Readme example: no method matching getindex

[jariji]

I ran the example from the Readme, using master cf5c96e but I got an error.

julia> let 

           df = dataset("datasets", "iris")
           df.binary = zeros(Float64, size(df,1))
           df[df.SepalLength .> 5.0,:binary] .= 1.0
           df.SpeciesStr = string.(df.Species)
           idx = rand(1:3,size(df,1),1)
           a = ["A","B","C"]
           df.Random = vec([a[i] for i in idx])

           m = @formula binary ~ SepalWidth + fe(Species)
           x = nlreg(df, m, Binomial(), LogitLink(), start = [0.2] )

           m = @formula binary ~ SepalWidth + PetalLength + fe(Species)
           nlreg(df, m, Binomial(), LogitLink(), Vcov.cluster(:SpeciesStr,:Random) , start = [0.2, 0.2] )    

       end
ERROR: MethodError: no method matching getindex(::Nothing, ::Int64)
Stacktrace:
 [1] S_hat(x::VcovData{LinearAlgebra.Cholesky{Float64, Matrix{Float64}}, 1}, v::Vcov.ClusterCovariance)
   @ Vcov ~/.julia/packages/Vcov/uszbF/src/covarianceestimators/vcovcluster.jl:84
 [2] vcov(x::VcovData{LinearAlgebra.Cholesky{Float64, Matrix{Float64}}, 1}, v::Vcov.ClusterCovariance)
   @ Vcov ~/.julia/packages/Vcov/uszbF/src/covarianceestimators/vcovcluster.jl:115
 [3] nlreg(df::Any, formula::FormulaTerm, distribution::Binomial{Float64}, link::LogitLink, vcov::CovarianceEstimator; weights::Union{Nothing, Symbol}, subset::Union{Nothing, AbstractVector}, start::Union{Nothing, AbstractVector}, maxiter_center::Int64, maxiter::Int64, contrasts::Dict{Symbol, Any}, dof_add::Int64, save::Vector{Symbol}, method::Symbol, drop_singletons::Bool, double_precision::Bool, dev_tol::Float64, rho_tol::Float64, step_tol::Float64, center_tol::Float64, separation::Vector{Symbol}, separation_mu_lbound::Float64, separation_mu_ubound::Float64, separation_ReLU_tol::Float64, separation_ReLU_maxiter::Int64, vcovformula::Union{Nothing, Expr, Symbol}, subsetformula::Union{Nothing, Expr, Symbol}, verbose::Bool)
   @ GLFixedEffectModels ~/.julia/packages/GLFixedEffectModels/Va8dj/src/fit.jl:526

[jmboehm]

I have tried but failed to replicate your error message (it all runs smoothly here). Is this in a fresh Julia session? If so, could you show the output of ] status?

[jariji]

Fresh session:

[bafb0ae5] GLFixedEffectModels v0.4.0 `https://github.com/jmboehm/GLFixedEffectModels.jl.git#master`
[ce6b1742] RDatasets v0.7.7
[ec2bfdc2] Vcov v0.5.2

julia> using GLFixedEffectModels, Vcov, RDatasets

     let
         df = dataset("datasets", "iris")
         df.binary = zeros(Float64, size(df,1))
         df[df.SepalLength .> 5.0,:binary] .= 1.0
         df.SpeciesStr = string.(df.Species)
         idx = rand(1:3,size(df,1),1)
         a = ["A","B","C"]
         df.Random = vec([a[i] for i in idx])
         m = @formula binary ~ SepalWidth + PetalLength + fe(Species)
         nlreg(df, m, Binomial(), LogitLink(), Vcov.cluster(:SpeciesStr,:Random) , start = [0.2, 0.2] )
     end
ERROR: MethodError: no method matching getindex(::Nothing, ::Int64)
Stacktrace:
[1] S_hat(x::VcovData{LinearAlgebra.Cholesky{Float64, Matrix{Float64}}, 1}, v::Vcov.ClusterCovariance)
 @ Vcov ~/.julia/packages/Vcov/uszbF/src/covarianceestimators/vcovcluster.jl:84
[2] vcov(x::VcovData{LinearAlgebra.Cholesky{Float64, Matrix{Float64}}, 1}, v::Vcov.ClusterCovariance)
 @ Vcov ~/.julia/packages/Vcov/uszbF/src/covarianceestimators/vcovcluster.jl:115
[3] nlreg(df::Any, formula::FormulaTerm, distribution::Binomial{Float64}, link::LogitLink, vcov::CovarianceEstimator; weights::Union{Nothing, Symbol}, subset::Union{Nothing, AbstractVector}, start::Union{Nothing, AbstractVector}, maxiter_center::Int64, maxiter::Int64, contrasts::Dict{Symbol, Any}, dof_add::Int64, save::Vector{Symbol}, method::Symbol, drop_singletons::Bool, double_precision::Bool, dev_tol::Float64, rho_tol::Float64, step_tol::Float64, center_tol::Float64, separation::Vector{Symbol}, separation_mu_lbound::Float64, separation_mu_ubound::Float64, separation_ReLU_tol::Float64, separation_ReLU_maxiter::Int64, vcovformula::Union{Nothing, Expr, Symbol}, subsetformula::Union{Nothing, Expr, Symbol}, verbose::Bool)
 @ GLFixedEffectModels ~/.julia/packages/GLFixedEffectModels/Va8dj/src/fit.jl:526
[4] top-level scope

[jmboehm]

Yeah, that's the issue. You're using Vcov.jl, which is incompatible with the fact that GLFixedEffectModels exports its own module VCov. In fact the example from the readme does not do using Vcov.

It would be desirable to move to Vcov.jl (see #16 ) but I'm not sure when I'll have time to implement this. PR's are welcome.

Closing this as duplicate of #16.