Update deprecated functions.

doc/examples/epil.jl

@@ -81,9 +81,9 @@ epil = Dict{Symbol, Any}(
 epil[:N] = size(epil[:y], 1)
 epil[:T] = size(epil[:y], 2)
 
-epil[:logBase4] = log(epil[:Base] / 4)
+epil[:logBase4] = log.(epil[:Base] / 4)
 epil[:BT] = epil[:logBase4] .* epil[:Trt]
-epil[:logAge] = log(epil[:Age])
+epil[:logAge] = log.(epil[:Age])
 map(key -> epil[Symbol(string(key, "bar"))] = mean(epil[key]),
     [:logBase4, :Trt, :BT, :logAge, :V4])
 

doc/examples/leuk.jl

@@ -58,7 +58,7 @@ model = Model(
   ),
 
   S0 = Logical(1,
-    dL0 -> exp(-cumsum(dL0)),
+    dL0 -> exp.(-cumsum(dL0)),
     false
   ),
 

doc/examples/line_amwg_slice.jl

@@ -19,7 +19,7 @@ logf = function(x::DenseVector)
    b1 = x[2]
    logs2 = x[3]
    (-0.5 * length(data[:y]) - 0.001) * logs2 -
-     (0.5 * sumabs2(data[:y] - b0 - b1 * data[:x]) + 0.001) / exp(logs2) -
+     (0.5 * sum(abs2, data[:y] - b0 - b1 * data[:x]) + 0.001) / exp(logs2) -
      0.5 * b0^2 / 1000 - 0.5 * b1^2 / 1000
 end
 
@@ -37,6 +37,6 @@ logs2 = SliceMultivariate([0.0], 5.0, logf_logs2)
 for i in 1:n
   sample!(beta, adapt = (i <= burnin))
   sample!(logs2)
-  sim[i, :, 1] = [beta; exp(logs2)]
+  sim[i, :, 1] = [beta; exp.(logs2)]
 end
 describe(sim)

doc/examples/magnesium.jl

@@ -56,7 +56,7 @@ model = Model(
   ),
 
   OR = Logical(1,
-    mu -> exp(mu)
+    mu -> exp.(mu)
   ),
 
   tau = Logical(1,

doc/examples/mice.jl

@@ -44,7 +44,7 @@ model = Model(
   ),
 
   median = Logical(1,
-    (beta, r) -> exp(-beta / r) * log(2)^(1 / r)
+    (beta, r) -> exp.(-beta / r) * log(2)^(1 / r)
   ),
 
   veh_control = Logical(

doc/examples/pollution.jl

@@ -113,7 +113,7 @@ Gibbs_sigma2 = Sampler([:sigma2],
   (mu, sigma2, y) ->
     begin
       a = length(y) / 2.0 + shape(sigma2.distr)
-      b = sumabs2(y - mu) / 2.0 + scale(sigma2.distr)
+      b = sum(abs2, y - mu) / 2.0 + scale(sigma2.distr)
       rand(InverseGamma(a, b))
     end
 )

doc/mcmc/newmultivardist.jl

@@ -19,12 +19,12 @@
 
   ## Logical indicating whether x is in the support
   function insupport{T<:Real}(d::NewMultivarDist, x::AbstractVector{T})
-    length(d) == length(x) && all(isfinite(x))
+    length(d) == length(x) && all(isfinite.(x))
   end
 
   ## Normalized or unnormalized log-density value
   function _logpdf{T<:Real}(d::NewMultivarDist, x::AbstractVector{T})
-    -length(x) * log(d.sigma) - 0.5 * sumabs2(x - d.mu) / d.sigma^2
+    -length(x) * log(d.sigma) - 0.5 * sum(abs2, x - d.mu) / d.sigma^2
   end
 
 end

doc/samplers/abc.rst

@@ -13,7 +13,7 @@ Model-Based Constructor
 .. function:: ABC(params::ElementOrVector{Symbol}, \
                   scale::ElementOrVector{T<:Real}, summary::Function, \
                   epsilon::Real; kernel::KernelDensityType=SymUniform, \
-                  dist::Function=(Tsim, Tobs) -> sqrt(sumabs2(Tsim - Tobs)), \
+                  dist::Function=(Tsim, Tobs) -> sqrt(sum(abs2, Tsim - Tobs)), \
                   proposal::SymDistributionType=Normal, maxdraw::Integer=1, \
                   nsim::Integer=1, decay::Real=1.0, randeps::Bool=false, \
                   args...)

