Merge branch 'development'

Conflicts:

	opentox-ruby-api-wrapper.gemspec

9 files changed, 166 insertions, 116 deletions

diff --git a/Rakefile b/Rakefile
index 41832bb..3846bd1 100644
--- a/Rakefile
+++ b/Rakefile
@@ -34,7 +34,8 @@ begin
 			'dm-timestamps',
 			'dm-types',
 			'dm-migrations',
-			"dm-mysql-adapter"
+			"dm-mysql-adapter",
+      "dm-validations",
 		].each {|dep| gem.add_dependency dep, ">= 1" }
 		gem.add_dependency "haml", ">=3"
 		['cucumber','jeweler'].each { |dep| gem.add_development_dependency dep }
diff --git a/VERSION b/VERSION
index 266146b..9edc58b 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-1.6.3
+1.6.4
diff --git a/lib/algorithm.rb b/lib/algorithm.rb
index 2fde0eb..d7b57af 100644
--- a/lib/algorithm.rb
+++ b/lib/algorithm.rb
@@ -64,6 +64,18 @@ module OpenTox
           0.0
         end
       end
+      def self.euclidean(prop_a,prop_b)
+        common_properties = prop_a.keys & prop_b.keys
+        if common_properties.size > 1
+          dist_sum = 0
+          common_properties.each do |p|
+            dist_sum += (prop_a[p] - prop_b[p])**2
+          end
+          1/(1+Math.sqrt(dist_sum))
+        else
+          nil
+        end
+      end
     end
 
   end
diff --git a/lib/dataset.rb b/lib/dataset.rb
index 0be804c..2eb2206 100644
--- a/lib/dataset.rb
+++ b/lib/dataset.rb
@@ -66,8 +66,8 @@ module OpenTox
       raise "no new compounds selected" unless new_compounds and new_compounds.size>0
       
       # load require features 
-      if ((defined? @dirty_features) && (@dirty_features - new_features).size > 0)
-        (@dirty_features - new_features).each{|f| load_feature_values(f)}
+      if ((defined? @dirty_features) && (@dirty_features & new_features).size > 0)
+        (@dirty_features & new_features).each{|f| load_feature_values(f)}
       end
       
       dataset = OpenTox::Dataset.new
@@ -202,16 +202,19 @@ module OpenTox
     
     # overwrite to yaml:
     # in case dataset is loaded from owl:
-    # * load all values
-    # * set @owl to nil (not necessary in yaml) 
+    # * load all values 
     def to_yaml
       # loads all features  
       if ((defined? @dirty_features) && @dirty_features.size > 0)
         load_feature_values
       end
-      @owl = nil
       super
     end
+    
+    # * remove @owl from yaml, not necessary
+    def to_yaml_properties
+      super - ["@owl"]
+    end
 
     # saves (changes) as new dataset in dataset service
     # returns uri
diff --git a/lib/environment.rb b/lib/environment.rb
index 80899c5..cfc875d 100644
--- a/lib/environment.rb
+++ b/lib/environment.rb
@@ -25,7 +25,7 @@ end
 
 # database
 if @@config[:database]
-	['dm-core', 'dm-serializer', 'dm-timestamps', 'dm-types', 'dm-migrations' ].each{|lib| require lib }
+	['dm-core', 'dm-serializer', 'dm-timestamps', 'dm-types', 'dm-migrations', 'dm-validations' ].each{|lib| require lib }
 	case @@config[:database][:adapter]
 	when /sqlite/i
 		db_dir = File.join(basedir, "db")
diff --git a/lib/model.rb b/lib/model.rb
index e8f6048..e36b538 100644
--- a/lib/model.rb
+++ b/lib/model.rb
@@ -1,13 +1,13 @@
 module OpenTox
-	module Model
+  module Model
 
