Replace broadcasting over distributions with broadcasting with partially applied functions

src/deprecates.jl

@@ -34,13 +34,13 @@ for fun in [:pdf, :logpdf,
     fun! = Symbol(fun, '!')
 
     @eval begin
-        @deprecate ($_fun!)(r::AbstractArray{<:Real}, d::UnivariateDistribution, X::AbstractArray{<:Real}) r .= ($fun).(d, X) false
-        @deprecate ($fun!)(r::AbstractArray{<:Real}, d::UnivariateDistribution, X::AbstractArray{<:Real}) r .= ($fun).(d, X) false
-        @deprecate ($fun)(d::UnivariateDistribution, X::AbstractArray{<:Real}) ($fun).(d, X)
+        @deprecate ($_fun!)(r::AbstractArray{<:Real}, d::UnivariateDistribution, X::AbstractArray{<:Real}) r .= Base.Fix1($fun, d).(X) false
+        @deprecate ($fun!)(r::AbstractArray{<:Real}, d::UnivariateDistribution, X::AbstractArray{<:Real}) r .= Base.Fix1($fun, d).(X) false
+        @deprecate ($fun)(d::UnivariateDistribution, X::AbstractArray{<:Real}) map(Base.Fix1($fun, d), X)
     end
 end
 
-@deprecate pdf(d::DiscreteUnivariateDistribution) pdf.(Ref(d), support(d))
+@deprecate pdf(d::DiscreteUnivariateDistribution) map(Base.Fix1(pdf, d), support(d))
 
 # Wishart constructors
 @deprecate Wishart(df::Real, S::AbstractPDMat, warn::Bool) Wishart(df, S)

src/mixtures/mixturemodel.jl

@@ -297,7 +297,7 @@ function _mixpdf!(r::AbstractArray, d::AbstractMixtureModel, x)
         pi = p[i]
         if pi > 0.0
             if d isa UnivariateMixture
-                t .= pdf.(component(d, i), x)
+                t .= Base.Fix1(pdf, component(d, i)).(x)
             else
                 pdf!(t, component(d, i), x)
             end
@@ -326,7 +326,7 @@ function _mixlogpdf!(r::AbstractArray, d::AbstractMixtureModel, x)
             lp_i = view(Lp, :, i)
             # compute logpdf in batch and store
             if d isa UnivariateMixture
-                lp_i .= logpdf.(component(d, i), x)
+                lp_i .= Base.Fix1(logpdf, component(d, i)).(x)
             else
                 logpdf!(lp_i, component(d, i), x)
             end
@@ -398,7 +398,7 @@ function _cwise_pdf!(r::AbstractMatrix, d::AbstractMixtureModel, X)
     size(r) == (n, K) || error("The size of r is incorrect.")
     for i = 1:K
         if d isa UnivariateMixture
-            view(r,:,i) .= pdf.(Ref(component(d, i)), X)
+            view(r,:,i) .= Base.Fix1(pdf, component(d, i)).(X)
         else
             pdf!(view(r,:,i),component(d, i), X)
         end
@@ -412,7 +412,7 @@ function _cwise_logpdf!(r::AbstractMatrix, d::AbstractMixtureModel, X)
     size(r) == (n, K) || error("The size of r is incorrect.")
     for i = 1:K
         if d isa UnivariateMixture
-            view(r,:,i) .= logpdf.(Ref(component(d, i)), X)
+            view(r,:,i) .= Base.Fix1(logpdf, component(d, i)).(X)
         else
             logpdf!(view(r,:,i), component(d, i), X)
         end

src/multivariate/jointorderstatistics.jl

@@ -162,7 +162,7 @@ function _rand!(rng::AbstractRNG, d::JointOrderStatistics, x::AbstractVector{<:R
         else
             s += randexp(rng, T)
         end
-        x .= quantile.(d.dist, x ./ s)
+        x .= Base.Fix1(quantile, d.dist).(x ./ s)
     end
     return x
 end

src/qq.jl

@@ -35,14 +35,14 @@ function qqbuild(x::AbstractVector, d::UnivariateDistribution)
     n = length(x)
     grid = ppoints(n)
     qx = quantile(x, grid)
-    qd = quantile.(Ref(d), grid)
+    qd = map(Base.Fix1(quantile, d), grid)
     return QQPair(qx, qd)
 end
 
