Breaking: refactor

.github/workflows/CI.yml

@@ -25,12 +25,12 @@ jobs:
         arch:
           - x64
     steps:
-      - uses: actions/checkout@v3
-      - uses: julia-actions/setup-julia@v1
+      - uses: actions/checkout@v4
+      - uses: julia-actions/setup-julia@v2
         with:
           version: ${{ matrix.version }}
           arch: ${{ matrix.arch }}
-      - uses: julia-actions/cache@v1
+      - uses: julia-actions/cache@v2
       - uses: julia-actions/julia-buildpkg@v1
       - uses: julia-actions/julia-runtest@v1
       - uses: julia-actions/julia-processcoverage@v1
@@ -44,8 +44,8 @@ jobs:
       contents: write
       statuses: write
     steps:
-      - uses: actions/checkout@v3
-      - uses: julia-actions/setup-julia@v1
+      - uses: actions/checkout@v4
+      - uses: julia-actions/setup-julia@v2
         with:
           version: '1'
       - name: Configure doc environment
@@ -64,3 +64,9 @@ jobs:
             using UnfoldStats
             DocMeta.setdocmeta!(UnfoldStats, :DocTestSetup, :(using UnfoldStats); recursive=true)
             doctest(UnfoldStats)'
+      - name: Update UnfoldDocs
+        uses: peter-evans/repository-dispatch@v3
+        with:
+          token: ${{ secrets.UNFOLDDOCSPAT }}
+          repository: unfoldtoolbox/UnfoldDocs
+          event-type: individual-docs-updated

.gitignore

@@ -6,3 +6,4 @@
 /docs/Manifest.toml
 /docs/build/
 /docs/src/generated
+/docs/dev

Project.toml

@@ -1,15 +1,37 @@
 name = "UnfoldStats"
 uuid = "96fd419a-8306-4ce8-ba5b-cd907cb7647c"
 authors = ["Judith Schepers", "Benedikt V. Ehinger"]
-version = "1.0.0-DEV"
+version = "0.2.0"
 
 [deps]
 BSplineKit = "093aae92-e908-43d7-9660-e50ee39d5a0a"
+ClusterDepth = "c8d8bbfa-f476-4995-adff-2987f04015d1"
+Logging = "56ddb016-857b-54e1-b83d-db4d58db5568"
+MixedModels = "ff71e718-51f3-5ec2-a782-8ffcbfa3c316"
+MixedModelsPermutations = "647c4018-d7ef-4d03-a0cc-8889a722319e"
+Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+StatsAPI = "82ae8749-77ed-4fe6-ae5f-f523153014b0"
 StatsModels = "3eaba693-59b7-5ba5-a881-562e759f1c8d"
 Unfold = "181c99d8-e21b-4ff3-b70b-c233eddec679"
 
+[weakdeps]
+ClusterDepth = "c8d8bbfa-f476-4995-adff-2987f04015d1"
+MixedModelsPermutations = "647c4018-d7ef-4d03-a0cc-8889a722319e"
+UnfoldMixedModels = "019ae9e0-8363-565c-86e5-97a5a2fe84f4"
+
+[extensions]
+UnfoldStatsMixedModelsExt = ["UnfoldMixedModels"]
+UnfoldStatsMixedModelsPermutationsExt = ["UnfoldMixedModels", "ClusterDepth", "MixedModelsPermutations"]
+
 [compat]
-julia = "1"
+BSplineKit = "0.16,0.17"
+ClusterDepth = "0.2"
+MixedModelsPermutations = "0.2"
+StatsAPI = "1"
+StatsModels = "0.7"
+Unfold = "0.8"
+UnfoldMixedModels = "0.1"
+julia = "1.9,1.10,1.11"
 
 [extras]
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"

benchmark/sim_and_fit.jl

@@ -1,5 +1,5 @@
 using UnfoldSim
-using Unfold
+using Unfold, UnfoldMixedModels
 using StableRNGs
 using DataFrames
 
@@ -51,7 +51,7 @@ function sim_and_fit(
     data, events = simulate_data(sim_type, simulation, return_epoched, seed)
 
     # Create event dict containing basis function(s)/times and formula(s) for all events
-    event_dict = create_event_dict(sim_type, model_type, simulation)
+    event_dict = create_event_vector(sim_type, model_type, simulation)
 
