Merge remote branch 'helma/master' into development

5 files changed, 213 insertions, 21 deletions

diff --git a/application.rb b/application.rb
index 6229178..ad13b0d 100644
--- a/application.rb
+++ b/application.rb
@@ -1,5 +1,5 @@
 require 'rubygems'
-gem "opentox-ruby-api-wrapper", "= 1.5.7"
+gem "opentox-ruby-api-wrapper", "= 1.6.0"
 require 'opentox-ruby-api-wrapper'
 LOGGER.progname = File.expand_path(__FILE__)
 
@@ -17,6 +17,21 @@ DataMapper.auto_upgrade!
 
 require 'lazar.rb'
 
+
+helpers do
+	def activity(a)
+		case a.to_s
+		when "true"
+			act = "active"
+		when "false"
+			act = "inactive"
+		else
+			act = "not available"
+		end
+		act
+	end
+end
+
 get '/?' do # get index of models
 	response['Content-Type'] = 'text/uri-list'
 	Model.all(params).collect{|m| m.uri}.join("\n") + "\n"
diff --git a/config.ru b/config.ru
index 489932f..6f5db40 100644
--- a/config.ru
+++ b/config.ru
@@ -1,5 +1,6 @@
 require 'rubygems'
 require 'opentox-ruby-api-wrapper'
 require 'config/config_ru'
+set :app_file, __FILE__ # to get the view path right
 run Sinatra::Application
 
diff --git a/environment.rb b/environment.rb
deleted file mode 100644
index 067fc18..0000000
--- a/environment.rb
+++ /dev/null
@@ -1,5 +0,0 @@
-require 'rubygems'
-require 'opentox-ruby-api-wrapper'
-
-set :default_content, :yaml
-load 'models.rb'
diff --git a/lazar.rb b/lazar.rb
index 3bc441e..7e0bb98 100755
--- a/lazar.rb
+++ b/lazar.rb
@@ -3,6 +3,7 @@
 # avoids compiling R with X
 R = nil
 require ("rinruby") 
+require ("haml") 
 
 class Lazar < Model
 
@@ -21,6 +22,8 @@ class Lazar < Model
 
 		conf = 0.0
 		similarities = {}
+		activities = {}
+		fragments = { :activating => {}, :deactivating => {} }
 		regression = nil
 
 		regr_occurrences = [] # occurrence vector with {0,1} entries
@@ -36,6 +39,8 @@ class Lazar < Model
 			lazar.activities[uri].each do |act|
 				if sim > 0.3
 					similarities[uri] = sim
+					activities[uri] = [] unless activities[uri]
+					activities[uri] << act
 					conf += OpenTox::Utils.gauss(sim)
 					sims << OpenTox::Utils.gauss(sim)
 					#TODO check for 0 s
@@ -102,11 +107,13 @@ class Lazar < Model
 			prediction.compounds << compound_uri
 			prediction.features << feature_uri 
 			prediction.data[compound_uri] = [] unless prediction.data[compound_uri]
+			compound_matches.each { |m| fragments[lazar.effects[m].to_sym][m] = lazar.p_values[m] }
 			tuple = { 
 					File.join(@@config[:services]["opentox-model"],"lazar#regression") => regression,
-					File.join(@@config[:services]["opentox-model"],"lazar#confidence") => conf
+					File.join(@@config[:services]["opentox-model"],"lazar#confidence") => conf,
 					#File.join(@@config[:services]["opentox-model"],"lazar#similarities") => similarities,
-					#File.join(@@config[:services]["opentox-model"],"lazar#features") => compound_matches
+					#File.join(@@config[:services]["opentox-model"],"lazar#activities") => activities,
+					#File.join(@@config[:services]["opentox-model"],"lazar#features") => fragments
 			}
 			prediction.data[compound_uri] << {feature_uri => tuple}
 		end
@@ -124,6 +131,9 @@ class Lazar < Model
 
 		conf = 0.0
 		similarities = {}
+		#activities = {}
+		fragments = { :activating => {}, :deactivating => {} }
+		neighbors = {}
 		classification = nil
 
