intermediary commit

6 files changed, 118 insertions, 59 deletions

diff --git a/algorithm.gemspec b/algorithm.gemspec
index c3119e6..1a94225 100644
--- a/algorithm.gemspec
+++ b/algorithm.gemspec
@@ -20,7 +20,8 @@ Gem::Specification.new do |s|
   # specify any dependencies here; for example:
   #s.add_runtime_dependency "opentox-server"
   s.add_runtime_dependency "opentox-client"
-  s.add_runtime_dependency 'rinruby'#, "~>2.0.2"
+  s.add_runtime_dependency 'rserve-client'#, "~>2.0.2"
+  #s.add_runtime_dependency 'rinruby'#, "~>2.0.2"
   s.add_runtime_dependency 'nokogiri'#, "~>1.4.4"
   s.add_runtime_dependency 'statsample'#, "~>1.1"
   s.add_runtime_dependency 'gsl'#, "~>1.14"
diff --git a/lib/classification.rb b/lib/classification.rb
index f6c9b11..127fa28 100644
--- a/lib/classification.rb
+++ b/lib/classification.rb
@@ -17,10 +17,10 @@ module OpenTox
 
         $logger.debug "Weighted Majority Vote Classification."
 
-        values = neighbors.collect{|n| n[1]}.uniq
+        values = neighbors.collect{|n| n[2]}.uniq
         neighbors.each do |neighbor|
-          neighbor_weight = neighbor[2]
-          activity = values.index(neighbor[1]) + 1 # map values to integers > 1
+          neighbor_weight = neighbor[1]
+          activity = values.index(neighbor[2]) + 1 # map values to integers > 1
           neighbor_contribution += activity * neighbor_weight
           if values.size == 2 # AM: provide compat to binary classification: 1=>false 2=>true
             case activity
@@ -46,9 +46,38 @@ module OpenTox
         $logger.debug "Prediction: '" + prediction.to_s + "'." unless prediction.nil?
         confidence = (confidence_sum/neighbors.size).abs 
         $logger.debug "Confidence: '" + confidence.to_s + "'." unless prediction.nil?
-        return {:prediction => prediction, :confidence => confidence.abs}
+        [prediction, confidence.abs]
       end
 
+      # Local support vector regression from neighbors 
+      # @param [Hash] params Keys `:props, :activities, :sims, :min_train_performance` are required
+      # @return [Numeric] A prediction value.
+      def self.local_svm_classification(params)
+
+        confidence = 0.0
+        prediction = nil
+
+        $logger.debug "Local SVM."
+        if params[:activities].size>0
+          if params[:props]
+            n_prop = params[:props][0].collect.to_a
+            q_prop = params[:props][1].collect.to_a
+            props = [ n_prop, q_prop ]
+          end
+          activities = params[:activities].collect.to_a
+          activities = activities.collect{|v| "Val" + v.to_s} # Convert to string for R to recognize classification
+          prediction = local_svm_prop( props, activities, params[:min_train_performance]) # params[:props].nil? signals non-prop setting
+          prediction = prediction.sub(/Val/,"") if prediction # Convert back
+          confidence = 0.0 if prediction.nil?
+          #$logger.debug "Prediction: '" + prediction.to_s + "' ('#{prediction.class}')."
+          confidence = get_confidence({:sims => params[:sims][1], :activities => params[:activities]})
+        end
+        {:prediction => prediction, :confidence => confidence}
+
+      end
+
+
+
     end
 
   end
diff --git a/lib/descriptor.rb b/lib/descriptor.rb
index 1d43d7d..8ec7480 100644
--- a/lib/descriptor.rb
+++ b/lib/descriptor.rb
@@ -1,7 +1,6 @@
 require 'digest/md5'
 ENV["JAVA_HOME"] ||= "/usr/lib/jvm/java-7-openjdk" 
 BABEL_3D_CACHE_DIR = File.join(File.dirname(__FILE__),"..",'/babel_3d_cache')
-# TODO store 3D structures in mongodb
 # TODO store descriptors in mongodb
 
