thera 0.0.1

data/lib/quarry_rb/classifier/naive_bayes_classifier.rb

@@ -0,0 +1,10 @@
+module Quarry
+  module Classifier
+    class NaiveBayesClassifier < Classifier
+      def initialize(data_set)
+        @classifier = Quarry::Classifier::ImplNaiveBayesClassifier.new(data_set.data_set)
+        super(data_set)
+      end
+    end
+  end
+end

data/lib/quarry_rb/confusion_matrix.rb

@@ -0,0 +1,58 @@
+module Quarry
+  class ConfusionMatrix
+    attr_reader :confusion_matrix
+    attr_reader :data_set
+    def initialize(data_set, matrix=nil)
+      @data_set = data_set
+      if matrix
+        @confusion_matrix = matrix
+      else
+        @confusion_matrix = Quarry::ImplConfusionMatrix.new(data_set.data_set)
+      end
+    end
+    
+    def add(predicted, actual)
+      @confusion_matrix.add(predicted.category, actual.category)
+    end
+    
+    def accuracy
+      @confusion_matrix.accuracy
+    end
+    
+    def error
+      @confusion_matrix.error
+    end
+    
+    def tp(category)
+      @confusion_matrix.tp(category.category)
+    end
+    
+    def fp(category)
+      @confusion_matrix.fp(category.category)
+    end
+    
+    def tn(category)
+      @confusion_matrix.tn(category.category)
+    end
+    
+    def fn(category)
+      @confusion_matrix.fn(category.category)
+    end
+    
+    def precision(category)
+      @confusion_matrix.precision(category)
+    end
+    
+    def recall(category)
+      @confusion_matrix.recall(category)
+    end
+    
+    def fscore(category)
+      @confusion_matrix.fscore(category)
+    end
+    
+    def print_summary
+      @confusion_matrix.print_summary
+    end
+  end
+end

data/lib/quarry_rb/data_set/data_set.rb

@@ -0,0 +1,42 @@
+module Quarry
+  module DataSet
+    class DataSet
+      attr_reader :data_set
+      def initialize(data_set=nil)
+        @data_set = data_set || ImplDataSet.new()
+      end
+      
+      def name
+        @data_set.get_name
+      end
+      
+      def name=(new_name)
+        @data_set.set_name(new_name)
+      end
+      
+      def examples
+        @examples ||= EnumerableHelper.new(self, @data_set, Example, :examples_size, :get_example_by_index)
+      end
+      
+      def categories
+        @categories ||= EnumerableHelper.new(self, @data_set, Category, :categories_size, :get_category_by_index)
+      end
+      
+      def features
+        @features ||= EnumerableHelper.new(self, @data_set, Feature, :features_size, :get_feature_by_index)
+      end
+      
+      def stratify(classifier, folds, skip_fold)
+        @data_set.stratify(classifier.classifier, folds, skip_fold)
+      end
+      
+      def classify_fold(classifier, folds, fold)
+        ConfusionMatrix.new(self, @data_set.classify_fold(classifier.classifier, folds, fold))
+      end
+      
+      def cross_fold_validation(classifier, folds)
+        ConfusionMatrix.new(self, @data_set.cross_fold_validation(classifier.classifier, folds))
+      end
+    end
+  end
+end

data/lib/quarry_rb/data_set/example.rb

@@ -0,0 +1,33 @@
+module Quarry
+  module DataSet
+    class Example
+      attr_reader :example
+      
+      def initialize(example, data_set)
+        @data_set = data_set
+        @example  = example
+      end
+      
+      def [](index)
+        @example.get_value(index)
+      end
+      
+      def []=(index, value)
+        @example.set_value(index, value)
+      end
+      
+      def category
+        @data_set.categories[@example.get_category_index]
+      end
+      
+      def category_index
+        @example.get_category_index
+      end
+      
+      def category=(new_category)
+        raise "new_category must be an instance of Quarry::DataSet::Category" unless new_category.is_a?(Category)
+        @example.set_category(new_category.category)
+      end
+    end
+  end
+end

data/lib/quarry_rb/data_set/feature.rb

@@ -0,0 +1,28 @@
+module Quarry
+  module DataSet
+    class Feature
+      attr_reader :feature
+      
+      def initialize(feature, data_set)
+        @data_set = data_set
+        @feature  = feature
+      end
+      
+      def name
+        @feature.get_name
+      end
+      
+      def name=(new_name)
+        @feature.set_name(new_name)
+      end
+      
+      def type
+        @feature.get_type
+      end
+      
+      def type=(new_type)
+        @feature.set_type(new_type)
+      end
+    end
+  end
+end