 		lazar.fingerprints.each do |uri,matches|
@@ -131,7 +141,22 @@ class Lazar < Model
 			sim = OpenTox::Algorithm::Similarity.weighted_tanimoto(compound_matches,matches,lazar.p_values)
 			if sim > 0.3
 				similarities[uri] = sim
+				neighbors[uri] = {:similarity => sim}
+				neighbors[uri][:features] = { :activating => [], :deactivating => [] } unless neighbors[uri][:features]
+				matches.each do |m|
+					if lazar.effects[m] == 'activating'
+						neighbors[uri][:features][:activating] << {:smarts => m, :p_value => lazar.p_values[m]}
+					elsif lazar.effects[m] == 'deactivating'
+						neighbors[uri][:features][:deactivating] << {:smarts => m, :p_value => lazar.p_values[m]}
+					end
+				end
+				#neighbors[uri][:features] = [] unless neighbors[uri][:features]
+				#neighbors[uri][:features] << matches
 				lazar.activities[uri].each do |act|
+					#activities[uri] = [] unless activities[uri]
+					#activities[uri] << act
+					neighbors[uri][:activities] = [] unless neighbors[uri][:activities]
+					neighbors[uri][:activities] << act
 					case act.to_s
 					when 'true'
 						conf += OpenTox::Utils.gauss(sim)
@@ -153,11 +178,13 @@ class Lazar < Model
 			prediction.compounds << compound_uri
 			prediction.features << feature_uri 
 			prediction.data[compound_uri] = [] unless prediction.data[compound_uri]
+			compound_matches.each { |m| fragments[lazar.effects[m].to_sym][m] = lazar.p_values[m] }
+			#fragments[:activating] = fragments[:activating].sort{|a,b| b[1] <=> a[1]}
 			tuple = { 
 					File.join(@@config[:services]["opentox-model"],"lazar#classification") => classification,
-					File.join(@@config[:services]["opentox-model"],"lazar#confidence") => conf
-					#File.join(@@config[:services]["opentox-model"],"lazar#similarities") => similarities,
-					#File.join(@@config[:services]["opentox-model"],"lazar#features") => compound_matches
+					File.join(@@config[:services]["opentox-model"],"lazar#confidence") => conf,
+					#File.join(@@config[:services]["opentox-model"],"lazar#neighbors") => neighbors,
+					#File.join(@@config[:services]["opentox-model"],"lazar#features") => fragments
 			}
 			prediction.data[compound_uri] << {feature_uri => tuple}
 		end
@@ -278,10 +305,10 @@ post '/:id/?' do # create prediction
 	halt 404, "Model #{params[:id]} does not exist." unless lazar
 	halt 404, "No compound_uri or dataset_uri parameter." unless compound_uri = params[:compound_uri] or dataset_uri = params[:dataset_uri]
 
-	prediction = OpenTox::Dataset.new 
-	prediction.creator = lazar.uri
+	@prediction = OpenTox::Dataset.new 
+	@prediction.creator = lazar.uri
 	dependent_variable = YAML.load(lazar.yaml).dependentVariables
-	prediction.title = URI.decode(dependent_variable.split(/#/).last) 
+	@prediction.title = URI.decode(dependent_variable.split(/#/).last) 
 	case dependent_variable
 	when /classification/
 		prediction_type = "classification"
@@ -292,18 +319,39 @@ post '/:id/?' do # create prediction
 	if compound_uri
 		# AM: switch here between regression and classification
 		begin
-			eval "lazar.#{prediction_type}(compound_uri,prediction) unless lazar.database_activity?(compound_uri,prediction)"
+			eval "lazar.#{prediction_type}(compound_uri,@prediction) unless lazar.database_activity?(compound_uri,@prediction)"
 		rescue
 			LOGGER.error "#{prediction_type} failed for #{compound_uri} with #{$!} "
 			halt 500, "Prediction of #{compound_uri} failed."
 		end
 		case request.env['HTTP_ACCEPT']
 		when /yaml/ 
-			prediction.to_yaml
+			@prediction.to_yaml
 		when 'application/rdf+xml'
-			prediction.to_owl
-		else
-			halt 404, "Content type #{request.env['HTTP_ACCEPT']} not available."
+			@prediction.to_owl
+		when /html/
+			@compound = OpenTox::Compound.new(:uri => compound_uri)
+			@title = @prediction.title
+			if @prediction.data[@compound.uri]
+				if @prediction.creator.to_s.match(/model/) # real prediction
+					p = @prediction.data[@compound.uri].first.values.first
+					if !p[File.join(@@config[:services]["opentox-model"],"lazar#classification")].nil?
+						feature = File.join(@@config[:services]["opentox-model"],"lazar#classification")
+					elsif !p[File.join(@@config[:services]["opentox-model"],"lazar#regression")].nil?
+						feature = File.join(@@config[:services]["opentox-model"],"lazar#regression")
+					end
+					@activity = p[feature]
+					@confidence = p[File.join(@@config[:services]["opentox-model"],"lazar#confidence")]
+					@neighbors = p[File.join(@@config[:services]["opentox-model"],"lazar#neighbors")]#.sort{|a,b| b.last[:similarity] <=> a.last[:similarity]}
+					#@training_activities = p[File.join(@@config[:services]["opentox-model"],"lazar#activities")]
+					@features = p[File.join(@@config[:services]["opentox-model"],"lazar#features")]
+				else # database value
+					@measured_activities = @prediction.data[@compound.uri].first.values
+				end
+			else
+				@activity = "not available (no similar compounds in the training dataset)"
+			end
+			haml :prediction
 		end
 
