path: root/lib/model.rb

module OpenTox

  module Model

    include OpenTox

    # Run a model with parameters
    # @param [Hash] params Parameters for OpenTox model
    # @param [optional,OpenTox::Task] waiting_task (can be a OpenTox::Subtask as well), progress is updated accordingly
    # @return [text/uri-list] Task or resource URI
    def run( params, accept_header=nil, waiting_task=nil )
      unless accept_header
        if CONFIG[:yaml_hosts].include?(URI.parse(@uri).host)
          accept_header = 'application/x-yaml' 
        else
          accept_header = 'application/rdf+xml'
        end
      end
      LOGGER.info "running model "+@uri.to_s+", params: "+params.inspect+", accept: "+accept_header.to_s
      RestClientWrapper.post(@uri,params,{:accept => accept_header},waiting_task).to_s
    end

    # Generic OpenTox model class for all API compliant services
    class Generic
      include Model

      # Find Generic Opentox Model via URI, and loads metadata, could raise NotFound/NotAuthorized error 
      # @param [String] uri Model URI
      # @return [OpenTox::Model::Generic] Model instance
      def self.find(uri,subjectid=nil)
        return nil unless uri
        model = Generic.new(uri)
        model.load_metadata(subjectid)
        raise "could not load model metadata '"+uri.to_s+"'" if model.metadata==nil or model.metadata.size==0
        model
      end

      # provides feature type, possible types are "regression" or "classification"
      # @return [String] feature type, "unknown" if type could not be estimated
      def feature_type(subjectid=nil)
        unless @feature_type
          load_predicted_variables( subjectid ) unless @predicted_variable
          @feature_type = OpenTox::Feature.find( @predicted_variable, subjectid ).feature_type
        end
        @feature_type
      end
    
      def predicted_variable( subjectid )
        load_predicted_variables( subjectid ) unless @predicted_variable
        @predicted_variable
      end

      def predicted_confidence( subjectid )
        load_predicted_variables( subjectid ) unless @predicted_confidence
        @predicted_confidence
      end
  
      private
      def load_predicted_variables( subjectid=nil )
        load_metadata(subjectid) if @metadata==nil or @metadata.size==0 or (@metadata.size==1 && @metadata.values[0]==@uri)
        if @metadata[OT.predictedVariables]
          predictedVariables = @metadata[OT.predictedVariables]
          if predictedVariables.is_a?(Array)
            if (predictedVariables.size==1)
              @predicted_variable = predictedVariables[0]
            elsif (predictedVariables.size==2)
              # PENDING identify confidence
              conf_index = -1
              predictedVariables.size.times do |i|
                f = OpenTox::Feature.find(predictedVariables[i])
                conf_index = i if f.metadata[DC.title]=~/(?i)confidence/
              end
              raise "could not estimate predicted variable from model: '"+uri.to_s+
                "', number of predicted-variables==2, but no confidence found" if conf_index==-1
              @predicted_variable = predictedVariables[1-conf_index]
              @predicted_confidence = predictedVariables[conf_index]
            else
              raise "could not estimate predicted variable from model: '"+uri.to_s+"', number of predicted-variables > 2"  
            end
          else
            raise "could not estimate predicted variable from model: '"+uri.to_s+"', predicted-variables is no array"
          end        
        end
        raise "could not estimate predicted variable from model: '"+uri.to_s+"'" unless @predicted_variable
      end
    end

    # Lazy Structure Activity Relationship class
    class Lazar

      include Model
      include Algorithm

      attr_accessor :compound, :prediction_dataset, :features, :effects, :activities, :p_values, :fingerprints, :feature_calculation_algorithm, :similarity_algorithm, :prediction_algorithm, :min_sim, :subjectid, :prop_kernel, :value_map

      def initialize(uri=nil)

        if uri
          super uri
        else
          super CONFIG[:services]["opentox-model"]
        end

        @metadata[OT.algorithm] = File.join(CONFIG[:services]["opentox-algorithm"],"lazar")

        @features = []
        @effects = {}
        @activities = {}
        @p_values = {}
        @fingerprints = {}
        @value_map = {}

        @feature_calculation_algorithm = "Substructure.match"
        @similarity_algorithm = "Similarity.tanimoto"
        @prediction_algorithm = "Neighbors.weighted_majority_vote"