data/lib/quarry_rb/enumerable_helper.rb

@@ -0,0 +1,32 @@
+module Quarry
+  class EnumerableHelper
+    include Enumerable
+    attr_reader :size
+    
+    def initialize(data_set, container, klass, size, get_fn)
+      @size = container.send(size)
+      @container = container
+      @data_set = data_set
+      @get_fn = get_fn
+      @klass = klass
+      
+      # because of the way the Ruby GC works, it's easier to store
+      # references to enumerated objects here than from the C++ side.
+      # by keeping a reference to returned objects in this object,
+      # iterated objects that shouldn't be released (e.g Examples)
+      # won't be until the data set is released.
+      @objects = Hash.new {|hash, index| hash[index] = @klass.new(@container.send(@get_fn, index), @data_set)}
+    end
+    
+    def [](index)
+      return nil if (index < 0) || (index >= @size)
+      @objects[index]
+    end
+    
+    def each
+      (0...@size).each do |index|
+        yield @objects[index]
+      end
+    end
+  end
+end

data/lib/quarry_rb/model/model.rb

@@ -0,0 +1,56 @@
+module Quarry
+  class Model
+    attr_reader :model
+    def initialize(m = nil)
+      @model = m || Quarry::ImplModel.new
+    end
+    
+    def data_set
+      Quarry::DataSet::DataSet.new(@model.get_data_set)
+    end
+    
+    def data_set=(ds)
+      @model.set_data_set(ds.data_set)
+    end
+    
+    # def classifier
+    #   @model.get_classifier
+    # end
+
+    def classifier=(c)
+      @model.set_classifier(c.classifier)
+    end
+    
+    def text_pipeline
+      TextPipeline.new(@model.get_text_pipeline)
+    end
+    
+    def text_pipeline=(t)
+      @model.set_text_pipeline(t.text_pipeline)
+    end
+    
+    def train(example)
+      @model.train(example)
+    end
+    
+    def train_text(text)
+      @model.train_text(text)
+    end
+    
+    def classify(example)
+      @model.classify(example)
+    end
+    
+    def classify_text(text)
+      @model.classify_text(text)
+    end
+    
+    def rank(example)
+      @model.rank(example)
+    end
+    
+    def rank_text(text)
+      @model.rank_text(text)
+    end
+  end
+end

data/lib/quarry_rb/storage/arff.rb

@@ -0,0 +1,11 @@
+module Quarry
+  class ARFF
+    def initialize(path)
+      @arff = Quarry::ImplARFF.new(path)
+    end
+    
+    def read
+      Quarry::DataSet::DataSet.new(@arff.read)
+    end
+  end
+end

data/lib/quarry_rb/storage/binary.rb

@@ -0,0 +1,23 @@
+module Quarry
+  class Binary
+    def initialize(path)
+      @binary = Quarry::ImplBinary.new(path)
+    end
+    
+    def read
+      Quarry::DataSet::DataSet.new(@binary.read)
+    end
+    
+    def write(data_set)
+      @binary.write(data_set.data_set)
+    end
+    
+    def read_model
+      Quarry::Model.new(@binary.read_model)
+    end
+    
+    def write_model(model)
+      @binary.write_model(model.model)
+    end
+  end
+end

data/lib/quarry_rb/storage/folders.rb

@@ -0,0 +1,11 @@
+module Quarry
+  class Folders
+    def initialize(path, pipeline)
+      @folders = Quarry::ImplFolders.new(path, pipeline.text_pipeline)
+    end
+    
+    def read
+      Quarry::DataSet::DataSet.new(@folders.read)
+    end
+  end
+end

data/lib/quarry_rb/text_pipeline.rb

@@ -0,0 +1,16 @@
+module Quarry
+  class TextPipeline
+    attr_reader :text_pipeline
+    def initialize(tp = nil)
+      @text_pipeline = tp || Quarry::ImplTextPipeline.new
+    end
+    
+    def process_text(data_set, text)
+      Example.new(@text_pipeline.process_text(data_set, text))
+    end
+    
+    def self.standard_pipeline
+      new(Quarry::Preprocessing::Text.standard_pipeline)
+    end
+  end
+end

