polish docs of paintshop, coloring and matching (#13)

* new paintshop problem

* save

* update paint shop doc

* update

* update

* update

* fix docs

* fix docs

* fix one doc

Author: GiggleLiu

docs/make.jl

@@ -23,7 +23,7 @@ makedocs(;
     repo="https://github.com/Happy-Diode/GraphTensorNetworks.jl/blob/{commit}{path}#{line}",
     sitename="GraphTensorNetworks.jl",
     format=Documenter.HTML(;
-        prettyurls=false,
+        prettyurls=get(ENV, "CI", "false") == "true",
         canonical="https://Happy-Diode.github.io/GraphTensorNetworks.jl",
         assets=String[indigo],
     ),

docs/src/ref.md

@@ -35,6 +35,11 @@ mis_compactify!
 
 cut_size
 cut_assign
+
+num_paint_shop_color_switch
+paint_shop_coloring_from_config
+
+is_good_vertex_coloring
 ```
 
 ## Properties
@@ -90,10 +95,12 @@ MergeGreedy
 #### Graph
 ```@docs
 show_graph
+spring_layout
 
 diagonal_coupled_graph
 square_lattice_graph
 unit_disk_graph
+line_graph
 
 random_diagonal_coupled_graph
 random_square_lattice_graph

examples/Coloring.jl

@@ -47,4 +47,27 @@ show_graph(graph; locs=locations)
 # We construct the tensor network for the 3-coloring problem as
 problem = Coloring{3}(graph);
 
-# ## Solving properties
+# ## Solving properties
+# ##### counting all possible coloring
+num_of_coloring = solve(problem, CountingAll())[]
+
+# ##### finding one best coloring
+single_solution = solve(problem, SingleConfigMax())[]
+
+is_good_vertex_coloring(graph, single_solution.c.data)
+
+vertex_color_map = Dict(0=>"red", 1=>"green", 2=>"blue")
+
+show_graph(graph; locs=locations, vertex_colors=[vertex_color_map[Int(c)] for c in single_solution.c.data])
+
+# Let us try to solve the same issue on its line graph, a graph that generated by mapping an edge to a vertex and two edges sharing a common vertex will be connected.
+linegraph = line_graph(graph)
+
+show_graph(linegraph; locs=[0.5 .* (locations[e.src] .+ locations[e.dst]) for e in edges(graph)])
+
+# Let us construct the tensor network and see if there are solutions.
+lineproblem = Coloring{3}(linegraph);
+
+num_of_coloring = solve(lineproblem, CountingAll())[]
+
+# You will see a zero printed, meaning no solution for the 3-coloring on edges of a Petersen graph.

examples/IndependentSet.jl

@@ -89,7 +89,7 @@ max_config = solve(problem, SingleConfigMax(; bounded=false))[]
 # The return value contains a bit string, and one should read this bit string from left to right.
 # Having value 1 at i-th bit means vertex ``i`` is in the maximum independent set.
 # One can visualize this MIS with the following function.
-show_graph(graph; locs=locations, colors=[iszero(max_config.c.data[i]) ? "white" : "red"
+show_graph(graph; locs=locations, vertex_colors=[iszero(max_config.c.data[i]) ? "white" : "red"
                                  for i=1:nv(graph)])
 
 # ##### enumeration of all MISs
@@ -101,12 +101,12 @@ using Compose
 
 m = length(all_max_configs.c)
 
-imgs = ntuple(k->(context((k-1)/m, 0.0, 1.2/m, 1.0), show_graph(graph;
+imgs = ntuple(k->show_graph(graph;
                             locs=locations, scale=0.25,
-                            colors=[iszero(all_max_configs.c[k][i]) ? "white" : "red"
-                                 for i=1:nv(graph)])), m)
+                            vertex_colors=[iszero(all_max_configs.c[k][i]) ? "white" : "red"
+                            for i=1:nv(graph)]), m);
 
-Compose.set_default_graphic_size(18cm, 4cm); Compose.compose(context(), imgs...)
+Compose.set_default_graphic_size(18cm, 4cm); Compose.compose(context(), ntuple(k->(context((k-1)/m, 0.0, 1.2/m, 1.0), imgs[k]), m)...)
 