 function qqbuild(d::UnivariateDistribution, x::AbstractVector)
     n = length(x)
     grid = ppoints(n)
-    qd = quantile.(Ref(d), grid)
+    qd = map(Base.Fix1(quantile, d), grid)
     qx = quantile(x, grid)
     return QQPair(qd, qx)
 end

src/univariate/continuous/uniform.jl

@@ -154,7 +154,7 @@ Base.:*(c::Real, d::Uniform) = Uniform(minmax(c * d.a, c * d.b)...)
 rand(rng::AbstractRNG, d::Uniform) = d.a + (d.b - d.a) * rand(rng)
 
 _rand!(rng::AbstractRNG, d::Uniform, A::AbstractArray{<:Real}) =
-    A .= quantile.(d, rand!(rng, A))
+    A .= Base.Fix1(quantile, d).(rand!(rng, A))
 
 
 #### Fitting

src/univariate/discrete/betabinomial.jl

@@ -103,12 +103,15 @@ for f in (:ccdf, :logcdf, :logccdf)
     end
 end
 
-entropy(d::BetaBinomial) = entropy(Categorical(pdf.(Ref(d),support(d))))
-median(d::BetaBinomial) = median(Categorical(pdf.(Ref(d),support(d)))) - 1
-mode(d::BetaBinomial) = argmax(pdf.(Ref(d),support(d))) - 1
-modes(d::BetaBinomial) = modes(Categorical(pdf.(Ref(d),support(d)))) .- 1
+# Shifted categorical distribution corresponding to `BetaBinomial`
+_categorical(d::BetaBinomial) = Categorical(map(Base.Fix1(pdf, d), support(d)))
 
-quantile(d::BetaBinomial, p::Float64) = quantile(Categorical(pdf.(Ref(d), support(d))), p) - 1
+entropy(d::BetaBinomial) = entropy(_categorical(d))
+median(d::BetaBinomial) = median(_categorical(d)) - 1
+mode(d::BetaBinomial) = mode(_categorical(d)) - 1
+modes(d::BetaBinomial) = modes(_categorical(d)) .- 1
+
+quantile(d::BetaBinomial, p::Float64) = quantile(_categorical(d), p) - 1
 
 #### Sampling
 

src/univariate/discrete/hypergeometric.jl

@@ -75,7 +75,7 @@ function kurtosis(d::Hypergeometric)
     a/b
 end
 
-entropy(d::Hypergeometric) = entropy(pdf.(Ref(d), support(d)))
+entropy(d::Hypergeometric) = entropy(map(Base.Fix1(pdf, d), support(d)))
 
 ### Evaluation & Sampling
 

src/univariate/discrete/noncentralhypergeometric.jl

@@ -256,9 +256,12 @@ end
 Base.convert(::Type{WalleniusNoncentralHypergeometric{T}}, d::WalleniusNoncentralHypergeometric{T}) where {T<:Real} = d
 
 # Properties
-mean(d::WalleniusNoncentralHypergeometric) = sum(support(d) .* pdf.(Ref(d), support(d)))
-var(d::WalleniusNoncentralHypergeometric)  = sum((support(d) .- mean(d)).^2 .* pdf.(Ref(d), support(d)))
-mode(d::WalleniusNoncentralHypergeometric) = support(d)[argmax(pdf.(Ref(d), support(d)))]
+function _discretenonparametric(d::WalleniusNoncentralHypergeometric)
+    return DiscreteNonParametric(support(d), map(Base.Fix1(pdf, d), support(d)))
+end
+mean(d::WalleniusNoncentralHypergeometric) = mean(_discretenonparametric(d))
+var(d::WalleniusNoncentralHypergeometric)  = var(_discretenonparametric(d))
+mode(d::WalleniusNoncentralHypergeometric) = mode(_discretenonparametric(d))
 
