regression plot in 2 different scales

2 files changed, 41 insertions, 25 deletions

diff --git a/report/plot_factory.rb b/report/plot_factory.rb
index d6c7d35..26a9c67 100644
--- a/report/plot_factory.rb
+++ b/report/plot_factory.rb
@@ -52,11 +52,12 @@ module Reports
   
   module PlotFactory
     
-    def self.create_regression_plot( out_files, validation_set, name_attribute )
+    def self.create_regression_plot( out_files, validation_set, name_attribute, logscale=true )
       
       out_files = [out_files] unless out_files.is_a?(Array)
       LOGGER.debug "Creating regression plot, out-file:"+out_files.to_s
       
+      omit_count = 0
       names = []
       x = []
       y = []
@@ -64,25 +65,32 @@ module Reports
         x_i = v.get_predictions.predicted_values
         y_i = v.get_predictions.actual_values
         
-        # filter out nil-predictions
-        not_nil_indices = []
+        # filter out nil-predictions and <=0 predictions if log-scale wanted
+        valid_indices = []
         x_i.size.times do |i|
-          not_nil_indices << i if x_i[i]!=nil && y_i[i]!=nil
+          if x_i[i]!=nil and y_i[i]!=nil and (!logscale or (x_i[i]>0 and y_i[i]>0))
+            valid_indices << i 
+          else
+            omit_count += 1
+          end
         end
-        if not_nil_indices.size < x_i.size
-          x_i = not_nil_indices.collect{ |i| x_i[i] }
-          y_i = not_nil_indices.collect{ |i| y_i[i] }
+        if valid_indices.size < x_i.size
+          x_i = valid_indices.collect{ |i| x_i[i] }
+          y_i = valid_indices.collect{ |i| y_i[i] }
         end
 
         names << ( name_attribute==:crossvalidation_fold ? "fold " : "" ) + v.send(name_attribute).to_s
         x << x_i
         y << y_i
       end
-  
-      raise "no predictions performed" if x.size==0 || x[0].size==0
+      names = [""] if names.size==1 
+
+      omit_str = omit_count>0 ? " ("+omit_count.to_s+" predictions omitted)" : ""
+      raise "no predictions performed"+omit_str if x.size==0 || x[0].size==0
       out_files.each do |out_file|
-        RubyPlot::regression_point_plot(out_file, "Regression plot", "Predicted values", "Actual values", names, x, y )
+        RubyPlot::regression_point_plot(out_file, "Regression plot", "Predicted values", "Actual values", names, x, y, logscale)
       end
+      omit_count
     end
     
     
diff --git a/report/report_content.rb b/report/report_content.rb
index ea2ad5a..8c437a8 100755
--- a/report/report_content.rb
+++ b/report/report_content.rb
@@ -141,31 +141,39 @@ class Reports::ReportContent
                             name_attribute,
                             section_title="Regression Plot",
                             section_text=nil,
-                            image_title=nil)
+                            image_title="Regression Plot")
                             
-    image_title = "Regression plot" unless image_title
     #section_regr = @xml_report.add_section(@current_section, section_title)
     section_regr = @current_section
     prediction_set = validation_set.collect{ |v| v.get_predictions }
         
     if prediction_set.size>0
       
-      section_text += "\nWARNING: regression plot information not available for all validation results" if prediction_set.size!=validation_set.size
-      @xml_report.add_paragraph(section_regr, section_text) if section_text
-      begin
-        plot_png = add_tmp_file("regr_plot", "png")
-        plot_svg = add_tmp_file("regr_plot", "svg")
-        Reports::PlotFactory.create_regression_plot( [plot_png[:path], plot_svg[:path]], prediction_set, name_attribute )
-        @xml_report.add_imagefigure(section_regr, image_title,  plot_png[:name], "PNG", 100, plot_svg[:name])
-      rescue Exception => ex
-        LOGGER.error("Could not create regression plot: "+ex.message)
-        rm_tmp_file(plot_png[:name])
-        rm_tmp_file(plot_svg[:name])
-        @xml_report.add_paragraph(section_regr, "could not create regression plot: "+ex.message)
-      end  
+      [true, false].each do |log|
+        scale_str = (log ? " (logarithmic scale)" : " (linear scale)")
+        image_title_2 = image_title + scale_str
+        section_title_2 = section_title + scale_str
+      
+        section_text += "\nWARNING: regression plot information not available for all validation results" if prediction_set.size!=validation_set.size
+        @xml_report.add_paragraph(section_regr, section_text) if section_text
+        begin
+          log_str = (log ? "_log" : "")
+          plot_png = add_tmp_file("regr_plot"+log_str, "png")
+          plot_svg = add_tmp_file("regr_plot"+log_str, "svg")
+          omit_count = Reports::PlotFactory.create_regression_plot( [plot_png[:path], plot_svg[:path]], prediction_set, name_attribute, log )
+          image_title_2 += " ("+omit_count.to_s+" datapoints omitted)" if omit_count>0
+          @xml_report.add_imagefigure(section_regr, image_title_2,  plot_png[:name], "PNG", 100, plot_svg[:name])
+        rescue Exception => ex
+          LOGGER.error("Could not create regression plot: "+ex.message)
+          rm_tmp_file(plot_png[:name])
+          rm_tmp_file(plot_svg[:name])
+          @xml_report.add_paragraph(section_regr, "could not create regression plot: "+ex.message)
+        end  
+      end
     else
       @xml_report.add_paragraph(section_regr, "No prediction info for regression available.")
     end
+    align_last_two_images section_title+" in logarithmic and linear scale (values <= 0 are omitted in logarithmic scale)"
   end
   
   def add_roc_plot( validation_set,