doc/tutorial.rst

@@ -179,7 +179,7 @@ whose form is inverse gamma with :math:`n / 2 + \alpha_\pi` shape and :math:`(\b
       (mu, s2, y) ->
         begin
           a = length(y) / 2.0 + shape(s2.distr)
-          b = sumabs2(y - mu) / 2.0 + scale(s2.distr)
+          b = sum(abs2, y - mu) / 2.0 + scale(s2.distr)
           rand(InverseGamma(a, b))
         end
     )

doc/tutorial/line.jl

@@ -39,7 +39,7 @@ Gibbs_s2 = Sampler([:s2],
   (mu, s2, y) ->
     begin
       a = length(y) / 2.0 + shape(s2.distr)
-      b = sumabs2(y - mu) / 2.0 + scale(s2.distr)
+      b = sum(abs2, y - mu) / 2.0 + scale(s2.distr)
       rand(InverseGamma(a, b))
     end
 )

src/distributions/pdmats2.jl

@@ -96,12 +96,12 @@ module PDMats2
     length(r) == n || throw(ArgumentError("incompatible array length"))
     wx = whiten(a, x)
     for j = 1:n
-      r[j] = sumabs2(wx[:, j])
+      r[j] = sum(abs2, wx[:, j])
     end
     r
   end
 
-  invquad(a::PBDiagMat, x::Vector{Float64}) = sumabs2(whiten(a, x))
+  invquad(a::PBDiagMat, x::Vector{Float64}) = sum(abs2, whiten(a, x))
 
 
   #################### Auxiliary Functions ####################

src/output/chainsummary.jl

@@ -55,7 +55,7 @@ function Base.show(io::IO, s::ChainSummary)
   cnwid = map(length, s.colnames)
   ## s.value as right alignable strings
   charv = mapslices(showoff, s.value, 1)
-  colwid = 1 + max(cnwid, vec(maximum(map(length, charv), [1, 3])))
+  colwid = 1 + max.(cnwid, vec(maximum(map(length, charv), [1, 3])))
   m, n, f = size(charv)
   for k in 1:f
     ## write the column headers centered on the column widths

src/output/gelmandiag.jl

@@ -56,5 +56,5 @@ function gelmandiag(c::AbstractChains; alpha::Real=0.05, mpsrf::Bool=false,
   end
 
   hdr = header(c) * "\nGelman, Rubin, and Brooks Diagnostic:"
-  ChainSummary(round(psrf, 3), psrf_names, psrf_labels, hdr)
+  ChainSummary(round.(psrf, 3), psrf_names, psrf_labels, hdr)
 end

src/output/plot.jl

@@ -204,7 +204,7 @@ function mixeddensityplot(c::AbstractChains;
   plots = Array{Plot}(size(c, 2))
   discrete = indiscretesupport(c, barbounds)
   plots[discrete] = plot(c[:, discrete, :], :bar; args...)
-  plots[!discrete] = plot(c[:, !discrete, :], :density; args...)
+  plots[.!discrete] = plot(c[:, .!discrete, :], :density; args...)
   return plots
 end
 

src/output/stats.jl

@@ -34,7 +34,7 @@ function changerate(c::AbstractChains)
       r_mv += any(delta)
     end
   end
-  vals = round([r; r_mv] / (m * (n - 1)), 3)
+  vals = round.([r; r_mv] / (m * (n - 1)), 3)
   ChainSummary(vals, [c.names; "Multivariate"], ["Change Rate"], header(c))
 end
 
@@ -89,6 +89,6 @@ function summarystats(c::AbstractChains; etype=:bm, args...)
     mapslices(x -> f(x), c.value, [1, 3]),
     [2, 1, 3]
   )
-  stats = [vals  min((vals[:, 2] ./ vals[:, 4]).^2, size(c.value, 1))]
+  stats = [vals  min.((vals[:, 2] ./ vals[:, 4]).^2, size(c.value, 1))]
   ChainSummary(stats, c.names, labels, header(c))
 end

src/samplers/abc.jl

@@ -23,7 +23,7 @@ end
 function ABC{T<:Real}(params::ElementOrVector{Symbol},
                       scale::ElementOrVector{T}, summary::Function,
                       epsilon::Real; kernel::KernelDensityType=SymUniform,
-                      dist::Function=(Tsim, Tobs) -> sqrt(sumabs2(Tsim - Tobs)),
+                      dist::Function=(Tsim, Tobs) -> sqrt(sum(abs2, Tsim - Tobs)),
                       proposal::SymDistributionType=Normal, maxdraw::Integer=1,
                       nsim::Integer=1, decay::Real=1.0, randeps::Bool=false,
                       args...)

