Merge pull request #17 from JuliaAI/row-table-datasets

Add row table datasets

Author: ablaom

README.md

@@ -54,11 +54,11 @@ failures, summary = MLJTestInterface.test(
 The following commands generate small datasets of the form `(X, y)` suitable for interface
 tests:
 
-- `MLJTestInterface.make_binary` 
+- `MLJTestInterface.make_binary(; row_table=false)` 
 
-- `MLJTestInterface.make_multiclass` 
+- `MLJTestInterface.make_multiclass(; row_table=false)` ` 
 
-- `MLJTestInterface.make_regression` 
+- `MLJTestInterface.make_regression(; row_table=false)` ` 
 
-- `MLJTestInterface.make_count` 
+- `MLJTestInterface.make_count(; row_table=false)` ` 
 

src/attemptors.jl

@@ -95,7 +95,7 @@ function operations(fitted_machine, data...; throw=false, verbosity=1)
         methods = MLJBase.implemented_methods(fitted_machine.model)
         if model isa Static && !(:transform in methods)
             push!(methods, :transform)
-        end 
+        end
         _, test = MLJBase.partition(1:MLJBase.nrows(first(data)), 0.5)
         if :predict in methods
             predict(fitted_machine, first(data))

src/datasets.jl

@@ -1,47 +1,59 @@
 """
-    make_binary()
+    make_binary(; row_table=false)
 
 Return data `(X, y)` for the crabs dataset, restricted to the two features `:FL`,
 `:RW`. Target is `Multiclass{2}`.
 
+The table `X` is a named tuple of vectors. For a vector of named tuples, set
+`row_table=true`.
+
 """
-function make_binary()
+function make_binary(; row_table=false)
     data = MLJBase.load_crabs()
     y_, X = unpack(data, ==(:sp), col->col in [:FL, :RW])
     y = coerce(y_, MLJBase.OrderedFactor)
-    return X, y
+    row_table ? (MLJBase.Tables.rowtable(X), y) : (X, y)
 end
 
 """
-    make_multiclass()
+    make_multiclass(; row_table=false)
 
 Return data `(X, y)` for the unshuffled iris dataset. Target is `Multiclass{3}`.
 
 """
-make_multiclass() = MLJBase.@load_iris
+function make_multiclass(; row_table=false)
+    X, y = MLJBase.@load_iris
+    row_table ? (MLJBase.Tables.rowtable(X), y) : (X, y)
+end
 
 """
-    make_regression()
+    make_regression(; row_table=false)
 
 Return data `(X, y)` for the Boston dataset, restricted to the two features `:LStat`,
 `:Rm`. Target is `Continuous`.
 
+The table `X` is a named tuple of vectors. For a vector of named tuples, set
+`row_table=true`.
+
 """
-function make_regression()
+function make_regression(; row_table=false)
     data = MLJBase.load_boston()
     y, X = unpack(data, ==(:MedV), col->col in [:LStat, :Rm])
-    return X, y
+    row_table ? (MLJBase.Tables.rowtable(X), y) : (X, y)
 end
 
 """
-    make_count()
+    make_count(; row_table=false)
 
 Return data `(X, y)` for the Boston dataset, restricted to the two features `:LStat`,
 `:Rm`, with the `Continuous` target converted to `Count` (integer).
 
+The table `X` is a named tuple of vectors. For a vector of named tuples, set
+`row_table=true`.
+
 """
-function make_count()
+function make_count(; row_table=false)
     X, y_ = make_regression()
     y = map(η -> round(Int, η), y_)
-    return X, y
+    row_table ? (MLJBase.Tables.rowtable(X), y) : (X, y)
 end

test/datasets.jl

@@ -0,0 +1,39 @@
+@testset "loading of datasets" begin
+    X, y = MTI.make_binary()
+    @test X isa NamedTuple
+    @test first(X) isa AbstractVector{Float64}
+    @test MLJBase.scitype(y) == AbstractVector{MLJBase.OrderedFactor{2}}
+    Xr, yr = MTI.make_binary(row_table=true)
+    @test Xr isa AbstractVector
+    @test MLJBase.Tables.rowtable(X) == Xr
+    @test yr == y
+
+    X, y = MTI.make_multiclass()
+    @test X isa NamedTuple
+    @test first(X) isa AbstractVector{Float64}
+    @test MLJBase.scitype(y) == AbstractVector{MLJBase.Multiclass{3}}
+    Xr, yr = MTI.make_multiclass(row_table=true)
+    @test Xr isa AbstractVector
+    @test MLJBase.Tables.rowtable(X) == Xr
+    @test yr == y
+
+    X, y = MTI.make_regression()
+    @test X isa NamedTuple
+    @test first(X) isa AbstractVector{Float64}
+    @test MLJBase.scitype(y) == AbstractVector{MLJBase.Continuous}
+    Xr, yr = MTI.make_regression(row_table=true)
+    @test Xr isa AbstractVector
+    @test MLJBase.Tables.rowtable(X) == Xr
+    @test yr == y
+
+    X, y = MTI.make_count()
+    @test X isa NamedTuple
+    @test first(X) isa AbstractVector{Float64}
+    @test MLJBase.scitype(y) == AbstractVector{MLJBase.Count}
+    Xr, yr = MTI.make_count(row_table=true)
+    @test Xr isa AbstractVector
+    @test MLJBase.Tables.rowtable(X) == Xr
+    @test yr == y
+end
+
+true

test/runtests.jl

@@ -20,6 +20,10 @@ macro conditional_testset(name, expr)
     end)
 end
 
+@conditional_testset "datasets" begin
+    include("datasets.jl")
+end
+
 @conditional_testset "attemptors" begin
     include("attemptors.jl")
 end