Merge branch 'main' into db/RegressionTests

.github/dependabot.yml

@@ -0,0 +1,7 @@
+# https://docs.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/" # Location of package manifests
+    schedule:
+      interval: "weekly"

.github/workflows/benchmark.yml

@@ -21,10 +21,10 @@ jobs:
           fail-fast: false
         steps:
           - uses: actions/checkout@v4
-          - uses: julia-actions/setup-julia@v1
+          - uses: julia-actions/setup-julia@v2
             with:
               version: "1"
-          - uses: julia-actions/cache@v1
+          - uses: julia-actions/cache@v2
           - uses: julia-actions/julia-buildpkg@v1
           - name: Benchmark run
             run: |

.github/workflows/current.yml

@@ -27,10 +27,10 @@ jobs:
          - macOS-14 # apple silicon!
     steps:
       - uses: actions/checkout@v4
-      - uses: julia-actions/setup-julia@v1
+      - uses: julia-actions/setup-julia@v2
         with:
           version: ${{ matrix.julia-version }}
-      - uses: julia-actions/cache@v1
+      - uses: julia-actions/cache@v2
         with:
           cache-compiled: "true"
       - uses: julia-actions/julia-buildpkg@v1

.github/workflows/documenter.yml

@@ -22,10 +22,10 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - uses: actions/checkout@v4
-      - uses: julia-actions/setup-julia@v1
+      - uses: julia-actions/setup-julia@v2
         with:
           version: 1.8
-      - uses: julia-actions/cache@v1
+      - uses: julia-actions/cache@v2
         with:
             cache-compiled: "true"
       - uses: julia-actions/julia-buildpkg@latest

.github/workflows/minimum.yml

@@ -25,10 +25,10 @@ jobs:
         os: [ubuntu-22.04, macos-11, windows-2019]
     steps:
       - uses: actions/checkout@v4
-      - uses: julia-actions/setup-julia@v1
+      - uses: julia-actions/setup-julia@v2
         with:
           version: ${{ matrix.julia-version }}
-      - uses: julia-actions/cache@v1
+      - uses: julia-actions/cache@v2
         with:
           cache-compiled: "true"
       - uses: julia-actions/julia-buildpkg@v1

.github/workflows/nightly.yml

@@ -25,10 +25,10 @@ jobs:
         os: [ubuntu-latest]
     steps:
       - uses: actions/checkout@v4
-      - uses: julia-actions/setup-julia@v1
+      - uses: julia-actions/setup-julia@v2
         with:
           version: ${{ matrix.julia-version }}
-      - uses: julia-actions/cache@v1
+      - uses: julia-actions/cache@v2
         with:
             cache-compiled: true
       - uses: julia-actions/julia-runtest@v1

NEWS.md

@@ -1,3 +1,11 @@
+MixedModels v4.24.1 Release Notes
+==============================
+Add type notations in `pwrss(::LinearMixedModel)` and `logdet(::LinearMixedModel)` to enhance type inference. [#773]
+
+MixedModels v4.24.0 Release Notes
+==============================
+* Properties for `GeneralizedLinearMixedModel` now default to delegation to the internal weighted `LinearMixedModel` when that property is not explicitly handled by `GeneralizedLinearMixedModel`. Previously, properties were delegated on an explicit basis, which meant that they had to be added manually as use cases were discovered. The downside to the new approach is that it is now possible to access properties whose definition in the LMM case doesn't match the GLMM definition when the GLMM definition hasn't been explicitly been implemented. [#767]
+
 MixedModels v4.23.1 Release Notes
 ==============================
 * Fix for `simulate!` when only the estimable coefficients for a rank-deficient model are provided. [#756]
@@ -516,3 +524,5 @@ Package dependencies
 [#748]: https://github.com/JuliaStats/MixedModels.jl/issues/748
 [#755]: https://github.com/JuliaStats/MixedModels.jl/issues/755
 [#756]: https://github.com/JuliaStats/MixedModels.jl/issues/756
+[#767]: https://github.com/JuliaStats/MixedModels.jl/issues/767
+[#773]: https://github.com/JuliaStats/MixedModels.jl/issues/773

