Draft of using GH Actions (#258)

* Use GH Actions:  Replaces Travis and AppVeyor
* Test only with current release
* Two tiers support
* Documenter compatibility

Co-authored-by: Dave Kleinschmidt <[email protected]>
Co-authored-by: Douglas Bates <[email protected]>
Co-authored-by: Phillip Alday <[email protected]>

Author: 4 people

.github/workflows/Tier1.yml

@@ -0,0 +1,43 @@
+name: Tier1
+on:
+  push:
+    branches:
+      - master
+    paths-ignore:
+      - 'LICENSE.md'
+      - 'README.md'
+  pull_request:
+    branches:
+      - master
+    paths-ignore:
+      - 'LICENSE.md'
+      - 'README.md'
+  release:
+    types:
+      - created
+jobs:
+  ci:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        julia-version: [1.3]
+        julia-arch: [x64]
+        os: [ubuntu-18.04, macos-10.15, windows-2019]
+    steps:
+      - uses: actions/checkout@v2
+      - uses: julia-actions/setup-julia@v1
+        with:
+          version: ${{ matrix.julia-version }}
+      - uses: julia-actions/[email protected]
+      - uses: julia-actions/[email protected]
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      - uses: julia-actions/[email protected]
+        if: startsWith(matrix.os, 'Ubuntu')
+        env:
+          CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+      - uses: julia-actions/julia-docdeploy@releases/v1
+        if: startsWith(matrix.os, 'Ubuntu')
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          DOCUMENTER_KEY: ${{ secrets.DOCUMENTER_KEY }}

.github/workflows/Tier2.yml

@@ -0,0 +1,40 @@
+name: Tier2
+on:
+  push:
+    branches:
+      - master
+    paths-ignore:
+      - 'LICENSE.md'
+      - 'README.md'
+  pull_request:
+    branches:
+      - master
+    paths-ignore:
+      - 'LICENSE.md'
+      - 'README.md'
+  release:
+    types:
+      - created
+jobs:
+  ci:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        julia-version: [1.3, 1.4.0-rc1]
+        julia-arch: [x64, x86]
+        os: [ubuntu-18.04, macos-10.15, windows-2019]
+        exclude:
+          - julia-version: 1.3
+            julia-arch: x64
+          - os: macos-10.15
+            julia-arch: x86
+    steps:
+      - uses: actions/checkout@v2
+      - uses: julia-actions/setup-julia@v1
+        with:
+          version: ${{ matrix.julia-version }}
+      - uses: julia-actions/[email protected]
+      - uses: julia-actions/[email protected]
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.travis.yml

@@ -2,7 +2,7 @@
 
 |**Documentation**|**Citation**|**Build Status**|**Code Coverage**|
 |:-:|:-:|:-:|:-:|
-|[![][docs-stable-img]][docs-stable-url] [![][docs-latest-img]][docs-latest-url] | [![][doi-img]][doi-url] | [![][travis-img]][travis-url] [![][appveyor-img]][appveyor-url] [![PkgEval][pkgeval-img]][pkgeval-url] | [![][codecov-img]][codecov-url]|
+|[![][docs-stable-img]][docs-stable-url] [![][docs-latest-img]][docs-latest-url] | [![][doi-img]][doi-url] | [![T1-url][T1-img]][pkgeval-url] [![T2-url][T2-img]][pkgeval-url] [![PkgEval][pkgeval-img]][pkgeval-url] | [![][codecov-img]][codecov-url]|
 
 [doi-img]: https://zenodo.org/badge/9106942.svg
 [doi-url]: https://zenodo.org/badge/latestdoi/9106942
@@ -13,15 +13,15 @@
 [docs-stable-img]: https://img.shields.io/badge/docs-stable-blue.svg
 [docs-stable-url]: https://juliastats.github.io/MixedModels.jl/stable
 
