Merge pull request #229 from alan-turing-institute/dev

Migration of code out to MLJBase and MLJModels; update to new versions of these
JuliaAI · Sep 11, 2019 · 99ef588 · 99ef588
2 parents c755afa + 2acd52f
commit 99ef588
Show file tree

Hide file tree

Showing 58 changed files with 780 additions and 3,231 deletions.
diff --git a/.travis.yml b/.travis.yml
@@ -34,7 +34,7 @@ jobs:
         - |
           julia --color=yes --project=docs/ -e'
             using Pkg
-            Pkg.add(PackageSpec(name="MLJBase", rev="master"));Pkg.add(PackageSpec(name="MLJModels", rev="master"));Pkg.develop(PackageSpec(path=pwd()))
+            Pkg.pin(PackageSpec(name="Missings", version="0.4.1"));Pkg.add(PackageSpec(name="MLJBase", rev="master"));Pkg.add(PackageSpec(name="MLJModels", rev="master"));Pkg.develop(PackageSpec(path=pwd()))
             Pkg.instantiate()
             include("docs/make.jl")
           '
diff --git a/CONTRIBUTE.md b/CONTRIBUTE.md
@@ -6,8 +6,7 @@ channel](https://slackinvite.julialang.org), #MLJ.
 
 
 - [List of presently implemented
-  models](src/Registry/Models.toml)
-  excluding built-in models. For full list, do `models()` in MLJ.
+  models](https://github.com/alan-turing-institute/MLJModels.jl/tree/master/src/registry/Models.toml). Or, do `using MLJ; models()`.
 
 - [Enhancement requests](https://github.com/alan-turing-institute/MLJ.jl/issues?utf8=✓&q=is%3Aissue+is%3Aopen+label%3A%22enhancement%22)
 
@@ -17,7 +16,7 @@ While new model implementations are a priority at present, help adding
 core functionality to MLJ is also welcome. If you are interested in
 contributing, please read the this rest of this document. A guide to
 implementing the MLJ interface for new models is
-[here](docs/src/adding_models_for_general_use.md).
+[here](https://alan-turing-institute.github.io/MLJ.jl/dev/adding_models_for_general_use/).
 
 
 ### Brief design overview

diff --git a/Project.toml b/Project.toml
@@ -1,7 +1,7 @@
 name = "MLJ"
 uuid = "add582a8-e3ab-11e8-2d5e-e98b27df1bc7"
 authors = ["Anthony D. Blaom <anthony.blaom@gmail.com>"]
-version = "0.3.0"
+version = "0.4.0"
 
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
@@ -26,8 +26,8 @@ StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [compat]
-MLJBase = "0.4"
-MLJModels = "0.3"
+MLJBase = "0.5"
+MLJModels = "0.4"
 julia = "1"
 
 [extras]

diff --git a/README.md b/README.md
@@ -2,7 +2,7 @@
 
 A pure Julia machine learning framework.
 
-[MLJ News](https://github.com/alan-turing-institute/MLJ.jl/blob/master/docs/src/NEWS.md)
+[MLJ News](https://github.com/alan-turing-institute/MLJ.jl/blob/master/docs/src/NEWS.md) | [MLJ Cheatsheet](docs/src/mlj_cheatsheet.md)
 
 
 ## `join!(MLJ, YourModel)`
@@ -20,7 +20,7 @@ and the core team is happy to respond to [issue requests](https://github.com/ala
 assistance. Please click [here](CONTRIBUTE.md) for more details on
 contributing.
 
-MLJ is presently supported by a small Alan Turing Institute grant and is looking for new funding sources to grow the project.
+MLJ is presently supported by a small Alan Turing Institute grant and is looking for new funding sources to grow and maintain the project.
 
 [![Build Status](https://travis-ci.com/alan-turing-institute/MLJ.jl.svg?branch=master)](https://travis-ci.com/alan-turing-institute/MLJ.jl)
 [![Slack Channel mlj](https://img.shields.io/badge/chat-on%20slack-yellow.svg)](https://slackinvite.julialang.org/)
@@ -36,12 +36,8 @@ scientific programming language, [Julia](https://julialang.org).
 
 The MLJ project is partly inspired by [MLR](https://mlr.mlr-org.com/index.html).
 
-A list of models in external packages that can be used with MLJ:
-[Models.toml](src/registry/Models.toml)
-
-[MLJ Cheatsheet](docs/src/mlj_cheatsheet.md)
-
-
+[List of presently implemented models](https://github.com/alan-turing-institute/MLJModels.jl/tree/master/src/registry/Models.toml)
+
 
 ### Installation
 
@@ -53,7 +49,7 @@ Pkg.add("MLJ")
 Pkg.add("MLJModels")
 ```
 
-To obtain a list of all registered models, keyed on package name:
+To obtain a list of all registered models:
 
 ```julia
 using MLJ
@@ -106,6 +102,8 @@ available.
 