     # Fit an Unfold model for each subject
     subject_list = unique(events.subject)
@@ -81,30 +81,32 @@ simulate_data(sim_type::T, simulation, return_epoched, seed) where {T} =
 function simulate_data(::SingleEventType, simulation, return_epoched, seed)
 
     # Simulate data
-    data, events = simulate(StableRNG(seed), simulation; return_epoched = return_epoched)
+    data, events =
+        UnfoldSim.simulate(StableRNG(seed), simulation; return_epoched = return_epoched)
 
     return data, events
 end
 
 function simulate_data(::MultipleEventTypes, simulation, return_epoched, seed)
 
     # Simulate data
-    data, events = simulate(StableRNG(seed), simulation; return_epoched = return_epoched)
+    data, events =
+        UnfoldSim.simulate(StableRNG(seed), simulation; return_epoched = return_epoched)
 
     # Add an event column to the events df and assign each event to an event type
     events[!, :event] = repeat(["stim", "fix"], size(events, 1) ÷ 2)
 
     return data, events
 end
 
-create_event_dict(sim_type::T, model_type, simulation) where {T} = create_event_dict(
+create_event_vector(sim_type::T, model_type, simulation) where {T} = create_event_vector(
     EventStyle(T),
     PredictorStyle(T),
     model_type::Type{<:UnfoldModel},
     simulation,
 )
 
-function create_event_dict(
+function create_event_vector(
     ::MultipleEventTypes,
     ::ManyPredictors,
     model_type::Type{<:UnfoldModel},
@@ -117,8 +119,8 @@ function create_event_dict(
         t_stim = times
         t_fix = times
     else # UnfoldLinearModelContinuousTime
-        t_stim = firbasis(τ = (-0.1, 1.5), sfreq = 100, name = "stimulus")
-        t_fix = firbasis(τ = (-0.1, 1), sfreq = 100, name = "fixation")
+        t_stim = firbasis(τ = (-0.1, 1.5), sfreq = 100, name = "stim")
+        t_fix = firbasis(τ = (-0.1, 1), sfreq = 100, name = "fix")
     end
 
     # Define formula(s)
@@ -131,7 +133,7 @@ function create_event_dict(
     return event_dict
 end
 
-function create_event_dict(
+function create_event_vector(
     ::MultipleEventTypes,
     ::OnlySplines,
     model_type::Type{<:UnfoldModel},
@@ -144,21 +146,21 @@ function create_event_dict(
         t_stim = times
         t_fix = times
     else # UnfoldLinearModelContinuousTime
-        t_stim = firbasis(τ = (-0.1, 1.5), sfreq = 100, name = "stimulus")
-        t_fix = firbasis(τ = (-0.1, 1), sfreq = 100, name = "fixation")
+        t_stim = firbasis(τ = (-0.1, 1.5), sfreq = 100, name = "stim")
+        t_fix = firbasis(τ = (-0.1, 1), sfreq = 100, name = "fix")
     end
 
     # Define formula(s)
     f_stim = @formula 0 ~ 1
     f_fix = @formula 0 ~ 1 + spl(continuous, 4)
 
     # Combine basis function(s)/times and formula(s) with the corresponding event
-    event_dict = Dict("stim" => (f_stim, t_stim), "fix" => (f_fix, t_fix))
+    event_vec = ["stim" => (f_stim, t_stim), "fix" => (f_fix, t_fix)]
 
-    return event_dict
+    return event_vec
 end
 
-function create_event_dict(
+function create_event_vector(
     ::SingleEventType,
     ::ManyPredictors,
     model_type::Type{<:UnfoldModel},
@@ -176,12 +178,12 @@ function create_event_dict(
     f = @formula 0 ~ 1 + spl(continuous, 4) + continuous + condition * pet
 
     # Combine basis function(s)/times and formula(s) with the corresponding event
-    event_dict = Dict(Any => (f, t))
+    event_vec = [Any => (f, t)]
 
-    return event_dict
+    return event_vec
 end
 
-function create_event_dict(
+function create_event_vector(
     ::SingleEventType,
     ::OnlySplines,
     model_type::Type{<:UnfoldModel},
@@ -199,9 +201,9 @@ function create_event_dict(
     f = @formula 0 ~ 1 + spl(continuous, 4)
 
     # Combine basis function(s)/times and formula(s) with the corresponding event
-    event_dict = Dict(Any => (f, t))
+    event_vec = [Any => (f, t)]
 