 entropy(d::WalleniusNoncentralHypergeometric) = 1
 

test/censored.jl

@@ -207,7 +207,7 @@ end
             end
             @test @inferred(median(d)) ≈ clamp(median(d0), l, u)
             @inferred quantile(d, 0.5)
-            @test quantile.(d, 0:0.01:1) ≈ clamp.(quantile.(d0, 0:0.01:1), l, u)
+            @test Base.Fix1(quantile, d).(0:0.01:1) ≈ clamp.(Base.Fix1(quantile, d0).(0:0.01:1), l, u)
             # special-case pdf/logpdf/loglikelihood since when replacing Dirac(μ) with
             # Normal(μ, 0), they are infinite
             if lower === nothing || !isfinite(lower)
@@ -253,7 +253,7 @@ end
                 @test f(d) ≈ f(dmix)
             end
             @test median(d) ≈ clamp(median(d0), l, u)
-            @test quantile.(d, 0:0.01:1) ≈ clamp.(quantile.(d0, 0:0.01:1), l, u)
+            @test Base.Fix1(quantile, d).(0:0.01:1) ≈ clamp.(Base.Fix1(quantile, d0).(0:0.01:1), l, u)
             # special-case pdf/logpdf/loglikelihood since when replacing Dirac(μ) with
             # Normal(μ, 0), they are infinite
             if lower === nothing
@@ -311,7 +311,7 @@ end
             end
             @test @inferred(median(d)) ≈ clamp(median(d0), l, u)
             @inferred quantile(d, 0.5)
-            @test quantile.(d, 0:0.01:1) ≈ clamp.(quantile.(d0, 0:0.01:1), l, u)
+            @test Base.Fix1(quantile, d).(0:0.01:1) ≈ clamp.(Base.Fix1(quantile, d0).(0:0.01:1), l, u)
             # rand
             x = rand(d, 10_000)
             @test all(x -> insupport(d, x), x)
@@ -346,7 +346,7 @@ end
                 @test f(d, 5) ≈ f(dmix, 5)
             end
             @test median(d) ≈ clamp(median(d0), l, u)
-            @test quantile.(d, 0:0.01:0.99) ≈ clamp.(quantile.(d0, 0:0.01:0.99), l, u)
+            @test Base.Fix1(quantile, d).(0:0.01:0.99) ≈ clamp.(Base.Fix1(quantile, d0).(0:0.01:0.99), l, u)
             x = rand(d, 100)
             @test loglikelihood(d, x) ≈ loglikelihood(dmix, x)
             # rand

test/matrixvariates.jl

@@ -117,7 +117,7 @@ function test_convert(d::MatrixDistribution)
     @test d == deepcopy(d)
     for elty in (Float32, Float64, BigFloat)
         del1 = convert(distname{elty}, d)
-        del2 = convert(distname{elty}, getfield.(Ref(d), fieldnames(typeof(d)))...)
+        del2 = convert(distname{elty}, (Base.Fix1(getfield, d)).(fieldnames(typeof(d)))...)
         @test del1 isa distname{elty}
         @test del2 isa distname{elty}
         @test partype(del1) == elty

test/mixture.jl

@@ -40,7 +40,7 @@ function test_mixture(g::UnivariateMixture, n::Int, ns::Int,
     for i = 1:n
         @test @inferred(cdf(g, X[i])) ≈ cf[i]
     end
-    @test cdf.(g, X) ≈ cf
+    @test Base.Fix1(cdf, g).(X) ≈ cf
 
     # evaluation
     P0 = zeros(T, n, K)

test/multivariate/product.jl

@@ -67,8 +67,7 @@ end
 
     for a in ([0, 1], [-0.5, 0.5])
         # Construct independent distributions and `Product` distribution from these.
-        support = fill(a, N)
-        ds = DiscreteNonParametric.(support, Ref([0.5, 0.5]))
+        ds = [DiscreteNonParametric(copy(a), [0.5, 0.5]) for _ in 1:N]
         x = rand.(ds)
         d_product = product_distribution(ds)
         @test d_product isa Product

test/product.jl

@@ -93,7 +93,7 @@ end
 
     for a in ([0, 1], [-0.5, 0.5])
         # Construct independent distributions and `ProductDistribution` from these.
-        ds1 = DiscreteNonParametric.(fill(a, N), Ref([0.5, 0.5]))
+        ds1 = [DiscreteNonParametric(copy(a), [0.5, 0.5]) for _ in 1:N]
         # Replace with
         # d_product1 = @inferred(product_distribution(ds1))
         # when `Product` is removed