 - Option to tune hyperparameters using gradient descent and **automatic
 	differentiation** (for learning algorithms written in Julia).
+
+- Option to tune hyperaparameters using **Bayesian optimisation**
 
 - **Data agnostic**: Train models on any data supported by the Tables.jl 
 [interface](https://github.com/JuliaData/Tables.jl). &#10004;
@@ -114,52 +112,48 @@ available.
   **learning networks** .&#10004;
 
 - Learning networks can be exported as self-contained **composite models** &#10004;, but
-  common networks (e.g., linear pipelines, stacks) come ready to plug-and-play.
+  common networks (e.g., linear **pipelines** &#10004;, **stacks**) come ready to plug-and-play.
 
 - Performant parallel implementation of large homogeneous **ensembles**
   of arbitrary models (e.g., random forests). &#10004;
 
-- **Task** interface matches machine learning problem to available models. &#10004; 
+- Model **registry** and facility to **match models** to machine learning
+  tasks. &#10004;
 
 - **Benchmarking** a battery of assorted models for a given task.
 
 - Automated estimates of cpu and memory requirements for given task/model.
 
+- Friendly interface for handling **probabilistic** prediction. &#10004;
+
 
 ### Frequently Asked Questions
 
 See [here](docs/src/frequently_asked_questions.md).
 
 
-### Known issues
-
-- The ScikitLearn SVM models will not work under Julia 1.0.3 but do work under Julia 1.1 due to [Issue #29208](https://github.com/JuliaLang/julia/issues/29208)
-
-- When MLJRegistry is updated with new models you may need to force a new
-  precompilation of MLJ to make new models available.
-
-
 ### Getting started
 
 Get started
-[here](https://alan-turing-institute.github.io/MLJ.jl/dev/),
+[here](https://alan-turing-institute.github.io/MLJ.jl/stable/),
 or take the MLJ [tour](/examples/tour/tour.ipynb).
 
 
 ### History
 
-Predecessors of the current package are
-[AnalyticalEngine.jl](https://github.com/tlienart/AnalyticalEngine.jl)
-and [Orchestra.jl](https://github.com/svs14/Orchestra.jl), and
-[Koala.jl](https://github.com/ablaom/Koala.jl). Work
-continued as a research study group at the University of Warwick,
+Antecedents for the current package are
+[AnalyticalEngine.jl](https://github.com/tlienart/AnalyticalEngine.jl),
+[Orchestra.jl](https://github.com/svs14/Orchestra.jl), and
+[Koala.jl](https://github.com/ablaom/Koala.jl). Development was also
+guided by a research study group at the University of Warwick,
 beginning with a review of existing ML Modules that were available in
-Julia at the time ([in-depth](https://github.com/dominusmi/Julia-Machine-Learning-Review/tree/master/Educational),
+Julia at the time
+([in-depth](https://github.com/dominusmi/Julia-Machine-Learning-Review/tree/master/Educational),
 [overview](https://github.com/dominusmi/Julia-Machine-Learning-Review/tree/master/Package%20Review)).
 
 ![alt text](material/packages.jpg)
 
 Further work culminated in the first MLJ
 [proof-of-concept](https://github.com/alan-turing-institute/MLJ.jl/tree/poc)
 
-For administrators: [Applying requests to register new models](REGISTRY.md).
+For administrators: [Implementing requests to register new models](REGISTRY.md).
diff --git a/REGISTRY.md b/REGISTRY.md
@@ -1,21 +1,23 @@
+# Instructions for updating the MLJ Model Registry
+
 To register all the models in GreatNewPackage with MLJ:
 
-- In a clone of the master branch of MLJ, change to the
-  `/src/registry/` directory and, in Julia, activate the environment
-  specified by the Project.toml there, after checking the [compat]
-  conditions thre are up to date.
+- In a clone of the master branch of
+  [MLJModels](https://github.com/alan-turing-institute/MLJModels.jl),
+  change to the `/src/registry/` directory and, in Julia, activate the
+  environment specified by the Project.toml there, after checking the
+  [compat] conditions there are up to date.
 
 - Add `GreatNewPackage` to the environment.
 
-- Activate a new environment in which your MLJ clone has been
-  `dev`ed. Execute `using MLJ; MLJ.Registry.@update`. This updates
-  `/Metadata.toml` and `/Models.toml` (the latter is generated for
-  convenience and not used by MLJ).
+- In some environment in which your MLJModels clone has been added
+  using `Pkg.dev`, execute `using MLJmodels; @update`. This updates
+  `src/registry/Metadata.toml` and `src/registry/Models.toml` (the
+  latter is generated for convenience and not used by MLJ).
 
 -  Quit your REPL session, whose namespace is now polluted.
 
