weka 0.5.0-java → 0.6.0-java

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA256:
-  metadata.gz: 07d0cbed2d245de34e10101d0597017194804735e3b5ea6bc972201282c60d5d
-  data.tar.gz: b304c03a4552b766f56e60e52b765068fdb0de0350d1e5c541882f085e90824b
+SHA1:
+  metadata.gz: 17c3a647cf8e23d2c81a4f456c7e3e576a8e768c
+  data.tar.gz: 90419f5ac130f3a1e44c860947b68ed8fd7e1fc3
 SHA512:
-  metadata.gz: 942621fa83a7670384adccb717610ac19dc5d961fd9de5ab579a331c40119ef8e0e73cb0b0962aaff9fe937628335ee453be2ef1ed0669edf5421dadc075acdc
-  data.tar.gz: ee68c277c9a40f6349fdea303897d44fd93ded9311d11ffba8c1cd3909764ea51d22638cc7240bb375570615a9acaa4d98ba591b8de3102c41c3dcabf550283b
+  metadata.gz: 0e940c23d7234b6a3dad46e9625040912122d046916bb1db31f1c57abc0eacd14b2ee08178010340a12e55a3fe99039f348ea944067ec1b1c6996e0adb4954fd
+  data.tar.gz: ad512593515521dbe0dc5627628df1ccfa2e8a9279acf1359a75f2321911c65bfe2e5ef32527ffa872743514e48554e73fbf88744e8989f6ae93449c40266df1

data/.gitignore

@@ -8,3 +8,4 @@
 /spec/reports/
 /tmp/
 jars
+/lib/*_jars.rb

data/.rubocop.yml

@@ -5,6 +5,7 @@ AllCops:
     - '*.gemspec'
     - 'Gemfile'
     - 'Gemfile.lock'
+    - 'lib/weka_jars.rb'
 
 Style/Copyright:
   Enabled: false

data/.travis.yml

@@ -2,14 +2,6 @@ sudo: false
 language: ruby
 
 rvm:
-  - jruby-9000
-
-cache:
-  - bundler
-
-before_install:
-  - rvm get head
-  - rvm use jruby-9.0.1.0 --install
-  - gem install bundler
-
-script: bundle exec rake spec
+  - jruby-9.0.0.0
+  - jruby-9.1.0.0
+  - jruby-head

data/CHANGELOG.md

@@ -0,0 +1,68 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
+and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+## [0.6.0] – 2017-12-17
+### Added
+- #copy method for Weka::Core::Instances
+
+### Changed
+- Load Jars with jar-dependencies instead of lock_jar gem
+- Make Weka::Core::Instances#instance_from public
+
+### Fixed
+- Weka::UnassignedTrainingInstancesError when running #classify/#cluster and
+  #distribution_for on deserialized classfiers and clusterers
+
+
+## [0.5.0] – 2017-06-17
+### Added
+- #to_m (to Matrix) method for Weka::Core::Instances
+- Curve classes in Weka::Classifiers::Evaluation module
+- Allow including additional modules on class building
+- Rubocop config for project
+
+### Changed
+- Allow passing a hash to Weka::Core::Instances#add_instance
+
+### Removed
+- Block argument in Weka::Core::Instances#initialize
+
+
+## [0.4.0] – 2016-12-22
+### Added
+- C45 converters
+- Full support for string attributes
+
+### Removed
+- ActiveSupport as dependency
+
+
+## [0.3.0] – 2016-02-10
+### Added
+- Allow adding Instances with missing values
+- Allow creating DenseInstances with missing values
+- #merge method for Weka::Core:Instances
+
+
+## [0.2.0] – 2016-01-19
+### Added
+- Serialization/deserialization functionality
+- #apply_filters method for Weka::Core::Instances
+
+
+## [0.1.0] – 2015-12-26
+Initial release
+
+[Unreleased]: https://github.com/paulgoetze/weka-jruby/compare/v0.6.0...HEAD
+[0.6.0]: https://github.com/paulgoetze/weka-jruby/compare/v0.5.0...v0.6.0
+[0.5.0]: https://github.com/paulgoetze/weka-jruby/compare/v0.4.0...v0.5.0
+[0.4.0]: https://github.com/paulgoetze/weka-jruby/compare/v0.3.0...v0.4.0
+[0.3.0]: https://github.com/paulgoetze/weka-jruby/compare/v0.2.0...v0.3.0
+[0.2.0]: https://github.com/paulgoetze/weka-jruby/compare/v0.1.0...v0.2.0
+[0.1.0]: https://github.com/paulgoetze/weka-jruby/compare/ce6a985017c28ea755290a9baba4d81acddc2d20...v0.1.0