-		class Generic
+    class Generic
 
       MODEL_ATTRIBS = [:uri, :title, :creator, :date, :format, :predictedVariables, :independentVariables, :dependentVariables, :trainingDataset, :algorithm]
       MODEL_ATTRIBS.each{ |a| attr_accessor(a) }
 
-			def self.find(uri)
-				owl = OpenTox::Owl.from_uri(uri, "Model")
+      def self.find(uri)
+        owl = OpenTox::Owl.from_uri(uri, "Model")
         return self.new(owl)
       end
       
@@ -28,86 +28,116 @@ module OpenTox
         if ENV['RACK_ENV'] =~ /test|debug/
           begin
             raise "uri invalid" unless Utils.is_uri?(@uri)
-            raise "no algorithm" unless @algorithm and @algorithm.size>0
-            raise "no dependent variables" unless @dependentVariables and @dependentVariables.size>0 
-            raise "no indenpendent variables" unless @independentVariables
             raise "no predicted variables" unless @predictedVariables and @predictedVariables.size>0
           rescue => ex
             RestClientWrapper.raise_uri_error "invalid model: '"+ex.message+"'\n"+self.to_yaml+"\n",@uri.to_s    
           end
+          LOGGER.warn "model has no dependent variable" unless @dependentVariables and @dependentVariables.size>0
+          LOGGER.warn "model has no algorithm" unless @algorithm and @algorithm.size>0
+          LOGGER.warn "model has no indenpendent variables" unless @independentVariables
         end
-			end
-	 end
+      end
+    end
   
-   class PredictionModel < Generic
+    class PredictionModel < Generic
+      
+      def self.build( algorithm_uri, algorithm_params )
+         
+        LOGGER.debug "Build model, algorithm_uri:"+algorithm_uri.to_s+", algorithm_parms: "+algorithm_params.inspect.to_s
+        uri = OpenTox::RestClientWrapper.post(algorithm_uri,algorithm_params).to_s
+        LOGGER.debug "Build model done: "+uri.to_s
+        RestClientWrapper.raise_uri_error("Invalid build model result: '"+uri.to_s+"'", algorithm_uri, algorithm_params ) unless Utils.model_uri?(uri)
+        return PredictionModel.find(uri)
+      end
      
-     def self.build( algorithm_uri, algorithm_params )
-        
-       LOGGER.debug "Build model, algorithm_uri:"+algorithm_uri.to_s+", algorithm_parms: "+algorithm_params.inspect.to_s
-       uri = OpenTox::RestClientWrapper.post(algorithm_uri,algorithm_params).to_s
-       LOGGER.debug "Build model done: "+uri.to_s
-       RestClientWrapper.raise_uri_error("Invalid build model result: '"+uri.to_s+"'", algorithm_uri, algorithm_params ) unless Utils.model_uri?(uri)
-       return PredictionModel.find(uri)
-     end
-    
-     def predict_dataset( dataset_uri )
-
-       LOGGER.debug "Predict dataset: "+dataset_uri.to_s+" with model "+@uri.to_s
-       uri = RestClientWrapper.post(@uri, {:accept => "text/uri-list", :dataset_uri=>dataset_uri})
-       RestClientWrapper.raise_uri_error("Prediciton result no dataset uri: "+uri.to_s, @uri, {:dataset_uri=>dataset_uri} ) unless Utils.dataset_uri?(uri)
-       uri
-     end
-    
-     def classification?
-       #HACK replace with request to ontology server
-       if @title =~ /(?i)classification/
-         return true
-       elsif @title =~ /(?i)regression/
-         return false
-       elsif @uri =~/ntua/ and @title =~ /mlr/
-         return false
-       elsif @uri =~/tu-muenchen/ and @title =~ /regression|M5P|GaussP/
-         return false
-       elsif @uri =~/ambit2/ and @title =~ /pKa/ || @title =~ /Regression|Caco/
-         return false
-       elsif @uri =~/majority/
-         return (@uri =~ /class/) != nil
-       else
-         raise "unknown model, uri:'"+@uri.to_s+"' title:'"+@title.to_s+"'"
-       end
-     end
-   end
-  
+      def predict_dataset( dataset_uri )
+ 
+        LOGGER.debug "Predict dataset: "+dataset_uri.to_s+" with model "+@uri.to_s
+        uri = RestClientWrapper.post(@uri, {:accept => "text/uri-list", :dataset_uri=>dataset_uri})
+        RestClientWrapper.raise_uri_error("Prediciton result no dataset uri: "+uri.to_s, @uri, {:dataset_uri=>dataset_uri} ) unless Utils.dataset_uri?(uri)
+        uri
+      end
+     
+      def classification?
+        #HACK replace with request to ontology server
+        if @title =~ /(?i)classification/
+          return true
+        elsif @title =~ /(?i)regression/
+          return false
+        elsif @uri =~/ntua/ and @title =~ /mlr/
+          return false
+        elsif @uri =~/tu-muenchen/ and @title =~ /regression|M5P|GaussP/
+          return false
+        elsif @uri =~/ambit2/ and @title =~ /pKa/ || @title =~ /Regression|Caco/
+          return false
+        elsif @uri =~/majority/
+          return (@uri =~ /class/) != nil
+        else
+          raise "unknown model, uri:'"+@uri.to_s+"' title:'"+@title.to_s+"'"
+        end
+      end
+    end
    
