dirty speedup hack for parse csv

2 files changed, 22 insertions, 15 deletions

diff --git a/lib/parser.rb b/lib/parser.rb
index 2193cf4..7bf46cf 100644
--- a/lib/parser.rb
+++ b/lib/parser.rb
@@ -373,7 +373,8 @@ module OpenTox
           }
         }
 
-        input.each_with_index { |row, i| 
+        first_data_row = true
+        input.each_with_index { |row, i|
           drop=false
           row = split_row(row)
           raise "Entry has size #{row.size}, different from headers (#{headers.size})" if row.size != headers.size
@@ -382,7 +383,10 @@ module OpenTox
             drop=true
             drop_missing=true if (row.count("") == row.size-1) 
           end
-          add_values(row, regression_features) unless (drop_missing && drop)
+          unless (drop_missing && drop)
+            add_values(row, regression_features, first_data_row)
+            first_data_row = false
+          end 
           if (drop_missing && drop) 
             @format_errors << "Row #{i} not added" 
           end
@@ -458,7 +462,7 @@ module OpenTox
       # @param Array A row split up as an array
       # @param Array Indicator for regression for each field
       # @param Array Indices for duplicate features
-      def add_values(row, regression_features)
+      def add_values(row, regression_features, add_features=true)
 
         id = row.shift
         case id
@@ -475,6 +479,8 @@ module OpenTox
         @duplicates[compound.inchi] = [] unless @duplicates[compound.inchi]
         @duplicates[compound.inchi] << id+", "+row.join(", ")
 
+        @dataset.add_compound(compound.uri)
+  
         feature_idx = 0
         row.each_index do |i|
 
@@ -502,7 +508,8 @@ module OpenTox
             feature_idx += 1
   
             if val != nil 
-              @dataset.add(compound.uri, feature, val)
+              @dataset.add_feature(feature) if add_features
+              @dataset.add(compound.uri, feature, val, true)
               if @feature_types[feature].include? OT.NominalFeature
                 @dataset.features[feature][OT.acceptValue] = [] unless @dataset.features[feature][OT.acceptValue]
                 @dataset.features[feature][OT.acceptValue] << val unless @dataset.features[feature][OT.acceptValue].include?(val)
diff --git a/lib/r-util.rb b/lib/r-util.rb
index 8357d0f..bbf98f3 100644
--- a/lib/r-util.rb
+++ b/lib/r-util.rb
@@ -61,11 +61,11 @@ module OpenTox
     # <0 -> array1 << array2
     # 0  -> no significant difference
     # >0 -> array2 >> array1
-    def paired_ttest(array1, array2, significance_level=0.95)
+    def ttest(array1, array2, paired, significance_level=0.95)
       @r.assign "v1",array1
       @r.assign "v2",array2
-      paired = array1.size==array2.size ? "T" : "F"
-      @r.eval "ttest = t.test(as.numeric(v1),as.numeric(v2),paired=#{paired})"
+      raise if paired && array1.size!=array2.size
+      @r.eval "ttest = t.test(as.numeric(v1),as.numeric(v2),paired=#{paired ? "T" : "F"})"
       t = @r.pull "ttest$statistic"
       p = @r.pull "ttest$p.value"
       if (1-significance_level > p)
@@ -83,7 +83,7 @@ module OpenTox
     end
         
     
-    def ttest(array1, value2, significance_level=0.95)
+    def ttest_single_value(array1, value2, significance_level=0.95)
       @r.assign "v1",array1
       @r.eval "ttest = t.test(as.numeric(v1),conf.level=#{significance_level})"
       min = @r.pull "ttest$conf.int[1]"
@@ -151,7 +151,7 @@ module OpenTox
       hlines << [max_median,2+max_median_idx] 
       hlines << [min_median,2+min_median_idx]
       plot_to_files(files, hlines) do |file|
-        @r.eval "boxplot(boxdata,main='#{title}',col=rep(2:#{data.size+1})#{param_str})"
+        @r.eval "superboxplot(boxdata,main='#{title}',col=rep(2:#{data.size+1})#{param_str})"
       end
     end
 
@@ -522,20 +522,20 @@ module OpenTox
       begin File.delete(tmp); rescue; end
     end
     
-    @svg_plot_width = 14
-    @svg_plot_height = 10
-    
+    @@svg_plot_width = 12
+    @@svg_plot_height = 8
+        
     public
     def set_svg_plot_size(width,height)
-      @svg_plot_width = width
-      @svg_plot_height = height
+      @@svg_plot_width = width
+      @@svg_plot_height = height
     end
     
     private
     def plot_to_files(files,hlines=nil)
       files.each do |file|
         if file=~/(?i)\.svg/
-          @r.eval("svg('#{file}',#{@svg_plot_width},#{@svg_plot_height})")
+          @r.eval("svg('#{file}',#{@@svg_plot_width},#{@@svg_plot_height})")
         elsif file=~/(?i)\.png/
           @r.eval("png('#{file}')")
         else