-- Commit and make a PR request to merge your clone with master. Once
-  merged, the new metadata is available to users of MLJ#master.
+- Push your changes to an appropriate branch of MLJModels to make
+  the updated metadata available to users of the next MLJModels tagged
+  release.
 
-- Consider registering an new tagged version of MLJ. 
-
diff --git a/docs/Project.toml b/docs/Project.toml
@@ -9,6 +9,7 @@ LossFunctions = "30fc2ffe-d236-52d8-8643-a9d8f7c094a7"
 MLJ = "add582a8-e3ab-11e8-2d5e-e98b27df1bc7"
 MLJBase = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
 MLJModels = "d491faf4-2d78-11e9-2867-c94bc002c0b7"
+Missings = "e1d29d7a-bbdc-5cf2-9ac0-f12de2c33e28"
 MultivariateStats = "6f286f6a-111f-5878-ab1e-185364afe411"
 RDatasets = "ce6b1742-4840-55fa-b093-852dadbb1d8b"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
@@ -17,4 +18,3 @@ TypedTables = "9d95f2ec-7b3d-5a63-8d20-e2491e220bb9"
 
 [compat]
 Documenter = "~0.22"
-Missings = "<0.4.2"
diff --git a/docs/make.jl b/docs/make.jl
@@ -2,48 +2,47 @@ if Base.HOME_PROJECT[] !== nothing
     Base.HOME_PROJECT[] = abspath(Base.HOME_PROJECT[])
 end
 using Pkg
-#Pkg.add("Documenter")
-#Pkg.clone("https://github.com/alan-turing-institute/MLJBase.jl")
-#Pkg.clone("https://github.com/alan-turing-institute/MLJModels.jl")
-#Pkg.clone("https://github.com/alan-turing-institute/MLJ.jl")
 using Documenter
 using MLJ
 using MLJBase
-using MLJ.Transformers
-using MLJ.Constant
+using MLJModels.Transformers
+using MLJModels.Constant
+using MLJModels.KNN
 using MLJModels
 using ScientificTypes
 
-#prettyurls to be changed
+pages = Any["Getting Started"=>"index.md",
+            "Evaluating model performance"=>"evaluating_model_performance.md",
+            "Performance Measures"=> "performance_measures.md",
+            "Tuning models"=>"tuning_models.md",
+            "Built-in Transformers" => "built_in_transformers.md",
+            "Composing Models" => "composing_models.md",
+            "Homogeneous Ensembles" => "homogeneous_ensembles.md",
+            "Simple User Defined Models" => "simple_user_defined_models.md",
+            "Adding Models for General Use" => "adding_models_for_general_use.md",
+            "Benchmarking" => "benchmarking.md",
+            "Working with tasks" => "working_with_tasks.md",
+            "Internals"=>"internals.md",
+            "Glossary"=>"glossary.md",
+            "API"=>"api.md",
+            "MLJ Cheatsheet" => "mlj_cheatsheet.md",
+            "MLJ News"=>"NEWS.md",
+            "FAQ" => "frequently_asked_questions.md",
+            "Julia BlogPost"=>"julia_blogpost.md"]
+
+for p in pages
+    println(first(p))
+end
+
 makedocs(
     sitename = "MLJ",
     format = Documenter.HTML(),
-    modules = [MLJ, MLJBase, MLJModels, MLJ.Transformers, ScientificTypes],
-    pages = Any["Getting Started"=>"index.md",
-                "Evaluating model performance"=>"evaluating_model_performance.md",
-                "Performance Measures"=> "performance_measures.md",
-                "Tuning models"=>"tuning_models.md",
-                "Built-in Transformers" => "built_in_transformers.md",
-                "Composing Models" => "composing_models.md",
-                "Homogeneous Ensembles" => "homogeneous_ensembles.md",
-                "Simple User Defined Models" => "simple_user_defined_models.md",
-                "Adding Models for General Use" => "adding_models_for_general_use.md",
-                "Working with Tasks" => "working_with_tasks.md",
-                "Benchmarking" => "benchmarking.md",
-                "Internals"=>"internals.md",
-                "Glossary"=>"glossary.md",
-                "API"=>"api.md",
-                "MLJ Cheatsheet" => "mlj_cheatsheet.md",
-                "MLJ News"=>"NEWS.md",
-                "FAQ" => "frequently_asked_questions.md",
-                 "Julia BlogPost"=>"julia_blogpost.md"]
-)
+    modules = [MLJ, MLJBase, MLJModels, MLJModels.Transformers, MLJModels.Constant,
+               MLJModels.KNN, ScientificTypes],
+    pages=pages)
 
 deploydocs(
     repo = "github.com/alan-turing-institute/MLJ.jl.git"
 )
 
-#    modules = [MLJ]
-# Documenter can also automatically deploy documentation to gh-pages.
-# See "Hosting Documentation" and deploydocs() in the Documenter manual
-# for more MLJBase.information.
+