src/samplers/amm.jl

@@ -18,8 +18,8 @@ type AMMTune <: SamplerTune
   function AMMTune{T<:Real}(x::Vector, Sigma::Matrix{T},
                             logf::Nullable{Function}; beta::Real=0.05,
                             scale::Real=2.38)
-    new(logf, false, beta, 0, Vector{Float64}(), Matrix{Float64}(), scale,
-        cholfact(Sigma)[:L], Matrix{Float64}())
+    new(logf, false, beta, 0, Vector{Float64}(), Matrix{Float64}(0, 0), scale,
+        cholfact(Sigma)[:L], Matrix{Float64}(0, 0))
   end
 end
 

src/samplers/bhmc.jl

@@ -55,12 +55,12 @@ function sample!(v::BHMCVariate, logf::Function)
   totaltime = 0.0                     ## time the particle already moved
 
   n = length(v)                       ## length of binary vector
-  S = sign(tune.position)
+  S = sign.(tune.position)
 
   while true
     a = tune.velocity[:]
     b = tune.position[:]
-    phi = atan2(b, a)
+    phi = atan2.(b, a)
 
     ## time to hit or cross wall
     walltime = -phi
@@ -117,6 +117,6 @@ function sample!(v::BHMCVariate, logf::Function)
   end
 
   ## convert from (-/+1) to (0/1)
-  v[:] = (sign(tune.position) + ones(n)) / 2.0
+  v[:] = (sign.(tune.position) + ones(n)) / 2.0
   v
 end

src/samplers/hmc.jl

@@ -100,8 +100,8 @@ function sample!(v::HMCVariate, logfgrad::Function)
 
   ## Evaluate potential and kinetic energies at start and end of trajectory
   SigmaLinv = inv(tune.SigmaL)
-  Kp0 = 0.5 * sumabs2(SigmaLinv * p0)
-  Kp1 = 0.5 * sumabs2(SigmaLinv * p1)
+  Kp0 = 0.5 * sum(abs2, SigmaLinv * p0)
+  Kp1 = 0.5 * sum(abs2, SigmaLinv * p1)
 
   if rand() < exp((logf1 - Kp1) - (logf0 - Kp0))
     v[:] = x1

src/samplers/mala.jl

@@ -75,8 +75,8 @@ function sample!(v::MALAVariate, logfgrad::Function)
   y = v + M2 * grad0 + L * randn(length(v))
   logf1, grad1 = logfgrad(y)
 
-  q0 = -0.5 * sumabs2(Linv * (v - y - M2 * grad1))
-  q1 = -0.5 * sumabs2(Linv * (y - v - M2 * grad0))
+  q0 = -0.5 * sum(abs2, Linv * (v - y - M2 * grad1))
+  q1 = -0.5 * sum(abs2, Linv * (y - v - M2 * grad0))
 
   if rand() < exp((logf1 - q1) - (logf0 - q0))
     v[:] = y

src/samplers/miss.jl

@@ -17,7 +17,7 @@ function MISSTune(d::Distribution, v)
 end
 
 function MISSTune(D::Array{UnivariateDistribution}, v::Array)
-  inds = find(isnan(v))
+  inds = find(isnan.(v))
   MISSTune(dims(D), inds, inds)
 end
 

src/samplers/sampler.jl

@@ -107,7 +107,7 @@ function logpdfgrad!{T<:Real}(block::SamplingBlock, x::AbstractVector{T},
                               dtype::Symbol)
   grad = gradlogpdf!(block, x, dtype)
   logf = logpdf!(block, x)
-  (logf, ifelse(isfinite(grad), grad, 0.0))
+  (logf, ifelse.(isfinite.(grad), grad, 0.0))
 end
 
 function unlist(block::SamplingBlock)