new params for new r-util funcitons

3 files changed, 21 insertions, 13 deletions

diff --git a/validation/validation_application.rb b/validation/validation_application.rb
index 689241c..eaa3ce2 100755
--- a/validation/validation_application.rb
+++ b/validation/validation_application.rb
@@ -9,8 +9,8 @@ require 'validation/validation_service.rb'
 helpers do
   def check_stratified(params)
     params[:stratified] = "false" unless params[:stratified]
-    raise OpenTox::BadRequestError.new "stratified != true|false|super|super4|anti, is #{params[:stratified]}" unless
-      params[:stratified]=~/^(true|false|super|super4|anti)$/
+    raise OpenTox::BadRequestError.new "stratified != true|false|super|super4|super5|anti, is #{params[:stratified]}" unless
+      params[:stratified]=~/^(true|false|super|super4|super5|anti)$/
   end
 end
 
@@ -586,7 +586,8 @@ post '/plain_training_test_split' do
     split_features = nil
     split_features = params[:split_features].split(";") if params[:split_features]
     result = Validation::Util.train_test_dataset_split(params[:dataset_uri], params[:prediction_feature], @subjectid,
-      params[:stratified], params[:split_ratio], params[:random_seed], params[:missing_values], task, split_features)
+      params[:stratified], params[:split_ratio], params[:random_seed], params[:missing_values], task, split_features, 
+        params[:store_split_clusters])
     content_type "text/uri-list"
     res = result[:training_dataset_uri]+"\n"+result[:test_dataset_uri]+"\n"
     LOGGER.info "plain training test split done #{res.to_s.gsub("\n"," \\n ")}"      
diff --git a/validation/validation_service.rb b/validation/validation_service.rb
index 1b29630..a2e97d0 100755
--- a/validation/validation_service.rb
+++ b/validation/validation_service.rb
@@ -500,7 +500,7 @@ module Validation
             meta, self.subjectid ).uri
           test_dataset_uris << test_dataset_uri
         end
-      when /^(true|super|super4)$/
+      when /^(true|super|super4|super5)$/
         raise "DO NOT USED SUPER-STRATIFICATION FOR VAL-EXPERIMENTS AND CV, IF SO SOLVE _MISSING_VAULE_NA_ PROBLEM" if stratified=~/super/
         if stratified=="true"
           features = [ self.prediction_feature ] 
@@ -627,7 +627,7 @@ module Validation
     # splits a dataset into test and training dataset
     # returns map with training_dataset_uri and test_dataset_uri
     def self.train_test_dataset_split( orig_dataset_uri, prediction_feature, subjectid, stratified="false",
-      split_ratio=nil, random_seed=nil, missing_values=nil, task=nil, features=nil )
+      split_ratio=nil, random_seed=nil, missing_values=nil, task=nil, features=nil, store_split_clusters=false )
       
       split_ratio=0.67 unless split_ratio
       split_ratio = split_ratio.to_f
@@ -651,7 +651,8 @@ module Validation
       meta = { DC.creator => $url_provider.url_for('/training_test_split',:full) }
       
       case stratified
-      when /^(true|super|super4|anti)$/
+      when /^(true|super|super4|super5|anti)$/
+        raise "store split clusters not available for true stratified splits" if store_split_clusters and stratified=="true"
         if stratified=="true"
           raise OpenTox::BadRequestError.new "prediction feature required for stratified splits" unless prediction_feature
           LOGGER.warn "split features are ignored for stratified splits (use super instead)" if features
@@ -660,10 +661,12 @@ module Validation
           LOGGER.warn "prediction feature is ignored for super- or anti-stratified splits" if prediction_feature
         end
         r_util = OpenTox::RUtil.new 
-        train, test = r_util.stratified_split( orig_dataset, meta, missing_values, split_ratio, @subjectid, random_seed, features, stratified )
+        train, test = r_util.stratified_split( orig_dataset, meta, missing_values, split_ratio, 
+          @subjectid, random_seed, features, stratified, store_split_clusters )
         r_util.quit_r
         result = {:training_dataset_uri => train.uri, :test_dataset_uri => test.uri}
       when "false"
+        raise "store split clusters not available for random splits" if store_split_clusters
         LOGGER.warn "split features are ignored for non-stratified splits (use super instead)" if features
         compounds = orig_dataset.compounds
         raise OpenTox::BadRequestError.new "Cannot split datset, num compounds in dataset < 2 ("+compounds.size.to_s+")" if compounds.size<2
diff --git a/validation/validation_test.rb b/validation/validation_test.rb
index 5a7c2aa..1a41881 100755
--- a/validation/validation_test.rb
+++ b/validation/validation_test.rb
@@ -85,12 +85,16 @@ class ValidationTest < Test::Unit::TestCase
 #        {:validation_uris=>"http://local-ot/validation/389,http://local-ot/validation/390,http://local-ot/validation/391,http://local-ot/validation/392",
 #         :identifier=>"split1,split1,split2,split2"}
       
-#      post "/training_test_validation",{:prediction_feature=>"http://opentox.informatik.uni-freiburg.de/dataset/2/feature/MTP", 
-#        :training_dataset_uri=>"http://opentox.informatik.uni-freiburg.de/dataset/4755", 
-#        :algorithm_uri=>"http://opentox.informatik.uni-freiburg.de/superservice", 
-#        :test_dataset_uri=>"http://opentox.informatik.uni-freiburg.de/dataset/3556", 
-#        :algorithm_params=>"prediction_algorithm=http://opentox.informatik.uni-freiburg.de/weka/M5P;create_bbrc_features=false;ad_algorithm=http://opentox.informatik.uni-freiburg.de/appdomain/EuclideanDistance"}
-#      exit
+      post "/training_test_validation",{:prediction_feature=>"http://local-ot/dataset/9264/feature/endpoint", 
+        :training_dataset_uri=>"http://local-ot/dataset/119127", 
+        :algorithm_uri=>"http://local-ot/weka/RandomForest", 
+        :test_dataset_uri=>"http://local-ot/dataset/119128",
+        :test_target_dataset_uri=>"http://local-ot/dataset/9264",  
+        :algorithm_params=>""}
+      uri = last_response.body
+      rep = wait_for_task(uri)
+      puts rep
+      exit
 #        
 #        #post "/report/validation",{:validation_uris=>"http://local-ot/validation/22849",:min_confidence=>0.5}
 #      get "/22849",{:min_confidence=>0.5}