-		class Lazar < Generic
+    class Lazar < Generic
       
       attr_accessor :feature_dataset_uri, :effects, :activities, :p_values, :fingerprints, :features
       
-			def initialize
-				@source = "http://github.com/helma/opentox-model"
-				@algorithm = File.join(@@config[:services]["opentox-algorithm"],"lazar")
-				#@independent_variables = File.join(@@config[:services]["opentox-algorithm"],"fminer#BBRC_representative")
-				@features = []
-				@effects = {}
-				@activities = {}
-				@p_values = {}
-				@fingerprints = {}
-			end
+      def initialize
+        @source = "http://github.com/helma/opentox-model"
+        @algorithm = File.join(@@config[:services]["opentox-algorithm"],"lazar")
+        #@independent_variables = File.join(@@config[:services]["opentox-algorithm"],"fminer#BBRC_representative")
+        @features = []
+        @effects = {}
+        @activities = {}
+        @p_values = {}
+        @fingerprints = {}
+      end
+
+      def save
+        @features.uniq!
+        resource = RestClient::Resource.new(@@config[:services]["opentox-model"], :user => @@users[:users].keys[0], :password => @@users[:users].values[0])
+        resource.post(self.to_yaml, :content_type => "application/x-yaml").chomp.to_s
+      end
+
+      def self.find_all
+        RestClientWrapper.get(@@config[:services]["opentox-model"]).chomp.split("\n")
+      end
+
+      def self.predict(compound_uri,model_uri)
+        #RestClientWrapper.post(model_uri,{:compound_uri => compound_uri, :accept => 'application/x-yaml'})
+        `curl -X POST -d 'compound_uri=#{compound_uri}' -H 'Accept:application/x-yaml' #{model_uri}`
+      end
+    end
+   
+    class PropertyLazar < Generic
+      
+      attr_accessor :feature_dataset_uri, :properties, :features, :activities#, :effects, :p_values
+      
+      def initialize
+        @source = "http://github.com/helma/opentox-model"
+        @algorithm = File.join(@@config[:services]["opentox-algorithm"],"property_lazar")
+        #@independent_variables = File.join(@@config[:services]["opentox-algorithm"],"fminer#BBRC_representative")
+        @features = []
+        #@effects = {}
+        @activities = {}
+        #@p_values = {}
+        @properties = {}
+      end
 