-[travis-img]: https://travis-ci.org/JuliaStats/MixedModels.jl.svg?branch=master
-[travis-url]: https://travis-ci.org/JuliaStats/MixedModels.jl
-
-[appveyor-img]: https://ci.appveyor.com/api/projects/status/github/JuliaStats/MixedModels.jl?svg=true
-[appveyor-url]: https://ci.appveyor.com/project/JuliaStats/mixedmodels-jl
-
 [codecov-img]: https://codecov.io/github/JuliaStats/MixedModels.jl/badge.svg?branch=master
 [codecov-url]: https://codecov.io/github/JuliaStats/MixedModels.jl?branch=master
 
+[T1-img]: https://github.com/JuliaStats/MixedModels.jl/workflows/Tier1/badge.svg
+[T1-url]: https://github.com/JuliaStats/MixedModels.jl/actions?workflow=Tier1
+
+[T2-img]: https://github.com/JuliaStats/MixedModels.jl/workflows/Tier2/badge.svg
+[T2-url]: https://github.com/JuliaStats/MixedModels.jl/actions?workflow=Tier2
+
 [pkgeval-img]: https://juliaci.github.io/NanosoldierReports/pkgeval_badges/M/MixedModels.svg
 [pkgeval-url]: https://juliaci.github.io/NanosoldierReports/pkgeval_badges/report.html
 
@@ -45,6 +45,21 @@ In a `LinearMixedModel` the conditional mean, 𝛍 = 𝔼[𝓨|𝓑 = 𝐛], is
 In a `GeneralizedLinearMixedModel`, the conditional mean, 𝔼[𝓨|𝓑 = 𝐛], is related to the linear predictor via a _link function_.
 Typical distribution forms are _Bernoulli_ for binary data or _Poisson_ for count data.
 
+## Currently Supported Platforms
+
+|OS|OS Version|Arch|Julia|Tier|
+|:-:|:-:|:-:|:-:|:-:|
+|Linux|Ubuntu 18.04|x64|v1.3|1|
+|macOS|Catalina 10.15|x64|v1.3|1|
+|Windows|Server 2019|x64|v1.3|1|
+|Linux|Ubuntu 18.04|x86|v1.3|2|
+|Windows|Server 2019|x84|v1.3|2|
+|Linux|Ubuntu 18.04|x64|v1.4|2|
+|macOS|Catalina 10.15|x64|v1.4|2|
+|Windows|Server 2019|x64|v1.4|2|
+|Linux|Ubuntu 18.04|x86|v1.4|2|
+|Windows|Server 2019|x84|v1.4|2|
+
 ## Version 2.0.0
 
 Version 2.0.0 contains some user-visible changes and many changes in the underlying code.

README.md

@@ -10,4 +10,11 @@ StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 Weave = "44d3d7a6-8a23-5bf8-98c5-b353f8df5ec9"
 
 [compat]
-Documenter = "~0.20"
+DataFrames = "0.20.0"
+Documenter = "0.24"
+FreqTables = "0.3"
+Gadfly = "1"
+PooledArrays = "0.5"
+RData = "0.7"
+StatsBase = "0.32"
+Weave = "0.9"

