Print a module that exports the name rather than the module that defines the name

closes #24

On this branch I get:

julia> print_explicit_imports(MixedModels)
WARNING: both Distributions and BSplineKit export "support"; uses of it in module MixedModels must be qualified
Module MixedModels is relying on implicit imports for 162 names. These could be explicitly imported as follows:

```julia
using Arrow: Arrow
using BSplineKit: BSplineKit
using BSplineKit: BSplineOrder
using BSplineKit: Natural
using BSplineKit: Derivative
using BSplineKit: SplineInterpolation
using BSplineKit: interpolate
using DataAPI: DataAPI
using Distributions: Distributions
using GLM: Bernoulli
using GLM: Binomial
using Distributions: Chisq
using Distributions: Distribution
using GLM: Gamma
using GLM: InverseGaussian
using GLM: Normal
using GLM: Poisson
using Distributions: ccdf
using Distributions: estimate
using GLM: GLM
using GLM: GeneralizedLinearModel
using GLM: IdentityLink
using GLM: InverseLink
using GLM: LinearModel
using GLM: LogLink
using GLM: LogitLink
using GLM: ProbitLink
using GLM: SqrtLink
using GLM: glm
using JSON3: JSON3
using LinearAlgebra: LinearAlgebra
using LinearAlgebra: Adjoint
using LinearAlgebra: BLAS
using LinearAlgebra: ColumnNorm
using LinearAlgebra: Diagonal
using LinearAlgebra: Hermitian
using LinearAlgebra: I
using LinearAlgebra: LAPACK
using LinearAlgebra: LowerTriangular
using LinearAlgebra: SVD
using LinearAlgebra: SymTridiagonal
using LinearAlgebra: Symmetric
using LinearAlgebra: UpperTriangular
using LinearAlgebra: cond
using LinearAlgebra: diag
using LinearAlgebra: diagind
using LinearAlgebra: dot
using LinearAlgebra: eigen
using LinearAlgebra: isdiag
using LinearAlgebra: ldiv!
using LinearAlgebra: lmul!
using LinearAlgebra: logdet
using LinearAlgebra: mul!
using LinearAlgebra: norm
using LinearAlgebra: normalize
using LinearAlgebra: normalize!
using LinearAlgebra: qr
using LinearAlgebra: rank
using LinearAlgebra: rdiv!
using LinearAlgebra: rmul!
using LinearAlgebra: svd
using LinearAlgebra: tr
using LinearAlgebra: tril!
using Markdown: Markdown
using NLopt: NLopt
using NLopt: ftol_abs
using NLopt: ftol_rel
using NLopt: initial_step
using NLopt: maxtime
using NLopt: xtol_abs
using NLopt: xtol_rel
using PooledArrays: PooledArrays
using PooledArrays: PooledArray
using PrecompileTools: PrecompileTools
using PrecompileTools: @compile_workload
using PrecompileTools: @setup_workload
using ProgressMeter: ProgressMeter
using ProgressMeter: Progress
using ProgressMeter: ProgressUnknown
using ProgressMeter: finish!
using ProgressMeter: next!
using Random: Random
using Random: AbstractRNG
using Random: randn!
using SparseArrays: SparseArrays
using SparseArrays: SparseMatrixCSC
using SparseArrays: SparseVector
using SparseArrays: dropzeros!
using SparseArrays: nnz
using SparseArrays: nonzeros
using SparseArrays: nzrange
using SparseArrays: rowvals
using SparseArrays: sparse
using StaticArrays: StaticArrays
using StaticArrays: SVector
using Statistics: Statistics
using StatsBase: mean
using StatsBase: quantile
using StatsBase: std
using StatsAPI: StatsAPI
using StatsBase: aic
using StatsBase: aicc
using StatsBase: bic
using StatsBase: coef
using StatsModels: coefnames
using StatsBase: coeftable
using StatsBase: confint
using StatsBase: deviance
using StatsBase: dof
using StatsBase: dof_residual
using StatsBase: fit
using StatsBase: fit!
using StatsBase: fitted
using StatsBase: isfitted
using StatsBase: islinear
using StatsBase: leverage
using StatsBase: loglikelihood
using StatsBase: meanresponse
using StatsModels: modelmatrix
using StatsBase: nobs
using StatsBase: predict
using StatsBase: r2
using StatsBase: residuals
using StatsModels: response
using StatsBase: responsename
using StatsBase: stderror
using StatsBase: vcov
using StatsBase: weights
using StatsBase: StatsBase
using StatsBase: CoefTable
using StatsBase: model_response
using StatsBase: summarystats
using StatsModels: StatsModels
using StatsModels: @formula
using StatsModels: AbstractContrasts
using StatsModels: AbstractTerm
using StatsModels: CategoricalTerm
using StatsModels: ConstantTerm
using StatsModels: DummyCoding
using StatsModels: EffectsCoding
using StatsModels: FormulaTerm
using StatsModels: FunctionTerm
using StatsModels: HelmertCoding
using StatsModels: HypothesisCoding
using StatsModels: InteractionTerm
using StatsModels: InterceptTerm
using StatsModels: MatrixTerm
using StatsModels: SeqDiffCoding
using StatsModels: Term
using StatsModels: apply_schema
using StatsModels: drop_term
using StatsModels: formula
using StatsModels: modelcols
using StatsModels: term
using StructTypes: StructTypes
using Tables: Tables
using Tables: columntable
using TypedTables: rows
using TypedTables: TypedTables
using TypedTables: DictTable
using TypedTables: FlexTable
using TypedTables: Table

Additionally, module MixedModels has stale explicit imports for these unused names:

BlasReal
copytri!
dataset
datasets
linkfun

In particular, note the

using StaticArrays: StaticArrays
using StaticArrays: SVector

and

using BSplineKit: BSplineKit
using BSplineKit: BSplineOrder
using BSplineKit: Natural
using BSplineKit: Derivative
using BSplineKit: SplineInterpolation
using BSplineKit: interpolate

which fixes the two issues @palday called out in https://github.com/JuliaStats/MixedModels.jl/pull/748

ericphanson / ExplicitImports.jl

Print a module that exports the name rather than the module that defines the name #25

Codecov Report