 # ##### enumeration of all IS configurations
 all_independent_sets = solve(problem, ConfigsAll())[]
@@ -129,7 +129,7 @@ problem = IndependentSet(graph; weights=collect(1:10))
 
 max_config_weighted = solve(problem, SingleConfigMax())[]
 
-show_graph(graph; locs=locations, colors=
+show_graph(graph; locs=locations, vertex_colors=
           [iszero(max_config_weighted.c.data[i]) ? "white" : "red" for i=1:nv(graph)])
 
 # The following code computes the MIS tropical tensor (reference to be added) with open vertices 1 and 2.

examples/Matching.jl

@@ -48,9 +48,25 @@ show_graph(graph; locs=locations)
 problem = Matching(graph);
 
 # ## Solving properties
+# ### Maximum matching
+# ### Configuration properties
+max_matching = solve(problem, SizeMax())[]
+# The largest number of matching is 5, which means we have a perfect matching (vertices are all paired).
+
+# ##### matching polynomial
 # The graph polynomial defined for the independence problem is known as the matching polynomial.
 # Here, we adopt the first definition in the [wiki page](https://en.wikipedia.org/wiki/Matching_polynomial).
 # ```math
 # M(G, x) = \sum\limits_{k=1}^{|V|/2} c_k x^k,
 # ```
-# where ``k`` is the number of matches, and coefficients ``c_k`` are the corresponding counting.
+# where ``k`` is the number of matches, and coefficients ``c_k`` are the corresponding counting.
+
+matching_poly = solve(problem, GraphPolynomial())[]
+
+# ## Configuration properties
+
+# ##### one of the perfect matches
+match_config = solve(problem, SingleConfigMax())[]
+
+# Let us show the result by coloring the matched edges to red
+show_graph(graph; locs=locations, edge_colors=[isone(match_config.c.data[i]) ? "red" : "black" for i=1:ne(graph)])

examples/MaxCut.jl

@@ -71,5 +71,5 @@ max_cut_size_verify = cut_size(graph, max_vertex_config)
 
 # You should see a consistent result as above `max_cut_size`.
 
-show_graph(graph; locs=locations, colors=[
+show_graph(graph; locs=locations, vertex_colors=[
         iszero(max_vertex_config[i]) ? "white" : "red" for i=1:nv(graph)])

examples/MaximalIS.jl

@@ -9,7 +9,7 @@
 #     * how to compute weighted graphs and handle open vertices.
 
 # ## Problem definition
-using GraphTensorNetworks, Graphs
+using GraphTensorNetworks, Graphs, Compose
 
 # In graph theory, a [maximal independent set](https://en.wikipedia.org/wiki/Maximal_independent_set) is an independent set that is not a subset of any other independent set.
 # It is different from maximum independent set because it does not require the set to have the max size.
@@ -26,6 +26,7 @@ locations = [[rot15(0.0, 1.0, i) for i=0:4]..., [rot15(0.0, 0.6, i) for i=0:4]..
 show_graph(graph; locs=locations)
 
 # ## Tensor network representation
+# Type [`MaximalIS`](@ref) can be used for constructing the tensor network with optimized contraction order for solving a maximal independent set problem.
 # For a vertex ``v\in V``, we define a boolean degree of freedom ``s_v\in\{0, 1\}``.
 # We defined the restriction on its neighbourhood ``N[v]``:
 # ```math
@@ -58,9 +59,16 @@ max_config = solve(problem, GraphPolynomial())[]
 
 # ### Configuration properties
 # ##### finding all maximal independent set
-max_edge_config = solve(problem, ConfigsAll())[]
+maximal_configs = solve(problem, ConfigsAll())[]
+
+imgs = ntuple(k->show_graph(graph;
+                            locs=locations, scale=0.25,
+                            vertex_colors=[iszero(maximal_configs[k][i]) ? "white" : "red"
+                            for i=1:nv(graph)]), length(maximal_configs));
+
+Compose.set_default_graphic_size(18cm, 12cm); Compose.compose(context(), ntuple(k->(context((mod1(k,5)-1)/5, ((k-1)÷5)/3, 1.2/5, 1.0/3), imgs[k]), 15)...)
 
 # This result should be consistent with that given by the [Bron Kerbosch algorithm](https://en.wikipedia.org/wiki/Bron%E2%80%93Kerbosch_algorithm) on the complement of Petersen graph.
-maximal_cliques = maximal_cliques(complement(graph))
+cliques = maximal_cliques(complement(graph))
 
 # For sparse graphs, the generic tensor network approach is usually much faster and memory efficient.

examples/PaintShop.jl

@@ -9,10 +9,96 @@
 #     * how to compute weighted graphs and handle open vertices.
 
