documentation; small fix; addleaf! returns leaf; ready for v0.16.4 (JuliaPhylo#212)

cecileane · web-flow · commit 4a2facecf11a · 2024-06-04T13:02:00.000-05:00
fix: r2 corrected in ftest
addleaf! output new leaf node, not net: breaking change, but internal function

documentation:
- workflow
- badges
- workshop link
- ftest warnings explained
- typo fix in doc for slurm
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -7,7 +7,13 @@ on:
   push:
     branches:
       - master
-    tags: '*'
+    tags: ['*']
+
+concurrency:
+  # Skip intermediate builds: always.
+  # Cancel intermediate builds: only if it is a pull request build.
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ startsWith(github.ref, 'refs/pull/') }}
 
 jobs:
   test:
@@ -17,53 +23,24 @@ jobs:
       fail-fast: false
       matrix:
         version:
-          - '1' # will expand to the latest stable 1.x release of Julia.
+          - '1'
         os:
           - ubuntu-latest
           - macos-latest
           - windows-latest
         arch:
           - x64
     steps:
-      - uses: actions/checkout@v2
-      - uses: julia-actions/setup-julia@v1
+      - uses: actions/checkout@v4
+      - uses: julia-actions/setup-julia@latest
         with:
           version: ${{ matrix.version }}
           arch: ${{ matrix.arch }}
-      - uses: actions/cache@v1
-        env:
-          cache-name: cache-artifacts
-        with:
-          path: ~/.julia/artifacts
-          key: ${{ runner.os }}-test-${{ env.cache-name }}-${{ hashFiles('**/Project.toml') }}
-          restore-keys: |
-            ${{ runner.os }}-test-${{ env.cache-name }}-
-            ${{ runner.os }}-test-
-            ${{ runner.os }}-
+      - uses: julia-actions/cache@v2
       - uses: julia-actions/julia-buildpkg@v1
       - uses: julia-actions/julia-runtest@v1
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v2
-        with:
-          file: lcov.info
-  docs:
-    name: Documentation
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v2
-      - uses: r-lib/actions/setup-r@v2
-      - run: echo "LD_LIBRARY_PATH=$(R RHOME)/lib" >> $GITHUB_ENV
-      - uses: julia-actions/setup-julia@v1
+      - uses: codecov/codecov-action@v4
         with:
-          version: '1'
-      - name: install dependencies
-        run: |
-          julia --project=docs -e '
-            using Pkg
-            Pkg.develop(PackageSpec(path=pwd()))
-            Pkg.instantiate()'
-      - name: make docs
-        run: julia --project=docs docs/make.jl
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          DOCUMENTER_KEY: ${{ secrets.DOCUMENTER_KEY }}
+          files: lcov.info
+          token: ${{ secrets.CODECOV_TOKEN }}
diff --git a/.github/workflows/doccleanup.yml b/.github/workflows/doccleanup.yml
@@ -1,28 +1,33 @@
-name: Doc Preview Cleanup
+name: doc preview cleanup
 
 on:
   pull_request:
     types: [closed]
 