Project.toml

@@ -1,7 +1,7 @@
 name = "MixedModels"
 uuid = "ff71e718-51f3-5ec2-a782-8ffcbfa3c316"
 author = ["Phillip Alday <[email protected]>", "Douglas Bates <[email protected]>", "Jose Bayoan Santiago Calderon <[email protected]>"]
-version = "4.23.1"
+version = "4.24.1"
 
 [deps]
 Arrow = "69666777-d1a9-59fb-9406-91d4454c9d45"

docs/src/bootstrap.md

@@ -163,13 +163,20 @@ For computers with many processors (as opposed to a single processor with severa
 
 ```@example Main
 using Distributed
-using ProgressMeter
 # you already have 1 proc by default, so add the number of additional cores with `addprocs`
 # you need at least as many RNGs as cores you want to use in parallel
 # but you shouldn't use all of your cores because nested within this
 # is the multithreading of the linear algebra
+# addprocs(1)
 @info "Currently using $(nprocs()) processors total and $(nworkers()) for work"
 
+# Load the necessary packages on all workers
+# For clusters, you will also need to make sure that the Julia
+# environment (Project.toml) is set up and activated on each worker.
+@everywhere begin
+    using ProgressMeter
+    using MixedModels
+end
 # copy everything to workers
 @showprogress for w in workers()
     remotecall_fetch(() -> coefnames(m2), w)

src/MixedModels.jl

@@ -6,7 +6,7 @@ using BSplineKit: BSplineKit, BSplineOrder, Natural, Derivative, SplineInterpola
 using BSplineKit: interpolate
 using DataAPI: DataAPI, levels, refpool, refarray, refvalue
 using Distributions: Distributions, Bernoulli, Binomial, Chisq, Distribution, Gamma
-using Distributions: InverseGaussian, Normal, Poisson, ccdf, estimate
+using Distributions: InverseGaussian, Normal, Poisson, ccdf
 using GLM: GLM, GeneralizedLinearModel, IdentityLink, InverseLink, LinearModel
 using GLM: Link, LogLink, LogitLink, ProbitLink, SqrtLink
 using GLM: canonicallink, glm, linkinv, dispersion, dispersion_parameter
@@ -16,10 +16,10 @@ using LinearAlgebra: Diagonal, Hermitian, HermOrSym, I, LAPACK, LowerTriangular
 using LinearAlgebra: PosDefException, SVD, SymTridiagonal, Symmetric
 using LinearAlgebra: UpperTriangular, cond, diag, diagind, dot, eigen, isdiag
 using LinearAlgebra: ldiv!, lmul!, logdet, mul!, norm, normalize, normalize!, qr
-using LinearAlgebra: rank, rdiv!, rmul!, svd, tr, tril!
+using LinearAlgebra: rank, rdiv!, rmul!, svd, tril!
 using Markdown: Markdown
 using MixedModelsDatasets: dataset, datasets
-using NLopt: NLopt, Opt, ftol_abs, ftol_rel, initial_step, maxtime, xtol_abs, xtol_rel
+using NLopt: NLopt, Opt, ftol_abs, ftol_rel, initial_step, xtol_abs, xtol_rel
 using PooledArrays: PooledArrays, PooledArray
 using PrecompileTools: PrecompileTools, @setup_workload, @compile_workload
 using ProgressMeter: ProgressMeter, Progress, ProgressUnknown, finish!, next!
@@ -38,9 +38,9 @@ using StatsModels: StatsModels, AbstractContrasts, AbstractTerm, CategoricalTerm
 using StatsModels: ConstantTerm, DummyCoding, EffectsCoding, FormulaTerm, FunctionTerm
 using StatsModels: HelmertCoding, HypothesisCoding, InteractionTerm, InterceptTerm
 using StatsModels: MatrixTerm, SeqDiffCoding, TableRegressionModel, Term
-using StatsModels: apply_schema, drop_term, formula, modelcols, term, @formula
+using StatsModels: apply_schema, drop_term, formula, lrtest, modelcols, term, @formula
 using StructTypes: StructTypes
-using Tables: Tables, columntable, rows
+using Tables: Tables, columntable
 using TypedTables: TypedTables, DictTable, FlexTable, Table
 
 export @formula,
@@ -114,6 +114,7 @@ export @formula,
     logdet,
     loglikelihood,
     lowerbd,
+    lrtest,
     meanresponse,
     modelmatrix,
     model_response,

