scoruby 0.2.8 → 0.2.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: f78c07d033cba7c4f13531ae30e51f9c00b33d11
-  data.tar.gz: d487653a5f8604e669db89a8766b4e84d3e1950b
+SHA256:
+  metadata.gz: 3af72ce9c41b6f116e34af71b5ad3264d8043c6fd1aaa3544a0a362bff096c0f
+  data.tar.gz: de77039296cfd39f2ef036372e90ad2b4028180381a153380d77164ccf02a743
 SHA512:
-  metadata.gz: d64f07f42472a9ae54435d26d738f05381a3ae2542c4ad2fdfdf879696e0af7c6bd8df6d946fb4aafed2e316fcc00c09c75b164c6a1d56b11ea5cd2ed48b82a5
-  data.tar.gz: 5ffde48fb242ef57f3c06f089aecfe73b5b9b8d9038a75caae516a368619aba3f1d597ddf664783346c538f26fbe9756efc02d20036900d2d94abd0fb7219204
+  metadata.gz: c827de777f0b1b0d83af165a6079a6a2cff83d2612758c78171328cc2811cf718532fef6ff628a04a38f6ef9179c60dae529ee70a7daa90e5aed0e38462cf370
+  data.tar.gz: eff6fcc12bbe064b5a1af1a272c8b544af87c8ae9aa48e7f59a506abcaf292d3b72a235c835caec524eb5ae04a66e2f37ca12691735ce9c1bda19489c1fff385

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    scoruby (0.2.7)
+    scoruby (0.2.8)
       nokogiri (~> 1.7)
 
 GEM

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

@@ -0,0 +1,45 @@
+# frozen_string_literal: true
+
+module Scoruby
+  module Models
+    module RandomForest
+      class Data
+        RF_FOREST_XPATH = 'PMML/MiningModel/Segmentation/Segment'
+        FEATURES_XPATH = 'PMML/DataDictionary/DataField'
+
+        def initialize(xml)
+          @xml = xml
+        end
+
+        def decision_trees
+          @xml.xpath(RF_FOREST_XPATH).map do |xml_tree|
+            DecisionTree.new(xml_tree)
+          end
+        end
+
+        def categorical_features
+          categorical_predicates.each_with_object(Hash.new([])) do |xml, res|
+            predicate = Predicates::SimpleSetPredicate.new(xml)
+            res[predicate.field] = res[predicate.field] | predicate.array
+          end
+        end
+
+        def continuous_features
+          continuous_predicates.map do |xml|
+            Predicates::SimplePredicate.new(xml).field
+          end.uniq
+        end
+
+        private
+
+        def categorical_predicates
+          @xml.xpath('//SimpleSetPredicate')
+        end
+
+        def continuous_predicates
+          @xml.xpath('//SimplePredicate')
+        end
+      end
+    end
+  end
+end

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

@@ -0,0 +1,44 @@
+# frozen_string_literal: true
+
+require 'scoruby/models/random_forest/data'
+
+module Scoruby
+  module Models
+    module RandomForest
+      class Model
+        def initialize(xml)
+          @decision_trees = Data.new(xml).decision_trees
+        end
+
+        def predict(features)
+          decisions_count = decisions_count(features)
+          decision = decisions_count.max_by { |_, v| v }
+          {
+            label: decision[0],
+            score: decision[1] / decisions_count.values.reduce(0, :+).to_f
+          }
+        end
+
+        def decisions_count(features)
+          formatted_features = Features.new(features).formatted
+          decisions = traverse_trees(formatted_features)
+          aggregate_decisions(decisions)
+        end
+
+        private
+
+        def traverse_trees(formatted_features)
+          @decision_trees.map do |decision_tree|
+            decision_tree.decide(formatted_features).score
+          end
+        end
+
+        def aggregate_decisions(decisions)
+          decisions.each_with_object(Hash.new(0)) do |score, counts|
+            counts[score] += 1
+          end
+        end
+      end
+    end
+  end
+end

data/lib/scoruby/models_factory.rb

@@ -2,7 +2,7 @@
 
 require 'scoruby/models/decision_tree'
 require 'scoruby/models/gbm'
-require 'scoruby/models/random_forest'
+require 'scoruby/models/random_forest/model'
 require 'scoruby/models/naive_bayes/model'
 
 module Scoruby
@@ -12,7 +12,7 @@ module Scoruby
     MODEL_NOT_SUPPORTED_ERROR = 'model not supported'
 
     def self.factory_for(xml)
-      return Models::RandomForest.new(xml) if random_forest?(xml)
+      return Models::RandomForest::Model.new(xml) if random_forest?(xml)
       return Models::Gbm.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/predicates/simple_set_predicate.rb

@@ -5,7 +5,7 @@ module Scoruby
     class SimpleSetPredicate
       IS_IN = 'isIn'
 
-      attr_reader :field
+      attr_reader :field, :array
 
       def initialize(pred_xml)
         attributes = pred_xml.attributes

data/lib/scoruby/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Scoruby
-  VERSION = '0.2.8'
+  VERSION = '0.2.9'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: scoruby
 version: !ruby/object:Gem::Version
-  version: 0.2.8
+  version: 0.2.9
 platform: ruby
 authors:
 - Asaf Schers
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-12-02 00:00:00.000000000 Z
+date: 2017-12-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -134,7 +134,8 @@ files:
 - lib/scoruby/models/gbm.rb
 - lib/scoruby/models/naive_bayes/model.rb
 - lib/scoruby/models/naive_bayes/model_data.rb
-- lib/scoruby/models/random_forest.rb
+- lib/scoruby/models/random_forest/data.rb
+- lib/scoruby/models/random_forest/model.rb
 - lib/scoruby/models_factory.rb
 - lib/scoruby/node.rb
 - lib/scoruby/predicate_factory.rb
@@ -168,9 +169,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.6.13
+rubygems_version: 2.7.3
 signing_key: 
 specification_version: 4
 summary: Ruby Scoring API for PMML.
 test_files: []
-has_rdoc: 

data/lib/scoruby/models/random_forest.rb

@@ -1,45 +0,0 @@
-# frozen_string_literal: true
-
-module Scoruby
-  module Models
-    class RandomForest
-      RF_FOREST_XPATH = 'PMML/MiningModel/Segmentation/Segment'
-
-      def initialize(xml)
-        xml_trees       = xml.xpath(RF_FOREST_XPATH)
-        @decision_trees = xml_trees.map do |xml_tree|
-          DecisionTree.new(xml_tree)
-        end
-      end
-
-      def predict(features)
-        decisions_count = decisions_count(features)
-        decision = decisions_count.max_by { |_, v| v }
-        {
-          label: decision[0],
-          score: decision[1] / decisions_count.values.reduce(0, :+).to_f
-        }
-      end
-
-      def decisions_count(features)
-        formatted_features = Features.new(features).formatted
-        decisions = traverse_trees(formatted_features)
-        aggregate_decisions(decisions)
-      end
-
-      private
-
-      def traverse_trees(formatted_features)
-        @decision_trees.map do |decision_tree|
-          decision_tree.decide(formatted_features).score
-        end
-      end
-
-      def aggregate_decisions(decisions)
-        decisions.each_with_object(Hash.new(0)) do |score, counts|
-          counts[score] += 1
-        end
-      end
-    end
-  end
-end