+# ensure that only one "doc preview cleanup" workflow is force-pushing at a time
+concurrency:
+  group: doc-preview-cleanup
+  cancel-in-progress: false
+
 jobs:
   doc-preview-cleanup:
     runs-on: ubuntu-latest
+    permissions:
+      contents: write
     steps:
-      - name: Checkout gh-pages branch
-        uses: actions/checkout@v2
+      - name: checkout gh-pages branch
+        uses: actions/checkout@v4
         with:
           ref: gh-pages
-
-      - name: Delete preview and history
+      - name: delete preview and history + push changes
         run: |
-            git config user.name "Documenter.jl"
-            git config user.email "documenter@juliadocs.github.io"
-            git rm -rf "previews/PR$PRNUM"
-            git commit -m "delete preview"
-            git branch gh-pages-new $(echo "delete history" | git commit-tree HEAD^{tree})
+          if [ -d "${preview_dir}" ]; then
+              git config user.name "Documenter.jl"
+              git config user.email "documenter@juliadocs.github.io"
+              git rm -rf "${preview_dir}"
+              git commit -m "delete preview"
+              git branch gh-pages-new $(echo "delete history" | git commit-tree HEAD^{tree})
+              git push --force origin gh-pages-new:gh-pages
+          fi
         env:
-            PRNUM: ${{ github.event.number }}
-
-      - name: Push changes
-        run: |
-            git push --force origin gh-pages-new:gh-pages
+          preview_dir: previews/PR${{ github.event.number }}
diff --git a/.github/workflows/documentation.yml b/.github/workflows/documentation.yml
@@ -0,0 +1,30 @@
+name: Documenter
+
+on:
+  push:
+    branches:
+      - main
+    tags: '*'
+  pull_request:
+
+jobs:
+  Documenter:
+    name: documentation
+    permissions:
+      contents: write
+      pull-requests: read
+      statuses: write
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: r-lib/actions/setup-r@v2
+      - run: echo "LD_LIBRARY_PATH=$(R RHOME)/lib:$LD_LIBRARY_PATH" >> $GITHUB_ENV
+      - uses: julia-actions/setup-julia@v2
+        with:
+          version: '1'
+      - uses: julia-actions/cache@v2
+      - uses: julia-actions/julia-buildpkg@latest
+      - uses: julia-actions/julia-docdeploy@latest
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          DOCUMENTER_KEY: ${{ secrets.DOCUMENTER_KEY }}
diff --git a/Project.toml b/Project.toml
@@ -1,7 +1,7 @@
 name = "PhyloNetworks"
 uuid = "33ad39ac-ed31-50eb-9b15-43d0656eaa72"
 license = "MIT"
-version = "0.16.3"
+version = "0.16.4"
 
 [deps]
 BioSequences = "7e6ae17a-c86d-528c-b3b9-7f778a29fe59"
diff --git a/README.md b/README.md
@@ -1,9 +1,12 @@
 # PhyloNetworks: analysis for phylogenetic networks <img src="docs/src/logo_text.png" align=right>
 
+[![doc stable](https://img.shields.io/badge/docs-stable-blue.svg)](https://crsl4.github.io/PhyloNetworks.jl/stable)
+[![doc dev](https://img.shields.io/badge/docs-dev-blue.svg)](https://crsl4.github.io/PhyloNetworks.jl/dev)
 [![Build status](https://github.com/crsl4/PhyloNetworks.jl/workflows/CI/badge.svg?branch=master)](https://github.com/crsl4/PhyloNetworks.jl/actions/workflows/ci.yml)
-[![](https://img.shields.io/badge/docs-stable-blue.svg)](https://crsl4.github.io/PhyloNetworks.jl/stable)
-[![](https://img.shields.io/badge/docs-dev-blue.svg)](https://crsl4.github.io/PhyloNetworks.jl/dev)
-[![codecov](https://codecov.io/gh/crsl4/PhyloNetworks.jl/branch/master/graph/badge.svg)](https://codecov.io/gh/crsl4/PhyloNetworks.jl)
+[![coverage](https://codecov.io/gh/crsl4/PhyloNetworks.jl/branch/master/graph/badge.svg)](https://codecov.io/gh/crsl4/PhyloNetworks.jl)
+[![Code Style: Blue](https://img.shields.io/badge/code%20style-blue-4495d1.svg)](https://github.com/invenia/BlueStyle)
+[![ColPrac: Contributor's Guide on Collaborative Practices for Community Packages](https://img.shields.io/badge/ColPrac-Contributor's%20Guide-blueviolet)](https://github.com/SciML/ColPrac)
+[![PkgEval](https://JuliaCI.github.io/NanosoldierReports/pkgeval_badges/P/PhyloNetworks.svg)](https://JuliaCI.github.io/NanosoldierReports/pkgeval_badges/report.html)
 