src/generalizedlinearmixedmodel.jl

@@ -351,7 +351,7 @@ function GeneralizedLinearMixedModel(
     tbl,
     d::Distribution,
     l::Type;
-    kwargs...
+    kwargs...,
 )
     throw(ArgumentError("Expected a Link instance (`$l()`), got a type (`$l`)."))
 end
@@ -376,7 +376,7 @@ function GeneralizedLinearMixedModel(
     tbl::Tables.ColumnTable,
     d::Normal,
     l::IdentityLink;
-    kwargs...
+    kwargs...,
 )
     return throw(
         ArgumentError("use LinearMixedModel for Normal distribution with IdentityLink")
@@ -489,12 +489,12 @@ function Base.getproperty(m::GeneralizedLinearMixedModel, s::Symbol)
         σs(m)
     elseif s == :σρs
         σρs(m)
-    elseif s ∈ (:A, :L, :optsum, :reterms, :Xymat, :feterm, :formula, :parmap)
-        getfield(m.LMM, s)
-    elseif s ∈ (:dims, :λ, :lowerbd, :corr, :PCA, :rePCA, :X)
-        getproperty(m.LMM, s)
     elseif s == :y
         m.resp.y
+    elseif !hasfield(GeneralizedLinearMixedModel, s) && s ∈ propertynames(m.LMM, true)
+        # automatically delegate as much as possible to the internal local linear approximation
+        # NB: the !hasfield call has to be first since we're calling getproperty() with m.LMM...
+        getproperty(m.LMM, s)
     else
         getfield(m, s)
     end

src/likelihoodratiotest.jl

@@ -219,7 +219,7 @@ function _iscomparable(m::LinearMixedModel...)
     if any(getproperty.(getproperty.(m, :optsum), :REML))
         isconstant(coefnames.(m)) || throw(
             ArgumentError(
-                "Likelihood-ratio tests for REML-fitted models are only valid when the fixed-effects specifications are identical",
+                "Likelihood-ratio tests for REML-fitted models are only valid when the fixed-effects specifications are identical"
             ),
         )
     end

src/linalg/logdet.jl

@@ -25,11 +25,11 @@ lower Cholesky factor.
 """
 function LinearAlgebra.logdet(m::LinearMixedModel{T}) where {T}
     L = m.L
-    @inbounds s = sum(j -> LD(L[kp1choose2(j)]), axes(m.reterms, 1))
+    @inbounds s = sum(j -> LD(L[kp1choose2(j)])::T, axes(m.reterms, 1))
     if m.optsum.REML
-        lastL = last(L)
+        lastL = last(L)::Matrix{T}
         s += LD(lastL)        # this includes the log of sqrtpwrss
         s -= log(last(lastL)) # so we need to subtract it from the sum
     end
-    return s + s  # multiply by 2 b/c the desired det is of the symmetric mat, not the factor
+    return (s + s)::T  # multiply by 2 b/c the desired det is of the symmetric mat, not the factor
 end

src/linalg/rankUpdate.jl

@@ -12,7 +12,7 @@ function rankUpdate! end
 
 function rankUpdate!(C::AbstractMatrix, a::AbstractArray, α, β)
     return error(
-        "We haven't implemented a method for $(typeof(C)), $(typeof(a)). Please file an issue on GitHub.",
+        "We haven't implemented a method for $(typeof(C)), $(typeof(a)). Please file an issue on GitHub."
     )
 end
 

src/linearmixedmodel.jl

@@ -50,7 +50,7 @@ function LinearMixedModel(
 end
 
 const _MISSING_RE_ERROR = ArgumentError(
-    "Formula contains no random effects; this isn't a mixed model. Perhaps you want to use GLM.jl?",
+    "Formula contains no random effects; this isn't a mixed model. Perhaps you want to use GLM.jl?"
 )
 
 function LinearMixedModel(
@@ -62,7 +62,7 @@ function LinearMixedModel(
     fvars ⊆ tvars ||
         throw(
             ArgumentError(
-                "The following formula variables are not present in the table: $(setdiff(fvars, tvars))",
+                "The following formula variables are not present in the table: $(setdiff(fvars, tvars))"
             ),
         )
 
@@ -226,7 +226,7 @@ end
