Merge pull request #210 from JuliaAI/dev

Update docs

README.md

@@ -7,7 +7,7 @@ machine learning models into
 
 | Linux | Coverage |
 | :-----------: | :------: |
-| [![Build Status](https://github.com/JuliaAI/MLJModelInterface.jl/workflows/CI/badge.svg)](https://github.com/JuliaAI/MLJModelInterface.jl/actions) | [![codecov.io](http://codecov.io/github/JuliaAI/MLJModelInterface.jl/coverage.svg?branch=master)](http://codecov.io/github/JuliaAI/MLJModelInterface.jl?branch=master) |
+| [![Build Status](https://github.com/JuliaAI/MLJModelInterface.jl/workflows/CI/badge.svg)](https://github.com/JuliaAI/MLJModelInterface.jl/actions) | [![codecov](https://codecov.io/github/JuliaAI/MLJModelInterface.jl/graph/badge.svg?token=rkvwHku1dW)](https://codecov.io/github/JuliaAI/MLJModelInterface.jl)
 
 [![Stable](https://img.shields.io/badge/docs-stable-blue.svg)](https://juliaai.github.io/MLJModelInterface.jl/dev/)
 

docs/src/model_wrappers.md

@@ -27,3 +27,21 @@ EnsembleModel(model=tree, n=100)
 
 This is the only case in MLJ where positional arguments in a model constructor are
 allowed.
+
+## Handling generic constructors
+
+Model wrappers frequently have a public facing constructor with a name different from that
+of the model type constructed. For example, `TunedModel(model, ...)` is a constructor that
+will construct either an instance of `DeterministicTunedModel` or
+`ProbabilisticTunedModel`, depending on the type of `model`. In this case it is necessary
+to overload the `constructor` trait, which in that case looks like this:
+
+```julia
+MLJModelInterface.constructor(::Type{<:Union{
+    DeterministicTunedModel,
+	ProbabilisticTunedModel,
+	}}) = TunedModel
+```
+
+This allows the MLJ Model Registry to correctly associate model metadata to the
+constructor, rather than the (private) types.