-    return event_dict
+    return event_vec
 end
 
 get_conditions(sim_type::T) where {T} = get_conditions(PredictorStyle(T))

docs/Project.toml

@@ -1,15 +1,20 @@
 [deps]
 CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"
 Chain = "8be319e6-bccf-4806-a6f7-6fae938471bc"
+ClusterDepth = "c8d8bbfa-f476-4995-adff-2987f04015d1"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
+Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
 Glob = "c27321d9-0574-5035-807b-f59d2c89b15c"
 HypothesisTests = "09f84164-cd44-5f33-b23f-e6b0d136a0d5"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
 LiveServer = "16fef848-5104-11e9-1b77-fb7a48bbb589"
+MixedModelsPermutations = "647c4018-d7ef-4d03-a0cc-8889a722319e"
 StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
+StatsModels = "3eaba693-59b7-5ba5-a881-562e759f1c8d"
 Unfold = "181c99d8-e21b-4ff3-b70b-c233eddec679"
 UnfoldMakie = "69a5ce3b-64fb-4f22-ae69-36dd4416af2a"
+UnfoldMixedModels = "019ae9e0-8363-565c-86e5-97a5a2fe84f4"
 UnfoldSim = "ed8ae6d2-84d3-44c6-ab46-0baf21700804"
 UnfoldStats = "96fd419a-8306-4ce8-ba5b-cd907cb7647c"

docs/literate/tutorial/lmm_clusterdepth.jl

@@ -0,0 +1,82 @@
+# !!! important
+#       This functionality is not tested. While it returns reasonable results, we haven't written any unit-tests, nor tested the type-1 error probability yet
+# get some data
+using UnfoldSim
+using UnfoldMixedModels
+
+using MixedModelsPermutations, ClusterDepth # both necessary to activate correct extension!
+using UnfoldStats
+using StatsModels
+using Random
+
+srate = 25
+design = MultiSubjectDesign(;
+    n_subjects = 30,
+    n_items = 40,
+    items_between = Dict(:stimtype => ["car", "face"]),
+)
+#both_within = Dict(:condition=>["scrambled","intact"]))
+contrasts = Dict(:stimtype => DummyCoding())
+p1 = MixedModelComponent(;
+    basis = UnfoldSim.p100(; sfreq = srate),
+    formula = @formula(dv ~ 1 + (1 | subject) + (1 | item)),
+    β = [5.0],
+    σs = Dict(:subject => [0.0], :item => [0.0]),
+    contrasts = contrasts,
+);
+
+n1 = MixedModelComponent(;
+    basis = UnfoldSim.n170(; sfreq = srate),
+    formula = @formula(dv ~ 1 + stimtype + (1 + stimtype | subject) + (1 | item)),
+    β = [1.0, 4], # n170-basis is negative
+    σs = Dict(:subject => [2.0, 0.25], :item => [0.25]),
+    contrasts = contrasts,
+);
+
+p3 = MixedModelComponent(;
+    basis = UnfoldSim.p300(; sfreq = srate),
+    formula = @formula(dv ~ 1 + (1 | subject) + (1 + stimtype | item)),
+    β = [4.0],
+    σs = Dict(:subject => [1.0], :item => [0.5, 2]),
+    contrasts = contrasts,
+);
+
+
+
+data_e, events = UnfoldSim.simulate(
+    design,
+    [p1, n1, p3],
+    UniformOnset(srate * 2, 10),
+    PinkNoise(; noiselevel = 1);
+    return_epoched = true,
+) # 18
+times = range(-0.1, 0.5, length = size(data_e, 1))
+data_e = reshape(data_e, 1, size(data_e, 1), :)
+#events.latency .+= repeat(range(0,length=size(data,2),step=size(data,1)),inner=size(events[events.subject.=="S01",:],1))
+
+
+
+# # Fit LMM 
+m = fit(
+    UnfoldModel,
+    [
+        Any => (
+            @formula(0 ~ 1 + stimtype + (1 + stimtype | item) + (1 + stimtype | subject)),
+            times,
+        ),
+    ],
+    events,
+    data_e,
+);
+
+
+# # Cluster Permute :)
+coefficient = 2
+pvalue(
+    MersenneTwister(1),
+    m,
+    data_e,
+    coefficient;
+    n_permutations = 10,
+    clusterforming_threshold = 1.8,
+)