        @min_sim = 0.3
        @prop_kernel = false

      end

      # Get URIs of all lazar models
      # @return [Array] List of lazar model URIs
      def self.all(subjectid=nil)
        RestClientWrapper.get(CONFIG[:services]["opentox-model"], :subjectid => subjectid).to_s.split("\n")
      end

      # Find a lazar model
      # @param [String] uri Model URI
      # @return [OpenTox::Model::Lazar] lazar model
      def self.find(uri, subjectid=nil)
        YAML.load RestClientWrapper.get(uri,{:accept => 'application/x-yaml', :subjectid => subjectid})
      end

      # Create a new lazar model
      # @param [optional,Hash] params Parameters for the lazar algorithm (OpenTox::Algorithm::Lazar)
      # @return [OpenTox::Model::Lazar] lazar model
      def self.create(params)
        subjectid = params[:subjectid]
        lazar_algorithm = OpenTox::Algorithm::Generic.new File.join( CONFIG[:services]["opentox-algorithm"],"lazar")
        model_uri = lazar_algorithm.run(params)
        OpenTox::Model::Lazar.find(model_uri, subjectid)      
      end

      # Get a parameter value
      # @param [String] param Parameter name
      # @return [String] Parameter value
      def parameter(param)
        @metadata[OT.parameters].collect{|p| p[OT.paramValue] if p[DC.title] == param}.compact.first
      end

      # Predict a dataset
      # @param [String] dataset_uri Dataset URI
      # @param [optional,subjectid] 
      # @param [optional,OpenTox::Task] waiting_task (can be a OpenTox::Subtask as well), progress is updated accordingly
      # @return [OpenTox::Dataset] Dataset with predictions
      def predict_dataset(dataset_uri, subjectid=nil, waiting_task=nil)
        @prediction_dataset = Dataset.create(CONFIG[:services]["opentox-dataset"], subjectid)
        @prediction_dataset.add_metadata({
          OT.hasSource => @uri,
          DC.creator => @uri,
          DC.title => URI.decode(File.basename( @metadata[OT.dependentVariables] )),
          OT.parameters => [{DC.title => "dataset_uri", OT.paramValue => dataset_uri}]
        })
        d = Dataset.new(dataset_uri,subjectid)
        d.load_compounds(subjectid)
        count = 0
        d.compounds.each do |compound_uri|
          begin
            predict(compound_uri,false,subjectid)
            count += 1
            waiting_task.progress( count/d.compounds.size.to_f*100.0 ) if waiting_task
          rescue => ex
            LOGGER.warn "prediction for compound "+compound_uri.to_s+" failed: "+ex.message
          end
        end
        @prediction_dataset.save(subjectid)
        @prediction_dataset
      end

      # Predict a compound
      # @param [String] compound_uri Compound URI
      # @param [optinal,Boolean] verbose Verbose prediction (output includes neighbors and features)
      # @return [OpenTox::Dataset] Dataset with prediction
      def predict(compound_uri,verbose=false,subjectid=nil)

        @compound = Compound.new compound_uri
        features = {}

        unless @prediction_dataset
          @prediction_dataset = Dataset.create(CONFIG[:services]["opentox-dataset"], subjectid)
          @prediction_dataset.add_metadata( {
            OT.hasSource => @uri,
            DC.creator => @uri,
            DC.title => URI.decode(File.basename( @metadata[OT.dependentVariables] )),
            OT.parameters => [{DC.title => "compound_uri", OT.paramValue => compound_uri}]
          } )
        end

        return @prediction_dataset if database_activity(subjectid)

        load_metadata(subjectid)
        case OpenTox::Feature.find(metadata[OT.dependentVariables]).feature_type
        when "classification"
          # AM: Balancing, see http://www.maunz.de/wordpress/opentox/2011/balanced-lazar
          l = Array.new # larger 
          s = Array.new # smaller fraction

          raise "no fingerprints in model" if @fingerprints.size==0

          @fingerprints.each do |training_compound,training_features|
            @activities[training_compound].each do |act|
              case act.to_s
              when "false" 
                l << training_compound
              when "true"  
                s << training_compound
              else
                LOGGER.warn "BLAZAR: Activity #{act.to_s} should not be reached."
              end
            end
          end
          if s.size > l.size then 
            l,s = s,l # happy swapping
            LOGGER.info "BLAZAR: |s|=#{s.size}, |l|=#{l.size}."
          end
          # determine ratio
          modulo = l.size.divmod(s.size)# modulo[0]=ratio, modulo[1]=rest
          LOGGER.info "BLAZAR: Balance: #{modulo[0]}, rest #{modulo[1]}."