 module OpenTox
@@ -59,6 +58,7 @@ module OpenTox
         bad_request_error "Compounds for smarts_match are empty" unless compounds
         bad_request_error "Smarts for smarts_match are empty" unless smarts
         parse compounds
+        @count = count
         obconversion = OpenBabel::OBConversion.new
         obmol = OpenBabel::OBMol.new
         obconversion.set_in_format('inchi')
@@ -100,13 +100,19 @@ module OpenTox
             @data_entries
           end
         when "OpenTox::Dataset"
-          dataset = OpenTox::Dataset.new(:compound_ids => @compounds.collect{|c| c.id})
+          dataset = OpenTox::DescriptorDataset.new(:compound_ids => @compounds.collect{|c| c.id})
           if @smarts
             dataset.feature_ids = @smarts.collect{|smart| Smarts.find_or_create_by(:smarts => smart).id}
+            @count ? algo = "count" : algo = "match"
+            dataset.feature_calculation_algorithm = "#{self}.smarts_#{algo}"
+            
           elsif @physchem_descriptors
             dataset.feature_ids = @physchem_descriptors.collect{|d| PhysChemDescriptor.find_or_create_by(:name => d, :creator => __FILE__).id}
             dataset.data_entries = @data_entries
+            dataset.feature_calculation_algorithm = "#{self}.physchem"
+            #TODO params?
           end
+          dataset.save_all
           dataset
         end
       end
diff --git a/lib/lazar.rb b/lib/lazar.rb
index 19f8cdd..399f5c1 100644
--- a/lib/lazar.rb
+++ b/lib/lazar.rb
@@ -23,8 +23,8 @@ module OpenTox
       field :training_dataset_id, type: BSON::ObjectId
       field :feature_dataset_id, type: BSON::ObjectId
       # algorithms
-      field :feature_generation, type: String
-      field :feature_calculation_algorithm, type: String
+      #field :feature_generation, type: String
+      #field :feature_calculation_algorithm, type: String
       field :prediction_algorithm, type: String
       field :similarity_algorithm, type: String
       # prediction features
@@ -34,7 +34,7 @@ module OpenTox
       # parameters
       field :nr_hits, type: Boolean
       field :min_sim, type: Float
-      field :propositionalized, type:Boolean
+      #field :propositionalized, type:Boolean
       field :min_train_performance, type: Float
 
       attr_accessor :prediction_dataset
@@ -54,7 +54,6 @@ module OpenTox
         bad_request_error "No features found in feature dataset #{feature_dataset.id}." if feature_dataset.features.empty?
         lazar.feature_dataset_id = feature_dataset.id
         @training_dataset = training_dataset
-        #@training_dataset = OpenTox::Dataset.find(feature_dataset.parameters.select{|p| p["title"] == "dataset_id"}.first["paramValue"])
         bad_request_error "Training dataset compounds do not match feature dataset compounds. Please ensure that they are in the same order." unless @training_dataset.compounds == feature_dataset.compounds
         lazar.training_dataset_id = @training_dataset.id
 
@@ -73,31 +72,26 @@ module OpenTox
           lazar.prediction_algorithm = params[:prediction_algorithm]
         end
 
-        unless lazar.prediction_algorithm
-          lazar.prediction_algorithm = "OpenTox::Algorithm::Classification.weighted_majority_vote" if prediction_feature.nominal
-          lazar.prediction_algorithm = "OpenTox::Algorithm::Regression.local_svm_regression" if prediction_feature.numeric
+        unless lazar.prediction_algorithm # set defaults
+          # TODO consider params
+          if prediction_feature.nominal
+            lazar.prediction_algorithm = "OpenTox::Algorithm::Classification.weighted_majority_vote"
+            lazar.similarity_algorithm = "OpenTox::Algorithm::Similarity.tanimoto"
+            lazar.min_sim = 0.3 unless lazar.min_sim
+          elsif prediction_feature.numeric
+            lazar.prediction_algorithm = "OpenTox::Algorithm::Regression.local_svm_regression" 
+            lazar.similarity_algorithm = "OpenTox::Algorithm::Similarity.cosine"
+            # cosine similartiy is default
+            lazar.min_sim = 0.7 unless lazar.min_sim
+          end
         end
-        lazar.prediction_algorithm =~ /majority_vote/ ? lazar.propositionalized = false :  lazar.propositionalized = true
+        #lazar.prediction_algorithm =~ /majority_vote/ ? lazar.propositionalized = false :  lazar.propositionalized = true
 