 ## Overview
 
@@ -18,7 +21,7 @@ PhyloNetworks is a [Julia](http://julialang.org) package with utilities to:
   (Robinson-Foulds distance on trees)
 - summarize samples of bootstrap networks (or trees)
   with edge and node support
-- estimate species networks from multilocus data (see below)
+- estimate species networks from multilocus data: SNaQ
 - phylogenetic comparative methods for continuous trait evolution
   on species networks / trees
 - plot networks (and trees), via the companion package
@@ -28,7 +31,9 @@ To get help, check
 
 - the [latest documentation](https://crsl4.github.io/PhyloNetworks.jl/dev)
 - the [wiki](https://github.com/crsl4/PhyloNetworks.jl/wiki) for a step-by-step tutorial
-  (July 2018) with background on networks
+  with background on networks (last revised 2022)
+- [tutorial](https://cecileane.github.io/networkPCM-workshop/) for
+  comparative methods, including network calibration (2023 workshop)
 - the [google group](https://groups.google.com/forum/#!forum/phylonetworks-users)
   for common questions. Join the group to post/email your questions,
   or to receive information on new versions, bugs fixed, etc.
diff --git a/docs/make.jl b/docs/make.jl
@@ -13,7 +13,11 @@ makedocs(
     modules = [PhyloNetworks], # to list methods from PhyloNetworks only, not from Base etc.
     format = Documenter.HTML(
         prettyurls = get(ENV, "CI", nothing) == "true", # easier local build
-        size_threshold = 600 * 2^10, size_threshold_warn = 500 * 2^10), # 600 KiB
+        size_threshold = 600 * 2^10,
+        size_threshold_warn = 500 * 2^10, # 600 KiB
+        canonical="https://crsl4.github.io/PhyloNetworks.jl/stable/",
+        edit_link="master",
+    ),
     # exception, so warning-only for :missing_docs. List all others:
     warnonly = Documenter.except(:autodocs_block, :cross_references, :docs_block,
         :doctest, :eval_block, :example_block, :footnote, :linkcheck_remotes,
@@ -46,4 +50,5 @@ makedocs(
 deploydocs(
     repo = "github.com/crsl4/PhyloNetworks.jl.git",
     push_preview = true,
+    devbranch = "master",
 )
diff --git a/docs/readme.md b/docs/readme.md
@@ -1,13 +1,13 @@
 # notes to maintain documentation
 
-- built with [Documenter](https://juliadocs.github.io/Documenter.jl).
+- built with [Documenter](https://documenter.juliadocs.org/stable/).
 - deployed [here](https://crsl4.github.io/PhyloNetworks.jl/)
   (go to `dev/` or `stable/`)
   using GitHub and files committed to the `gh-pages` branch.
 
 ## how it works: overview
 
-- `.github/workflows/ci.yml` asks to start the doc project
+- `.github/workflows/documentation.yml` asks to start the doc project
   (installs R and dependencies like `PhyloPlots` & `Documenter`) and
   run `./docs/make.jl` after a successful test & build.
 - the julia script `docs/make.jl` has these steps:
diff --git a/docs/src/index.md b/docs/src/index.md
@@ -10,8 +10,10 @@ and their use for trait evolution.
 **How to get help**
 
