Fix cycling inferrability

Author: gdalle

docs/src/dev.md

@@ -69,5 +69,5 @@ SparseMatrixColorings.efficient_fig_4
 ## Misc
 
 ```@docs
-SparseMatrixColorings.cycle_until
-```
+SparseMatrixColorings.cycle_range
+```

src/structured.jl

@@ -6,13 +6,16 @@ Question: when decompressing, should we always assume that the coloring was opti
 =#
 
 """
-    cycle_until(iterator, max_length::Integer)
+    cycle_range(k::Integer, n::Integer)
 
-Concatenate copies of `iterator` to fill a vector of length `max_length` (with one partial copy allowed at the end).
+Concatenate copies of `1:k` to fill a vector of length `n` (with one partial copy allowed at the end).
 """
-function cycle_until(iterator, max_length::Integer)
-    a = repeat(iterator, div(max_length, length(iterator)) + 1)
-    return resize!(a, max_length)
+function cycle_range(k::Integer, n::Integer)
+    color = Vector{Int}(undef, n)
+    for i in eachindex(color)
+        color[i] = 1 + (i - 1) % k
+    end
+    return color
 end
 
 ## Diagonal
@@ -67,7 +70,7 @@ function coloring(
     algo::GreedyColoringAlgorithm;
     kwargs...,
 )
-    color = cycle_until(1:2, size(A, 2))
+    color = cycle_range(1:2, size(A, 2))
     bg = BipartiteGraph(A)
     return ColumnColoringResult(A, bg, color)
 end
@@ -78,7 +81,7 @@ function coloring(
     algo::GreedyColoringAlgorithm;
     kwargs...,
 )
-    color = cycle_until(1:2, size(A, 1))
+    color = cycle_range(1:2, size(A, 1))
     bg = BipartiteGraph(A)
     return RowColoringResult(A, bg, color)
 end
@@ -123,7 +126,7 @@ function coloring(
     algo::GreedyColoringAlgorithm;
     kwargs...,
 )
-    color = cycle_until(1:3, size(A, 2))
+    color = cycle_range(1:3, size(A, 2))
     bg = BipartiteGraph(A)
     return ColumnColoringResult(A, bg, color)
 end
@@ -134,7 +137,7 @@ function coloring(
     algo::GreedyColoringAlgorithm;
     kwargs...,
 )
-    color = cycle_until(1:3, size(A, 1))
+    color = cycle_range(1:3, size(A, 1))
     bg = BipartiteGraph(A)
     return RowColoringResult(A, bg, color)
 end

test/structured.jl

@@ -1,12 +1,24 @@
 using LinearAlgebra
 using SparseMatrixColorings
+using SparseMatrixColorings: cycle_range
 using Test
 
 column_problem = ColoringProblem(; structure=:nonsymmetric, partition=:column)
 row_problem = ColoringProblem(; structure=:nonsymmetric, partition=:row)
 
 algo = GreedyColoringAlgorithm()
 
+@testset "Utils" begin
+    @test cycle_range(2, 3) == [1, 2, 1]
+    @test cycle_range(2, 4) == [1, 2, 1, 2]
+    @test cycle_range(2, 5) == [1, 2, 1, 2, 1]
+    @test cycle_range(3, 5) == [1, 2, 3, 1, 2]
+    @test cycle_range(3, 6) == [1, 2, 3, 1, 2, 3]
+    @test cycle_range(2, 1) == [1]
+    @test cycle_range(3, 1) == [1]
+    @test cycle_range(3, 2) == [1, 2]
+end
+
 @testset "Diagonal" begin
     for n in (1, 2, 10, 100, 1000)
         A = Diagonal(rand(n))