data/README.md

@@ -39,8 +39,12 @@ detailed information about how to use weka with JRuby and some examplary code sn
 
 ## Development
 
-After checking out the repo, run `bin/setup` to install dependencies.
-To install this gem onto your local machine, run `bundle exec rake install`.
+1. Check out the repo with `git clone git@github.com:paulgoetze/weka-jruby.git`.
+2. Set a local environment variable `export JARS_VENDOR=false`.  
+   This will prevent
+   compiling the jars into your repo’s /lib directory and will load them from your local maven repository instead.
+   See the  [jar-dependencies README](https://github.com/mkristian/jar-dependencies#for-development-you-do-not-need-to-vendor-the-jars-at-all) for more information.
+3. Run `bin/setup` or `bundle install` to install the dependencies.
 
 Then, run `rake spec` to run the tests. You can also run `bin/console` or `rake irb` for an interactive prompt that will allow you to experiment.
 

data/Rakefile

@@ -3,17 +3,7 @@ require 'rspec/core/rake_task'
 
 RSpec::Core::RakeTask.new(:spec)
 
-task default: :prepare
-task install: :prepare
-
-desc 'Install weka jars & dependencies'
-task :prepare do
-  require 'lock_jar'
-  lib_path = File.expand_path('.', File.dirname(__FILE__))
-  jars_dir = File.join(lib_path, 'jars')
-
-  LockJar.install('Jarfile.lock', local_repo: jars_dir)
-end
+task default: :spec
 
 desc 'Start an irb session with the gem loaded'
 task :irb do

data/lib/weka.rb

@@ -1,11 +1,9 @@
 require 'java'
-require 'weka/jars'
 require 'weka/version'
 require 'weka/exceptions'
+require Dir[File.join(File.dirname(__FILE__), '*_jars.rb')].first
 
 module Weka
-  include Jars
-
   class << self
     def require_all(type)
       files        = Dir[File.expand_path("../weka/#{type}/**/*.rb", __FILE__)]

data/lib/weka/class_builder.rb

@@ -103,7 +103,7 @@ module Weka
       end
 
       def constantize(module_names)
-        Object.module_eval("::#{module_names}")
+        Object.module_eval("::#{module_names}", __FILE__, __LINE__)
       end
 
       def utils

data/lib/weka/classifiers/evaluation.rb

@@ -1,4 +1,5 @@
 require 'weka/class_builder'
+require 'weka/concerns'
 
 module Weka
   module Classifiers
@@ -6,6 +7,7 @@ module Weka
 
     class Evaluation
       include ClassBuilder
+      include Weka::Concerns::Persistent
 
       # Use both nomenclatures f_measure and fmeasure for consistency
       # due to jruby's auto method generation of 'fMeasure' to 'f_measure' and

data/lib/weka/classifiers/utils.rb

@@ -19,7 +19,7 @@ module Weka
 
           error   = 'Class attribute is not assigned for Instances.'
           hint    = 'You can assign a class attribute with #class_attribute=.'
-          message = "#{error} #{hint}"
+          message = "#{error}\n#{hint}"
 
           raise UnassignedClassError, message
         end
@@ -29,27 +29,49 @@ module Weka
 
           error   = 'Classifier is not trained with Instances.'
           hint    = 'You can set the training instances with #train_with_instances.'
-          message = "#{error} #{hint}"
+          message = "#{error}\n#{hint}"
 