data/lib/thera.rb

@@ -0,0 +1,20 @@
+require 'quarry_toolkit'
+Dir.chdir(File.dirname(__FILE__)) do
+  require './quarry_rb/text_pipeline'
+  
+  require './quarry_rb/model/model'
+  
+  require './quarry_rb/storage/arff'
+  require './quarry_rb/storage/binary'
+  require './quarry_rb/storage/folders'
+  
+  require './quarry_rb/enumerable_helper'
+  require './quarry_rb/confusion_matrix'
+  
+  require './quarry_rb/data_set/data_set'
+  require './quarry_rb/data_set/example'
+  require './quarry_rb/data_set/feature'
+  
+  require './quarry_rb/classifier/classifier'
+  require './quarry_rb/classifier/naive_bayes_classifier'
+end

data/test/helper.rb

@@ -0,0 +1,19 @@
+require 'rubygems'
+require 'bundler'
+begin
+  Bundler.setup(:default, :development)
+rescue Bundler::BundlerError => e
+  $stderr.puts e.message
+  $stderr.puts "Run `bundle install` to install missing gems"
+  exit e.status_code
+end
+require 'test/unit'
+require 'shoulda'
+
+$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'ext'))
+$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
+$LOAD_PATH.unshift(File.dirname(__FILE__))
+require 'quarry'
+
+class Test::Unit::TestCase
+end

data/test/test_quarry.rb

@@ -0,0 +1,33 @@
+require 'helper'
+
+class TestQuarry < Test::Unit::TestCase
+  should 'load' do
+    require 'quarry'
+    Quarry
+  end
+  
+  should 'expose ARFF' do
+    Quarry::ARFF
+  end
+  
+  should 'allow creating ARFF objects' do
+    a = Quarry::ARFF.new("/Users/will/dev/classifier/src/categories.arff")
+  end
+  
+  should 'read ARFF objects' do
+    a = Quarry::ARFF.new("/Users/will/Desktop/test_reduced.arff")
+    ds = a.read
+    
+    puts "\nFeatures: #{ds.features.size}, examples: #{ds.examples.size}"
+    classifier = Quarry::Classifier::NaiveBayesClassifier.new(ds)
+    print "Created classifier\n"
+    cm = ds.cross_fold_validation(classifier, 2)
+    print "Cross fold done\n"
+    cm.print_summary
+    
+    #ds.stratify(classifier, 10, 0)
+    #classifier.classify(ds.examples[0])
+    #cm = ds.classify_fold(classifier, 10, 0)
+    #cm.print_summary
+  end
+end

data/thera.gemspec

@@ -0,0 +1,21 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path('../lib', __FILE__)
+
+Gem::Specification.new do |s|
+  s.name = "thera"
+  s.homepage = "http://github.com/willcannings/quarry"
+  s.license = "Public Domain"
+  s.summary = "Ruby Data Mining Library"
+  s.description = "C++ Data Mining Library for Ruby"
+  s.email = "me@willcannings.com"
+  s.authors = ["Will Cannings"]
+  s.version = '0.0.1'
+  s.extensions = ["ext/extconf.rb"]
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = ['lib', 'ext']
+  
+  s.add_runtime_dependency 'rice'
+end

