Expose feature names

[ablaom]

Context: https://github.com/bensadeghi/DecisionTree.jl/issues/147

Addresses #13:

using MLJ
tree = (@load DecisionTreeClassifier pkg=DecisionTree)()
X, y = @load_iris
mach = machine(tree, X, y) |> fit!

julia> fitted_params(mach)
(tree = Decision Tree
Leaves: 9
Depth:  5,
 encoding = Dict{CategoricalArrays.CategoricalValue{String, UInt32}, UInt32}("virginica" => 0x00000003, "setosa" => 0x00000001, "versicolor" => 0x00000002),
 features = [:sepal_length, :sepal_width, :petal_length, :petal_width],)

julia> MLJ.report(mach)
(classes_seen = CategoricalArrays.CategoricalValue{String, UInt32}["setosa", "versicolor", "virginica"],
 print_tree = TreePrinter object (call with display depth),
 features = [:sepal_length, :sepal_width, :petal_length, :petal_width],)

[ablaom]

@roland-KA Could you confirm this meets requirements? If you have the time, a review would be nice.

[roland-KA]

That looks good 👍. If the index values of features correspond to the featid values of a DecisionTree.Node, then it's exactly what we need.

And with a short test with two different data sets I could confirm that.