           raise UnassignedTrainingInstancesError, message
         end
+
+        def ensure_valid_instances_structure!(instances)
+          unless instances.is_a?(Weka::Core::Instances)
+            message = 'Instances has to be a Weka::Core::Instances object.'
+            raise ArgumentError, message
+          end
+
+          return if training_instances.nil?
+          return if instances.equal_headers(training_instances)
+
+          message = 'The passed instances need to have the same structure as ' \
+                    'the classifier’s training instances.'
+
+          raise InvalidInstancesStructureError, message
+        end
+
+        def ensure_instances_structure_available!
+          return unless instances_structure.nil?
+
+          error   = 'Classifier does not have any instances structure info.'
+          hint    = 'You probably tried to classify values with a ' \
+                    'classifier that is untrained or doesn’t have an ' \
+                    'instances_structure set. Please run ' \
+                    '#train_with_instances, try serializing and ' \
+                    'deserializing your classifier again in case you used a ' \
+                    'deserialized classifier or set its instances_structure.'
+          message = "#{error}\n#{hint}"
+
+          raise MissingInstancesStructureError, message
+        end
       end
 
       module Transformers
         private
 
         def classifiable_instance_from(instance_or_values)
-          attributes = training_instances.attributes
-          instances  = Weka::Core::Instances.new(attributes: attributes)
+          ensure_instances_structure_available!
 
-          class_attribute = training_instances.class_attribute
-          class_index     = training_instances.class_index
-          instances.insert_attribute_at(class_attribute, class_index)
-
-          instances.class_index = training_instances.class_index
-          instances.add_instance(instance_or_values)
-
-          instance = instances.first
+          instance = instances_structure.instance_from(instance_or_values)
           instance.set_class_missing
           instance
         end
@@ -60,16 +82,24 @@ module Weka
         include Checks
 
         attr_reader :training_instances
+        attr_reader :instances_structure
 
         def train_with_instances(instances)
           ensure_class_attribute_assigned!(instances)
 
           @training_instances = instances
+          @instances_structure = instances.string_free_structure
+
           build_classifier(instances)
 
           self
         end
 
+        def instances_structure=(instances)
+          ensure_valid_instances_structure!(instances)
+          @instances_structure = instances.string_free_structure
+        end
+
         def cross_validate(folds: 3)
           ensure_trained_with_instances!
 
@@ -101,8 +131,6 @@ module Weka
         include Transformers
 
         def classify(instance_or_values)
-          ensure_trained_with_instances!
-
           instance = classifiable_instance_from(instance_or_values)
           index    = classify_instance(instance)
 
@@ -112,7 +140,7 @@ module Weka
         private
 
         def class_value_of_index(index)
-          training_instances.class_attribute.value(index)
+          instances_structure.class_attribute.value(index)
         end
       end
 
@@ -136,8 +164,6 @@ module Weka
         include Transformers
 
         def distribution_for(instance_or_values)
-          ensure_trained_with_instances!
-
           instance      = classifiable_instance_from(instance_or_values)
           distributions = distribution_for_instance(instance)
 
@@ -147,7 +173,7 @@ module Weka
         private
 
         def class_distributions_from(distributions)
-          class_values = training_instances.class_attribute.values
+          class_values = instances_structure.class_attribute.values
 
           distributions.each_with_object({}).with_index do |(distribution, result), index|
             class_value = class_values[index]

data/lib/weka/clusterers/utils.rb

@@ -27,21 +27,39 @@ module Weka
 
           error   = 'Clusterer is not trained with Instances.'
           hint    = 'You can set the training instances with #train_with_instances.'
-          message = "#{error} #{hint}"
+          message = "#{error}\n#{hint}"
 
           raise UnassignedTrainingInstancesError, message
         end
-      end
 
-      module Transformers
-        private
+        def ensure_valid_instances_structure!(instances)
+          unless instances.is_a?(Weka::Core::Instances)
+            message = 'Instances has to be a Weka::Core::Instances object.'
+            raise ArgumentError, message
+          end
+
+          return if training_instances.nil?
+          return if instances.equal_headers(training_instances)
 
-        def clusterable_instance_from(instance_or_values)
-          attributes = training_instances.attributes
-          instances  = Weka::Core::Instances.new(attributes: attributes)
+          message = 'The passed instances need to have the same structure as ' \
+                    'the clusterers training instances.'
 