appveyor.yml

@@ -30,7 +30,7 @@ More formally, a `k`th order rule is exact when `f` is a `k-1` order polynomial.
 In the [*Golub-Welsch algorithm*](https://en.wikipedia.org/wiki/Gaussian_quadrature#The_Golub-Welsch_algorithm) the abscissae for a particular Gaussian quadrature rule are determined as the eigenvalues of a symmetric tri-diagonal matrix and the weights are derived from the squares of the first row of the matrix of eigenvectors.
 For a `k`th order normalized Gauss-Hermite rule the tridiagonal matrix has zeros on the diagonal and the square roots of `1:k-1` on the super- and sub-diagonal, e.g.
 ```{julia;term=true}
-using LinearAlgebra, Gadfly
+using DataFrames, LinearAlgebra, Gadfly
 sym3 = SymTridiagonal(zeros(3), sqrt.(1:2))
 ev = eigen(sym3);
 show(ev.values)
@@ -91,24 +91,20 @@ Several covariates were recorded including the woman's age (centered at the mean
 The version of the data used here is that used in review of multilevel modeling software conducted by the Center for Multilevel Modelling, currently at University of Bristol (http://www.bristol.ac.uk/cmm/learning/mmsoftware/data-rev.html).
 These data are available as the `Contraception` data frame in the test data for the `MixedModels` package.
 ```{julia;term=true}
-using DataFrames, PooledArrays, RData
-const dat = Dict(Symbol(k)=>v for (k,v) in
-    load(joinpath(dirname(pathof(MixedModels)), "..", "test", "dat.rda")));
-contra = dat[:Contraception];
-contra.urbdist = PooledArray(string.(contra.urb, contra.d));
+contra = MixedModels.dataset(:contra)
 describe(contra)
 ```
 
 A smoothed scatterplot of contraception use versus age
 ```{julia;echo=false;fig_cap="Scatterplot smooth of contraception use versus age"}
-plot(contra, x=:a, y=:use, Geom.smooth, Guide.xlabel("Centered age (yr)"),
+plot(contra, x=:age, y=:use, Geom.smooth, Guide.xlabel("Centered age (yr)"),
     Guide.ylabel("Contraception use"))
 ```
 shows that the proportion of women using artificial contraception is approximately quadratic in age.
 
 A model with fixed-effects for age, age squared, number of live children and urban location and with random effects for district, is fit as
 ```{julia;term=true}
-const form1 = @formula use ~ 1 + a + abs2(a) + l + urb + (1|d);
+const form1 = @formula use ~ 1 + age + abs2(age) + livch + urban + (1|dist);
 m1 = fit!(GeneralizedLinearMixedModel(form1, contra,
     Bernoulli()), fast=true)
 ```
@@ -136,7 +132,7 @@ One thing to notice is that, even on the deviance scale, the contributions of di
 This is primarily due to different sample sizes in the different districts.
 ```{julia;term=true}
 using FreqTables
-freqtable(contra, :d)'
+freqtable(contra, :dist)'
 ```
 
 Because the first district has one of the largest sample sizes and the third district has the smallest sample size, these two will be used for illustration.

docs/Project.toml

@@ -25,14 +25,12 @@ For example, a simple linear mixed-effects model for the `Dyestuff` data in the
 package for [`R`](https://www.r-project.org) is fit by
 
 ```{julia;term=true}
-using DataFrames, Gadfly, MixedModels, Random, RData
-datf = joinpath(dirname(pathof(MixedModels)), "..", "test", "dat.rda");
-const dat = Dict(Symbol(k)=>v for (k,v) in load(datf));
+using DataFrames, Gadfly, MixedModels, Random
 ```
 
 ```{julia;term=true}
-ds = rename!(dat[:Dyestuff], [:Batch, :Yield])  # the Dyestuff data
-m1 = fit(MixedModel, @formula(Yield ~ 1 + (1 | Batch)), ds)
+dyestuff = MixedModels.dataset(:dyestuff)
+m1 = fit(MixedModel, @formula(yield ~ 1 + (1 | batch)), dyestuff)
 ```
 
 To bootstrap the model parameters, first initialize a random number generator
@@ -88,10 +86,15 @@ However, it is not as straightforward to detect singularity in vector-valued ran
 
 For example, if we bootstrap a model fit to the `sleepstudy` data
 ```{julia;term=true}
-m2 = fit(MixedModel, @formula(Y ~ 1+U+(1+U|G)), dat[:sleepstudy])
+sleepstudy = MixedModels.dataset(:sleepstudy)
+m2 = fit(
+    MixedModel,
+    @formula(reaction ~ 1+days+(1+days|subj)), 
+    sleepstudy,
+)
 ```
 ```{julia;term=true}
-samp2 = parametricbootstrap(rng, 10_000, m2);
+samp2 = parametricbootstrap(rng, 10_000, m2, use_threads=true);
 ```
 the singularity can be exhibited as a standard deviation of zero or as a correlation of $\pm1$.
 The `σρs` property of the sample is a vector of named tuples