Merge pull request #468 from alan-turing-institute/dev

Last minute addition to 0.16.2 release

src/MLJBase.jl

@@ -1,4 +1,4 @@
-module MLJBase 
+module MLJBase
 
 # ===================================================================
 # IMPORTS
@@ -134,7 +134,7 @@ export params # note this is *not* an extension of StatsBase.params
 export partition, unpack, complement, restrict, corestrict
 
 # utilities.jl:
-export @set_defaults, flat_values, recursive_setproperty!,
+export flat_values, recursive_setproperty!,
        recursive_getproperty, pretty, unwind
 
 # show.jl
@@ -162,7 +162,7 @@ export machine, Machine, fit!, report, fit_only!
 export make_blobs, make_moons, make_circles, make_regression
 
 # composition:
-export machines, sources, anonymize!, @from_network, @pipeline,
+export machines, sources, @from_network, @pipeline,
     glb, @tuple, node, @node, sources, origins, return!,
     nrows_at_source, machine,
     rebind!, nodes, freeze!, thaw!, models, Node, AbstractNode,
@@ -284,7 +284,7 @@ export DWDMarginLoss, ExpLoss, L1HingeLoss, L2HingeLoss, L2MarginLoss,
     SmoothedL1HingeLoss, ZeroOneLoss, HuberLoss, L1EpsilonInsLoss,
     L2EpsilonInsLoss, LPDistLoss, LogitDistLoss, PeriodicLoss,
     QuantileLoss
- 
+
 # -------------------------------------------------------------------
 # re-export from Random, StatsBase, Statistics, Distributions,
 # OrderedCollections, CategoricalArrays, InvertedIndices:

src/composition/learning_networks/inspection.jl

@@ -102,7 +102,7 @@ function machines(W::Node, model=nothing)
                     (machines(arg) for arg in W.args)...,
                     (machines(arg) for arg in W.machine.args)...) |> unique
     end
-    model == nothing && return machs
+    model === nothing && return machs
     return filter(machs) do mach
         mach.model == model
     end

src/measures/finite.jl

@@ -4,7 +4,7 @@
 # -----------------------------------------------------
 # LogLoss
 
-struct LogLoss{R} <: Measure where R <: Real
+struct LogLoss{R <: Real} <: Measure
     tol::R
 end
 LogLoss(;eps=eps(), tol=eps) = LogLoss(tol)
@@ -107,12 +107,12 @@ function (::BrierScore)(ŷ::Vec{<:UnivariateFinite},
                         y::Vec,
                         w::Union{Nothing,Vec{<:Real}}=nothing)
     check_dimensions(ŷ, y)
-    w == nothing || check_dimensions(w, y)
+    w === nothing || check_dimensions(w, y)
 
     check_pools(ŷ, y)
     unweighted = broadcast(_brier_score, ŷ, y)
 
-    if w == nothing
+    if w === nothing
         return unweighted
     end
     return w.*unweighted
@@ -125,7 +125,7 @@ function (::BrierScore)(
     w::Union{Nothing,Vec{<:Real}}=nothing) where {S,V,R,P<:Real}
 
     check_dimensions(ŷ, y)
-    w == nothing || check_dimensions(w, y)
+    w === nothing || check_dimensions(w, y)
 
     isempty(y) && return P(0)
 
@@ -136,7 +136,7 @@ function (::BrierScore)(
 
     unweighted = P(2) .* broadcast(pdf, ŷ, y) .- offset
 
-    if w == nothing
+    if w === nothing
         return unweighted
     end
     return w.*unweighted

src/resampling.jl

@@ -445,7 +445,7 @@ function _process_accel_settings(accel::CPUThreads)
       throw(ArgumentError("`n`used in `acceleration = CPUThreads(n)`must" *
                         "be an instance of type `T<:Signed`"))
       accel.settings > 0 ||
-            throw(error("Can't create $(acceleration.settings) tasks)"))
+            throw(error("Can't create $(accel.settings) tasks)"))
       _accel = accel
     end
     return _accel
@@ -790,7 +790,7 @@ function evaluate!(mach::Machine, resampling, weights,
         fit!(mach; rows=train, verbosity=verbosity - 1, force=force)
         Xtest = selectrows(X, test)
         ytest = selectrows(y, test)
-        if weights == nothing
+        if weights === nothing
             wtest = nothing
         else
             wtest = weights[test]

src/univariate_finite/arrays.jl

@@ -96,7 +96,7 @@ for func in [:pdf, :logpdf]
         function Distributions.$func(
             u::AbstractArray{UnivariateFinite{S,V,R,P},N},
             C::AbstractVector{<:Union{V, CategoricalValue{V,R}}}) where {S,V,R,P,N}
-        
+
             #ret = Array{P,N+1}(undef, size(u)..., length(C))
             ret = zeros(P, size(u)..., length(C))
             for i in eachindex(C)
@@ -117,9 +117,9 @@ end
 # v - a vector of CategoricalArrays
 
