Corrected verbose for tests

Author: theogf

Project.toml

@@ -3,9 +3,6 @@ uuid = "b1c2f168-d918-11e8-38ee-334f3e5b0caa"
 authors = ["Theo Galy-Fajou <theo.galyfajou@gmail.com>"]
 version = "0.3.1"
 
-[compat]
-julia = "0.7"
-
 [deps]
 Clustering = "aaaa29a8-35af-508c-8bc3-b662a17a0fe5"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
@@ -22,8 +19,12 @@ Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 ValueHistories = "98cad3c8-aec3-5f06-8e41-884608649ab7"
 
+[compat]
+julia = "1.0"
+
 [extras]
-Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
+Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
+
 [targets]
-test = ["Test","Pkg"]
+test = ["Test", "Pkg"]

src/models/StudentT_Functions.jl

@@ -1,7 +1,7 @@
 # Functions related to the Student T likelihood (StudentT)
 "Update the local variational parameters of the full batch GP StudentT"
 function local_update!(model::BatchStudentT)
-    model.β = 0.5.*(diag(model.Σ)+(model.μ.-model.y).^2+model.ν)
+    model.β = 0.5.*(diag(model.Σ)+(model.μ.-model.y).^2 .+model.ν)
     model.θ = 0.5.*(model.ν.+1.0)./model.β
 end
 
@@ -14,7 +14,7 @@ end
 
 "Update the local variational parameters of the sparse GP StudentT"
 function local_update!(model::SparseStudentT)
-    model.β = 0.5.*(model.Ktilde+sum((model.κ*model.Σ).*model.κ,dims=2)[:]+(model.κ*model.μ.-model.y[model.MBIndices]).^2.+model.ν)
+    model.β = 0.5.*(model.Ktilde+sum((model.κ*model.Σ).*model.κ,dims=2)[:]+(model.κ*model.μ.-model.y[model.MBIndices]).^2 .+model.ν)
     model.θ = 0.5.*(model.ν.+1.0)./model.β
 
 end

test/test_IO.jl

@@ -25,13 +25,15 @@ kernel = RBFKernel([3.0],dim=N_dim)
 fullm = true
 sparsem = true
 ps = []; t_full = 0; t_sparse = 0; t_stoch = 0;
+println("Testing the save/load model function")
 if fullm
     println("Testing the full model")
     t_full = @elapsed fullmodel = AugmentedGaussianProcesses.BatchXGPC(X,y,noise=noise,kernel=kernel,verbose=verbose,Autotuning=true)
     t_full += @elapsed fullmodel.train(iterations=20)
     y_full = fullmodel.predictproba(X_test); acc_full = 1-sum(abs.(sign.(y_full.-0.5)-y_test))/(2*length(y_test))
     save_trained_model("fullXGPC_test.jld2",fullmodel)
     fmodel = load_trained_model("fullXGPC_test.jld2")
+    rm("fullXGPC_test.jld2")
     y_full2 = fmodel.predictproba(X_test)
     if mean(abs.(y_full-y_full2))>1e-5
         return false
@@ -46,6 +48,7 @@ if sparsem
     y_sparse = sparsemodel.predictproba(X_test); acc_sparse = 1-sum(abs.(sign.(y_sparse.-0.5)-y_test))/(2*length(y_test))
     save_trained_model("sparseXGPC_test.jld2",sparsemodel)
     smodel = load_trained_model("sparseXGPC_test.jld2")
+    rm("sparseXGPC_test.jld2")
     y_sparse2 = smodel.predictproba(X_test)
     if mean(abs.(y_sparse-y_sparse2))>1e-5
         return false

test/test_Regression.jl

@@ -31,6 +31,7 @@ ps = []; t_full = 0; t_sparse = 0; t_stoch = 0;
 
 kernel = RBFKernel(2.0)
 autotuning=false
+println("Testing the regression model")
 
 println("Testing the full model")
 t_full = @elapsed fullmodel = AugmentedGaussianProcesses.GPRegression(X,y,noise=noise,Autotuning=autotuning,kernel=kernel,verbose=verbose)

test/test_StudentT.jl

@@ -37,6 +37,7 @@ ps = []; t_full = 0; t_sparse = 0; t_stoch = 0;
 fullm = true
 sparsem =true
 stochm = true
+println("Testing the StudentT model")
 
 if fullm
     println("Testing the full model")

test/test_XGPC.jl

@@ -13,7 +13,7 @@ if doPlots
     using Plots
     pyplot()
 end
-println("Testing the XGPC")
+println("Testing the XGPC model")
 ### TESTING WITH TOY XOR DATASET
     N_data = 1000;N_test = 20
     N_indpoints = 20; N_dim = 2