(re)moved MixedModelsExt traces (#243)

* (re)moved MixedModelsExt traces

* fix julia1.10 bug, it was non-stable rng unittest

* remove lmm docs in make.jl

* feel free to include this test, once UnfoldSim.jl is through review...

* fix ref error

Author: behinger

.github/workflows/Docs.yml

@@ -57,6 +57,7 @@ jobs:
           GKSwstype: "100" # https://discourse.julialang.org/t/generation-of-documentation-fails-qt-qpa-xcb-could-not-connect-to-display/60988
 
       - name: Update UnfoldDocs
+        if: github.ref == 'refs/heads/main'
         uses: peter-evans/repository-dispatch@v3
         with:
           token: ${{ secrets.UNFOLDDOCSPAT }}

Project.toml

@@ -1,7 +1,7 @@
 name = "Unfold"
 uuid = "181c99d8-e21b-4ff3-b70b-c233eddec679"
 authors = ["Benedikt Ehinger <[email protected]>"]
-version = "0.7.9"
+version = "0.8.0-dev"
 
 [deps]
 BSplineKit = "093aae92-e908-43d7-9660-e50ee39d5a0a"
@@ -22,7 +22,6 @@ LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 Logging = "56ddb016-857b-54e1-b83d-db4d58db5568"
 MLBase = "f0e99cf1-93fa-52ec-9ecc-5026115318e0"
 Missings = "e1d29d7a-bbdc-5cf2-9ac0-f12de2c33e28"
-MixedModels = "ff71e718-51f3-5ec2-a782-8ffcbfa3c316"
 OrderedCollections = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 PooledArrays = "2dfb63ee-cc39-5dd5-95bd-886bf059d720"
 ProgressMeter = "92933f4c-e287-5a05-a399-4b506db050ca"
@@ -44,14 +43,12 @@ TypedTables = "9d95f2ec-7b3d-5a63-8d20-e2491e220bb9"
 BSplineKit = "093aae92-e908-43d7-9660-e50ee39d5a0a"
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 Krylov = "ba0b0d4f-ebba-5204-a429-3ac8c609bfb7"
-MixedModels = "ff71e718-51f3-5ec2-a782-8ffcbfa3c316"
 RobustModels = "d6ea1423-9682-4bbd-952f-b1577cbf8c98"
 
 [extensions]
 UnfoldBSplineKitExt = "BSplineKit"
 UnfoldCUDAExt = "CUDA"
 UnfoldKrylovExt = ["Krylov", "CUDA"]
-UnfoldMixedModelsExt = "MixedModels"
 UnfoldRobustModelsExt = "RobustModels"
 
 [compat]
@@ -73,7 +70,6 @@ LinearAlgebra = "1"
 Logging = "1"
 MLBase = "0.9"
 Missings = "1"
-MixedModels = "4"
 OrderedCollections = "1"
 PooledArrays = "1"
 ProgressMeter = "1"

docs/make.jl

@@ -27,8 +27,6 @@ makedocs(
         "Tutorials" => [
             "rERP (mass univariate)" => "tutorials/lm_mu.md",
             "rERP (overlap correction)" => "tutorials/lm_overlap.md",
-            "lmmERP (mass univariate)" => "tutorials/lmm_mu.md",
-            "lmmERP (overlap correction)" => "tutorials/lmm_overlap.md",
         ],
         "HowTo" => [
             "Multiple events" => "HowTo/multiple_events.md",
@@ -37,7 +35,6 @@ makedocs(
             "Marginal effects (must read re: splines)" => "generated/HowTo/effects.md",
             "Alternative Solvers (Robust, GPU, B2B)" => "HowTo/custom_solvers.md",
             #"Time domain basis functions"=>"generated/HowTo/timesplines.md",
-            "P-values for mixedModels" => "HowTo/lmm_pvalues.md",
             "Save and load Unfold models" => "generated/HowTo/unfold_io.md",
             "Duration-scaled basisfunctions (Hassall-style)" => "generated/HowTo/FIRduration.md",
             "🐍 Import EEG with PyMNE.jl" => "HowTo/pymne.md",

docs/src/HowTo/lmm_pvalues.md

@@ -1,6 +1,6 @@
 # Unfold Documentation
 
-If you want to follow the **tutorials**, best to start with the [mass-univariate approach](@ref lm_massunivariate), which should be familiar to you if you did ERPs before. Then the [overlap-correction tutorial](@ref lm_overlap), [mixed mass univariate](@ref lmm_massunivariate), [mixed overlap (tricky!)](@ref lmm_overlap). If you are then not satisfied, check out more advanced topics: [effects-interface (aka what to do after fitting)](@ref effects), or non-linear effects.
+If you want to follow the **tutorials**, best to start with the [mass-univariate approach](@ref lm_massunivariate), which should be familiar to you if you did ERPs before. Then the [overlap-correction tutorial](@ref lm_overlap). If you are then not satisfied, check out more advanced topics: [effects-interface (aka what to do after fitting)](@ref effects), or non-linear effects.
 
 In case you want to understand the tools better, check out our **explanations**.
 
@@ -17,8 +17,32 @@ There are four main model types
 1. Timeexpansion **No**, Mixed **Yes** : `fit(UnfoldModel, [Any=>(fLMM, -0.1:0.01:0.5)], evts, data_epoch)`
 1. Timeexpansion **Yes**, Mixed **Yes**: `fit(UnfoldModel, [Any=>(fLMM, basisfunction)], evts, data)`
 
+## rERP model
+
 ```julia
+using Unfold
+using UnfoldSim
+data, evts = UnfoldSim.predef_eeg()
+
 f = @formula 0 ~ 1 + condition
+basisfunction = firbasis(τ = (-0.1,0.5), sfreq = 100)
+fit(UnfoldModel, [Any=>(f, basisfunction)], evts, data)
+nothing #hide
+```
+
+## MixedModels
+
+It is also possible to fit Linear Mixed Models using the sister-package `UnfoldMixedModels.jl`
+
+```julia
+using UnfoldMixedModels
+using UnfoldSim
+data, evts = UnfoldSim.predef_eeg(10;return_epoched=true) # 10 subjects
+data = reshape(data,size(data,1),:) # concatenate subjects
+
+times = range(-0.1,0.5,size(data,1)) # arbitrary time-vector
+
 fLMM = @formula 0 ~ 1 + condition + (1|subject) + (1|item)
-basisfunction = firbasis(τ = (-0.1,0.5), sfreq = 100))
+fit(UnfoldModel, [Any=>(f, times)], evts, data)
+nothing #hide
 ```

docs/src/index.md

@@ -3,15 +3,6 @@
 In  Julia 1.9 Package Extensions were introduced. Unfold.jl is making use of them in four ways.
 Prior to using some functionality, you have to add + load specific package(s) for the functionality to be available. The reason for this is, that if you don't need e.g. GPU-support, you also will not need to install it.
 
-## MixedModels
-
-To use formulas like `@formula(0~1+condition+(1+condition|subject))` you have to load MixedModels. e.g.
-
-```julia
-using MixedModels
-using Unfold
-```
-
 ## GPU: Krylov,CUDA
 
 To use gpu support as described in @Ref(custom_solvers) you have to: