Merge pull request #4 from MichielStock/main

Major refactor PR

src/Xicor.jl

@@ -2,9 +2,39 @@ module Xicor
 
 using Distributions, Random, StatsBase
 
+"""
+    xicor(X, Y; rank=tiedrank, noties=false)
+
+Computes the asymmetric ξ correlation coefficient sbetween two vectors `X` and `Y`.
+`rank` species how the rank is computed, default uses `tiedrank`. If there
+are no ties in `Y`, `noties` can be set to `true`, which speeds up computation.
+
+`ξ` is an alias for this function.
+"""
+function xicor(X::AbstractVector, Y::AbstractVector; rank=denserank, noties=false)
+    @assert length(X) == length(Y) "length `X` and `Y` missmatch"
+    n = length(Y)
+    xsortperm = sortperm(X)
+    # todo: carfully check the rank function
+    r = rank(xsortperm, by=i->Y[i])
+    if !noties
+        l = rank(xsortperm, by=i->Y[i], rev=true)
+    end
+    ξn = 0.0
+    for i in 1:n-1
+        ξn += abs(r[i+1] - r[i])
+    end
+    ξn *= noties ? 3 / (n^2 - 1) : n / 2sum(li->li*(n-li), l)
+    ξn = 1 - ξn
+    return ξn
+end
+
+ξ = xicor
+
+
 include("correlation.jl")
 include("rank.jl")
 
-export xicor
+export xicor, ξ
 
 end

src/rank.jl

@@ -2,29 +2,33 @@ using Random, StatsBase
 
 
 """
+    maxrank(x)
+
 A function that returns the sample ranks using maximum ranks 
 for equal values.
     - x : A Vector of Floats, Integers, or Bools
 """
-function maxrank(x)
+function maxrank(x::AbstractVector)
     sorted_x = sort(x)
     ranks = zeros(Int, length(x))
-    for (idx,i) in enumerate(x)
+    for (idx, i) in enumerate(x)
         ranks[idx] = findlast(isequal(i), sorted_x)
     end
     return ranks
 end
 
 
 """
+    minrank(x)
+
 A function that returns the sample ranks using minimum ranks 
 for equal values.
     - x : A Vector of Floats, Integers, or Bools
 """
 function minrank(x)
     sorted_x = sort(x)
     ranks = zeros(Int, length(x))
-    for (idx,i) in enumerate(x)
+    for (idx, i) in enumerate(x)
         ranks[idx] = findfirst(isequal(i), sorted_x)
     end
     return ranks
@@ -38,11 +42,8 @@ of equal values.
     - seed : either an Integer to seed the random number 
              generator or `missing` to use the default seed.
 """
-function randrank(x; seed=missing)
-    @assert typeof(seed) <: Integer || typeof(seed) <: Missing
-    if !ismissing(seed)
-        Random.seed!(seed)
-    end
+function randrank(x; seed::Union{Nothing,Integer}=nothing)
+    @assert seed isa Integer || isnothing(seed)
     sorted_x = sort(x)
     ranks = zeros(Int, length(x))
     for (idx,i) in enumerate(x)
@@ -62,9 +63,8 @@ A function that returns the sample ranks of the values in a vector.
 TO DO: the original R `rank` function has an argument that let's you choose the
        placement of the missing values.
 """
-function rank(x; ties_method::String="average", seed=missing)
+function rank(x; ties_method::String="average", seed::Union{Nothing,Integer}=missing)
     @assert ties_method ∈ ["average", "random", "max", "min"]
-    @assert typeof(seed) <: Integer || typeof(seed) <: Missing
 
     # get NA/missing
     missing_values = ismissing.(x)

test/runtests.jl

@@ -2,5 +2,9 @@ using Xicor
 using Test
 
 @testset "Xicor.jl" begin
-    # Write your tests here.
+    X = [1, 2, 3, 5]
+    Y = [1.0, 3.0, 8.0, 10.0]
+
+    @test ξ(X, Y) > 0 
+    @test ξ(X, Y, rank=denserank) ≈ ξ(X, Y, rank=denserank, noties=true)
 end