test/testutils.jl

@@ -128,7 +128,7 @@ function test_samples(s::Sampleable{Univariate, Discrete},      # the sampleable
     rmin = floor(Int,quantile(distr, 0.00001))::Int
     rmax = floor(Int,quantile(distr, 0.99999))::Int
     m = rmax - rmin + 1  # length of the range
-    p0 = pdf.(Ref(distr), rmin:rmax)  # reference probability masses
+    p0 = map(Base.Fix1(pdf, distr), rmin:rmax)  # reference probability masses
     @assert length(p0) == m
 
     # determine confidence intervals for counts:
@@ -233,7 +233,7 @@ function test_samples(s::Sampleable{Univariate, Continuous},    # the sampleable
     #
     clb = Vector{Int}(undef, nbins)
     cub = Vector{Int}(undef, nbins)
-    cdfs = cdf.(Ref(distr), edges)
+    cdfs = map(Base.Fix1(cdf, distr), edges)
 
     for i = 1:nbins
         pi = cdfs[i+1] - cdfs[i]
@@ -385,19 +385,19 @@ function test_range_evaluation(d::DiscreteUnivariateDistribution)
     rmin = round(Int, islowerbounded(d) ? vmin : quantile(d, 0.001))::Int
     rmax = round(Int, isupperbounded(d) ? vmax : quantile(d, 0.999))::Int
 
-    p0 = pdf.(Ref(d), collect(rmin:rmax))
-    @test pdf.(Ref(d), rmin:rmax) ≈ p0
+    p0 = map(Base.Fix1(pdf, d), collect(rmin:rmax))
+    @test map(Base.Fix1(pdf, d), rmin:rmax) ≈ p0
     if rmin + 2 <= rmax
-        @test pdf.(Ref(d), rmin+1:rmax-1) ≈ p0[2:end-1]
+        @test map(Base.Fix1(pdf, d), rmin+1:rmax-1) ≈ p0[2:end-1]
     end
 
     if isbounded(d)
-        @test pdf.(Ref(d), support(d)) ≈ p0
-        @test pdf.(Ref(d), rmin-2:rmax) ≈ vcat(0.0, 0.0, p0)
-        @test pdf.(Ref(d), rmin:rmax+3) ≈ vcat(p0, 0.0, 0.0, 0.0)
-        @test pdf.(Ref(d), rmin-2:rmax+3) ≈ vcat(0.0, 0.0, p0, 0.0, 0.0, 0.0)
+        @test map(Base.Fix1(pdf, d), support(d)) ≈ p0
+        @test map(Base.Fix1(pdf, d), rmin-2:rmax) ≈ vcat(0.0, 0.0, p0)
+        @test map(Base.Fix1(pdf, d), rmin:rmax+3) ≈ vcat(p0, 0.0, 0.0, 0.0)
+        @test map(Base.Fix1(pdf, d), rmin-2:rmax+3) ≈ vcat(0.0, 0.0, p0, 0.0, 0.0, 0.0)
     elseif islowerbounded(d)
-        @test pdf.(Ref(d), rmin-2:rmax) ≈ vcat(0.0, 0.0, p0)
+        @test map(Base.Fix1(pdf, d), rmin-2:rmax) ≈ vcat(0.0, 0.0, p0)
     end
 end
 
@@ -444,13 +444,13 @@ function test_evaluation(d::DiscreteUnivariateDistribution, vs::AbstractVector,
     end
 
     # consistency of scalar-based and vectorized evaluation
-    @test pdf.(Ref(d), vs)  ≈ p
-    @test cdf.(Ref(d), vs)  ≈ c
-    @test ccdf.(Ref(d), vs) ≈ cc
+    @test Base.Fix1(pdf, d).(vs)  ≈ p
+    @test Base.Fix1(cdf, d).(vs)  ≈ c
+    @test Base.Fix1(ccdf, d).(vs) ≈ cc
 
