Add Colorful scitype (#16)

* Add Colorful scitype

* Add tests

* Fix typo

* Update test/runtests.jl

---------

Co-authored-by: Elias Carvalho <[email protected]>

Author: juliohm

Project.toml

@@ -9,6 +9,7 @@ Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
 [weakdeps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 CoDa = "5900dafe-f573-5c72-b367-76665857777b"
+ColorTypes = "3da002f7-5984-5a60-b8a6-cbb66c0b333f"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 DynamicQuantities = "06fc5a27-2a28-4c7c-a15d-362465fb6821"
 Meshes = "eacbb407-ea5a-433e-ab97-5258b1ca43fa"
@@ -17,6 +18,7 @@ Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"
 [extensions]
 DataScienceTraitsCategoricalArraysExt = "CategoricalArrays"
 DataScienceTraitsCoDaExt = "CoDa"
+DataScienceTraitsColorTypesExt = "ColorTypes"
 DataScienceTraitsDistributionsExt = "Distributions"
 DataScienceTraitsDynamicQuantitiesExt = "DynamicQuantities"
 DataScienceTraitsMeshesExt = "Meshes"
@@ -25,6 +27,7 @@ DataScienceTraitsUnitfulExt = "Unitful"
 [compat]
 CategoricalArrays = "0.10"
 CoDa = "1.2"
+ColorTypes = "0.11"
 Dates = "1.9"
 Distributions = "0.25"
 DynamicQuantities = "0.7 - 0.14"

ext/DataScienceTraitsColorTypesExt.jl

@@ -0,0 +1,12 @@
+# -----------------------------------------------------------------
+# Licensed under the MIT License. See LICENSE in the project root.
+# -----------------------------------------------------------------
+
+module DataScienceTraitsColorTypesExt
+
+using DataScienceTraits
+using ColorTypes
+
+DataScienceTraits.scitype(::Type{<:Colorant}) = DataScienceTraits.Colorful
+
+end

src/DataScienceTraits.jl

@@ -27,24 +27,31 @@ Scientific type of categorical (a.k.a. discrete) variables.
 """
 abstract type Categorical <: SciType end
 
+"""
+    Colorful
+
+Scientific type of colorful data (see Colors.jl).
+"""
+abstract type Colorful <: SciType end
+
 """
     Compositional
 
-Scientific type of compositional data (See CoDa.jl).
+Scientific type of compositional data (see CoDa.jl).
 """
 abstract type Compositional <: SciType end
 
 """
     Distributional
 
-Scientific type of distributional data (See Distributions.jl)
+Scientific type of distributional data (see Distributions.jl)
 """
 abstract type Distributional <: SciType end
 
 """
     Geometrical
 
-Scientific type of geometrical data (See Meshes.jl)
+Scientific type of geometrical data (see Meshes.jl)
 """
 abstract type Geometrical <: SciType end
 
@@ -165,6 +172,7 @@ export
   # types
   Continuous,
   Categorical,
+  Colorful,
   Compositional,
   Distributional,
   Geometrical,

test/Project.toml

@@ -1,6 +1,7 @@
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 CoDa = "5900dafe-f573-5c72-b367-76665857777b"
+ColorTypes = "3da002f7-5984-5a60-b8a6-cbb66c0b333f"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 DynamicQuantities = "06fc5a27-2a28-4c7c-a15d-362465fb6821"

test/runtests.jl

@@ -1,13 +1,16 @@
 using DataScienceTraits
 using CategoricalArrays
+using ColorTypes
+using CoDa
 using Distributions
 using Meshes
-using CoDa
-import DynamicQuantities
-import Unitful
 using Dates
 using Test
 
+# import to avoid conflicts
+import DynamicQuantities
+import Unitful
+
 const DST = DataScienceTraits
 
 @testset "DataScienceTraits.jl" begin
@@ -148,16 +151,6 @@ const DST = DataScienceTraits
     @test elscitype(DST.coerce(DST.Categorical, [1.0, missing, 3.0])) <: DST.Categorical
   end
 
-  @testset "CoDa" begin
-    c1 = Composition(a=0.2, b=0.8)
-    c2 = Composition(a=0.5, b=0.5)
-    @test scitype(Composition{2,(:a, :b)}) <: DST.Compositional
-    @test scitype(c1) <: DST.Compositional
-    @test elscitype(Vector{Composition{2,(:a, :b)}}) <: DST.Compositional
-    @test elscitype([c1, c2]) <: DST.Compositional
-    @test elscitype([c1, missing, c2]) <: DST.Compositional
-  end
-
   @testset "Unitful" begin
     u = Unitful.u"m"
     q1 = 1 * u
@@ -247,6 +240,28 @@ const DST = DataScienceTraits
     @test DST.isordered(carr)
   end
 
+  @testset "Colors" begin
+    c1 = Gray(0)
+    c2 = RGB(0,0,0)
+    @test scitype(Gray) <: DST.Colorful
+    @test scitype(c1) <: DST.Colorful
+    @test scitype(RGB) <: DST.Colorful
+    @test scitype(c2) <: DST.Colorful
+    @test elscitype(Vector{Colorant}) <: DST.Colorful
+    @test elscitype([c1, c2]) <: DST.Colorful
+    @test elscitype([c1, missing, c2]) <: DST.Colorful
+  end
+
+  @testset "CoDa" begin
+    c1 = Composition(a=0.2, b=0.8)
+    c2 = Composition(a=0.5, b=0.5)
+    @test scitype(Composition{2,(:a, :b)}) <: DST.Compositional
+    @test scitype(c1) <: DST.Compositional
+    @test elscitype(Vector{Composition{2,(:a, :b)}}) <: DST.Compositional
+    @test elscitype([c1, c2]) <: DST.Compositional
+    @test elscitype([c1, missing, c2]) <: DST.Compositional
+  end
+
   @testset "Distributions" begin
     @test scitype(Normal()) <: DST.Distributional
     @test scitype(Exponential()) <: DST.Distributional