 function _offseterr()
     return throw(
         ArgumentError(
-            "Offsets are not supported in linear models. You can simply shift the response by the offset.",
+            "Offsets are not supported in linear models. You can simply shift the response by the offset."
         ),
     )
 end
@@ -372,7 +372,7 @@ function StatsBase.confint(m::MixedModel{T}; level=0.95) where {T}
     return DictTable(;
         coef=coefnames(m),
         lower=β .- cutoff .* std,
-        upper=β .+ cutoff .* std
+        upper=β .+ cutoff .* std,
     )
 end
 
@@ -895,7 +895,7 @@ end
 
 The penalized, weighted residual sum-of-squares.
 """
-pwrss(m::LinearMixedModel) = abs2(last(last(m.L)))
+pwrss(m::LinearMixedModel{T}) where {T} = abs2(last(last(m.L)::Matrix{T}))
 
 """
     ranef!(v::Vector{Matrix{T}}, m::MixedModel{T}, β, uscale::Bool) where {T}

src/mixedmodel.jl

@@ -80,7 +80,7 @@ function StatsAPI.fit(
     tbl,
     d::Type,
     args...;
-    kwargs...
+    kwargs...,
 )
     throw(ArgumentError("Expected a Distribution instance (`$d()`), got a type (`$d`)."))
 end
@@ -91,7 +91,7 @@ function StatsAPI.fit(
     tbl,
     d::Distribution,
     l::Type;
-    kwargs...
+    kwargs...,
 )
     throw(ArgumentError("Expected a Link instance (`$l()`), got a type (`$l`)."))
 end

src/predict.jl

@@ -115,7 +115,7 @@ function _predict(m::MixedModel{T}, newdata, β; new_re_levels) where {T}
             any(any(ismissing, Tables.getcolumn(newdata, col)) for col in respvars)
             throw(
                 ArgumentError(
-                    "Response column must be initialized to a non-missing numeric value.",
+                    "Response column must be initialized to a non-missing numeric value."
                 ),
             )
         end

src/profile/thetapr.jl

@@ -10,7 +10,7 @@ function optsumj(os::OptSummary, j::Integer)
     return OptSummary(
         deleteat!(copy(os.final), j),
         deleteat!(copy(os.lowerbd), j),
-        os.optimizer
+        os.optimizer,
     )
 end
 

src/randomeffectsterm.jl

@@ -174,7 +174,7 @@ function StatsModels.apply_schema(
     if !(typeof(first) <: CategoricalTerm)
         throw(
             ArgumentError(
-                "nesting terms requires categorical grouping term, got $first.  Manually specify $first as `CategoricalTerm` in hints/contrasts",
+                "nesting terms requires categorical grouping term, got $first.  Manually specify $first as `CategoricalTerm` in hints/contrasts"
             ),
         )
     end

src/remat.jl

@@ -586,7 +586,7 @@ end
 function copyscaleinflate!(
     Ljj::UniformBlockDiagonal{T},
     Ajj::UniformBlockDiagonal{T},
-    Λj::ReMat{T,S}
+    Λj::ReMat{T,S},
 ) where {T,S}
     λ = Λj.λ
     dind = diagind(S, S)
@@ -604,7 +604,7 @@ end
 function copyscaleinflate!(
     Ljj::Matrix{T},
     Ajj::UniformBlockDiagonal{T},
-    Λj::ReMat{T,S}
+    Λj::ReMat{T,S},
 ) where {T,S}
     copyto!(Ljj, Ajj)
     n = LinearAlgebra.checksquare(Ljj)

src/simulate.jl

@@ -236,7 +236,7 @@ function _simulate!(
         ismissing(σ) ||
         throw(
             ArgumentError(
-                "You must not specify a dispersion parameter for model families without a dispersion parameter",
+                "You must not specify a dispersion parameter for model families without a dispersion parameter"
             ),
         )
 

src/utilities.jl

@@ -136,7 +136,7 @@ function replicate(
 )
     use_threads && Base.depwarn(
         "use_threads is deprecated and will be removed in a future release",
-        :replicate
+        :replicate,
     )
     if !isnothing(hide_progress)
         Base.depwarn(