Complete colorvec transition

Author: ChrisRackauckas

README.md

@@ -44,7 +44,7 @@ sparsity_pattern = sparsity!(f,output,input)
 jac = Float64.(sparse(sparsity_pattern))
 ```
 
-Now we call `matrix_colors` to get the color vector for that matrix:
+Now we call `matrix_colors` to get the colorvec vector for that matrix:
 
 ```julia
 colors = matrix_colors(jac)
@@ -54,14 +54,14 @@ Since `maximum(colors)` is 3, this means that finite differencing can now
 compute the Jacobian in just 4 `f`-evaluations:
 
 ```julia
-DiffEqDiffTools.finite_difference_jacobian!(jac, f, rand(30), color=colors)
+DiffEqDiffTools.finite_difference_jacobian!(jac, f, rand(30), colorvec=colors)
 @show fcalls # 4
 ```
 
 In addition, a faster forward-mode autodiff call can be utilized as well:
 
 ```julia
-forwarddiff_color_jacobian!(jac, f, x, color = colors)
+forwarddiff_color_jacobian!(jac, f, x, colorvec = colors)
 ```
 
 If one only need to compute products, one can use the operators. For example,
@@ -102,7 +102,7 @@ requires an `f` call to `maximum(colors)+1`, or reduces automatic differentiatio
 to using `maximum(colors)` partials. Since normally these values are `length(x)`,
 this can be significant savings.
 
-The API for computing the color vector is:
+The API for computing the colorvec vector is:
 
 ```julia
 matrix_colors(A::AbstractMatrix,alg::ColoringAlgorithm = GreedyD1Color();
@@ -113,31 +113,31 @@ The first argument is the abstract matrix which represents the sparsity pattern
 of the Jacobian. The second argument is the optional choice of coloring algorithm.
 It will default to a greedy distance 1 coloring, though if your special matrix
 type has more information, like is a `Tridiagonal` or `BlockBandedMatrix`, the
-color vector will be analytically calculated instead. The keyword argument
+colorvec vector will be analytically calculated instead. The keyword argument
 `partition_by_rows` allows you to partition the Jacobian on the basis of rows instead
 of columns and generate a corresponding coloring vector which can be used for
 reverse-mode AD. Default value is false.
 
-The result is a vector which assigns a color to each column (or row) of the matrix.
+The result is a vector which assigns a colorvec to each column (or row) of the matrix.
 
-### Color-Assisted Differentiation
+### Colorvec-Assisted Differentiation
 
-Color-assisted differentiation for numerical differentiation is provided by
+Colorvec-assisted differentiation for numerical differentiation is provided by
 DiffEqDiffTools.jl and for automatic differentiation is provided by
 ForwardDiff.jl.
 
-For DiffEqDiffTools.jl, one simply has to use the provided `color` keyword
+For DiffEqDiffTools.jl, one simply has to use the provided `colorvec` keyword
 argument. See [the DiffEqDiffTools Jacobian documentation](https://github.com/JuliaDiffEq/DiffEqDiffTools.jl#jacobians) for more details.
 
-For forward-mode automatic differentiation, use of a color vector is provided
+For forward-mode automatic differentiation, use of a colorvec vector is provided
 by the following function:
 
