoff
+            # collect adjacent targets if more exist and add them to queue
+            rest_children = Set(children[next_childe_index: end])
+            state.queued = collect(setdiff(intersect(spi.vs, rest_children), Set(state.visited)))
+
+            if isempty(state.queued)
+                _stepback!(state)
+            end
+        else
+            result = if child in spi.vs
+                vcat(reverse(collect(state.visited)), child)
+            else
+                nothing
+            end
+
+            # update state variables
+            push!(state.visited, child) # move to child vertex
+            if !isempty(setdiff(spi.vs, state.visited)) # expand stack until all targets are found
+                push!(state.stack, [child, 1]) # add the child node as a parent for next iteration
+            else
+                pop!(state.visited) # step back and explore the remaining child nodes
+            end
+
+            if !isnothing(result) # found a new path, return it
+                return result, state
+            end
+        end
+    end
+end
diff --git a/test/runtests.jl b/test/runtests.jl
@@ -109,6 +109,7 @@ tests = [
     "traversals/randomwalks",
     "traversals/diffusion",
     "traversals/eulerian",
+    "traversals/all_simple_paths",
     "community/cliques",
     "community/core-periphery",
     "community/label_propagation",
diff --git a/test/traversals/all_simple_paths.jl b/test/traversals/all_simple_paths.jl
@@ -0,0 +1,127 @@
+@testset "All simple paths" begin
+    # single path
+    g = path_graph(4)
+    paths = all_simple_paths(g, 1, 4)
+    @test Set(p for p in paths) == Set([[1, 2, 3, 4]])
+    @test Set(collect(paths)) == Set([[1, 2, 3, 4]])
+    @test 1 == length(paths)
+
+
+    # single path with cutoff
+    @test collect(all_simple_paths(g, 1, 4; cutoff=2)) == [[1, 2, 4], [1, 3, 4], [1, 4]]
+
+    # two paths
+    g = path_graph(4)
+    add_vertex!(g)
+    add_edge!(g, 3, 5)
+    paths = all_simple_paths(g, 1, [4, 5])
+    @test Set(p for p in paths) == Set([[1, 2, 3, 4], [1, 2, 3, 5]])
+    @test Set(collect(paths)) == Set([[1, 2, 3, 4], [1, 2, 3, 5]])
+    @test 2 == length(paths)
+
+    # two paths with cutoff
+    g = path_graph(4)
+    add_vertex!(g)
+    add_edge!(g, 3, 5)
+    paths = all_simple_paths(g, 1, [4, 5], cutoff=3)
+    @test Set(p for p in paths) == Set([[1, 2, 3, 4], [1, 2, 3, 5]])
+
+    # two targets in line emits two paths
+    g = path_graph(4)
+    add_vertex!(g)
+    paths = all_simple_paths(g, 1, [3, 4])
+    @test Set(p for p in paths) == Set([[1, 2, 3], [1, 2, 3, 4]])
+
+    # two paths digraph
+    g = SimpleDiGraph(5)
+    add_edge!(g, 1, 2)
+    add_edge!(g, 2, 3)
+    add_edge!(g, 3, 4)
+    add_edge!(g, 3, 5)    
+    paths = all_simple_paths(g, 1, [4, 5])
+    @test Set(p for p in paths) == Set([[1, 2, 3, 4], [1, 2, 3, 5]])
+
+    # two paths digraph with cutoff
+    g = SimpleDiGraph(5)
+    add_edge!(g, 1, 2)
+    add_edge!(g, 2, 3)
+    add_edge!(g, 3, 4)
+    add_edge!(g, 3, 5)    
+    paths = all_simple_paths(g, 1, [4, 5], cutoff=3)
+    @test Set(p for p in paths) == Set([[1, 2, 3, 4], [1, 2, 3, 5]])
+
+    # digraph with a cycle
+    g = SimpleDiGraph(4)
+    add_edge!(g, 1, 2)
+    add_edge!(g, 2, 3)
+    add_edge!(g, 3, 1)
+    add_edge!(g, 2, 4)
+    paths = all_simple_paths(g, 1, 4)
+    @test Set(p for p in paths) == Set([[1, 2, 4]])
+
+    # digraph with a cycle. paths with two targets share a node in the cycle.
+    g = SimpleDiGraph(4)
+    add_edge!(g, 1, 2)
+    add_edge!(g, 2, 3)
+    add_edge!(g, 3, 1)
+    add_edge!(g, 2, 4)
+    paths = all_simple_paths(g, 1, [3, 4])
+    @test Set(p for p in paths) == Set([[1, 2, 3], [1, 2, 4]])
+
+    # source equals targets
+    g = SimpleGraph(4)
+    paths = all_simple_paths(g, 1, 1)
+    @test Set(p for p in paths) == Set([])
+
+    # cutoff prones paths
+    # Note, a path lenght is node - 1
+    g = complete_graph(4)
+    paths = all_simple_paths(g, 1, 2; cutoff=1)
+    @test Set(p for p in paths) == Set([[1, 2]])
+
+    paths = all_simple_paths(g, 1, 2; cutoff=2)
+    @test Set(p for p in paths) == Set([[1, 2], [1, 3, 2], [1, 4, 2]])
+
+    # non trivial graph
+    g = SimpleDiGraph(6)
+    add_edge!(g, 1, 2)
+    add_edge!(g, 2, 3)
+    add_edge!(g, 3, 4)
+    add_edge!(g, 4, 5)
+
+    add_edge!(g, 1, 6)
+    add_edge!(g, 2, 6)
+    add_edge!(g, 2, 4)
+    add_edge!(g, 6, 5)
+    add_edge!(g, 5, 3)
+    add_edge!(g, 5, 4)
+
+    paths = all_simple_paths(g, 2, [3, 4])
+    @test Set(p for p in paths) == Set([
+         [2, 3],
+         [2, 4, 5, 3],
+         [2, 6, 5, 3],
+         [2, 4],
+         [2, 3, 4],
+         [2, 6, 5, 4],
+         [2, 6, 5, 3, 4],
+    ])
+
+    paths = all_simple_paths(g, 2, [3, 4], cutoff=3)
+    @test Set(p for p in paths) == Set([
+         [2, 3],
+         [2, 4, 5, 3],
+         [2, 6, 5, 3],
+         [2, 4],
+         [2, 3, 4],
+         [2, 6, 5, 4],
+    ])
+
+    paths = all_simple_paths(g, 2, [3, 4], cutoff=2)
+    @test Set(p for p in paths) == Set([
+         [2, 3],
+         [2, 4],
+         [2, 3, 4],
+    ])
+
+end