-    @test logpdf.(Ref(d), vs)  ≈ lp
-    @test logcdf.(Ref(d), vs)  ≈ lc
-    @test logccdf.(Ref(d), vs) ≈ lcc
+    @test Base.Fix1(logpdf, d).(vs)  ≈ lp
+    @test Base.Fix1(logcdf, d).(vs)  ≈ lc
+    @test Base.Fix1(logccdf, d).(vs) ≈ lcc
 end
 
 
@@ -511,15 +511,15 @@ function test_evaluation(d::ContinuousUnivariateDistribution, vs::AbstractVector
 
     # consistency of scalar-based and vectorized evaluation
     if !isa(d, StudentizedRange)
-        @test pdf.(Ref(d), vs)  ≈ p
-        @test logpdf.(Ref(d), vs)  ≈ lp
+        @test Base.Fix1(pdf, d).(vs) ≈ p
+        @test Base.Fix1(logpdf, d).(vs) ≈ lp
     end
 
-    @test cdf.(Ref(d), vs)  ≈ c
-    @test ccdf.(Ref(d), vs) ≈ cc
+    @test Base.Fix1(cdf, d).(vs)  ≈ c
+    @test Base.Fix1(ccdf, d).(vs) ≈ cc
 
-    @test logcdf.(Ref(d), vs)  ≈ lc
-    @test logccdf.(Ref(d), vs) ≈ lcc
+    @test Base.Fix1(logcdf, d).(vs)  ≈ lc
+    @test Base.Fix1(logccdf, d).(vs) ≈ lcc
 end
 
 function test_nonfinite(distr::UnivariateDistribution)
@@ -550,7 +550,7 @@ function test_stats(d::DiscreteUnivariateDistribution, vs::AbstractVector)
     # using definition (or an approximation)
 
     vf = Float64[v for v in vs]
-    p = pdf.(Ref(d), vf)
+    p = Base.Fix1(pdf, d).(vf)
     xmean = dot(p, vf)
     xvar = dot(p, abs2.(vf .- xmean))
     xstd = sqrt(xvar)

test/univariate/continuous/johnsonsu.jl

@@ -10,10 +10,10 @@
     @test rand(d1) isa Float64
 
     @test median(d1) == quantile(d1, 0.5)
-    x = quantile.(d1, [0.25, 0.45, 0.60, 0.80, 0.90])
-    @test all(cdf.(d1, x) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
-    y = cquantile.(d1, [0.25, 0.45, 0.60, 0.80, 0.90])
-    @test all(ccdf.(d1, y) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
+    x = Base.Fix1(quantile, d1).([0.25, 0.45, 0.60, 0.80, 0.90])
+    @test all(Base.Fix1(cdf, d1).(x) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
+    y = Base.Fix1(cquantile, d1).([0.25, 0.45, 0.60, 0.80, 0.90])
+    @test all(Base.Fix1(ccdf, d1).(y) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
 
     @test mean(d1) ≈ 7.581281
     @test var(d1) ≈ 19.1969485

test/univariate/continuous/logitnormal.jl

@@ -43,7 +43,7 @@ function test_logitnormal(g::LogitNormal, n_tsamples::Int=10^6,
     for i = 1:min(100, n_tsamples)
         @test logpdf(g, X[i]) ≈ log(pdf(g, X[i]))
     end
-    @test logpdf.(g, X) ≈ log.(pdf.(g, X))
+    @test Base.Fix1(logpdf, g).(X) ≈ log.(Base.Fix1(pdf, g).(X))
     @test isequal(logpdf(g, 0),-Inf)
     @test isequal(logpdf(g, 1),-Inf)
     @test isequal(logpdf(g, -eps()),-Inf)

