Merge pull request #55 from alan-turing-institute/dev

Towards 0.7.0

Project.toml

@@ -1,12 +1,13 @@
 name = "MLJBase"
 uuid = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
 authors = ["Anthony D. Blaom <[email protected]>"]
-version = "0.6.0"
+version = "0.7.0"
 
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
+Missings = "e1d29d7a-bbdc-5cf2-9ac0-f12de2c33e28"
 OrderedCollections = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 Requires = "ae029012-a4dd-5104-9daa-d747884805df"
@@ -17,8 +18,8 @@ StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [compat]
+CategoricalArrays = "<0.5.3, 0.7"
 CSV = "0.5"
-CategoricalArrays = "<0.5.3"
 Requires = "^0.5.2"
 ScientificTypes = "0.2.0"
 Tables = "<0.1.19, >= 0.2"

src/MLJBase.jl

@@ -36,6 +36,7 @@ export is_feature_dependent                          # measures.jl
 export default_measure, value                        # measures.jl
 export mav, mae, rms, rmsl, rmslp1, rmsp, l1, l2     # measures.jl
 export misclassification_rate, cross_entropy         # measures.jl
+export BrierScore                                    # measures.jl
 
 # methods from other packages to be rexported:
 export pdf, mean, mode
@@ -66,6 +67,8 @@ import ScientificTypes: trait
 
 # to be extended:
 import StatsBase: fit, predict, fit!
+import Missings.levels
+
 
 # from Standard Library:
 using Statistics

src/data.jl

@@ -117,7 +117,7 @@ CategoricalElement{U} = Union{CategoricalValue{<:Any,U},CategoricalString{U}}
 """
     classes(x)
 
-All the categorical elements with the same pool as `x` (including `x`),
+All the categorical elements with in the same pool as `x` (including `x`),
 returned as a list, with an ordering consistent with the pool. Here
 `x` has `CategoricalValue` or `CategoricalString` type, and
 `classes(x)` is a vector of the same eltype.
@@ -139,10 +139,8 @@ true, `x in x.pool.levels` is not true.
      :c
 
 """
-function classes(x::CategoricalElement)
-    p = x.pool
-    return [p.valindex[p.invindex[v]] for v in p.levels]
-end
+classes(p::CategoricalPool) = [p.valindex[p.invindex[v]] for v in p.levels]
+classes(x::CategoricalElement) = classes(x.pool)
 
 """
    int(x)
@@ -176,11 +174,11 @@ Broadcasted versions of `int`.
 See also: [`decoder`](@ref).
 """
 int(x::CategoricalElement) = x.pool.order[x.pool.invindex[x]]
-int(A::AbstractArray{<:CategoricalElement}) = broadcast(int, A)
-# workaround for CategoricalArrays issue
-# https://github.com/JuliaData/CategoricalArrays.jl/issues/199:
-# function int(X::CategoricalArray)
-    
+int(A::AbstractArray) = broadcast(int, A)
+
+# get the integer representation of a level given pool (private
+# method):
+int(pool::CategoricalPool, level) =  pool.order[pool.invindex[level]] 
 
 struct CategoricalDecoder{T,R} # <: MLJType
     pool::CategoricalPool{T,R}
@@ -207,17 +205,13 @@ integer arrays, in which case `d` is broadcast over all elements.
     julia> d(int(v)) == v
     true
 
+*Warning:* It is *not* true that `int(d(u)) == u` always holds. 
+
 See also: [`int`](@ref), [`classes`](@ref).
 
 """
 decoder(element::CategoricalElement) =
     CategoricalDecoder(element.pool, sortperm(element.pool.order))
-## in the next lot need to skip the missing one
-# decoder(X::CategoricalArray) = CategoricalDecoder(X.pool)
-# function decoder(V::Array{<:CategoricalElement})
-#     isempty(V) && error("Unable to extract decoder from empty array. ")
-#     return X[1]
-# end
 
 (decoder::CategoricalDecoder{T,R})(i::Integer) where {T,R} =
     CategoricalValue{T,R}(decoder.invorder[i], decoder.pool)