Merge branch 'development' of github.com:opentox/opentox-test into development

7 files changed, 228 insertions, 84 deletions

diff --git a/test/data/EPAFHM.medi.csv b/test/data/EPAFHM.medi.csv
new file mode 100644
index 0000000..4428fa2
--- /dev/null
+++ b/test/data/EPAFHM.medi.csv
@@ -0,0 +1,97 @@
+"STRUCTURE_SMILES","LC50_mmol"
+"C1=CC(C=O)=CC(OC)=C1OCCCCCC",1.13E-02
+"C1(OC)=C([N+]([O-])=O)C(C=O)=CC(Br)=C1O",2.66E-01
+"CCCCCCCCOC(=O)C1=CC=CC(C(=O)OCCCCCCCC)=C1",
+"C1=CC(Cl)=CC=C1OC2=C([N+](=O)[O-])C=CC=C2",7.69E-03
+"CC1=C(NC=O)C=CC=C1Cl",2.75E-01
+"CCCCOC(=O)C1=CC=CC(C(=O)OCCCC)=C1",3.23E-03
+"C(C1=CC=CC=C1)(C2=CC=CC=C2)(O)C#C",5.33E-02
+"CCCSCCSCCC",4.22E-02
+"CCCCCCCCOC(=O)C1=CC=C(C(=O)OCCCCCCCC)C=C1",
+"OCCCCOC(=O)C1=CC=CC=C1C(=O)OCCCCOC(=O)C2=CC=CC=C2C(=O)OCCCCO",
+"CCCSCCCCSCCC",1.45E-02
+"C1([N+](=O)[O-])=CC=C(C)C=C1OP(=O)(OC2=C([N+](=O)[O-])C=CC(C)=C2)OC3=C([N+]([O-])=O)C=CC(C)=C3",
+"C1=C([N+]([O-])=O)C=CC=C1P(=O)(C2=CC([N+](=O)[O-])=CC=C2)C3=CC([N+](=O)[O-])=CC=C3",
+"ClCCOC(=O)NC1CCCCC1",1.70E-01
+"O=C1C(C2=CC=CC=C2)(C(=O)NC(=O)N1)CC",2.08E+00
+"OC1=C(C=C(C=C1)[N+](=O)[O-])[N+](=O)[O-]",5.92E-02
+"NC(=O)OCC",5.88E+01
+"[O-]C(C1=CC=CC=C1O)=O.[Na+]",1.25E+01
+"C1=CC=CC=C1C(=O)N",5.46E+00
+"CC[N+](CC)(CC)CC1(=CC=CC=C1).[Cl-]",7.07E-01
+"CN(C)N",1.31E-01
+"CC(C(C(NC([O-])=N1)=O)(C1=O)CC)CCC.[Na+]",1.99E-01
+"N1C(=O)C(CC)(CCC(C)C)C(=O)NC1=O",3.77E-01
+"O=C1C2=C(N=CN2C)N(C(=O)N1C)C",7.78E-01
+"C1=CC=C2C(=C1)C(=O)C(C)=CC2=O",6.39E-04
+"OC1=C(Cl)C(Cl)=C(Cl)C=C1Cl",4.44E-03
+"OC1=CC(C)=C(Cl)C=C1",3.84E-02
+"O=S(O)(O)=O.C1(=CC=CC=C1CC(N)C).C2=CC=CC=C2CC(N)C",7.82E-02
+"O(CC)CC",3.45E+01
+"NC1=CC=CC=C1",1.13E+00
+"O=C(OC1=C2C(=CC=C1)C=CC=C2)NC",4.35E-02
+"CCO",3.19E+02
+"C1(=NC=CC=C1C2CCCN2C).OS(O)(=O)=O",5.30E-02
+"C1(O)=CC=CC=C1C(=O)N",7.36E-01
+"O=C1NC(=O)NC=C1",
+"CCCCCC=O",1.75E-01
+"O=C1OC2=CC=CC=C2C(O)=C1CC3=C(O)C4=CC=CC=C4OC3=O",1.52E-02
+"C1(C=O)=CC=C(OC2=CC=CC=C2)C=C1",2.32E-02
+"CO",9.17E+02
+"OC(C)C",1.44E+02
+"CC(=O)C",1.23E+02
+"ClC(Cl)Cl",5.92E-01
+"CS(=O)C",4.35E+02
+"ClC(C(Cl)(Cl)Cl)(Cl)Cl",6.00E-03
+"OC1=C(C=C(C(=C1CC2=C(C(=CC(=C2Cl)Cl)Cl)O)Cl)Cl)Cl",5.16E-05
+"C1=CC(=CC=C1N)C(=O)CC",9.79E-01
+"OCCC",7.57E+01
+"CCCCO",2.33E+01
+"CCCCCO",5.36E+00
+"C1=CC=CC=C1",2.25E-01
+"CC(Cl)(Cl)Cl",3.55E-01
+"[S-]C1=NC(C(C(C)CCC)(CC)C(N1)=O)=O.[Na+]",9.91E-02
+"CC#N",4.01E+01
+"CC=O",7.67E-01
+"ClCCl",3.89E+00
+"IC(I)I",7.42E-03
+"[N+](C)(C)(C)C.[Cl-]",4.22E+00
+"CC(C)(C)O",8.65E+01
+"C(F)(F)(F)CO",1.19E+00
+"CC(=O)C(C)(C)C",8.69E-01
+"ClC(C(Cl)Cl)(Cl)Cl",3.72E-02
+"CC1(C)NC(=O)NC1=O",1.29E+02
+"CCC(O)(C)CC",6.58E+00
+"C#CC(O)(C)CC",1.24E+01
+"C1CCCC(C#C)(O)C1",2.06E+00
+"CCCCOCCOP(=O)(OCCOCCCC)OCCOCCCC",2.81E-02
+"OCC(C)C",1.93E+01
+"CC(Cl)CCl",1.12E+00
+"NCC(N)C",1.36E+01
+"CC(O)CC",4.95E+01
+"CCC(=O)C",4.47E+01
+"OC(C)CN",3.36E+01
+"ClC(CCl)Cl",6.12E-01
+"ClC(=CCl)Cl",3.36E-01
+"CC(=O)OC",4.82E+00
+"ClC(C(Cl)Cl)Cl",1.21E-01
+"C1(C)(C)CCCC(C)=C1C=CC(C)=O",2.65E-02
+"ClC1=C(O)C(Cl)=CC(=C1)C(C2=CC(Cl)=C(O)C(=C2)Cl)(C)C",3.63E-03
