Address review comments

Author: andreasnoack

src/Loess.jl

@@ -44,6 +44,10 @@ function loess(
     degree::Integer = 2,
     cell::AbstractFloat = 0.2
 ) where T<:AbstractFloat
+
+    Base.require_one_based_indexing(xs)
+    Base.require_one_based_indexing(ys)
+
     if size(xs, 1) != size(ys, 1)
         throw(DimensionMismatch("Predictor and response arrays must of the same length"))
     end
@@ -180,15 +184,15 @@ function predict(model::LoessModel, zs::AbstractVector)
         throw(ArgumentError("Multivariate blending not yet implemented"))
     end
 
-    predict.(Ref(model), zs)
+    return [predict(model, z) for z in zs]
 end
 
 function predict(model::LoessModel, zs::AbstractMatrix)
     if size(model.xs, 2) > 1
         throw(ArgumentError("Multivariate blending not yet implemented"))
     end
 
-    return [predict(model, z) for z in vec(zs)]
+    return predict(model, vec(zs))
 end
 
 """

src/kd.jl

@@ -256,7 +256,6 @@ function traverse(kdtree::KDTree{T}, x::NTuple{N,T}) where {N,T}
 
     for j in 1:N
         if x[j] < kdtree.bounds[1, j] || x[j] > kdtree.bounds[2, j]
-            @show x, kdtree.bounds
             error(
                   """
                   Loess cannot perform extrapolation. Predict can only be applied