 ```julia
 forwarddiff_color_jacobian!(J::AbstractMatrix{<:Number},
                             f,
                             x::AbstractArray{<:Number};
                             dx = nothing,
-                            color = eachindex(x),
+                            colorvec = eachindex(x),
                             sparsity = nothing)
 ```
 
@@ -147,7 +147,7 @@ cache, construct the cache in advance:
 ```julia
 ForwardColorJacCache(f,x,_chunksize = nothing;
                               dx = nothing,
-                              color=1:length(x),
+                              colorvec=1:length(x),
                               sparsity = nothing)
 ```
 

src/coloring/backtracking_coloring.jl

@@ -20,7 +20,7 @@ function color_graph(g::LightGraphs.AbstractGraph, ::BacktrackingColor)
 
     start = 1
 
-    #optimal color number
+    #optimal chromatic number
     opt = v + 1
 
     #current vertex to be colored
@@ -135,7 +135,7 @@ end
                 opt::Integer)
 
 Returns set of free colors of x which are less
-than optimal color number (opt)
+than optimal chromatic number (opt)
 
 Arguments:
 
@@ -192,7 +192,7 @@ end
     least_index(F::AbstractVector{<:Integer}, A::AbstractVector{<:Integer}, opt::Integer)
 
 Returns least index i such that color of vertex
-A[i] is equal to `opt` (optimal color number)
+A[i] is equal to `opt` (optimal chromatic number)
 """
 function least_index(F::AbstractVector{<:Integer}, A::AbstractVector{<:Integer}, opt::Integer)
     for i in eachindex(A)
@@ -218,7 +218,7 @@ end
     remove_higher_colors(U::AbstractVector{<:Integer}, opt::Integer)
 
 Remove all the colors which are greater than or
-equal to the `opt` (optimal color number) from
+equal to the `opt` (optimal chromatic number) from
 the set of colors U
 """
 function remove_higher_colors(U::AbstractVector{<:Integer}, opt::Integer)

src/coloring/greedy_star1_coloring.jl

@@ -23,26 +23,26 @@
 """
 function color_graph(g::LightGraphs.AbstractGraph, ::GreedyStar1Color)
     v = nv(g)
-    color = zeros(Int, v)
+    colorvec = zeros(Int, v)
 
     forbidden_colors = zeros(Int, v+1)
 
     for vertex_i = vertices(g)
 
         for w in inneighbors(g, vertex_i)
-            if color[w] != 0
-                forbidden_colors[color[w]] = vertex_i
+            if colorvec[w] != 0
+                forbidden_colors[colorvec[w]] = vertex_i
             end
 
             for x in inneighbors(g, w)
-                if color[x] != 0
-                    if color[w] == 0
-                        forbidden_colors[color[x]] = vertex_i
+                if colorvec[x] != 0
+                    if colorvec[w] == 0
+                        forbidden_colors[colorvec[x]] = vertex_i
                     else
                         for y in inneighbors(g, x)
-                           if color[y] != 0
-                                if y != w && color[y] == color[w]
-                                    forbidden_colors[color[x]] = vertex_i
+                           if colorvec[y] != 0
+                                if y != w && colorvec[y] == colorvec[w]
+                                    forbidden_colors[colorvec[x]] = vertex_i
                                     break
                                 end
                             end
@@ -52,10 +52,10 @@ function color_graph(g::LightGraphs.AbstractGraph, ::GreedyStar1Color)
             end
         end
 
-        color[vertex_i] = find_min_color(forbidden_colors, vertex_i)
+        colorvec[vertex_i] = find_min_color(forbidden_colors, vertex_i)
     end
 
-    color
+    colorvec
 end
 
 function find_min_color(forbidden_colors::AbstractVector, vertex_i::Integer)

src/coloring/greedy_star2_coloring.jl

@@ -26,32 +26,32 @@
 """
 function color_graph(g::LightGraphs.AbstractGraph, :: GreedyStar2Color)
     v = nv(g)
-    color = zeros(Int, v)
+    colorvec = zeros(Int, v)
 
     forbidden_colors = zeros(Int, v+1)
 
     for vertex_i = vertices(g)
 
         for w in inneighbors(g, vertex_i)
-            if color[w] != 0
-                forbidden_colors[color[w]] = vertex_i
+            if colorvec[w] != 0
+                forbidden_colors[colorvec[w]] = vertex_i
             end
 
             for x in inneighbors(g, w)
-                if color[x] != 0
-                    if color[w] == 0
-                        forbidden_colors[color[x]] = vertex_i
+                if colorvec[x] != 0
+                    if colorvec[w] == 0
+                        forbidden_colors[colorvec[x]] = vertex_i
                     else
-                        if color[x] < color[w]
