Illegal argument: line must be non-negative

mwe.jl

using DataFrames
using Turing
using Distributions
using StatsPlots
using Distances
using LinearAlgebra
using StatsModels
using AbstractGPs, KernelFunctions


df = DataFrame(
    subject = repeat(1:5, inner=3),
    obese = repeat(rand(Bool, 5), inner=3),
    timepoint = [1,2,3,1,3,4,1,2,5,1,4,5,1,3,5],
    bug = rand(Beta(0.9, 5), 15),
    nutrient = rand(Beta(0.9,5), 15)
)

sqexpkernel(alpha::Real, rho::Real) = 
    alpha^2 * AbstractGPs.transform(SEKernel(), 1/(rho*sqrt(2)))


# Define model.
@model GPRegression(y, X, jitter=1e-6) = begin
    # Priors.
    alpha ~ LogNormal(0.0, 0.1)
    rho ~ LogNormal(0.0, 1.0)
    sigma ~ LogNormal(0.0, 1.0)
    
    # Covariance function.
    kernel = sqexpkernel(alpha, rho)
    kernel += jitter * WhiteKernel()  

    # GP (implicit zero-mean).
    gp = GP(kernel)
    
    # Sampling Distribution
    y ~ gp(X, sigma^2)  # add jitter for numerical stability.
end;

gp = GPRegression(df.bug,
          Matrix(df[!,[:subject, :obese, :nutrient]]),
          Matrix{Float64}(df[!,[:timepoint]]))

@time chain = sample(gp, HMC(0.01, 100), 200)

st

Status `~/repos/gptool/Project.toml`
  [99985d1d] AbstractGPs v0.2.13
  [c7e460c6] ArgParse v1.1.0
  [336ed68f] CSV v0.7.7 `https://github.com/JuliaData/CSV.jl.git#master`
  [324d7699] CategoricalArrays v0.8.3
  [a93c6f00] DataFrames v0.21.8
  [b4f34e82] Distances v0.10.0
  [31c24e10] Distributions v0.23.12
  [ec8451be] KernelFunctions v0.8.7
  [91a5bcdd] Plots v1.7.3
  [3eaba693] StatsModels v0.6.15
  [f3b207a7] StatsPlots v0.14.15
  [fce5fe82] Turing v0.14.10
  [37e2e46d] LinearAlgebra
  [56ddb016] Logging
  [de0858da] Printf
  [10745b16] Statistics

stacktrace.txt

julia> @time chain = sample(gp, HMC(0.01, 100), 200)
ERROR: DimensionMismatch("Inconsistent array dimensions.")
Stacktrace:
  [1] loglikelihood
    @ ~/.julia/packages/Distributions/HjzA0/src/multivariates.jl:261 [inlined]
  [2] observe
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/context_implementations.jl:152 [inlined]
  [3] _tilde
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/context_implementations.jl:109 [inlined]
  [4] tilde
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/context_implementations.jl:67 [inlined]
  [5] tilde_observe(ctx::DynamicPPL.DefaultContext, sampler::DynamicPPL.SampleFromPrior, right::AbstractGPs.FiniteGP{GP{AbstractGPs.ZeroMean{Float64}, ScaledKernel{TransformedKernel{SqExponentialKernel, ScaleTransform{Float64}}, Float64}}, ColVecs{Float64, Matrix{Float64}, SubArray{Float64, 1, Matrix{Float64}, Tuple{Base.Slice{Base.OneTo{Int64}}, Int64}, true}}, Matrix{Float64}}, left::Vector{Float64}, vname::DynamicPPL.VarName{:y, Tuple{}}, vinds::Tuple{}, vi::DynamicPPL.UntypedVarInfo{DynamicPPL.Metadata{Dict{DynamicPPL.VarName, Int64}, Vector{Distribution}, Vector{DynamicPPL.VarName}, Vector{Real}, Vector{Set{DynamicPPL.Selector}}}, Float64})
    @ DynamicPPL ~/.julia/packages/DynamicPPL/jOFDR/src/context_implementations.jl:89
  [6] #15
    @ ~/repos/gptool/scratch.jl:39 [inlined]
  [7] (::var"#15#16")(_rng::Random._GLOBAL_RNG, _model::DynamicPPL.Model{var"#15#16", (:y, :X, :jitter), (:jitter,), , Tuple{Vector{Float64}, Matrix{Float64}, Matrix{Float64}}, Tuple{Float64}}, _varinfo::DynamicPPL.UntypedVarInfo{DynamicPPL.Metadata{Dict{DynamicPPL.VarName, Int64}, Vector{Distribution}, Vector{DynamicPPL.VarName}, Vector{Real}, Vector{Set{DynamicPPL.Selector}}}, Float64}, _sampler::DynamicPPL.SampleFromPrior, _context::DynamicPPL.DefaultContext, y::Vector{Float64}, X::Matrix{Float64}, jitter::Matrix{Float64})
    @ Main ./none:0
  [8] macro expansion
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/model.jl:0 [inlined]
  [9] _evaluate
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/model.jl:160 [inlined]
 [10] evaluate_threadunsafe
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/model.jl:130 [inlined]
 [11] Model
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/model.jl:92 [inlined]
 [12] Model
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/model.jl:98 [inlined]
 [13] VarInfo
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/varinfo.jl:110 [inlined]
 [14] VarInfo
    @ ~/.julia/packages/DynamicPPL/jOFDR/src/varinfo.jl:109 [inlined]
 [15] DynamicPPL.Sampler(alg::HMC{Turing.Core.ForwardDiffAD{40}, , AdvancedHMC.UnitEuclideanMetric}, model::DynamicPPL.Model{var"#15#16", (:y, :X, :jitter), (:jitter,), , Tuple{Vector{Float64}, Matrix{Float64}, Matrix{Float64}}, Tuple{Float64}}, s::DynamicPPL.Selector)
    @ Turing.Inference ~/.julia/packages/Turing/UsQlw/src/inference/hmc.jl:378
 [16] Sampler
    @ ~/.julia/packages/Turing/UsQlw/src/inference/hmc.jl:376 [inlined]
 [17] sample(rng::Random._GLOBAL_RNG, model::DynamicPPL.Model{var"#15#16", (:y, :X, :jitter), (:jitter,), , Tuple{Vector{Float64}, Matrix{Float64}, Matrix{Float64}}, Tuple{Float64}}, alg::HMC{Turing.Core.ForwardDiffAD{40}, , AdvancedHMC.UnitEuclideanMetric}, N::Int64; kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{, Tuple{}}})
    @ Turing.Inference ~/.julia/packages/Turing/UsQlw/src/inference/Inference.jl:164
 [18] sample
    @ ~/.julia/packages/Turing/UsQlw/src/inference/Inference.jl:164 [inlined]
 [19] #sample#1
    @ ~/.julia/packages/Turing/UsQlw/src/inference/Inference.jl:154 [inlined]
 [20] sample(model::DynamicPPL.Model{var"#15#16", (:y, :X, :jitter), (:jitter,), , Tuple{Vector{Float64}, Matrix{Float64}, Matrix{Float64}}, Tuple{Float64}}, alg::HMC{Turing.Core.ForwardDiffAD{40}, , AdvancedHMC.UnitEuclideanMetric}, N::Int64)
    @ Turing.Inference ~/.julia/packages/Turing/UsQlw/src/inference/Inference.jl:154
 [21] top-level scope
    @ ./timing.jl:192 [inlined]
 [22] top-level scope
    @ ./REPL[4]:0
 [23] eval
    @ ./boot.jl:360 [inlined]