+"C(C1C=CC(=CC=1)O)(CC)(C)C",1.58E-02
+"C1CC(CCC1(N)C)C(C)(N)C",3.83E-01
+"ClC(Cl)C1=C(Cl)C=CC=C1Cl",4.22E-03
+"C1=CC=C2C=CC=C3C2=C1CC3",1.12E-02
+"CC1=CNC2=C1C=CC=C2",6.74E-02
+"C1=CC=CC=C1OC(=O)C2=CC=CC=C2C(=O)OC3=CC=CC=C3",2.51E-04
+"O=C1C2=C(C=CC=C2)C(=O)C3=C1C=CC=C3",
+"CCOC(=O)C1=CC=CC=C1C(=O)OCC",1.43E-01
+"C1=CC=C(C(=O)OCCCC)C(=C1)C(=O)OCCCC",3.5900E-03
+"CCC1=C(Br)C(Br)=C(Br)C(Br)=C1Br",
+"O=C1C2=C(C=CC=C2)N=NN1CSP(=S)(OC)OC",2.02E-04
+"C1=CC=CC=C1NC(=O)C2=C(O)C=CC=C2",1.85E-02
+"Cl\C(Cl)=C(Cl)/C(Cl)=C(Cl)\Cl",3.45E-04
+"OC1=C(C(=C(C(=C1Cl)Cl)Cl)Cl)Cl",9.12E-04
+"OC1=C(C=C(C=C1Cl)Cl)Cl",2.48E-02
+"OC1=CC(C(F)(F)F)=C([N+]([O-])=O)C=C1",4.41E-02
+"C1(N)=CC=CC=C1C(=O)N",2.90E+00
+"OC1=C([N+]([O-])=O)C=CC=C1",1.15E+00
diff --git a/test/lazar-physchem-long.rb b/test/lazar-physchem-long.rb
new file mode 100644
index 0000000..0b2820d
--- /dev/null
+++ b/test/lazar-physchem-long.rb
@@ -0,0 +1,52 @@
+require_relative "setup.rb"
+
+class LazarPhyschemDescriptorTest < MiniTest::Test
+
+  def test_lazar_pc_descriptors
+
+    # check available descriptors
+    desc = OpenTox::Algorithm::Descriptor.physchem_descriptors.keys
+    assert_equal 111,desc.size,"wrong num physchem descriptors"
+    sum = 0
+    {"Openbabel"=>16,"Cdk"=>50,"Joelib"=>45}.each do |k,v|
+      assert_equal v,desc.select{|x| x=~/^#{k}\./}.size,"wrong num #{k} descriptors"
+      sum += v
+    end
+    assert_equal 111,sum
+
+    # select descriptors for test
+    desc.keep_if{|x| x=~/^Openbabel\./}
+    unless defined?($short_tests)
+      # the actual descriptor calculation is rather fast, computing 3D structures takes time
+      desc += ["Cdk.XLogP", "Cdk.WienerNumbers", "Joelib.LogP", "Joelib.count.HeteroCycles"]
+    end
+    puts "descriptors for modeling: #{desc}"
+
+    dataset = OpenTox::Dataset.new 
+    dataset.upload File.join(DATA_DIR,"EPAFHM.medi.csv")
+    assert_equal dataset.uri.uri?, true
+    puts dataset.uri
+
+    model_uri = OpenTox::Model::Lazar.create :dataset_uri => dataset.uri, :feature_generation_uri => File.join($algorithm[:uri],"descriptor","physchem"), :descriptors => desc
+    puts model_uri
+    model = OpenTox::Model::Lazar.new model_uri
+    assert_equal model_uri.uri?, true
+    puts model.predicted_variable
+
+    compound_uri = "#{$compound[:uri]}/InChI=1S/C13H8Cl2O2/c14-12-5-4-11(7-13(12)15)17-10-3-1-2-9(6-10)8-16/h1-8H"
+    prediction_uri = model.predict :compound_uri => compound_uri
+    prediction = OpenTox::Dataset.new prediction_uri
+    assert_equal prediction.uri.uri?, true
+    puts prediction.uri
+
+    assert prediction.features.collect{|f| f.uri}.include?(model.predicted_variable),"prediction feature #{model.predicted_variable} not included prediction dataset #{prediction.features.collect{|f| f.uri}}"
+    assert prediction.compounds.collect{|c| c.uri}.include?(compound_uri),"compound #{compound_uri} not included in prediction dataset #{prediction.compounds.collect{|c| c.uri}}"
