Merge branch 'dev' into devPaso

aclai-lab · Nov 20, 2024 · 2ac77d3 · 2ac77d3
2 parents 92a782a + 9ce30fc
commit 2ac77d3
Show file tree

Hide file tree

Showing 10 changed files with 155 additions and 149 deletions.
diff --git a/.JuliaFormatter.toml b/.JuliaFormatter.toml
@@ -1 +1,2 @@
 style = "sciml"
+trailing_comma = true
diff --git a/.cirrus.yml b/.cirrus.yml
@@ -1,5 +1,5 @@
 freebsd_instance:
-  image_family: freebsd-14-0
+  image_family: freebsd-14-1
 task:
   name: FreeBSD
   artifacts_cache:

diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -6,10 +6,10 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v2
-      - name: Set up Julia 1.9.0
+      - name: Set up Julia 1.9
         uses: julia-actions/setup-julia@v1
         with:
-          version: "1.9.0"
+          version: "1.9"
       - uses: julia-actions/julia-buildpkg@v1
       - uses: julia-actions/julia-runtest@v1
       - uses: julia-actions/julia-processcoverage@v1

diff --git a/README.md b/README.md
@@ -1,15 +1,17 @@
 <div align="center"><a href="https://github.com/aclai-lab/Sole.jl"><img src="logo.png" alt="" title="This package is part of Sole.jl" width="200"></a></div>
 
 # SoleBase.jl
+
  <!-- - Base layer for [*Sole.jl*](https://github.com/aclai-lab/Sole.jl) -->
 
 [![Stable](https://img.shields.io/badge/docs-stable-blue.svg)](https://aclai-lab.github.io/SoleBase.jl/)
 [![Dev](https://img.shields.io/badge/docs-dev-blue.svg)](https://aclai-lab.github.io/SoleBase.jl/dev)
 [![Build Status](https://api.cirrus-ci.com/github/aclai-lab/SoleBase.jl.svg?branch=main)](https://cirrus-ci.com/github/aclai-lab/SoleBase.jl)
 [![Coverage](https://codecov.io/gh/aclai-lab/SoleBase.jl/branch/main/graph/badge.svg?token=LT9IYIYNFI)](https://codecov.io/gh/aclai-lab/SoleBase.jl)
+
 <!-- [![Code Style: Blue](https://img.shields.io/badge/code%20style-blue-4495d1.svg)](https://github.com/invenia/BlueStyle) -->
 
-This package lays the foundations for [*Sole.jl*](https://github.com/aclai-lab/Sole.jl), an open-source framework for *symbolic machine learning*.
+This package lays the foundations for [_Sole.jl_](https://github.com/aclai-lab/Sole.jl), an open-source framework for _symbolic machine learning_.
 
 ## About
 

diff --git a/src/SoleBase.jl b/src/SoleBase.jl
@@ -1,7 +1,6 @@
 
 module SoleBase
 
-
 export AbstractDataset
 export humansize
 export ninstances
@@ -16,11 +15,19 @@ export moving_window, movingwindow, movingaverage
 
 Abstract supertype for all datasets.
 
+# Interface
+- [`concatdatasets`](@ref)
+- [`instances`](@ref)
+- [`ninstances`](@ref)
+- [`eachinstance`](@ref)
+
+# Utility functions
+- [`slicedataset`](@ref)
+
 See also [`ninstances`](@ref).
 """
 abstract type AbstractDataset end
 
-
 _doc_slicedataset = """
     slicedataset(
         dataset::D,
@@ -46,11 +53,11 @@ provide the following method:
 
 """$(_doc_slicedataset)"""
 function slicedataset(
-    dataset::D,
-    dataset_slice::Union{Colon,Integer,AbstractVector,Tuple};
-    allow_no_instances = false,
-    return_view = false,
-    kwargs...,
+        dataset::D,
+        dataset_slice::Union{Colon, Integer, AbstractVector, Tuple};
+        allow_no_instances = false,
+        return_view = false,
+        kwargs...,
 ) where {D}
     if dataset_slice isa Colon
         return deepcopy(dataset)
@@ -60,8 +67,8 @@ function slicedataset(
             return error("Cannot slice dataset with slice of type $(eltype(dataset_slice))")
         end
         if !(allow_no_instances ||
-            (!(dataset_slice isa Union{AbstractVector{<:Integer},Tuple{<:Integer}}) ||
-                length(dataset_slice) > 0))
+             (!(dataset_slice isa Union{AbstractVector{<:Integer}, Tuple{<:Integer}}) ||
+              length(dataset_slice) > 0))
             return error("Cannot apply empty slice to dataset.")
         end
         return instances(dataset, dataset_slice, Val(return_view); kwargs...)
@@ -70,21 +77,20 @@ end
 
 function concatdatasets(datasets::D...) where {D}
     return error("`concatdatasets` method not implemented for type "
-        * string(typejoin(typeof.(datasets)...))) * "."
+                 * string(typejoin(typeof.(datasets)...))) * "."
 end
 
 """$(_doc_slicedataset)"""
 function instances(
-    dataset::D,
-    inds::AbstractVector,
-    return_view::Union{Val{true},Val{false}};
-    kwargs...
+        dataset::D,
+        inds::AbstractVector,
+        return_view::Union{Val{true}, Val{false}};
+        kwargs...,
 ) where {D}
     return error("`instances` method not implemented for type "
-        * string(typeof(dataset))) * "."
+                 * string(typeof(dataset))) * "."
 end
 
-
 # -------------------------------------------------------------
 # AbstractDataset - ninstances
 
@@ -128,5 +134,4 @@ include("movingwindow.jl")
 
 @deprecate moving_window(args...; kwargs...) movingwindow(args...; kwargs...)
 
