Added ReadMe. Fixed typo in partial_dependence.

And added more docstring

Project.toml

@@ -11,11 +11,9 @@ Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 julia = "1"
 
 [extras]
-# Conda = "8f4d0f93-b110-5947-807f-2305c1781a2d"
 RDatasets = "ce6b1742-4840-55fa-b093-852dadbb1d8b"
 ScikitLearn = "3646fa90-6ef7-5e7e-9f22-8aca16db6324"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
-# PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"
 
 [targets]
-test = ["Test", "ScikitLearn", "RDatasets"]#, "Conda", "PyCall"
+test = ["Test", "ScikitLearn", "RDatasets"]

README.md

@@ -1,6 +1,52 @@
 # InterpretMe
 
-[![Build Status](https://travis-ci.com/mmahrouss/InterpretMe.jl.svg?branch=master)](https://travis-ci.com/mmahrouss/InterpretMe.jl)
-[![Build Status](https://ci.appveyor.com/api/projects/status/github/mmahrouss/InterpretMe.jl?svg=true)](https://ci.appveyor.com/project/mmahrouss/InterpretMe-jl)
-[![Codecov](https://codecov.io/gh/mmahrouss/InterpretMe.jl/branch/master/graph/badge.svg)](https://codecov.io/gh/mmahrouss/InterpretMe.jl)
-[![Coveralls](https://coveralls.io/repos/github/mmahrouss/InterpretMe.jl/badge.svg?branch=master)](https://coveralls.io/github/mmahrouss/InterpretMe.jl?branch=master)
+[![Build Status](https://travis-ci.com/mmahrouss/InterpretMe.jl.svg?branch=master)](https://travis-ci.com/mmahrouss/InterpretMe.jl) [![Build Status](https://ci.appveyor.com/api/projects/status/github/mmahrouss/InterpretMe.jl?svg=true)](https://ci.appveyor.com/project/mmahrouss/InterpretMe-jl) [![Codecov](https://codecov.io/gh/mmahrouss/InterpretMe.jl/branch/master/graph/badge.svg)](https://codecov.io/gh/mmahrouss/InterpretMe.jl) [![Coveralls](https://coveralls.io/repos/github/mmahrouss/InterpretMe.jl/badge.svg?branch=master)](https://coveralls.io/github/mmahrouss/InterpretMe.jl?branch=master)
+
+InterpretMe is Julia Package aimed at providing interpret-ability tools for Machine Learning models in Julia.
+
+## Installation
+
+```julia
+julia>]
+pkg> add InterpretMe
+```
+
+Please note that the package is still a Work In Progress and has not been registered yet. Thus, for now you should install it as follows:
+
+```julia
+julia>]
+pkg> add https://github.com/mmahrouss/InterpretMe.jl
+```
+
+## Example
+
+First import the package:
+
+```julia
+Using InterpretMe
+```
+
+Then train a model from [ScikitLearn.jl](https://github.com/cstjean/ScikitLearn.jl)'s examples. We are going to use the [Iris Example](https://github.com/cstjean/ScikitLearn.jl/blob/master/notebooks/Iris.ipynb)
+
+Load the packages and train the model:
+
+```julia
+import ScikitLearn
+import RDatasets
+
+ScikitLearn.@sk_import linear_model: LogisticRegression
+
+iris = RDatasets.dataset("datasets", "iris")
+
+X = convert(Array, iris[[:SepalLength, :SepalWidth, :PetalLength, :PetalWidth]])
+y = convert(Array, iris[:Species])
+
+model = ScikitLearn.fit!(LogisticRegression(), X, y)
+```
+
+Now we can calculate the partial_dependance function from InterpretMe.
+
+```julia
+pd, vals = partial_dependence(x, [1], x -> ScikitLearn.predict_proba(model,x),
+                              100, false)
+```

src/InterpretMe.jl

@@ -1,6 +1,5 @@
 module InterpretMe
 
-# greet() = print("Hello World!")
 include("pdp.jl")
-export partial_dependance
+export partial_dependence
 end # module

src/pdp.jl

@@ -1,17 +1,20 @@
 using LinearAlgebra
 using Statistics
 """
-    InterpretMe.partial_dependance(data::Array{Any,2},
-        features::Vector{<:Integer}
-        ,feature::T, estimate_fn, grid_size::Int, return_ice::Bool)
-         where {T<:AbstractString}
+    InterpretMe.partial_dependence(data::Array{<:Any,2},
+                                   features::Union{<:Integer,Vector{<:Integer}},
+                                   estimate_fn, grid_size::Int,
+                                   return_ice::Bool)
 Calculate the partial dependance of a machine learning model.
 
 `estimate_fn` is a function from the model that generates a vector of predictions from a matrix of inputs.
 `feature_names` should be the feature names corresponding to the features in `data`
 `return_ice` whether to return the ICE lines.
+Returns
+`pd` the partial dependancy results
+`grid_values` the grid values for all the selected features 
 """
-function partial_dependance(data::Array{<:Any,2},
+function partial_dependence(data::Array{<:Any,2},
                             features::Union{<:Integer,Vector{<:Integer}},
                             estimate_fn, grid_size::Int, return_ice::Bool)
     # Mainly inspired by Sk-learn's function.

test/runtests.jl

@@ -1,14 +1,10 @@
 using InterpretMe
-# using Conda
-using ScikitLearn
-using RDatasets
 using Test
-# import PyCall
-# PyCall.pyimport_conda("sklearn", "scikit-learn")
-# Conda.add("scikit-learn")
+import ScikitLearn
+import RDatasets
 
-@sk_import linear_model: LogisticRegression
-@sk_import inspection: partial_dependence
+ScikitLearn.@sk_import linear_model: LogisticRegression
+ScikitLearn.@sk_import inspection: partial_dependence
 
 @testset "InterpretMe.jl" begin
     # Write your own tests here.
@@ -17,11 +13,13 @@ using Test
         x = convert(Matrix, iris[!,[:SepalLength,:SepalWidth,
                                     :PetalLength,:PetalWidth ]])
         y = convert(Array, iris.Species)
-        model = fit!(LogisticRegression(), x, y)
+        model = ScikitLearn.fit!(LogisticRegression(), x, y)
         averaged_predictions, values = partial_dependence(model, x, [0],
                                                           grid_resolution=100)
-        pd, vals = partial_dependance(x, [1], x -> predict_proba(model,x),
-                                       100, false)
+        pd, vals = InterpretMe.partial_dependence(x, [1],
+                                                 x -> ScikitLearn.predict_proba(
+                                                                       model,x),
+                                                  100, false)
         @test all(pd .≈ transpose(averaged_predictions))
     end
 end