KeyedDistributions.jl provides thin wrappers of Distribution and Sampleable, to store keys and dimnames for the variates.
julia> using KeyedDistributions, Distributions, NamedDims;
julia> kd = KeyedDistribution(MvNormal(3, 1.0); id=[:x, :y, :z]);
julia> axiskeys(kd)
([:x, :y, :z],)
julia> dimnames(kd)
(:id,)
julia> distribution(kd)
ZeroMeanIsoNormal(
dim: 3
μ: 3-element Zeros{Float64}
Σ: [1.0 0.0 0.0; 0.0 1.0 0.0; 0.0 0.0 1.0]
)Methods for Distribution and Sampleable return KeyedArrays in place of regular Arrays, where applicable.
julia> mean(kd)
1-dimensional KeyedArray(NamedDimsArray(...)) with keys:
↓ id ∈ 3-element Vector{Symbol}
And data, 3-element Zeros{Float64}:
(:x) 0.0
(:y) 0.0
(:z) 0.0
julia> rand(kd, 2)
2-dimensional KeyedArray(...) with keys:
↓ id ∈ 3-element Vector{Symbol}
→ sample ∈ 2-element OneTo{Int}
And data, 3×2 Matrix{Float64}:
(1) (2)
(:x) -1.11227 -0.279841
(:y) 0.00784496 0.871718
(:z) -0.930186 -0.8922In this way, KeyedDistributions.jl extends the AxisKeys and NamedDims ecosystems to Distributions.jl.