-
 end
diff --git a/src/machine-learning-utils.jl b/src/machine-learning-utils.jl
@@ -14,21 +14,21 @@ Types for supervised machine learning labels (classification and regression).
 """$(doc_supervised_ml)"""
 const CLabel = Union{AbstractString, CategoricalValue, Integer} # TODO it is improper to consider Integer's as categorical labels. Should actually be RLabel's, and CategoricalValue's should be used, instead. However, atm some algorithms still rely on Int's labels to be intended as indices of a vector of class names...
 """$(doc_supervised_ml)"""
-const RLabel  = AbstractFloat # Real
+const RLabel = AbstractFloat # TODO: Real
 """$(doc_supervised_ml)"""
-const Label   = Union{CLabel,RLabel}
+const Label = Union{CLabel, RLabel}
 
 # Raw labels
 const _CLabel = Integer # (classification labels are internally represented as integers)
-const _Label  = Union{_CLabel,RLabel}
+const _Label = Union{_CLabel, RLabel}
 
 ############################################################################################
 
 # Convert a list of labels to categorical form
 Base.@propagate_inbounds @inline function get_categorical_form(Y::AbstractVector)
     class_names = unique(Y)
 
-    dict = Dict{eltype(Y),Int64}()
+    dict = Dict{eltype(Y), Int64}()
     @simd for i in 1:length(class_names)
         @inbounds dict[class_names[i]] = i
     end
@@ -62,16 +62,16 @@ See also
 [`Label`](@ref).
 """
 function bestguess(
-    labels::AbstractVector{<:Label},
-    weights::Union{Nothing,AbstractVector} = nothing;
-    suppress_parity_warning = false,
+        labels::AbstractVector{<:Label},
+        weights::Union{Nothing, AbstractVector} = nothing;
+        suppress_parity_warning = false,
 ) end
 
 # Classification: (weighted) majority vote
 function bestguess(
-    labels::AbstractVector{<:CLabel},
-    weights::Union{Nothing,AbstractVector} = nothing;
-    suppress_parity_warning = false,
+        labels::AbstractVector{<:CLabel},
+        weights::Union{Nothing, AbstractVector} = nothing;
+        suppress_parity_warning = false,
 )
     if length(labels) == 0
         return nothing
@@ -82,33 +82,33 @@ function bestguess(
             # return StatsBase.mode(labels) ..?
             countmap(labels)
         else
-            @assert length(labels) === length(weights) "Cannot compute " *
-             "best guess with mismatching number of votes " *
-             "$(length(labels)) and weights $(length(weights))."
+            @assert length(labels)===length(weights) "Cannot compute "*
+                                                     "best guess with mismatching number of votes "*
+                                                     "$(length(labels)) and weights $(length(weights))."
             countmap(labels, weights)
         end
     end
 
     if !suppress_parity_warning && sum(counts[argmax(counts)] .== values(counts)) > 1
         @warn "Parity encountered in bestguess! " *
-            "counts ($(length(labels)) elements): $(counts), " *
-            "argmax: $(argmax(counts)), " *
-            "max: $(counts[argmax(counts)]) (sum = $(sum(values(counts))))"
+              "counts ($(length(labels)) elements): $(counts), " *
+              "argmax: $(argmax(counts)), " *
+              "max: $(counts[argmax(counts)]) (sum = $(sum(values(counts))))"
     end
     argmax(counts)
 end
 
 # Regression: (weighted) mean (or other central tendency measure?)
 function bestguess(
-    labels::AbstractVector{<:RLabel},
-    weights::Union{Nothing,AbstractVector} = nothing;
-    suppress_parity_warning = false,
+        labels::AbstractVector{<:RLabel},
+        weights::Union{Nothing, AbstractVector} = nothing;
+        suppress_parity_warning = false,
 )
     if length(labels) == 0
         return nothing
     end
 
-    (isnothing(weights) ? StatsBase.mean(labels) : sum(labels .* weights)/sum(weights))
+    (isnothing(weights) ? StatsBase.mean(labels) : sum(labels .* weights) / sum(weights))
 end
 
 ############################################################################################
@@ -130,22 +130,22 @@ default_weights(Y::AbstractVector) = default_weights(length(Y))
 
 Return a class-rebalancing weight vector, given a label vector `Y`.
 """
-function balanced_weights(Y::AbstractVector{L}) where {L<:CLabel}
+function balanced_weights(Y::AbstractVector{L}) where {L <: CLabel}
     class_counts_dict = countmap(Y)
     if length(unique(values(class_counts_dict))) == 1 # balanced case
         default_weights(length(Y))
     else
         # Assign weights in such a way that the dataset becomes balanced
         tot = sum(values(class_counts_dict))
-        balanced_tot_per_class = tot/length(class_counts_dict)
-        weights_map = Dict{L,Float64}([class => (balanced_tot_per_class/n_instances)
-            for (class,n_instances) in class_counts_dict])
+        balanced_tot_per_class = tot / length(class_counts_dict)
+        weights_map = Dict{L, Float64}([class => (balanced_tot_per_class / n_instances)
+                                        for (class, n_instances) in class_counts_dict])
         W = [weights_map[y] for y in Y]
         W ./ sum(W)
     end
 end
 
 slice_weights(W::Ones{Int64}, inds::AbstractVector) = default_weights(length(inds))
-slice_weights(W::Any,         inds::AbstractVector) = @view W[inds]
+slice_weights(W::Any, inds::AbstractVector) = @view W[inds]
 slice_weights(W::Ones{Int64}, i::Integer) = 1
-slice_weights(W::Any,         i::Integer) = W[i]
+slice_weights(W::Any, i::Integer) = W[i]