-          instances.add_instance(instance_or_values)
-          instances.first
+          raise InvalidInstancesStructureError, message
+        end
+
+        def ensure_instances_structure_available!
+          return unless instances_structure.nil?
+
+          error   = 'Clusterer does not have any instances structure info.'
+          hint    = 'You probably tried to cluster values with a clusterer ' \
+                    'that is untrained or doesn’t have an ' \
+                    'instances_structure set. Please run ' \
+                    '#train_with_instances, try serializing and ' \
+                    'deserializing your clusterer again in case you used a ' \
+                    'deserialized clusterer or set its instances_structure.'
+          message = "#{error}\n#{hint}"
+
+          raise MissingInstancesStructureError, message
         end
       end
 
@@ -49,14 +67,22 @@ module Weka
         include Checks
 
         attr_reader :training_instances
+        attr_reader :instances_structure
 
         def train_with_instances(instances)
           @training_instances = instances
+          @instances_structure = instances.string_free_structure
+
           build_clusterer(instances)
 
           self
         end
 
+        def instances_structure=(instances)
+          ensure_valid_instances_structure!(instances)
+          @instances_structure = instances.string_free_structure
+        end
+
         def evaluate(test_instances)
           ensure_trained_with_instances!
 
@@ -85,12 +111,11 @@ module Weka
 
       module Clusterable
         include Checks
-        include Transformers
 
         def cluster(instance_or_values)
-          ensure_trained_with_instances!
+          ensure_instances_structure_available!
 
-          instance = clusterable_instance_from(instance_or_values)
+          instance = instances_structure.instance_from(instance_or_values)
           cluster_instance(instance)
         end
       end
@@ -112,12 +137,11 @@ module Weka
 
       module Distributable
         include Checks
-        include Transformers
 
         def distribution_for(instance_or_values)
-          ensure_trained_with_instances!
+          ensure_instances_structure_available!
 
-          instance = clusterable_instance_from(instance_or_values)
+          instance = instances_structure.instance_from(instance_or_values)
           distribution_for_instance(instance).to_a
         end
       end

data/lib/weka/core/instances.rb

@@ -47,6 +47,11 @@ module Weka
         super(relation_name.to_s, attribute_list, 0)
       end
 
+      def copy
+        constructor = Instances.java_class.declared_constructor(Instances)
+        constructor.new_instance(self).to_java(Instances)
+      end
+
       def instances
         enumerate_instances.to_a
       end
@@ -267,6 +272,33 @@ module Weka
         Matrix[*instances.map(&:values)]
       end
 
+      # Wrap the attribute values for the instance to be added with
+      #   an Instance object, if needed. The Instance object is
+      #   assigned with the given weight.
+      #
+      # @param [Instance, Array, Hash] instance_or_values either the
+      #   instance object to be added or the attribute values for it.
+      #   For the latter case, it accepts an array or a hash.
+      #
+      # @param [Float] weight the weight for the Instance to be added
+      #
+      # @return [Instance] the object that contains the given
+      #   attribute values.
+      def instance_from(instance_or_values, weight: 1.0)
+        dataset = string_free_structure
+
+        if instance_or_values.is_a?(Java::WekaCore::Instance)
+          instance = instance_or_values
+          instance.weight = weight
+        else
+          data = instance_data(instance_or_values)
+          instance = DenseInstance.new(data, weight: weight)
+        end
+
+        dataset.add(instance)
+        dataset.first
+      end
+
       private
 
       def add_attribute(attribute)
@@ -291,42 +323,18 @@ module Weka
         end
       end
 