test/univariate/continuous/rician.jl

@@ -14,14 +14,14 @@
     @test var(d1) ≈ var(d2)
     @test mode(d1) ≈ mode(d2)
     @test median(d1) ≈ median(d2)
-    @test quantile.(d1, [0.25, 0.45, 0.60, 0.80, 0.90]) ≈ quantile.(d2, [0.25, 0.45, 0.60, 0.80, 0.90])
-    @test pdf.(d1, 0.0:0.1:1.0) ≈ pdf.(d2, 0.0:0.1:1.0)
-    @test cdf.(d1, 0.0:0.1:1.0) ≈ cdf.(d2, 0.0:0.1:1.0)
+    @test Base.Fix1(quantile, d1).([0.25, 0.45, 0.60, 0.80, 0.90]) ≈ Base.Fix1(quantile, d2).([0.25, 0.45, 0.60, 0.80, 0.90])
+    @test Base.Fix1(pdf, d1).(0.0:0.1:1.0) ≈ Base.Fix1(pdf, d2).(0.0:0.1:1.0)
+    @test Base.Fix1(cdf, d1).(0.0:0.1:1.0) ≈ Base.Fix1(cdf, d2).(0.0:0.1:1.0)
 
     d1 = Rician(10.0, 10.0)
     @test median(d1) == quantile(d1, 0.5)
-    x = quantile.(d1, [0.25, 0.45, 0.60, 0.80, 0.90])
-    @test all(cdf.(d1, x) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
+    x = Base.Fix1(quantile, d1).([0.25, 0.45, 0.60, 0.80, 0.90])
+    @test all(Base.Fix1(cdf, d1).(x) .≈ [0.25, 0.45, 0.60, 0.80, 0.90])
 
     x = rand(Rician(5.0, 5.0), 100000)
     d1 = fit(Rician, x)

test/univariate/continuous/skewnormal.jl

@@ -25,7 +25,7 @@ import Distributions: normpdf, normcdf, normlogpdf, normlogcdf
     d4 = Normal(0.5, 2.2)
     #
     @test pdf(d3, 3.3) == Distributions.pdf(d4, 3.3)
-    @test pdf.(d3, 1:3) == Distributions.pdf.(d4, 1:3)
+    @test Base.Fix1(pdf, d3).(1:3) == Base.Fix1(pdf, d4).(1:3)
     a = mean(d3), var(d3), std(d3)
     b = Distributions.mean(d4), Distributions.var(d4), Distributions.std(d4)
     @test a == b

test/univariate/discrete/binomial.jl

@@ -9,14 +9,14 @@ Random.seed!(1234)
 for (p, n) in [(0.6, 10), (0.8, 6), (0.5, 40), (0.04, 20), (1., 100), (0., 10), (0.999999, 1000), (1e-7, 1000)]
     d = Binomial(n, p)
 
-    a = pdf.(d, 0:n)
+    a = Base.Fix1(pdf, d).(0:n)
     for t=0:n
         @test pdf(d, t) ≈ a[1+t]
     end
 
     li = rand(0:n, 2)
     rng = minimum(li):maximum(li)
-    b = pdf.(d, rng)
+    b = Base.Fix1(pdf, d).(rng)
     for t in rng
         @test pdf(d, t) ≈ b[t - first(rng) + 1]
     end

test/univariate/discrete/categorical.jl

@@ -56,8 +56,8 @@ for p in Any[
     @test iszero(ccdf(d, Inf))
     @test isnan(ccdf(d, NaN))
 
-    @test pdf.(d, support(d)) == p
-    @test pdf.(d, 1:k) == p
+    @test Base.Fix1(pdf, d).(support(d)) == p
+    @test Base.Fix1(pdf, d).(1:k) == p
 
     @test cf(d, 0) ≈ 1.0
     @test cf(d, 1) ≈ p' * cis.(1:length(p))

test/univariate/discrete/poissonbinomial.jl

@@ -104,9 +104,9 @@ for (n₁, n₂, n₃, p₁, p₂, p₃) in [(10, 10, 10, 0.1, 0.5, 0.9),
     b2 = Binomial(n₂, p₂)
     b3 = Binomial(n₃, p₃)
 
-    pmf1 = pdf.(b1, support(b1))
-    pmf2 = pdf.(b2, support(b2))
-    pmf3 = pdf.(b3, support(b3))
+    pmf1 = Base.Fix1(pdf, b1).(support(b1))
+    pmf2 = Base.Fix1(pdf, b2).(support(b2))
+    pmf3 = Base.Fix1(pdf, b3).(support(b3))
 
     @test @inferred(mean(d)) ≈ (mean(b1) + mean(b2) + mean(b3))
     @test @inferred(var(d))  ≈ (var(b1) + var(b2) + var(b3))