Add target_scitype to unsupervised model traits (#16)

Author: tlienart

Project.toml

@@ -1,7 +1,7 @@
 name = "MLJModelInterface"
 uuid = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 authors = ["Thibaut Lienart and Anthony Blaom"]
-version = "0.1.5"
+version = "0.1.6"
 
 [deps]
 ScientificTypes = "321657f4-b219-11e9-178b-2701a2544e81"
@@ -15,9 +15,9 @@ CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Distances = "b4f34e82-e78d-54a5-968a-f98e89d6e8f7"
 InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
+MLJScientificTypes = "2e2323e0-db8b-457b-ae0d-bdfb3bc63afd"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "Tables", "Distances", "CategoricalArrays", "InteractiveUtils",
-        "DataFrames"]
+test = ["Test", "Tables", "Distances", "CategoricalArrays", "InteractiveUtils", "DataFrames", "MLJScientificTypes"]

README.md

@@ -1,5 +1,24 @@
 # MLJModelInterface.jl
 
+A light-weight interface for developers wanting to integrate
+machine learning models into
+[MLJ](https://github.com/alan-turing-institute/MLJ.jl).
+
+
 | [MacOS/Linux] | Coverage |
 | :-----------: | :------: |
 | [![Build Status](https://travis-ci.org/alan-turing-institute/MLJModelInterface.jl.svg?branch=master)](https://travis-ci.org/alan-turing-institute/MLJModelInterface.jl) | [![codecov.io](http://codecov.io/github/alan-turing-institute/MLJModelInterface.jl/coverage.svg?branch=master)](http://codecov.io/github/alan-turing-institute/MLJModelInterface.jl?branch=master) |
+
+
+[MLJ](https://github.com/alan-turing-institute/MLJ.jl) is a framework
+for evaluating, combining and optimizing machine learning models in
+Julia. A third party package wanting to integrate their supervised or
+unsupervised machine learning models must import the module
+`MLJModelInterface` defined in this package. 
+
+### Instructions
+
+- [Quick-start guide](https://alan-turing-institute.github.io/MLJ.jl/dev/quick_start_guide_to_adding_models/) to adding models to MLJ
+
+- [Detailed API
+  specification](https://alan-turing-institute.github.io/MLJ.jl/dev/adding_models_for_general_use/)

src/data_utils.jl

@@ -83,7 +83,7 @@ int(::LightInterface, x) = errlight("int")
 """
     classes(x)
 
-All the categorical elements with in the same pool as `x` (including `x`),
+All the categorical elements with the same pool as `x` (including `x`),
 returned as a list, with an ordering consistent with the pool $REQUIRE.
 Here `x` has `CategoricalValue` or `CategoricalString` type, and `classes(x)`
 is a vector of the same eltype. Note that `x in classes(x)` is always true.
@@ -277,24 +277,22 @@ const UNIVARIATE_FINITE_DOCSTRING =
     """
         UnivariateFinite(classes, p)
 
-    A discrete univariate distribution whose finite support is the elements of
-    the vector `classes`, and whose corresponding probabilities are elements of
-    the vector `p`, which must sum to one $REQUIRE. Here `classes` must have
-    type `AbstractVector{<:CategoricalElement}` where
+A discrete univariate distribution whose finite support is the elements of the
+vector `classes`, and whose corresponding probabilities are elements of the
+vector `p`, which must sum to one $REQUIRE. Here `classes` must have type
+`AbstractVector{<:CategoricalElement}` where
 
         CategoricalElement = Union{CategoricalValue,CategoricalString}
 
-    and all classes are assumed to share the same categorical pool.
+and all classes are assumed to share the same categorical pool.
 
         UnivariateFinite(prob_given_class)
 
-    A discrete univariate distribution whose finite support is the set of keys
-    of the provided dictionary, `prob_given_class` $REQUIRE. The dictionary
-    keys must be of type `CategoricalElement` (see above) and the dictionary
-    values specify the corresponding probabilities.
-    """
-
-"""$UNIVARIATE_FINITE_DOCSTRING"""
+A discrete univariate distribution whose finite support is the set of keys of
+the provided dictionary, `prob_given_class` $REQUIRE. The dictionary keys must
+be of type `CategoricalElement` (see above) and the dictionary values specify
+the corresponding probabilities.
+"""
 UnivariateFinite(d::AbstractDict) = UnivariateFinite(get_interface_mode(), d)
 UnivariateFinite(c::AbstractVector, p) =
     UnivariateFinite(get_interface_mode(), c, p)

src/model_def.jl

@@ -8,7 +8,7 @@
 clean!(::Model) = ""
 
 """
-	_process_model_def(ex)
+    _process_model_def(ex)
 
 Take an expression defining a model (`mutable struct Model ...`) and unpack key
 elements for further processing:
@@ -88,7 +88,7 @@ function _process_model_def(ex)
 end
 
 """
-	_unpack!(ex, rep)
+    _unpack!(ex, rep)
 
 Internal function to allow to read a constraint given after a default value for
 a parameter and transform it in an executable condition (which is returned to
@@ -111,7 +111,7 @@ _unpack!(ex, _) = ex # when it's been unpacked, it's not an expression anymore
 
 
 """
-	_model_constructor(modelname, params, defaults)
+    _model_constructor(modelname, params, defaults)
 
 Build the expression of the keyword constructor associated with a model
 definition. When the constructor is called, the `clean!` function is called
@@ -131,7 +131,7 @@ end
 
 
 """
-	_model_cleaner(modelname, defaults, constraints)
+    _model_cleaner(modelname, defaults, constraints)
 
 Build the expression of the cleaner associated with the constraints specified
 in a model def.
@@ -163,7 +163,7 @@ function _model_cleaner(modelname, defaults, constraints)
 end
 
 """
-	@mlj_model
+    @mlj_model
 
 Macro to help define MLJ models with constraints on the default parameters.
 """

src/model_traits.jl

@@ -13,7 +13,7 @@ const MODEL_TRAITS = [
 const SUPERVISED_TRAITS = setdiff(MODEL_TRAITS, [:output_scitype])
 
 const UNSUPERVISED_TRAITS = setdiff(MODEL_TRAITS,
-    [:target_scitype, :prediction_type, :supports_weights])
+    [:prediction_type, :supports_weights])
 
 for trait in MODEL_TRAITS
     ex = quote

test/data_utils.jl

@@ -1,9 +1,9 @@
-@testset "categorical-light" begin
+@testset "cat-light" begin
     setlight()
     x = 1:5
     @test_throws M.InterfaceError M.categorical(x)
 end
-@testset "categorical-full" begin
+@testset "cat-full" begin
     setfull()
     M.categorical(::FI, a...; kw...) = categorical(a...; kw...)
     x = 1:5

test/runtests.jl

@@ -1,4 +1,5 @@
-using Test, MLJModelInterface, ScientificTypes
+using Test, MLJModelInterface
+using ScientificTypes, MLJScientificTypes
 using Tables, Distances, CategoricalArrays, InteractiveUtils
 import DataFrames: DataFrame