 # ## Problme Definition
-# The [binary paint shop problem](http://m-hikari.com/ams/ams-2012/ams-93-96-2012/popovAMS93-96-2012-2.pdf).
+# The [binary paint shop problem](http://m-hikari.com/ams/ams-2012/ams-93-96-2012/popovAMS93-96-2012-2.pdf) is defined as follows:
+# we are given a ``2m`` length sequence containing ``m`` cars, where each car appears twice.
+# Each car need to be colored red in one occurrence, and blue in the other.
+# We need to choose which occurrence for each car to color with which color — the goal is to minimize the number of times we need to change the current color.
 
-# In the following, we are going to defined a binary paint shop problem for the following string
+# In the following, we use a character to represent a car,
+# and defined a binary paint shop problem as a string that each character appear exactly twice.
 
 using GraphTensorNetworks, Graphs
 
-sequence = "abaccb"
+sequence = collect("iadgbeadfcchghebif")
+
+# We can visualize this paint shop problem as a graph
+rot(a, b, θ) = cos(θ)*a + sin(θ)*b, cos(θ)*b - sin(θ)*a
+
+locations = [rot(0.0, 1.0, -0.25π - 1.5*π*(i-0.5)/length(sequence)) for i=1:length(sequence)]
+
+graph = path_graph(length(sequence))
+for i=1:length(sequence) 
+    j = findlast(==(sequence[i]), sequence)
+    i != j && add_edge!(graph, i, j)
+end
+
+show_graph(graph; locs=locations, texts=string.(sequence), edge_colors=[sequence[e.src] == sequence[e.dst] ? "blue" : "black" for e in edges(graph)])
+
+# Vertices connected by blue edges must have different colors,
+# and the goal becomes a min-cut problem defined on black edges.
+
+# ## Tensor network representation
+# Type [`PaintShop`](@ref) can be used for constructing the tensor network with optimized contraction order for solving a binary paint shop problem.
+# To obtain its tensor network representation, we associating car ``c_i`` (the ``i``-th character in our example) with a degree of freedom ``s_{c_i} \in \{0, 1\}``,
+# where we use ``0`` to denote the first appearance of a car is colored red and ``1`` to denote the first appearance of a car is colored blue.
+# For each black edges ``(i, i+1)``, we define an edge tensor labeld by ``(s_{c_i}, s_{c_{i+1}})`` as follows:
+# If both cars on this edge are their first or second appearance
+# ```math
+# B^{\rm parallel} = \begin{matrix}
+# x & 1 \\
+# 1 & x \\
+# \end{matrix},
+#
+# otherwise,
+# B^{\rm anti-parallel} = B^{\rm 10} = \begin{matrix}
+# 1 & x \\
+# x & 1 \\
+# \end{matrix}.
+# ```
+# It can be understood as, when both cars are their first appearance,
+# they tend to have the same configuration so that the color is not changed.
+# Otherwise, they tend to have different configuration to keep the color unchanged.
+
+# Let us contruct the problem instance as bellow.
+problem = PaintShop(sequence);
+
+# ### Counting properties
+# ##### maximal independence polynomial
+# The graph polynomial defined for the maximal independent set problem is
+# ```math
+# I_{\rm max}(G, x) = \sum_{k=0}^{\alpha(G)} b_k x^k,
+# ```
+# where ``b_k`` is the number of maximal independent sets of size ``k`` in graph ``G=(V, E)``.
+
+max_config = solve(problem, GraphPolynomial())[]
+
+# Since it only counts the maximal independent sets, the first several coefficients are 0.
+
+# ### Counting properties
+# ##### graph polynomial
+# The graph polynomial of the binary paint shop problem in our convension is defined as
+# ```math
+# D(G, x) = \sum_{k=0}^{\delta(G)} d_k x^k 
+# ```
+# where ``2d_k`` is the number of possible coloring with number of color changes ``2m-1-k``.
+paint_polynomial = solve(problem, GraphPolynomial())[]
+
+# ### Configuration properties
+# ##### finding best solutions
+best_configs = solve(problem, ConfigsMax())[]
+
+painting1 = paint_shop_coloring_from_config(best_configs.c.data[1]; initial=false)
+
+show_graph(graph; locs=locations, texts=string.(sequence), edge_colors=[sequence[e.src] == sequence[e.dst] ? "blue" : "black" for e in edges(graph)],
+                vertex_colors=[isone(c) ? "red" : "black" for c in painting1], vertex_text_color="white")
+
+#
+
+painting2 = paint_shop_coloring_from_config(best_configs.c.data[2]; initial=false)
+
+show_graph(graph; locs=locations, texts=string.(sequence), edge_colors=[sequence[e.src] == sequence[e.dst] ? "blue" : "black" for e in edges(graph)],
+                vertex_colors=[isone(c) ? "red" : "black" for c in painting2], vertex_text_color="white")
+
+# Since we have different choices of initial color, the number of best solution is 4.
+# The following function will check the solution and return you the number of color switchs
+num_paint_shop_color_switch(sequence, painting2)