-			def save
-				@features.uniq!
-			  resource = RestClient::Resource.new(@@config[:services]["opentox-model"], :user => @@users[:users].keys[0], :password => @@users[:users].values[0])
-			  resource.post(self.to_yaml, :content_type => "application/x-yaml").chomp.to_s
-			end
+      def save
+        @features.uniq!
+        resource = RestClient::Resource.new(@@config[:services]["opentox-model"], :user => @@users[:users].keys[0], :password => @@users[:users].values[0])
+        resource.post(self.to_yaml, :content_type => "application/x-yaml").chomp.to_s
+      end
 
-			def self.find_all
-				RestClientWrapper.get(@@config[:services]["opentox-model"]).chomp.split("\n")
-			end
+      def self.find_all
+        RestClientWrapper.get(@@config[:services]["opentox-model"]).chomp.split("\n")
+      end
 
       def self.predict(compound_uri,model_uri)
         #RestClientWrapper.post(model_uri,{:compound_uri => compound_uri, :accept => 'application/x-yaml'})
-		`curl -X POST -d 'compound_uri=#{compound_uri}' -H 'Accept:application/x-yaml' #{model_uri}`
+        `curl -X POST -d 'compound_uri=#{compound_uri}' -H 'Accept:application/x-yaml' #{model_uri}`
       end
-		end
-	end
+    end
+  end
 end
diff --git a/lib/owl.rb b/lib/owl.rb
index a4c2e68..dcf26a5 100644
--- a/lib/owl.rb
+++ b/lib/owl.rb
@@ -518,6 +518,11 @@ module OpenTox
             value_node = value_nodes[0]
             
             compound_uri = get_value( @model.object(value_node, node('compound')) )
+            unless compound_uri
+               LOGGER.warn "'compound' missing for data-entry of feature "+feature_uri.to_s+
+                 ", value: "+@model.object(feature_value_node,node("value")).to_s
+               next
+            end
             
             value_node_type = @model.object(feature_value_node, RDF_TYPE)
             if (value_node_type == node('FeatureValue'))
diff --git a/lib/task.rb b/lib/task.rb
index 971478f..1ab3893 100644
--- a/lib/task.rb
+++ b/lib/task.rb
@@ -2,7 +2,7 @@ $self_task=nil
 
 module OpenTox
 
-	class Task
+  class Task
 
     # due_to_time is only set in local tasks 
     TASK_ATTRIBS = [ :uri, :date, :title, :creator, :description, :hasStatus, :percentageCompleted, :resultURI, :due_to_time ]
@@ -15,13 +15,13 @@ module OpenTox
     end
     
     # create is private now, use OpenTox::Task.as_task
-		def self.create(max_duration)
-      task_uri = RestClientWrapper.post(@@config[:services]["opentox-task"], {:max_duration => max_duration}, nil, false).to_s
-			Task.find(task_uri.chomp)
-		end
-
+    def self.create( params )
+      task_uri = RestClientWrapper.post(@@config[:services]["opentox-task"], params, nil, false).to_s
+      Task.find(task_uri.chomp)
+    end
+  
     public
-    def self.find( uri, accept_header='application/rdf+xml' )
+    def self.find( uri, accept_header=nil )
       task = Task.new(uri)
       task.reload( accept_header )
       return task
@@ -34,7 +34,14 @@ module OpenTox
       return task
     end
     
-    def reload( accept_header='application/rdf+xml' )
+    def reload( accept_header=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
       result = RestClientWrapper.get(uri, {:accept => accept_header}, false)#'application/x-yaml'})
       @http_code = result.code
       reload_from_data(result, result.content_type, uri)
@@ -58,20 +65,20 @@ module OpenTox
       raise "uri is null after loading" unless @uri and @uri.to_s.strip.size>0
     end
     
-		def cancel
+    def cancel
       RestClientWrapper.put(File.join(@uri,'Cancelled'))
       reload
-		end
+    end
 