 # dummy function
-# returns `x[i]` for `Array` inputs `x` 
+# returns `x[i]` for `Array` inputs `x`
 # For non-Array inputs returns `zero(dtype)`
-#This avoids using an if statement 
+#This avoids using an if statement
 _getindex(x::Array,i, dtype)=x[i]
 _getindex(::Nothing, i, dtype) = zero(dtype)
 
@@ -130,15 +130,15 @@ function Base.Broadcast.broadcasted(
     cv::CategoricalValue) where {S,V,R,P,N}
 
     cv in classes(u) || _err_missing_class(cv)
-    
-    f() = zeros(P, size(u)) #default caller function 
-    
+
+    f() = zeros(P, size(u)) #default caller function
+
     return Base.Broadcast.Broadcasted(
     	identity,
     	(get(f, u.prob_given_ref, int(cv)),)
     	)
 end
-        
+
 # pdf.(u, v)
 function Base.Broadcast.broadcasted(
     ::typeof(pdf),
@@ -149,14 +149,14 @@ function Base.Broadcast.broadcasted(
         "Arrays could not be broadcast to a common size; "*
         "got a dimension with lengths $(length(u)) and $(length(v))"))
     for cv in v
-        cv in classes(u) || _err_missing_class(c)
+        cv in classes(u) || _err_missing_class(cv)
     end
 
     # will use linear indexing:
     v_flat = ((v[i], i) for i in 1:length(v))
-    
+
     getter((cv, i), dtype) = _getindex(get(u.prob_given_ref, int(cv), nothing), i, dtype)
-    
+
     ret_flat = getter.(v_flat, P)
     return reshape(ret_flat, size(u))
 end
@@ -173,9 +173,9 @@ end
 
 # logpdf.(u::UniFinArr{S,V,R,P,N}, cv::CategoricalValue)
 # logpdf.(u::UniFinArr{S,V,R,P,N}, v::AbstractArray{<:CategoricalValue{V,R},N})
-# logpdf.(u::UniFinArr{S,V,R,P,N}, raw::AbstractArray{V,N}) 
-# logpdf.(u::UniFinArr{S,V,R,P,N}, raw::V) 
-for typ in (:CategoricalValue, 
+# logpdf.(u::UniFinArr{S,V,R,P,N}, raw::AbstractArray{V,N})
+# logpdf.(u::UniFinArr{S,V,R,P,N}, raw::V)
+for typ in (:CategoricalValue,
 	    :(AbstractArray{<:CategoricalValue{V,R},N}),
 	    :V,
 	    :(AbstractArray{V,N}))
@@ -188,11 +188,11 @@ for typ in (:CategoricalValue,
 
     	    # Start with the pdf array
     	    # take advantage of loop fusion
-    	    result = log.(pdf.(u, c)) 
+    	    result = log.(pdf.(u, c))
  	    return result
     	end
     	end)
-    	
+
   else
   	eval(quote
     	function Base.Broadcast.broadcasted(
@@ -201,18 +201,18 @@ for typ in (:CategoricalValue,
 	    	     		c::$typ) where {S,V,R,P,N}
 
     	    # Start with the pdf array
-    	    result = pdf.(u, c) 
-	    	    
+    	    result = pdf.(u, c)
+
 	    # Take the log of each entry in-place
   	    @simd for j in eachindex(result)
   	    	@inbounds result[j] = log(result[j])
 	    end
-		
+
  	    return result
     	end
     	end)
   end
-    
+
 end
 
 ## PERFORMANT BROADCASTING OF mode:

src/utilities.jl

@@ -112,7 +112,7 @@ function reduce_nested_field(ex)
     tail = ex.args[2]
     tail isa QuoteNode || throw(ArgumentError)
     field = tail.value
-    field isa Symbol || throw(ArgmentError)
+    field isa Symbol || throw(ArgumentError)
     subex = ex.args[1]
     return (subex, field)
 end