 - the package [wiki](https://github.com/crsl4/PhyloNetworks.jl/wiki) has a step-by-step
-  tutorial, done for the 2019 MBL workshop, with background on networks and
+  tutorial, done for the MBL workshop (last revised 2022), with background on networks and
   explanations.
+- [tutorial](https://cecileane.github.io/networkPCM-workshop/) for
+  comparative methods, including network calibration (2023 workshop)
 - the [google group](https://groups.google.com/forum/#!forum/phylonetworks-users)
   has answers to common questions.
 - the [Manual](@ref) below has a quick tutorial (navigation on the left).
diff --git a/docs/src/man/netmanipulation.md b/docs/src/man/netmanipulation.md
@@ -36,9 +36,10 @@ To modify a network, for example:
   [`deleteaboveLSA!`](@ref) (the "least stable ancestor" may be different from the root)
 - [`deleteHybridThreshold!`](@ref) to simplify a network by deleting edges with small γ's
 - [`removedegree2nodes!`](@ref), [`shrink3cycles!`](@ref), [`shrink2cycles!`](@ref),
-  [`PhyloNetworks.shrinkedge!`](@ref)]
+  [`PhyloNetworks.shrinkedge!`](@ref)
 - [`PhyloNetworks.addleaf!`](@ref),
-  [`PhyloNetworks.deletehybridedge!`](@ref), [`PhyloNetworks.addhybridedge!`](@ref)
+  [`PhyloNetworks.deletehybridedge!`](@ref),
+  [`PhyloNetworks.addhybridedge!`](@ref)
 - [`nni!`](@ref) (nearest neighbor interchange),
   [`PhyloNetworks.fliphybrid!`](@ref) to flip the direction of a hybrid edge
 - [`PhyloNetworks.unzip_canonical!`](@ref) to "unzip" (or zip down) all
diff --git a/docs/src/man/snaq_plot.md b/docs/src/man/snaq_plot.md
@@ -261,7 +261,7 @@ Then log out of the cluster and go for coffee.
 echo "slurm task ID = $SLURM_ARRAY_TASK_ID used as hmax"
 echo "start of SNaQ parallel runs on $(hostname)"
 # finally: launch the julia script, using Julia executable appropriate for slurm, with full paths:
-/workspace/software/bin/julia --history-file=no -- runSNaQ.jl $SLURM_ARRAY_TASK_ID 30 > net$SLURM_ARRAY_TASK_ID_30runs.screenlog 2>&1
+/workspace/software/bin/julia --history-file=no -- runSNaQ.jl $SLURM_ARRAY_TASK_ID 30 > net${SLURM_ARRAY_TASK_ID}_30runs.screenlog 2>&1
 echo "end of SNaQ run ..."
 ```
 
diff --git a/docs/src/man/trait_tree.md b/docs/src/man/trait_tree.md
@@ -508,10 +508,24 @@ complex cases, it is possible to do a Fisher F test, thanks to the `GLM`
 function `ftest`.
 ```@example tree_trait
 fit_null = phylolm(@formula(trait ~ 1), dat, truenet) # fit against the null (no shift)
-ftest(fit_sh, fit_null)  # nested models
+ftest(fit_null, fit_sh)  # nested models
 ```
 Here, this test is equivalent to the Fisher F test, and gives the same p-value.
 
+!!! note "Warnings from GLM"
+    A warning may appear, saying
+    "Starting from GLM.jl 1.8, null model is defined as having no predictor at all when a model without an intercept is passed."
+    - Why? `ftest` is inherited from the GLM package, which does not know that
+      the intercept term is not a column of ones after transformation to remove
+      the phylogenetic correlation. This is why `ftest` sends a warning for
+      each model, when multiple models are compared.
+    - These specific warnings can be ignored:
+      * F values and p-values are correct
+      * R² values are also correct: they are obtained with the
+        `r2` function for phylogenetic linear models.
+    A future version of the package will attempt to remove these warnings
+    specifically.
+
 Note that models need to be ordered by complexity, when given to `ftest`:
 either from most complex to most simple, or from most simple to most complex.
 In the output table, models are listed in the order in which they were given.
diff --git a/src/manipulateNet.jl b/src/manipulateNet.jl
@@ -567,6 +567,8 @@ Add a new external edge between `node` or between the "middle" of `edge`
 and a newly-created leaf, of name `leafname`.
 By default, the new edge length is missing (-1).
 
+output: newly created leaf node.
+
 # examples
 
 ```jldoctest
@@ -616,13 +618,12 @@ function addleaf!(net::HybridNetwork, speciesnode::Node, leafname::String, edgel
         net.numTaxa -= 1
     end
     pushNode!(net, newleaf) # push node into network (see auxillary.jl)
-    return net
+    return newleaf
 end
 
 function addleaf!(net::HybridNetwork, startingedge::Edge, leafname::String, edgelength::Float64=-1.0)
-    newnode, newedge = breakedge!(startingedge, net)
-    addleaf!(net, newnode, leafname, edgelength)
-    return net
+    newnode, _ = breakedge!(startingedge, net)
+    return addleaf!(net, newnode, leafname, edgelength)
 end
 
 
diff --git a/src/traits.jl b/src/traits.jl
@@ -3019,14 +3019,28 @@ isnested(::Union{PagelLambda,ScalingHybrid}, ::BM) = false
 isnested(::ScalingHybrid,::PagelLambda) = false
 isnested(::PagelLambda,::ScalingHybrid) = false
 
