scoruby 0.2.13 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 0594a9d84f41423a9e7b3c611a713dc7eca9a10c
-  data.tar.gz: 723ccca30622ea47a3a1f22b4cc1a00ca3ba931b
+  metadata.gz: fb645c7c145b4d3fe4e72f759243bf473bd901e9
+  data.tar.gz: e0e5fe27be6ed36ae84567a82fd31fe31f9ce6ea
 SHA512:
-  metadata.gz: 96e4a373f3884d8e834d3c22b86477738d6ae9cecbd6b16195adcc336c92e72c0320f20bfbad96ec933692a39c931555ee23f21f2e87f082924c1a1b81fa4eec
-  data.tar.gz: 64932827e238356bd6535e5b2f497ad175fe977a09dbda3c5aee2382827cb5db4c2268a23607de36534bbe975ba9ace631344a0c3209a306696f5ae1a0908e8c
+  metadata.gz: e4492e5f7f635d61983bac4d95c13e29c7d965e3b56dc7c68bf4d4128cb26a7385bd45098fe809da513afc56786f840308688fc2769c3e2316277ccc8fd58205
+  data.tar.gz: a7923b40d7591ca56a286c2c72fb232bc335fb2079399210d4d67fb76bb1a9af27d23af79a6f6b3659bf11d3bc0553333c7fa719d48bfd9d19dc8537bae5ff6c

data/.codeclimate.yml

@@ -0,0 +1,3 @@
+plugins:
+  rubocop:
+    enabled: true

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    scoruby (0.2.13)
+    scoruby (0.3.0)
       nokogiri (~> 1.7)
 
 GEM

data/README.md

@@ -51,7 +51,7 @@ features =  {
         Embarked: 'Q'       
     }
 
-random_forest.predict(features)
+random_forest.score(features)
 
 => {:label=>"0", :score=>0.882}
 

data/lib/scoruby/model_factory.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 require 'scoruby/models/decision_tree'
-require 'scoruby/models/gbm'
+require 'scoruby/models/gradient_boosted_model/model'
 require 'scoruby/models/random_forest/model'
 require 'scoruby/models/naive_bayes/model'
 
@@ -14,7 +14,7 @@ module Scoruby
 
     def self.factory_for(xml)
       return Models::RandomForest::Model.new(xml) if random_forest?(xml)
-      return Models::Gbm.new(xml) if gbm?(xml)
+      return Models::GradientBoostedModel::Model.new(xml) if gbm?(xml)
       return Models::DecisionTree.new(xml.child) if decision_tree?(xml)
       return Models::NaiveBayes::Model.new(xml) if naive_bayes?(xml)
 

data/lib/scoruby/models/gradient_boosted_model/data.rb

@@ -0,0 +1,64 @@
+# frozen_string_literal: true
+
+module Scoruby
+  module Models
+    module GradientBoostedModel
+      class Data
+        GBM_FOREST_XPATH = '//Segmentation[@multipleModelMethod="sum"]/Segment'
+        CONST_XPATH      = '//Target/@rescaleConstant'
+        CONST_XPATH_4_2  = '//Constant'
+
+        def initialize(xml)
+          @xml = xml
+        end
+
+        def decision_trees
+          @decision_trees ||= @xml.xpath(GBM_FOREST_XPATH).map do |xml_tree|
+            DecisionTree.new(xml_tree)
+          end
+        end
+
+        def const
+          @const ||= const_by_version
+        end
+
+        def continuous_features
+          @continuous_features ||= fetch_continuous_features
+        end
+
+        def categorical_features
+          @categorical_features ||= fetch_categorical_features
+        end
+
+        private
+
+        def fetch_continuous_features
+          @xml.xpath('//DataField')
+              .select { |xml| xml.attr('optype') == 'continuous' }
+              .map { |xml| xml.attr('name') }
+        end
+
+        def fetch_categorical_features
+          @xml.xpath('//DataField')
+              .select { |xml| xml.attr('optype') == 'categorical' }
+              .reject { |xml| xml.attr('name') == target }
+              .each_with_object(Hash.new([])) do |xml, res|
+            res[xml.attr('name').to_sym] = xml.xpath('Value')
+                                              .map { |xml| xml.attr('value') }
+          end
+        end
+
+        def target
+          @target ||= @xml.xpath('//MiningField')
+                          .find { |xml| xml.attr('usageType') == 'target' }
+                          .attr('name').to_s
+        end
+
+        def const_by_version
+          return Float(@xml.xpath(CONST_XPATH).to_s) if ModelFactory.gbm_4_3?(@xml)
+          Float(@xml.xpath(CONST_XPATH_4_2).first.content)
+        end
+      end
+    end
+  end
+end

