Bug: BoundsError: attempt to access 200-element Array{Float64,1} at index [201]

[Gobutah]

The following code leads to an error:

               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.5.1 (2020-08-25)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

(@v1.5) pkg> status Lasso
Status `~/.julia/environments/v1.5/Project.toml`
  [b4fcebef] Lasso v0.5.2

julia> using Lasso

julia> lambdas=[0.001];

julia> X=rand(100,13500);

julia> Y=rand(100);

julia> res=fit(LassoPath, X, Y;λ=lambdas)
ERROR: BoundsError: attempt to access 200-element Array{Float64,1} at index [201]
Stacktrace:
 [1] getindex at ./array.jl:809 [inlined]
 [2] update_coef! at /home/peter/.julia/packages/Lasso/HtY0O/src/coordinate_descent.jl:189 [inlined]
 [3] cycle!(::Lasso.SparseCoefficients{Float64}, ::NaiveCoordinateDescent{Float64,true,Array{Float64,2},Lasso.RandomCoefficientIterator,Nothing}, ::Float64, ::Bool) at /home/peter/.julia/packages/Lasso/HtY0O/src/coordinate_descent.jl:246
 [4] cdfit!(::Lasso.SparseCoefficients{Float64}, ::NaiveCoordinateDescent{Float64,true,Array{Float64,2},Lasso.RandomCoefficientIterator,Nothing}, ::Float64, ::Symbol) at /home/peter/.julia/packages/Lasso/HtY0O/src/coordinate_descent.jl:598
 [5] fit!(::LassoPath{LinearModel{GLM.LmResp{Array{Float64,1}},NaiveCoordinateDescent{Float64,true,Array{Float64,2},Lasso.RandomCoefficientIterator,Nothing}},Float64}; verbose::Bool, cd_maxiter::Int64, cd_tol::Float64, irls_tol::Float64, stopearly::Bool, criterion::Symbol, minStepFac::Float64) at /home/peter/.julia/packages/Lasso/HtY0O/src/coordinate_descent.jl:827
 [6] fit(::Type{LassoPath}, ::Array{Float64,2}, ::Array{Float64,1}, ::Normal{Float64}, ::IdentityLink; wts::Array{Float64,1}, offset::Array{Float64,1}, α::Float64, nλ::Int64, λminratio::Float64, λ::Array{Float64,1}, standardize::Bool, intercept::Bool, algorithm::Type{T} where T, dofit::Bool, irls_tol::Float64, randomize::Bool, maxncoef::Int64, penalty_factor::Nothing, standardizeω::Bool, fitargs::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{,Tuple{}}}) at /home/peter/.julia/packages/Lasso/HtY0O/src/Lasso.jl:491
 [7] top-level scope at REPL[6]:1

[mapi1]

Ran into the same Error, seems to derail for a certain observation regressor ratio lower than 0.5 for me.

[jolars]

For me, setting maxncoef to something large solves this issue. I'm guessing that the package hard-codes an array for storage of coefficients that's of dimension maxncoef but then for low regularization may end up picking more than maxncoef features to be nonzero (which I'm not sure should actually happen given the earlystopping criteria).