-## ANOVA using ftest from GLM - need version 0.8.1
+#= ANOVA using ftest from GLM - need version 0.8.1
+ As of GLM v1.8, ftest throws a warning on typical BM models, one per model:
+ "Starting from GLM.jl 1.8, null model is defined as having no predictor at all when a model without an intercept is passed."
+ This is because after transforming the data to de-correlate the residuals,
+ the transformed intercept vector is not proportional to the constant vector 1.
+ The warning is from: ftest → r2(phylomodel.lm) → nulldeviance(phylomodel.lm) → warning.
+ R² by GLM is wrong: assume *no* intercept, and are based on the transformed data.
+ R² corrected them below: r2(phylomodel) reimplemented here.
+ But nulldeviance(phylomodel) does *not* call nulldeviance(phylomodel.lm),
+ instead re-implemented here to use the intercept properly.
+ Keep the warnings: unless they can be suppressed with specificity
+ Ideally: modify `ftest` here or in GLM.
+=#
 function GLM.ftest(objs::PhyloNetworkLinearModel...)
     if !all( isa(o.evomodel,BM) && isnothing(o.model_within) for o in objs)
         throw(ArgumentError("""F test is only valid for the vanilla BM model.
         Use a likelihood ratio test instead with function `lrtest`."""))
     end
     objslm = [obj.lm for obj in objs]
-    return ftest(objslm...)
+    resGLM = ftest(objslm...)
+    resGLM.r2 = r2.(objs)
+    return resGLM
 end
 ## ANOVA: old version - kept for tests purposes - do not export
 """
diff --git a/test/test_lm.jl b/test/test_lm.jl
@@ -210,19 +210,21 @@ modnull = phylolm(@formula(trait ~ 1), dfr, net)
 @test sigma2_phylo(modnull) ≈ 0.6517876326943942 atol=1e-6 # using REML
 modhom = phylolm(@formula(trait ~ sum), dfr, net)
 modhet = phylolm(@formula(trait ~ sum + shift_8), dfr, net)
-table1 = ftest(modhet, modhom, modnull)
+#= 3 warnings thrown by ftest, one for each model, because after transforming the
+   data to de-correlate the results, the intercept vector is not ∝ 1.
+   Keep the warnings: because incorrect R² values in the ftest output
+=#
+table1 = redirect_stdio(stderr=devnull) do # to avoid seeing the warnings
+    ftest(modhet, modhom, modnull)
+end
 table2 = PhyloNetworks.anova(modnull, modhom, modhet)
 
 @test table1.fstat[2] ≈ table2[2,:F]
 @test table1.fstat[3] ≈ table2[1,:F]
 @test table1.pval[2] ≈ table2[2,Symbol("Pr(>F)")]
 @test table1.pval[3] ≈ table2[1,Symbol("Pr(>F)")]
 @test hasintercept(modnull) && hasintercept(modhom) && hasintercept(modhet)
-# ## Replace next 4 lines with previous ones when GLM.ftest available
-# @test table1[:F][2] ≈ table2[:F][2]
-# @test table1[:F][1] ≈ table2[:F][1]
-# @test table1[Symbol("Pr(>F)")][1] ≈ table2[Symbol("Pr(>F)")][1]
-# @test table1[Symbol("Pr(>F)")][2] ≈ table2[Symbol("Pr(>F)")][2]
+@test all(isapprox.(table1.r2, (0.8398130376214782, 0.006032952123011026, 0), atol=1e-15))
 
 # Check that it is the same as doing shift_8 + shift_17
 modhetbis = phylolm(@formula(trait ~ shift_8 + shift_17), dfr, net)
