adding prior_covariance and some general functions

Project.toml

@@ -1,7 +1,7 @@
 name = "GPLinearODEMaker"
 uuid = "27ef9b34-1325-4cec-ba33-00f2f4637873"
 authors = ["Chistian Gilbertson <[email protected]>"]
-version = "0.1.12"
+version = "0.1.13"
 
 [deps]
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"

src/general_functions.jl

@@ -254,3 +254,13 @@ function log_laplace_approximation(
     return logh - λ * g + 0.5 * (n * log(2 * π / λ) - logdet(H))
 
 end
+
+
+riffle(list_of_things::Vector{<:Vector{<:Real}}) = collect(Iterators.flatten(zip(list_of_things...)))
+unriffle(riffled_list::Vector{<:Real}, n_out::Int) = [riffled_list[i:n_out:end] for i in 1:n_out]
+
+
+remainder(vec, x) = [i > 0 ? i % x : (i % x) + x for i in vec]
+
+
+rounded(x::Real) = round(x, digits=2)

src/gp_functions.jl

@@ -1272,3 +1272,11 @@ function reconstruct_total_hyperparameters(
     return total_hyperparameters
 
 end
+
+
+function prior_covariance(kernel::Function, kernel_hyper::Vector{<:Real}, a0::Matrix{<:Real}, obs_xs::Vector{<:Real}; noise_list::Vector{<:Vector{<:Real}}=[zeros(length(obs_xs)) for i in axes(a0,1)], kwargs...)
+    total_hyper = append!(collect(Iterators.flatten(a0)), kernel_hyper)
+    glo = GLO(kernel, length(kernel_hyper), size(a0, 2), size(a0, 1), obs_xs, zeros(length(obs_xs) * size(a0, 1)); noise=collect(Iterators.flatten(zip(noise_list...))), a=a0, kwargs...)
+    Σ = Σ_observations(glo, total_hyper)
+    return Σ
+end