-      # Wrap the attribute values for the instance to be added with
-      #   an Instance object, if needed. The Instance object is
-      #   assigned with the given weight.
-      #
-      # @param [Instance, Array, Hash] instance_or_values either the
-      #   instance object to be added or the attribute values for it.
-      #   For the latter case, it accepts an array or a hash.
-      #
-      # @param [Float] weight the weight for the Instance to be added
-      #
-      # @return [Instance] the object that contains the given
-      #   attribute values.
-      def instance_from(instance_or_values, weight:)
-        if instance_or_values.is_a?(Java::WekaCore::Instance)
-          instance_or_values.weight = weight
-          instance_or_values
-        else
-          if instance_or_values.is_a?(Hash)
-            instance_or_values = attribute_values_from_hash(instance_or_values)
-          end
-
-          data = internal_values_of(instance_or_values)
-
-          if has_string_attribute?
-            data = check_string_attributes(data, instance_or_values)
-          end
-
-          DenseInstance.new(data, weight: weight)
-        end
-      end
-
       def map_attribute_type(type)
         return -1 unless Attribute::TYPES.include?(type.downcase.to_sym)
         Attribute.const_get(type.upcase)
       end
 
+      def instance_data(values)
+        values = attribute_values_from_hash(values) if values.is_a?(Hash)
+        data = internal_values_of(values)
+        data = check_string_attributes(data, values) if has_string_attribute?
+        data
+      end
+
       # Convert a hash whose keys are attribute names and values are attribute
       #   values into an array containing attribute values in the order
       #   of the Instances attributes.

data/lib/weka/core/serialization_helper.rb

@@ -3,9 +3,47 @@ module Weka
     java_import 'weka.core.SerializationHelper'
 
     class SerializationHelper
+      STRUCTURE_FILE_EXTENSION = 'structure'.freeze
+
       class << self
+        original_read = instance_method(:read)
+        original_write = instance_method(:write)
+
+        define_method(:read) do |filename|
+          object = original_read.bind(self).call(filename)
+
+          structure_filename = structure_file(filename)
+          structure_needed = object.respond_to?(:instances_structure)
+          structure_available = File.exist?(structure_filename)
+
+          if structure_needed && structure_available
+            structure = original_read.bind(self).call(structure_filename)
+            object.instances_structure = structure
+          end
+
+          object
+        end
+
+        define_method(:write) do |filename, object|
+          structure_needed = object.respond_to?(:instances_structure)
+
+          if structure_needed && object.instances_structure
+            structure_filename = structure_file(filename)
+            structure = object.instances_structure
+            original_write.bind(self).call(structure_filename, structure)
+          end
+
+          original_write.bind(self).call(filename, object)
+        end
+
         alias deserialize read
         alias serialize   write
+
+        private
+
+        def structure_file(filename)
+          "#{filename}.#{STRUCTURE_FILE_EXTENSION}"
+        end
       end
     end
   end

data/lib/weka/exceptions.rb

@@ -3,4 +3,6 @@ module Weka
 
   class UnassignedClassError < Error; end
   class UnassignedTrainingInstancesError < Error; end
+  class MissingInstancesStructureError < Error; end
+  class InvalidInstancesStructureError < Error; end
 end

data/lib/weka/version.rb

@@ -1,3 +1,3 @@
 module Weka
-  VERSION = '0.5.0'.freeze
+  VERSION = '0.6.0'.freeze
 end

data/weka.gemspec

@@ -23,10 +23,11 @@ Gem::Specification.new do |spec|
   spec.executables   = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
   spec.require_paths = ['lib']
 