 	elsif dataset_uri
@@ -313,13 +361,13 @@ post '/:id/?' do # create prediction
 			input_dataset.compounds.each do |compound_uri|
 				# AM: switch here between regression and classification
 				begin
-					eval "lazar.#{prediction_type}(compound_uri,prediction) unless lazar.database_activity?(compound_uri,prediction)"
+					eval "lazar.#{prediction_type}(compound_uri,@prediction) unless lazar.database_activity?(compound_uri,@prediction)"
 				rescue
 					LOGGER.error "#{prediction_type} failed for #{compound_uri} with #{$!} "
 				end
 			end
 			begin
-				uri = prediction.save.chomp
+				uri = @prediction.save.chomp
 			rescue
 				halt 500, "Could not save prediction dataset"
 			end
diff --git a/views/prediction.haml b/views/prediction.haml
new file mode 100644
index 0000000..0187f87
--- /dev/null
+++ b/views/prediction.haml
@@ -0,0 +1,133 @@
+.predictions
+  %table
+    %tr
+      %th= @title.gsub(/_lazar_.*$/,' ').capitalize
+      %th Prediction
+      %th 
+        %a{:href => "#", :id => "linkConfidence#{p.object_id}"} Confidence
+        :javascript
+          $("a#linkConfidence#{p.object_id}").click(function () {
+            $("dl#confidence").toggle();  
+          });
+      %th Relevant features
+    %tr
+      %th
+        %img{:src => @compound.image_uri, :alt => @compound.smiles}
+        %td
+          - if @measured_activities
+            %br
+            - @measured_activities.each do |a|
+              - if activity(a) == 'active'
+                .active
+                  = activity(a)
+              - elsif activity(a) == 'inactive'
+                .inactive
+                  = activity(a)
+              - else
+                = a
+            %br
+              (
+              %a{:href => "#", :id => "linkTrainingData#{p.object_id}"} Training data
+              :javascript
+                $("a#linkTrainingData#{p.object_id}").click(function () {
+                  $("dl#training_data").toggle();  
+                });
+              )
+
+          - else
+            - if activity(@activity) == 'active'
+              .active
+                = activity(@activity)
+            - elsif activity(@activity) == 'inactive'
+              .inactive
+                = activity(@activity)
+            - elsif @activity.is_a?(Float)
+              .other
+                = sprintf('%.03g', @activity)
+            - else
+              .other
+                %em= @activity.to_s
+        %td
+          = sprintf('%.03g', @confidence.to_f.abs) if @confidence
+        %td
+          %table
+            %tr
+              %th{:colspan => 2} activating
+              %th p value
+            - if @features[:activating]
+              - @features[:activating].sort{|a,b| b.last <=> a.last }.each do |f|
+                %tr
+                  %th= f[0]
+                  %td= f[1]
+            %tr
+              %th{:colspan => 2} deactivating
+              %th p value
+            - if @features[:deactivating]
+              - @features[:deactivating].sort{|a,b| b.last <=> a.last }.each do |f|
+                %tr
+                  %th= f[0]
+                  %td= f[1]
+    %tr
+      %th Neighbors
+      %th Activity
+      %th Similarity (activity specific)
+      %th Relevant features
+    - @neighbors.sort{|a,b| b.last[:similarity] <=> a.last[:similarity]}.each do |uri,data|
+      - c = OpenTox::Compound.new(:uri => uri)
+      %tr
+        %th 
+          %br= c.smiles
+          %br
+            %a{:href => c.image_uri, :target => "_blank"} Image
+          %br
+            %img{:src => c.image_uri, :alt => c.smiles}
+        %td
+          - data[:activities].each do |act|
+            - if activity(act) == 'active'
+              .active
+                = activity(act)
+            - elsif activity(act) == 'inactive'
+              .inactive
+                = activity(act)
+            - elsif act.is_a?(Float)
+              .other
+                = sprintf('%.03g', act)
+            - else
+              .other
+                %em= act.to_s
+        %td
+          = sprintf('%.03g', data[:similarity])
+        %td
+          %table
+            %tr
+              %th{:colspan => 2} activating
+              %th p value
+              -#%td= data[:features].inspect
+            -# data[:features][:activating].each do |f|
+            - data[:features][:activating].sort{|a,b| b.last[:p_value] <=> a.last[:p_value] }.each do |f|
+              -# f.inspect
+              %tr
+                %th= f[:smarts]
+                %td= f[:p_value]
+                %td=# f[:p_value]
+            %tr
+              %th{:colspan => 2} deactivating
+              %th p value
+            -# data[:features][:deactivating].sort{|a,b| b.last[:p_value] <=> a.last[:p_value] }.each do |f|
+            - data[:features][:deactivating].each do |f|
+              %tr
+                %th= f[:smarts]
+                %td= f[:p_value]
+
+
+
+%dl#confidence{ :style => "display: none;" }
+  %dt Confidence:
+  %dd Indicates the applicability domain of a model. Predictions with a high confidence can be expected to be more reliable than predictions with low confidence. Confidence values may take any value between 0 and 1. For most models confidence &gt; 0.025 is a sensible (hard) cutoff to distiguish between reliable and unreliable predictions.
+
+
+%dl#training_data{ :style => "display: none;" }
+  %dt Training data:
+  %dd Experimental result(s) from the training dataset are displayed here. 
+
+