-                            forbidden_colors[color[x]] = vertex_i
+                        if colorvec[x] < colorvec[w]
+                            forbidden_colors[colorvec[x]] = vertex_i
                         end
                     end
                 end
             end
         end
 
-        color[vertex_i] = find_min_color(forbidden_colors, vertex_i)
+        colorvec[vertex_i] = find_min_color(forbidden_colors, vertex_i)
     end
 
-    color
+    colorvec
 end

src/coloring/high_level.jl

@@ -8,7 +8,7 @@ struct GreedyStar2Color <: SparseDiffToolsColoringAlgorithm end
 """
     matrix_colors(A,alg::ColoringAlgorithm = GreedyD1Color())
 
-    Returns the color vector for the matrix A using the chosen coloring
+    Returns the colorvec vector for the matrix A using the chosen coloring
     algorithm. If a known analytical solution exists, that is used instead.
     The coloring defaults to a greedy distance-1 coloring.
 

test/runtests.jl

@@ -9,7 +9,7 @@ if GROUP == "All"
     @time @safetestset "Greedy distance-1 coloring" begin include("test_greedy_d1.jl") end
     @time @safetestset "Greedy star coloring" begin include("test_greedy_star.jl") end
     @time @safetestset "Matrix to graph conversion" begin include("test_matrix2graph.jl") end
-    @time @safetestset "AD using color vector" begin include("test_ad.jl") end
+    @time @safetestset "AD using colorvec vector" begin include("test_ad.jl") end
     @time @safetestset "Integration test" begin include("test_integration.jl") end
     @time @safetestset "Special matrices" begin include("test_specialmatrices.jl") end
     @time @safetestset "Jac Vecs and Hes Vecs" begin include("test_jaches_products.jl") end

test/test_ad.jl

@@ -33,31 +33,31 @@ _J = sparse(J)
 
 fcalls = 0
 _J1 = similar(_J)
-forwarddiff_color_jacobian!(_J1, f, x, color = repeat(1:3,10))
+forwarddiff_color_jacobian!(_J1, f, x, colorvec = repeat(1:3,10))
 @test _J1 ≈ J
 @test fcalls == 1
 
 fcalls = 0
-jac_cache = ForwardColorJacCache(f,x,color = repeat(1:3,10), sparsity = _J1)
+jac_cache = ForwardColorJacCache(f,x,colorvec = repeat(1:3,10), sparsity = _J1)
 forwarddiff_color_jacobian!(_J1, f, x, jac_cache)
 @test _J1 ≈ J
 @test fcalls == 1
 
 fcalls = 0
 _J1 = similar(_J)
 _denseJ1 = collect(_J1)
-forwarddiff_color_jacobian!(_denseJ1, f, x, color = repeat(1:3,10), sparsity = _J1)
+forwarddiff_color_jacobian!(_denseJ1, f, x, colorvec = repeat(1:3,10), sparsity = _J1)
 @test _denseJ1 ≈ J
 @test fcalls == 1
 
 fcalls = 0
 _J1 = similar(_J)
 _denseJ1 = collect(_J1)
-jac_cache = ForwardColorJacCache(f,x,color = repeat(1:3,10), sparsity = _J1)
+jac_cache = ForwardColorJacCache(f,x,colorvec = repeat(1:3,10), sparsity = _J1)
 forwarddiff_color_jacobian!(_denseJ1, f, x, jac_cache)
 @test _denseJ1 ≈ J
 @test fcalls == 1
 
 _Jt = similar(Tridiagonal(J))
-forwarddiff_color_jacobian!(_Jt, f, x, color = repeat(1:3,10), sparsity = _Jt)
+forwarddiff_color_jacobian!(_Jt, f, x, colorvec = repeat(1:3,10), sparsity = _Jt)
 @test _Jt ≈ J

test/test_gpu_ad.jl

@@ -13,6 +13,6 @@ x = cu(rand(30))
 CuArrays.allowscalar(false)
 forwarddiff_color_jacobian!(_denseJ1, f, x)
 forwarddiff_color_jacobian!(_denseJ1, f, x, sparsity = _J1)
-forwarddiff_color_jacobian!(_denseJ1, f, x, color = repeat(1:3,10), sparsity = _J1)
+forwarddiff_color_jacobian!(_denseJ1, f, x, colorvec = repeat(1:3,10), sparsity = _J1)
 _Jt = similar(Tridiagonal(_J1))
-@test_broken forwarddiff_color_jacobian!(_denseJ1, f, x, color = repeat(1:3,10), sparsity = _Jt)
+@test_broken forwarddiff_color_jacobian!(_denseJ1, f, x, colorvec = repeat(1:3,10), sparsity = _Jt)

test/test_integration.jl

@@ -39,13 +39,13 @@ J = finite_difference_jacobian(f, rand(30))
 #Jacobian computed with coloring vectors
 fcalls = 0
 _J = similar(true_jac)
-finite_difference_jacobian!(_J, f, rand(30), color = colors)
+finite_difference_jacobian!(_J, f, rand(30), colorvec = colors)
 @test fcalls == 4
 @test _J ≈ J
 
 fcalls = 0
 _J = similar(true_jac)
 _denseJ = collect(_J)
-finite_difference_jacobian!(_denseJ, f, rand(30), color = colors, sparsity=_J)
+finite_difference_jacobian!(_denseJ, f, rand(30), colorvec = colors, sparsity=_J)
 @test fcalls == 4
 @test _denseJ ≈ J

test/test_specialmatrices.jl

@@ -41,8 +41,8 @@ bandedblockbanded3 = BandedBlockBandedMatrix(Zeros(2 * 4 , 2 * 4), ([4, 4] ,[4,
 function _testvalidity(A)
     colorvec=matrix_colors(A)
     ncolor=maximum(colorvec)
-    for color in 1:ncolor
-        subA=A[:,findall(x->x==color,colorvec)]
+    for colorvec in 1:ncolor
+        subA=A[:,findall(x->x==colorvec,colorvec)]
         @test maximum(sum(subA,dims=2))<=1.0
     end
 end