          # AM: Balanced predictions
          addon = (modulo[1].to_f/modulo[0]).ceil # what will be added in each round 
          slack = (addon!=0 ? modulo[1].divmod(addon)[1] : 0) # what remains for the last round
          position = 0
          predictions = Array.new

          prediction_best=nil
          neighbors_best=nil

          begin
            for i in 1..modulo[0] do
              (i == modulo[0]) && (slack>0) ? lr_size = s.size + slack : lr_size = s.size + addon  # determine fraction
              LOGGER.info "BLAZAR: Neighbors round #{i}: #{position} + #{lr_size}."
              neighbors_balanced(s, l, position, lr_size) # get ratio fraction of larger part
              if @prop_kernel && @prediction_algorithm.include?("svm")
                props = get_props 
              else
                props = nil
              end
              prediction = eval("#{@prediction_algorithm}(@neighbors,{:similarity_algorithm => @similarity_algorithm, :p_values => @p_values}, props)")
              if prediction_best.nil? || prediction[:confidence].abs > prediction_best[:confidence].abs 
                prediction_best=prediction 
                neighbors_best=@neighbors
              end
              position = position + lr_size
            end
          rescue Exception => e
            LOGGER.error "BLAZAR failed in prediction: "+e.class.to_s+": "+e.message
          end

          prediction=prediction_best
          @neighbors=neighbors_best
          ### END AM balanced predictions

        else # AM: no balancing
          LOGGER.info "LAZAR: Unbalanced."
          neighbors
          if @prop_kernel && @prediction_algorithm.include?("svm")
           props = get_props 
          else
           props = nil
          end
          prediction = eval("#{@prediction_algorithm}(@neighbors,{:similarity_algorithm => @similarity_algorithm, :p_values => @p_values}, props)")
        end
        
        value_feature_uri = File.join( @uri, "predicted", "value")
        confidence_feature_uri = File.join( @uri, "predicted", "confidence")

        #prediction_feature_uris = {value_feature_uri => prediction[:prediction], confidence_feature_uri => prediction[:confidence]}
        #prediction_feature_uris[value_feature_uri] = nil if @neighbors.size == 0 or prediction[:prediction].nil?

        @prediction_dataset.metadata[OT.dependentVariables] = @metadata[OT.dependentVariables]
        @prediction_dataset.metadata[OT.predictedVariables] = [value_feature_uri, confidence_feature_uri]

        if OpenTox::Feature.find(metadata[OT.dependentVariables]).feature_type == "classification"
          @prediction_dataset.add @compound.uri, value_feature_uri, @value_map[prediction[:prediction]]
        else
          @prediction_dataset.add @compound.uri, value_feature_uri, prediction[:prediction]
        end
        @prediction_dataset.add @compound.uri, confidence_feature_uri, prediction[:confidence]
        #prediction_feature_uris.each do |prediction_feature_uri,value|
            #@prediction_dataset.add @compound.uri, prediction_feature_uri, @value_map[value]
        #end

        if verbose
          if @feature_calculation_algorithm == "Substructure.match"
            f = 0
            @compound_features.each do |feature|
              feature_uri = File.join( @prediction_dataset.uri, "feature", "descriptor", f.to_s)
              features[feature] = feature_uri
              @prediction_dataset.add_feature(feature_uri, {
                RDF.type => [OT.Substructure],
                OT.smarts => feature,
                OT.pValue => @p_values[feature],
                OT.effect => @effects[feature]
              })
              @prediction_dataset.add @compound.uri, feature_uri, true
              f+=1
            end
          else
            @compound_features.each do |feature|
              features[feature] = feature
              @prediction_dataset.add @compound.uri, feature, true
            end
          end
          n = 0
          @neighbors.each do |neighbor|
            neighbor_uri = File.join( @prediction_dataset.uri, "feature", "neighbor", n.to_s )
            @prediction_dataset.add_feature(neighbor_uri, {
              OT.compound => neighbor[:compound],
              OT.similarity => neighbor[:similarity],
              OT.measuredActivity => neighbor[:activity],
              RDF.type => [OT.Neighbor]
            })
            @prediction_dataset.add @compound.uri, neighbor_uri, true
            f = 0 unless f
            neighbor[:features].each do |feature|
              if @feature_calculation_algorithm == "Substructure.match"
                feature_uri = File.join( @prediction_dataset.uri, "feature", "descriptor", f.to_s) unless feature_uri = features[feature]
              else
                feature_uri = feature
              end
              @prediction_dataset.add neighbor[:compound], feature_uri, true
              unless features.has_key? feature
                features[feature] = feature_uri
                @prediction_dataset.add_feature(feature_uri, {
                  RDF.type => [OT.Substructure],
                  OT.smarts => feature,
                  OT.pValue => @p_values[feature],
                  OT.effect => @effects[feature]
                })
                f+=1
              end
            end
            n+=1
          end
        end
        #end