metadata

@@ -0,0 +1,148 @@
+--- !ruby/object:Gem::Specification
+name: thera
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+  prerelease: 
+platform: ruby
+authors:
+- Will Cannings
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-12-07 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rice
+  requirement: &70234812248380 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: *70234812248380
+description: C++ Data Mining Library for Ruby
+email: me@willcannings.com
+executables: []
+extensions:
+- ext/extconf.rb
+extra_rdoc_files: []
+files:
+- .document
+- .gitignore
+- Gemfile
+- Gemfile.lock
+- LICENSE.txt
+- README.rdoc
+- Rakefile
+- ext/Makefile
+- ext/extconf.rb
+- ext/quarry/quarry_toolkit.cpp
+- lib/quarry/Makefile.linux
+- lib/quarry/Makefile.osx
+- lib/quarry/Makefile.targets
+- lib/quarry/obj/.gitkeep
+- lib/quarry/src/classifier/aode/aode_classifier.cpp
+- lib/quarry/src/classifier/aode/aode_classifier.h
+- lib/quarry/src/classifier/centroid/centroid_classifier.cpp
+- lib/quarry/src/classifier/centroid/centroid_classifier.h
+- lib/quarry/src/classifier/classifier.cpp
+- lib/quarry/src/classifier/classifier.h
+- lib/quarry/src/classifier/knn/knn_classifier.cpp
+- lib/quarry/src/classifier/knn/knn_classifier.h
+- lib/quarry/src/classifier/multinomial_bayes/multinomial_bayes_classifier.cpp
+- lib/quarry/src/classifier/multinomial_bayes/multinomial_bayes_classifier.h
+- lib/quarry/src/classifier/naive_bayes/naive_bayes_classifier.cpp
+- lib/quarry/src/classifier/naive_bayes/naive_bayes_classifier.h
+- lib/quarry/src/data_set/data_set.cpp
+- lib/quarry/src/data_set/data_set.h
+- lib/quarry/src/data_set/dense/dense_data_set.h
+- lib/quarry/src/data_set/dense/dense_example.h
+- lib/quarry/src/data_set/example.cpp
+- lib/quarry/src/data_set/example.h
+- lib/quarry/src/data_set/feature.h
+- lib/quarry/src/data_set/features/nominal_feature.cpp
+- lib/quarry/src/data_set/features/nominal_feature.h
+- lib/quarry/src/data_set/features/numeric_feature.cpp
+- lib/quarry/src/data_set/features/numeric_feature.h
+- lib/quarry/src/data_set/sparse/sparse_data_set.h
+- lib/quarry/src/data_set/sparse/sparse_example.cpp
+- lib/quarry/src/data_set/sparse/sparse_example.h
+- lib/quarry/src/metrics/confusion_matrix.cpp
+- lib/quarry/src/metrics/confusion_matrix.h
+- lib/quarry/src/model/model.cpp
+- lib/quarry/src/model/model.h
+- lib/quarry/src/preprocessing/examples/example_preprocessor.h
+- lib/quarry/src/preprocessing/examples/weights/binary_weight.h
+- lib/quarry/src/preprocessing/examples/weights/local_weight.h
+- lib/quarry/src/preprocessing/text/example_generator/example_generator.h
+- lib/quarry/src/preprocessing/text/example_generator/token_counter.h
+- lib/quarry/src/preprocessing/text/inplace_processor/downcase.h
+- lib/quarry/src/preprocessing/text/inplace_processor/inplace_processor.h
+- lib/quarry/src/preprocessing/text/inplace_processor/porter_stemmer.h
+- lib/quarry/src/preprocessing/text/inplace_processor/porter_stemmer_original.cpp
+- lib/quarry/src/preprocessing/text/text_pipeline.cpp
+- lib/quarry/src/preprocessing/text/text_pipeline.h
+- lib/quarry/src/preprocessing/text/token_selector/pos_tag_selector.h
+- lib/quarry/src/preprocessing/text/token_selector/stop_words.cpp
+- lib/quarry/src/preprocessing/text/token_selector/stop_words.h
+- lib/quarry/src/preprocessing/text/token_selector/token_selector.h
+- lib/quarry/src/preprocessing/text/tokeniser/simple_tokeniser.cpp
+- lib/quarry/src/preprocessing/text/tokeniser/simple_tokeniser.h
+- lib/quarry/src/preprocessing/text/tokeniser/tokeniser.h
+- lib/quarry/src/quarry.cpp
+- lib/quarry/src/quarry.h
+- lib/quarry/src/storage/arff.cpp
+- lib/quarry/src/storage/arff.h
+- lib/quarry/src/storage/binary.cpp
+- lib/quarry/src/storage/binary.h
+- lib/quarry/src/storage/folders.cpp
+- lib/quarry/src/storage/folders.h
+- lib/quarry/src/storage/storage.h
+- lib/quarry/src/test.cpp
+- lib/quarry_rb/classifier/classifier.rb
+- lib/quarry_rb/classifier/naive_bayes_classifier.rb
+- lib/quarry_rb/confusion_matrix.rb
+- lib/quarry_rb/data_set/data_set.rb
+- lib/quarry_rb/data_set/example.rb
+- lib/quarry_rb/data_set/feature.rb
+- lib/quarry_rb/enumerable_helper.rb
+- lib/quarry_rb/model/model.rb
+- lib/quarry_rb/storage/arff.rb
+- lib/quarry_rb/storage/binary.rb
+- lib/quarry_rb/storage/folders.rb
+- lib/quarry_rb/text_pipeline.rb
+- lib/thera.rb
+- test/helper.rb
+- test/test_quarry.rb
+- thera.gemspec
+homepage: http://github.com/willcannings/quarry
+licenses:
+- Public Domain
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+- ext
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.10
+signing_key: 
+specification_version: 3
+summary: Ruby Data Mining Library
+test_files:
+- test/helper.rb
+- test/test_quarry.rb