Fixes #22

src/chains.jl

@@ -107,9 +107,9 @@ end
 #################### Indexing ####################
 
 function Base.getindex(c::Chains, window, names, chains)
-  inds1 = window2inds(c, window)
-  inds2 = names2inds(c, names)
-  Chains(c.value[inds1, inds2, chains],
+    inds1 = window2inds(c, window)
+    inds2 = names2inds(c, names)
+    return Chains(c.value[inds1, inds2, chains],
          start = first(c) + (first(inds1) - 1) * step(c),
          thin = step(inds1) * step(c), names = c.names[inds2],
          chains = c.chains[chains])

src/plot.jl

@@ -38,7 +38,12 @@ const supportedplots = push!(collect(keys(translationdict)), :mixeddensity)
 
     if st == :mixeddensity
         discrete = MCMCChain.indiscretesupport(c, barbounds)
-        st = (discrete[i] && colordim == :chain) ? :histogram : :density
+        st = if colordim == :chain
+            discrete[i] ? :histogram : :density
+        else
+            # NOTE: It might make sense to overlay histograms and density plots here.
+            :density
+        end
         seriestype := st
     end
 
@@ -57,14 +62,14 @@ end
 @recipe function f(p::_DensityPlot)
     xaxis --> "Sample value"
     yaxis --> "Density"
-    [collect(skipmissing(p.val[:,k])) for k in axes(p.val, 2)]
+    [collect(skipmissing(p.val[:,k])) for k in 1:size(p.val, 2)]
 end
 
 @recipe function f(p::_HistogramPlot)
     xaxis --> "Sample value"
     yaxis --> "Frequency"
     fillalpha --> 0.7
-    [collect(skipmissing(p.val[:,k])) for k in axes(p.val, 2)]
+    [collect(skipmissing(p.val[:,k])) for k in 1:size(p.val, 2)]
 end
 
 @recipe function f(p::_MeanPlot)
@@ -100,7 +105,7 @@ end
     ptypes = ptypes isa AbstractVector || ptypes isa Tuple ? ptypes : (ptypes,)
     @assert all(map(ptype -> ptype ∈ supportedplots, ptypes))
 
-    nrows, nvars, nchains = size(c.value)
+    nrows, nvars, nchains = size(c)
     ntypes = length(ptypes)
     N = colordim == :chain ? nvars : nchains
     layout := (N, ntypes)

test/plot_test.jl

@@ -11,32 +11,46 @@ val = hcat(val, rand(1:2, n_iter, 1, n_chain))
 
 chn = Chains(val)
 
-# plotting singe plotting types
-ps_trace = traceplot(chn, 1)
-@test isa(ps_trace, Plots.Plot)
+@testset "Plotting tests" begin
 
-ps_mean = meanplot(chn, 1)
-@test isa(ps_mean, Plots.Plot)
+    # plotting singe plotting types
+    ps_trace = traceplot(chn, 1)
+    @test isa(ps_trace, Plots.Plot)
 
-ps_density = density(chn, 1)
-@test isa(ps_density, Plots.Plot)
+    ps_mean = meanplot(chn, 1)
+    @test isa(ps_mean, Plots.Plot)
 
-ps_autocor = autocorplot(chn, 1)
-@test isa(ps_autocor, Plots.Plot)
+    ps_density = density(chn, 1)
+    @test isa(ps_density, Plots.Plot)
 
-#ps_contour = plot(chn, :contour)
+    ps_autocor = autocorplot(chn, 1)
+    @test isa(ps_autocor, Plots.Plot)
 
-ps_hist = histogram(chn, 1)
-@test isa(ps_hist, Plots.Plot)
+    #ps_contour = plot(chn, :contour)
 
-ps_mixed = mixeddensity(chn, 1)
-@test isa(ps_mixed, Plots.Plot)
+    ps_hist = histogram(chn, 1)
+    @test isa(ps_hist, Plots.Plot)
 
-# plotting combinations
-ps_trace_mean = plot(chn)
-@test isa(ps_trace_mean, Plots.Plot)
+    ps_mixed = mixeddensity(chn, 1)
+    @test isa(ps_mixed, Plots.Plot)
 
-savefig("demo-plot.png")
+    # plotting combinations
+    ps_trace_mean = plot(chn)
+    @test isa(ps_trace_mean, Plots.Plot)
 
-ps_mixed_auto = plot(chn, seriestype = (:mixeddensity, :autocorplot))
-@test isa(ps_mixed_auto, Plots.Plot)
+    savefig("demo-plot.png")
+
+    ps_mixed_auto = plot(chn, seriestype = (:mixeddensity, :autocorplot))
+    @test isa(ps_mixed_auto, Plots.Plot)
+
+    # Test plotting using colordim keyword
+    p_colordim = plot(chn, colordim = :parameter)
+    @test isa(p_colordim, Plots.Plot)
+
+    # Test if plotting a sub-set work.s
+    p_subset = plot(chn[:,2,:])
+    @test isa(p_subset, Plots.Plot)
+
+    p_subset_colordim = plot(chn[:,2,:], colordim = :parameter)
+    @test isa(p_subset_colordim, Plots.Plot)
+end