src/GraphTensorNetworks.jl

@@ -15,6 +15,7 @@ export GreedyMethod, TreeSA, SABipartite, KaHyParBipartite, MergeVectors, MergeG
 export StaticBitVector, StaticElementVector, @bv_str
 export is_commutative_semiring
 export Max2Poly, TruncatedPoly, Polynomial, Tropical, CountingTropical, StaticElementVector, Mod, ConfigEnumerator, onehotv, ConfigSampler
+export CountingTropicalF64, CountingTropicalF32, TropicalF64, TropicalF32
 
 # Lower level APIs
 export AllConfigs, SingleConfig
@@ -25,11 +26,13 @@ export contractx, contractf, graph_polynomial, max_size, max_size_count
 # Graphs
 export random_regular_graph, diagonal_coupled_graph, is_independent_set
 export square_lattice_graph, unit_disk_graph, random_diagonal_coupled_graph, random_square_lattice_graph
+export line_graph
 
 # Tensor Networks (Graph problems)
 export GraphProblem, IndependentSet, MaximalIS, Matching, Coloring, optimize_code, set_packing, MaxCut, PaintShop, paintshop_from_pairs, UnWeighted
 export flavors, symbols, nflavor, get_weights
-export mis_compactify!, cut_assign, cut_size
+export mis_compactify!, cut_assign, cut_size, num_paint_shop_color_switch, paint_shop_coloring_from_config
+export is_good_vertex_coloring
 
 # Interfaces
 export solve, SizeMax, CountingAll, CountingMax, GraphPolynomial, SingleConfigMax, ConfigsAll, ConfigsMax
@@ -38,7 +41,7 @@ export solve, SizeMax, CountingAll, CountingMax, GraphPolynomial, SingleConfigMa
 export save_configs, load_configs
 
 # Visualization
-export show_graph
+export show_graph, spring_layout
 
 project_relative_path(xs...) = normpath(joinpath(dirname(dirname(pathof(@__MODULE__))), xs...))
 

src/arithematics.jl

@@ -220,15 +220,18 @@ end
 Base.:(==)(x::ConfigSampler{N,S,C}, y::ConfigSampler{N,S,C}) where {N,S,C} = x.data == y.data
 
 function Base.:+(x::ConfigSampler{N,S,C}, y::ConfigSampler{N,S,C}) where {N,S,C}  # biased sampling: return `x`, maybe using random sampler is better.
-    return x
+    return randn() > 0.5 ? x : y
 end
 
 function Base.:*(x::ConfigSampler{L,S,C}, y::ConfigSampler{L,S,C}) where {L,S,C}
     ConfigSampler(x.data | y.data)
 end
 
-Base.zero(::Type{ConfigSampler{N,S,C}}) where {N,S,C} = ConfigSampler{N,S,C}(statictrues(StaticElementVector{N,S,C}))
-Base.one(::Type{ConfigSampler{N,S,C}}) where {N,S,C} = ConfigSampler{N,S,C}(staticfalses(StaticElementVector{N,S,C}))
+@generated function Base.zero(::Type{ConfigSampler{N,S,C}}) where {N,S,C}
+    ex = Expr(:call, :(StaticElementVector{$N,$S,$C}), Expr(:tuple, fill(typemax(UInt64), C)...))
+    :(ConfigSampler{N,S,C}($ex))
+end
+Base.one(::Type{ConfigSampler{N,S,C}}) where {N,S,C} = ConfigSampler{N,S,C}(zero(StaticElementVector{N,S,C}))
 Base.zero(::ConfigSampler{N,S,C}) where {N,S,C} = zero(ConfigSampler{N,S,C})
 Base.one(::ConfigSampler{N,S,C}) where {N,S,C} = one(ConfigSampler{N,S,C})