         lazar.min_sim = params[:min_sim].to_f if params[:min_sim] and params[:min_sim].numeric?
+        # TODO: get info from training_dataset
         lazar.nr_hits =  nr_hits
-        lazar.feature_generation = feature_dataset.training_algorithm
+        #lazar.feature_generation = feature_dataset.training_algorithm
         #lazar.parameters << {"title" => "feature_generation_uri", "paramValue" => params[:feature_generation_uri]}
-        if lazar.feature_generation =~ /fminer|bbrc|last/
-          if lazar[:nr_hits] 
-            lazar.feature_calculation_algorithm = "OpenTox::Algorithm::Descriptor.smarts_count"
-          else
-            lazar.feature_calculation_algorithm = "OpenTox::Algorithm::Descriptor.smarts_match"
-          end
-          lazar.similarity_algorithm = "OpenTox::Algorithm::Similarity.tanimoto"
-          lazar.min_sim = 0.3 unless lazar.min_sim
-        elsif lazar.feature_generation =~/descriptor/ or lazar.feature_generation.nil?
-          # cosine similartiy is default (e.g. used when no fetature_generation_uri is given and a feature_dataset_uri is provided instead)
-          lazar.similarity_algorithm = "OpenTox::Algorithm::Similarity.cosine"
-          lazar.min_sim = 0.7 unless lazar.min_sim
-        else
-          bad_request_error "unkown feature generation method #{lazar.feature_generation}"
-        end
 
         bad_request_error "Parameter min_train_performance is not numeric." if params[:min_train_performance] and !params[:min_train_performance].numeric?
         lazar.min_train_performance = params[:min_train_performance].to_f if params[:min_train_performance] and params[:min_train_performance].numeric?
@@ -107,7 +101,7 @@ module OpenTox
         lazar
       end
 
-      def predict params
+      def predict object
 
         # tailored for performance
         # all consistency checks should be done during model creation
@@ -131,20 +125,21 @@ module OpenTox
         @feature_dataset = OpenTox::Dataset.find(feature_dataset_id)
 
         compounds = []
-        if params[:compound]
-          compounds = [ params[:compound]] 
-        elsif params[:compounds]
-          compounds = params[:compounds]
-        elsif params[:dataset]
-          compounds = params[:dataset].compounds
+        case object.class.to_s
+        when "OpenTox::Compound"
+          compounds = [object] 
+        when "Array"
+          compounds = object
+        when "OpenTox::Dataset"
+          compounds = object.compounds
         else 
-          bad_request_error "Please provide one of the parameters: :compound, :compounds, :dataset"
+          bad_request_error "Please provide a OpenTox::Compound an Array of OpenTox::Compounds or an OpenTox::Dataset as parameter."
         end
 
         $logger.debug "Setup: #{Time.now-time}"
         time = Time.now
 
-        @query_fingerprint = Algorithm.run(feature_calculation_algorithm, compounds, @feature_dataset.features.collect{|f| f.smarts} )
+        @query_fingerprint = Algorithm.run(feature_dataset.feature_calculation_algorithm, compounds, @feature_dataset.features.collect{|f| f.name} )
 
         $logger.debug "Fingerprint calculation: #{Time.now-time}"
         time = Time.now
@@ -166,35 +161,59 @@ module OpenTox
             end
             next
           else
-
-            # TODO reintroduce for regression
-            #mtf = OpenTox::Algorithm::Transform::ModelTransformer.new(self)
-            #mtf.transform
-            #
+            t = Time.new
+
+            if prediction_algorithm =~ /Regression/
+              mtf = OpenTox::Algorithm::Transform::ModelTransformer.new(self)
+              mtf.transform
+              training_fingerprints = mtf.n_prop
+              training_activities = mtf.activities
+              p training_activities
+              query_fingerprint = mtf.q_prop
+              neighbors = [[nil,nil,nil,query_fingerprint]]
+            else
+              training_fingerprints = @feature_dataset.data_entries
+              # TODO fix for multi feature datasets
+              training_activities = @training_dataset.data_entries[i].first
+              query_fingerprint = @query_fingerprint[c]
+              neighbors = []
+            end
+            $logger.debug "Transform: #{Time.now-t}"
+            t = Time.new
+            
 