-		def completed(uri)
-			RestClientWrapper.put(File.join(@uri,'Completed'),{:resultURI => uri})
+    def completed(uri)
+      RestClientWrapper.put(File.join(@uri,'Completed'),{:resultURI => uri})
       reload
-		end
+    end
 
-		def error(description)
+    def error(description)
       RestClientWrapper.put(File.join(@uri,'Error'),{:description => description.to_s[0..2000]})
       reload
-		end
+    end
     
     def pid=(pid)
       RestClientWrapper.put(File.join(@uri,'pid'), {:pid => pid})
@@ -81,16 +88,16 @@ module OpenTox
       @hasStatus.to_s == 'Running'
     end
 
-		def completed?
-			@hasStatus.to_s == 'Completed'
-		end
+    def completed?
+      @hasStatus.to_s == 'Completed'
+    end
 
-		def error?
-			@hasStatus.to_s == 'Error'
-		end
+    def error?
+      @hasStatus.to_s == 'Error'
+    end
 
     # waits for a task, unless time exceeds or state is no longer running
-		def wait_for_completion(dur=0.3)
+    def wait_for_completion(dur=0.3)
       
       if (@uri.match(@@config[:services]["opentox-task"]))
         due_to_time = (@due_to_time.is_a?(Time) ? @due_to_time : Time.parse(@due_to_time))
@@ -102,17 +109,17 @@ module OpenTox
       end
       LOGGER.debug "start waiting for task "+@uri.to_s+" at: "+Time.new.to_s+", waiting at least until "+due_to_time.to_s
       
-			while self.running?
-				sleep dur
+      while self.running?
+        sleep dur
         reload
         check_state
         if (Time.new > due_to_time)
           raise "max wait time exceeded ("+running_time.to_s+"sec), task: '"+@uri.to_s+"'"
         end
-			end
+      end
       
       LOGGER.debug "Task '"+@hasStatus+"': "+@uri.to_s+", Result: "+@resultURI.to_s
-	  end
+    end
   
     def check_state
       begin
@@ -133,10 +140,11 @@ module OpenTox
   
     # returns the task uri
     # catches halts and exceptions, task state is set to error then
-    def self.as_task(max_duration=DEFAULT_TASK_MAX_DURATION)
+    def self.as_task( title, creator, max_duration=DEFAULT_TASK_MAX_DURATION, description=nil )
       #return yield nil
       
-      task = OpenTox::Task.create(max_duration)
+      params = {:title=>title, :creator=>creator, :max_duration=>max_duration, :description=>description }
+      task = OpenTox::Task.create(params)
       task_pid = Spork.spork(:logger => LOGGER) do
         LOGGER.debug "Task #{task.uri} started #{Time.now}"
         $self_task = task
@@ -163,6 +171,6 @@ module OpenTox
       LOGGER.debug "Started task: "+task.uri.to_s
       task.uri
     end  
-	end
+  end
 
 end
diff --git a/lib/validation.rb b/lib/validation.rb
index bd38488..89a2a0c 100644
--- a/lib/validation.rb
+++ b/lib/validation.rb
@@ -4,16 +4,7 @@ module OpenTox
 		attr_accessor :uri
 
 		def initialize(params)
-			#resource = RestClient::Resource.new(params[:uri], :user => @@users[:users].keys[0], :password => @@users[:users].values[0])
-			#@uri = resource.post(params).body
-			#LOGGER.debug "VALIDATION URI: " + @uri.to_s
-			call = "curl -X POST "
-			params.each do |k,v|
-				call += " -d "+k.to_s+"=\""+URI.encode(v.to_s)+"\"" unless k == :uri
-			end
-			call += " "+params[:uri]
-			LOGGER.debug call
-			@uri = `#{call}`
+      @uri = OpenTox::RestClientWrapper.post(File.join(@@config[:services]["opentox-validation"],"/crossvalidation"),params,nil,false)
 		end
 		
 		def self.crossvalidation(params)