        @prediction_dataset.save(subjectid)
        @prediction_dataset
      end

      # Calculate the propositionalization matrix aka instantiation matrix (0/1 entries for features)
      # Same for the vector describing the query compound
      def get_props
        matrix = Array.new
        begin 
          @neighbors.each do |n|
            n = n[:compound]
            row = []
            @features.each do |f|
              if ! @fingerprints[n].nil? 
                row << (@fingerprints[n].include?(f) ? 0.0 : @p_values[f])
              else
                row << 0.0
              end
            end
            matrix << row
          end
          row = []
          @features.each do |f|
            row << (@compound.match([f]).size == 0 ? 0.0 : @p_values[f])
          end
        rescue Exception => e
          LOGGER.debug "get_props failed with '" + $! + "'"
        end
        [ matrix, row ]
      end

      # Find neighbors and store them as object variable, access only a subset of compounds for that.
      def neighbors_balanced(s, l, start, offset)
        @compound_features = eval("#{@feature_calculation_algorithm}(@compound,@features)") if @feature_calculation_algorithm
        @neighbors = []
        [ l[start, offset ] , s ].flatten.each do |training_compound| # AM: access only a balanced subset
          training_features = @fingerprints[training_compound]
          add_neighbor training_features, training_compound
        end

      end

      # Find neighbors and store them as object variable, access all compounds for that.
      def neighbors
        @compound_features = eval("#{@feature_calculation_algorithm}(@compound,@features)") if @feature_calculation_algorithm
        @neighbors = []
        @fingerprints.each do |training_compound,training_features| # AM: access all compounds
          add_neighbor training_features, training_compound
        end
      end

      # Adds a neighbor to @neighbors if it passes the similarity threshold.
      def add_neighbor(training_features, training_compound)
        sim = eval("#{@similarity_algorithm}(@compound_features,training_features,@p_values)")
        if sim > @min_sim
          @activities[training_compound].each do |act|
            @neighbors << {
              :compound => training_compound,
              :similarity => sim,
              :features => training_features,
              :activity => act
            }
          end
        end
      end

      # Find database activities and store them in @prediction_dataset
      # @return [Boolean] true if compound has databasse activities, false if not
      def database_activity(subjectid)
        if @activities[@compound.uri]
          @activities[@compound.uri].each { |act| @prediction_dataset.add @compound.uri, @metadata[OT.dependentVariables], act }
          @prediction_dataset.add_metadata(OT.hasSource => @metadata[OT.trainingDataset])
          @prediction_dataset.save(subjectid)
          true
        else
          false
        end
      end

      def prediction_features
        [prediction_value_feature,prediction_confidence_feature]
      end

      def prediction_value_feature
        dependent_uri = @metadata[OT.dependentVariables].first
        feature = OpenTox::Feature.new File.join( @uri, "predicted", "value")
        feature.add_metadata( {
          RDF.type => [OT.ModelPrediction],
          OT.hasSource => @uri,
          DC.creator => @uri,
          DC.title => URI.decode(File.basename( dependent_uri )),
          OWL.sameAs => dependent_uri
        })
        feature
      end

      def prediction_confidence_feature
        dependent_uri = @metadata[OT.dependentVariables].first
        feature = OpenTox::Feature.new File.join( @uri, "predicted", "confidence")
        feature.add_metadata( {
          RDF.type => [OT.ModelPrediction],
          OT.hasSource => @uri,
          DC.creator => @uri,
          DC.title => "#{URI.decode(File.basename( dependent_uri ))} confidence"
        })
        feature
      end

      # Save model at model service
      def save(subjectid)
        self.uri = RestClientWrapper.post(@uri,self.to_yaml,{:content_type =>  "application/x-yaml", :subjectid => subjectid})
      end

      # Delete model at model service
      def delete(subjectid)
        RestClientWrapper.delete(@uri, :subjectid => subjectid) unless @uri == CONFIG[:services]["opentox-model"]
      end

    end
  end
end