data/lib/scoruby/models/gradient_boosted_model/model.rb

@@ -0,0 +1,34 @@
+# frozen_string_literal: true
+
+require 'scoruby/features'
+require 'forwardable'
+require 'scoruby/models/gradient_boosted_model/data'
+
+module Scoruby
+  module Models
+    module GradientBoostedModel
+      class Model
+        extend Forwardable
+        def_delegators :@data, :decision_trees, :const, :continuous_features,
+                       :categorical_features
+
+        def initialize(xml)
+          @data = Data.new(xml)
+        end
+
+        def score(features)
+          formatted_features = Features.new(features).formatted
+          scores = traverse_trees(formatted_features)
+          sum = scores.reduce(:+) + const
+          Math.exp(sum) / (1 + Math.exp(sum))
+        end
+
+        def traverse_trees(formatted_features)
+          decision_trees.map do |dt|
+            dt.decide(formatted_features).score.to_s.to_f
+          end
+        end
+      end
+    end
+  end
+end

data/lib/scoruby/models/random_forest/data.rb

@@ -12,7 +12,7 @@ module Scoruby
         end
 
         def decision_trees
-          @xml.xpath(RF_FOREST_XPATH).map do |xml_tree|
+          @decision_trees ||= @xml.xpath(RF_FOREST_XPATH).map do |xml_tree|
             DecisionTree.new(xml_tree)
           end
         end

data/lib/scoruby/models/random_forest/model.rb

@@ -8,13 +8,14 @@ module Scoruby
     module RandomForest
       class Model
         extend Forwardable
-        def_delegators :@data, :decision_trees, :categorical_features, :continuous_features
+        def_delegators :@data, :decision_trees, :categorical_features,
+                       :continuous_features
 
         def initialize(xml)
           @data = Data.new(xml)
         end
 
-        def predict(features)
+        def score(features)
           decisions_count = decisions_count(features)
           decision = decisions_count.max_by { |_, v| v }
           {

data/lib/scoruby/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Scoruby
-  VERSION = '0.2.13'
+  VERSION = '0.3.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: scoruby
 version: !ruby/object:Gem::Version
-  version: 0.2.13
+  version: 0.3.0
 platform: ruby
 authors:
 - Asaf Schers
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-01-27 00:00:00.000000000 Z
+date: 2018-02-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -115,6 +115,7 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- ".codeclimate.yml"
 - ".gitignore"
 - ".rspec"
 - ".rubocop.yml"
@@ -132,7 +133,8 @@ files:
 - lib/scoruby/features.rb
 - lib/scoruby/model_factory.rb
 - lib/scoruby/models/decision_tree.rb
-- lib/scoruby/models/gbm.rb
+- lib/scoruby/models/gradient_boosted_model/data.rb
+- lib/scoruby/models/gradient_boosted_model/model.rb
 - lib/scoruby/models/naive_bayes/model.rb
 - lib/scoruby/models/naive_bayes/model_data.rb
 - lib/scoruby/models/random_forest/data.rb

data/lib/scoruby/models/gbm.rb

@@ -1,41 +0,0 @@
-# frozen_string_literal: true
-
-require 'scoruby/models/decision_tree'
-require 'scoruby/features'
-
-module Scoruby
-  module Models
-    class Gbm
-      GBM_FOREST_XPATH = '//Segmentation[@multipleModelMethod="sum"]/Segment'
-      CONST_XPATH      = '//Target/@rescaleConstant'
-      CONST_XPATH_4_2  = '//Constant'
-
-      def initialize(xml)
-        @decision_trees = xml.xpath(GBM_FOREST_XPATH).map do |xml_tree|
-          DecisionTree.new(xml_tree)
-        end
-        @const = const(xml)
-      end
-
-      def tree_count
-        @decision_trees.count
-      end
-
-      def score(features)
-        formatted_features = Features.new(features).formatted
-        scores = @decision_trees.map do |dt|
-          dt.decide(formatted_features).score.to_s.to_f
-        end
-        sum = scores.reduce(:+) + @const
-        Math.exp(sum) / (1 + Math.exp(sum))
-      end
-
-      private
-
-      def const(xml)
-        return Float(xml.xpath(CONST_XPATH).to_s) if ModelFactory.gbm_4_3?(xml)
-        Float(xml.xpath(CONST_XPATH_4_2).first.content)
-      end
-    end
-  end
-end