             # find neighbors
-            neighbors = []
-            @feature_dataset.data_entries.each_with_index do |fingerprint, i|
-
-              sim = Algorithm.run(similarity_algorithm,fingerprint, @query_fingerprint[c])
-              # TODO fix for multi feature datasets
-              neighbors << [@feature_dataset.compounds[i],@training_dataset.data_entries[i].first,sim] if sim > self.min_sim
+            training_fingerprints.each_with_index do |fingerprint, i|
+
+              sim = Algorithm.run(similarity_algorithm,fingerprint, query_fingerprint)
+              if sim > self.min_sim
+                if prediction_algorithm =~ /Regression/
+                  neighbors << [@feature_dataset.compounds[i],sim,training_activities[i], fingerprint]
+                else
+                  neighbors << [@feature_dataset.compounds[i],sim,training_activities[i]] 
+                end
+              end
             end
 
-            prediction = Algorithm.run(prediction_algorithm, neighbors)
+            if prediction_algorithm =~ /Regression/
+              prediction = Algorithm.run(prediction_algorithm, neighbors, :min_train_performance => self.min_train_performance)
+            else
+              prediction = Algorithm.run(prediction_algorithm, neighbors)
+            end
 
             $logger.debug "Prediction time: #{Time.now-time}"
             time = Time.now
 
+            p prediction
             # AM: transform to original space (TODO)
-            confidence_value = ((confidence_value+1.0)/2.0).abs if similarity_algorithm =~ /cosine/
+            confidence_value = ((confidence_value+1.0)/2.0).abs if prediction.first and similarity_algorithm =~ /cosine/
 
 
-            $logger.debug "predicted value: #{prediction[:prediction]}, confidence: #{prediction[:confidence]}"
+            $logger.debug "predicted value: #{prediction[0]}, confidence: #{prediction[1]}"
           end
           prediction_dataset.compound_ids << compound
-          prediction_dataset[c,0] = prediction[:prediction]
-          prediction_dataset[c,1] = prediction[:confidence] 
+          prediction_dataset[c,0] = prediction[0]
+          prediction_dataset[c,1] = prediction[1]
 
         end 
         prediction_dataset
diff --git a/lib/opentox-algorithm.rb b/lib/opentox-algorithm.rb
index d768cfd..7743247 100644
--- a/lib/opentox-algorithm.rb
+++ b/lib/opentox-algorithm.rb
@@ -19,5 +19,6 @@ require_relative "fminer.rb"
 require_relative "lazar.rb"
 require_relative "transform.rb"
 require_relative "similarity.rb"
-require_relative "neighbors.rb"
+#require_relative "neighbors.rb"
 require_relative "classification.rb"
+require_relative "regression.rb"
diff --git a/lib/transform.rb b/lib/transform.rb
index fad9517..15b7b60 100644
--- a/lib/transform.rb
+++ b/lib/transform.rb
@@ -231,7 +231,7 @@ module OpenTox
       # Attaches transformations to an OpenTox::Model
       # Stores props, sims, performs similarity calculations
       class ModelTransformer
-        attr_accessor :model, :similarity_algorithm, :activities, :sims
+        attr_accessor :model, :similarity_algorithm, :activities, :sims, :n_prop, :q_prop
 
         # @params[OpenTox::Model] model Model to transform
         def initialize model
@@ -282,6 +282,7 @@ module OpenTox
           @ids = [] # surviving compounds become neighbors
           @sims = [] # calculated by neighbor routine
           
+=begin
           neighbors
           n_prop_tmp = []; @ids.each { |idx| n_prop_tmp << @n_prop[idx] }; @n_prop = n_prop_tmp # select neighbors from matrix
           acts_tmp = []; @ids.each { |idx| acts_tmp << @activities[idx] }; @activities = acts_tmp
@@ -315,6 +316,7 @@ module OpenTox
           $logger.debug "Sims: #{@sims.size}, Acts: #{@activities.size}"
 
           @sims = [ gram_matrix, @sims ] 
+=end
 
         end
 
@@ -393,7 +395,8 @@ module OpenTox
         # @param[Array] A propositionalized data entry
         # @return[Float] Similarity to query structure
         def similarity(training_props)
-          OpenTox::Algorithm::Similarity.send(@model.similarity_algorithm,training_props, @q_prop)
+          eval("#{@model.similarity_algorithm}(#{training_props}, #{@q_prop})")
+          #OpenTox::Algorithm::Similarity.send(@model.similarity_algorithm,training_props, @q_prop)
         end