-  spec.add_runtime_dependency     'lock_jar',         '~> 0.13'
-
   spec.add_development_dependency 'bundler',          '~> 1.6'
   spec.add_development_dependency 'rake',             '~> 10.0'
   spec.add_development_dependency 'rspec',            '~> 3.0'
   spec.add_development_dependency 'shoulda-matchers', '~> 3.0'
+
+  spec.add_runtime_dependency 'jar-dependencies', '~> 0.3.11'
+  spec.requirements << 'jar nz.ac.waikato.cms.weka, weka-dev, 3.7.13'
 end

metadata

@@ -1,29 +1,15 @@
 --- !ruby/object:Gem::Specification
 name: weka
 version: !ruby/object:Gem::Version
-  version: 0.5.0
+  version: 0.6.0
 platform: java
 authors:
 - Paul Götze
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2017-06-17 00:00:00.000000000 Z
+date: 2017-12-17 00:00:00.000000000 Z
 dependencies:
-- !ruby/object:Gem::Dependency
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.13'
-  name: lock_jar
-  prerelease: false
-  type: :runtime
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.13'
 - !ruby/object:Gem::Dependency
   requirement: !ruby/object:Gem::Requirement
     requirements:
@@ -80,6 +66,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '3.0'
+- !ruby/object:Gem::Dependency
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.3.11
+  name: jar-dependencies
+  prerelease: false
+  type: :runtime
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.3.11
 description: A JRuby wrapper for the Weka library (http://www.cs.waikato.ac.nz/ml/weka/)
 email:
 - paul.christoph.goetze@gmail.com
@@ -93,10 +93,9 @@ files:
 - ".rspec"
 - ".rubocop.yml"
 - ".travis.yml"
+- CHANGELOG.md
 - CODE_OF_CONDUCT.md
 - Gemfile
-- Jarfile
-- Jarfile.lock
 - MIT-LICENSE.txt
 - README.md
 - Rakefile
@@ -142,7 +141,6 @@ files:
 - lib/weka/filters/unsupervised/attribute.rb
 - lib/weka/filters/unsupervised/instance.rb
 - lib/weka/filters/utils.rb
-- lib/weka/jars.rb
 - lib/weka/version.rb
 - weka.gemspec
 homepage: https://github.com/paulgoetze/weka-jruby
@@ -163,9 +161,10 @@ required_rubygems_version: !ruby/object:Gem::Requirement
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
-requirements: []
+requirements:
+- jar nz.ac.waikato.cms.weka, weka-dev, 3.7.13
 rubyforge_project:
-rubygems_version: 2.6.11
+rubygems_version: 2.6.14
 signing_key:
 specification_version: 4
 summary: Machine Learning & Data Mining with JRuby.

data/Jarfile

@@ -1 +0,0 @@
-jar 'nz.ac.waikato.cms.weka:weka-dev:jar:3.7.13'

data/Jarfile.lock

@@ -1,17 +0,0 @@
----
-version: 0.13.0
-groups:
-  default:
-    dependencies:
-    - nz.ac.waikato.cms.weka.thirdparty:java-cup-11b-runtime:jar:2015.03.26
-    - nz.ac.waikato.cms.weka.thirdparty:java-cup-11b:jar:2015.03.26
-    - nz.ac.waikato.cms.weka:weka-dev:jar:3.7.13
-    - org.pentaho.pentaho-commons:pentaho-package-manager:jar:1.0.11
-    artifacts:
-    - jar:nz.ac.waikato.cms.weka:weka-dev:jar:3.7.13:
-        transitive:
-          nz.ac.waikato.cms.weka.thirdparty:java-cup-11b:jar:2015.03.26: {}
-          org.pentaho.pentaho-commons:pentaho-package-manager:jar:1.0.11: {}
-          nz.ac.waikato.cms.weka.thirdparty:java-cup-11b-runtime:jar:2015.03.26: {}
-remote_repositories:
-- http://repo1.maven.org/maven2/

data/lib/weka/jars.rb

@@ -1,16 +0,0 @@
-module Weka
-  module Jars
-    def self.included(base)
-      base.class_eval do
-        require 'lock_jar'
-
-        lib_path = File.expand_path('../../', File.dirname(__FILE__))
-        lockfile = File.join(lib_path, 'Jarfile.lock')
-        jars_dir = File.join(lib_path, 'jars')
-
-        LockJar.install(lockfile, local_repo: jars_dir)
-        LockJar.load(lockfile, local_repo: jars_dir)
-      end
-    end
-  end
-end