fix docs

Author: mmkim1210

.github/workflows/CI.yml

@@ -60,12 +60,12 @@ jobs:
       - uses: actions/checkout@v2
       - uses: julia-actions/setup-julia@v1
         with:
-          version: '1.8'
+          version: '1'
       - run: |
           julia --project=docs -e '
             using Pkg
-            Pkg.add("Documenter")
-            Pkg.add(url = "https://github.com/Hua-Zhou/MultiResponseVarianceComponentModels.jl.git")'
+            Pkg.develop(PackageSpec(path=pwd()))
+            Pkg.instantiate()'
       - run: |
           julia --project=docs -e '
             using Documenter: doctest

.gitignore

@@ -1,3 +1,4 @@
 Manifest.toml
 *.jl.*.mem
 .DS_Store
+docs/build

docs/Project.toml

@@ -0,0 +1,3 @@
+[deps]
+Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
+MultiResponseVarianceComponentModels = "95bd6349-d788-4b5f-978d-e543c2eb4d7d"

docs/make.jl

@@ -3,7 +3,7 @@ using MultiResponseVarianceComponentModels
 
 makedocs(
     modules = [MultiResponseVarianceComponentModels],
-    sitename = "MRVCs.jl",
+    sitename = "MRVCModels.jl",
     format = Documenter.HTML(
         prettyurls = get(ENV, "CI", nothing) == "true",
         assets = String[],

docs/src/examples.md

@@ -33,18 +33,20 @@ Y = X * B + reshape(Ωchol.L * randn(n * d), n, d)
 # Maximum likelihood estimation
 ```@repl 1
 model = MultiResponseVarianceComponentModel(Y, X, V)
-@time fit!(model, verbose = true)
+@timev fit!(model, verbose = true)
 ```
 
 For residual maximum likelihood estimation, you can instead type:
 ```julia
-@time fit!(model, reml = true, verbose = true)
+@timev fit!(model, reml = true, verbose = true)
 ```
 
 Then variance components and mean effects estimates can be accessed through
 ```@repl 1
 model.Σ
 model.B
+hcat(vec(B), vec(model.B))
+reduce(hcat, [hcat(vec(Σ[i]), vec(model.Σ[i])) for i in 1:m])
 ```
 
 # Standard errors

src/MultiResponseVarianceComponentModels.jl

@@ -42,16 +42,16 @@ struct MultiResponseVarianceComponentModel{T <: BlasReal}
     storage_d_d_1    :: Matrix{T}
     storage_d_d_2    :: Matrix{T}
     storage_d_d_3    :: Matrix{T}
-    storage_d_d_4    :: Matrix{T}
-    storage_d_d_5    :: Matrix{T}
-    storage_d_d_6    :: Matrix{T}
-    storage_d_d_7    :: Matrix{T}
+    storage_d_d_4    :: Matrix{T} # for manopt.jl
+    storage_d_d_5    :: Matrix{T} # for manopt.jl
+    storage_d_d_6    :: Matrix{T} # for manopt.jl
+    storage_d_d_7    :: Matrix{T} # for manopt.jl
     storage_p_p      :: Matrix{T}
     storage_nd_nd    :: Matrix{T}
     storage_pd_pd    :: Matrix{T}
-    storages_nd_nd   :: Vector{Matrix{T}}
-    Bcov             :: Matrix{T}
-    Σcov             :: Matrix{T}
+    storages_nd_nd   :: Vector{Matrix{T}} # for fisher_Σ!
+    Bcov             :: Matrix{T} # for fisher_B!
+    Σcov             :: Matrix{T} # for fisher_Σ!
     logl             :: Vector{T}
 end
 
@@ -93,16 +93,16 @@ function MultiResponseVarianceComponentModel(
     storage_d_d_1    = Matrix{T}(undef, d, d)
     storage_d_d_2    = Matrix{T}(undef, d, d)
     storage_d_d_3    = Matrix{T}(undef, d, d)
-    storage_d_d_4    = Matrix{T}(undef, d, d) # for manopt.jl
-    storage_d_d_5    = Matrix{T}(undef, d, d) # for manopt.jl
-    storage_d_d_6    = Matrix{T}(undef, d, d) # for manopt.jl
-    storage_d_d_7    = Matrix{T}(undef, d, d) # for manopt.jl
+    storage_d_d_4    = Matrix{T}(undef, d, d)
+    storage_d_d_5    = Matrix{T}(undef, d, d)
+    storage_d_d_6    = Matrix{T}(undef, d, d)
+    storage_d_d_7    = Matrix{T}(undef, d, d) 
     storage_p_p      = Matrix{T}(undef, p, p)
     storage_nd_nd    = Matrix{T}(undef, nd, nd)
     storage_pd_pd    = Matrix{T}(undef, pd, pd)
-    storages_nd_nd   = [Matrix{T}(undef, nd, nd) for _ in 1:m] # for fisher_Σ!
-    Bcov             = Matrix{T}(undef, pd, pd) # for fisher_B!
-    Σcov             = Matrix{T}(undef, m * (binomial(d, 2) + d), m * (binomial(d, 2) + d)) # for fisher_Σ!
+    storages_nd_nd   = [Matrix{T}(undef, nd, nd) for _ in 1:m]
+    Bcov             = Matrix{T}(undef, pd, pd)
+    Σcov             = Matrix{T}(undef, m * (binomial(d, 2) + d), m * (binomial(d, 2) + d))
     logl             = zeros(T, 1)
     MultiResponseVarianceComponentModel{T}(
         Y, Xmat, V,