+    assert_equal 1,prediction.compound_indices(compound_uri).size,"compound should only be once in the dataset"
+
+    predicted_value = prediction.data_entry_value(prediction.compound_indices(compound_uri).first,model.predicted_variable)
+    puts predicted_value
+    assert predicted_value > 0.005,"predicted values should be above 0.005, is #{predicted_value}"
+    assert predicted_value < 0.1,"predicted values should be below 0.1, is #{predicted_value}"
+  end
+
+end
diff --git a/test/lazar-physchem-short.rb b/test/lazar-physchem-short.rb
new file mode 100644
index 0000000..5b00231
--- /dev/null
+++ b/test/lazar-physchem-short.rb
@@ -0,0 +1,3 @@
+$short_tests = true
+
+require File.join(File.expand_path(File.dirname(__FILE__)),"lazar-physchem-long.rb")
diff --git a/test/lazar-physchem.rb b/test/lazar-physchem.rb
deleted file mode 100644
index 066460a..0000000
--- a/test/lazar-physchem.rb
+++ /dev/null
@@ -1,22 +0,0 @@
-require_relative "setup.rb"
-
-class LazarPhyschemDescriptorTest < MiniTest::Test
-
-  def test_lazar_pc_descriptors
-    dataset = OpenTox::Dataset.new 
-    dataset.upload File.join(DATA_DIR,"EPAFHM.mini.csv")
-    assert_equal dataset.uri.uri?, true
-
-    model_uri = OpenTox::Model::Lazar.create :dataset_uri => dataset.uri, :feature_generation_uri => File.join($algorithm[:uri],"descriptor","physchem"), :descriptors => [ "Openbabel.atoms", "Openbabel.bonds", "Openbabel.dbonds", "Openbabel.HBA1", "Openbabel.HBA2", "Openbabel.HBD", "Openbabel.MP", "Openbabel.MR", "Openbabel.MW", "Openbabel.nF", "Openbabel.sbonds", "Openbabel.tbonds", "Openbabel.TPSA"]
-
-    puts model_uri
-    model = OpenTox::Model::Lazar.new model_uri
-    assert_equal model_uri.uri?, true
-    prediction_uri = model.predict :compound_uri => "#{$compound[:uri]}/InChI=1S/C6H6/c1-2-4-6-5-3-1/h1-6H"
-    prediction = OpenTox::Dataset.new prediction_uri
-    assert_equal prediction.uri.uri?, true
-    #TODO check correct prediction
-    puts prediction.uri
-  end
-
-end
diff --git a/test/task.rb b/test/task.rb
index 4ee07c1..df78b1a 100644
--- a/test/task.rb
+++ b/test/task.rb
@@ -189,6 +189,7 @@ class TaskTest < MiniTest::Test
         assert ex.is_a?(test[:error]),"error type should be a #{test[:error]}, but is a #{ex.class}"
         assert ex.message=~/#{test[:msg]}/,"message should be #{test[:msg]}, but is #{ex.message}"
         assert ex.error_cause=~/test.rb:#{test[:line]}/,"code line number test.rb:#{test[:line]} is lost or wrong: #{ex.error_cause}"
+        assert ex.uri==test[:uri]
       end
     end
   end
diff --git a/test/validation-long.rb b/test/validation-long.rb
index 1a458a2..88ee315 100644
--- a/test/validation-long.rb
+++ b/test/validation-long.rb
@@ -26,21 +26,30 @@ DATA = []
 #      :feat => "http://apps.ideaconsult.net:8080/ambit2/feature/533748",
 #      :info => "http://apps.ideaconsult.net:8080/ambit2/dataset/435293?page=0&pagesize=300" }
 
-HAMSTER_CV_FEATURE_TYPES = ["bbrc"]     
 FILES = {
   File.new(File.join(test_path,"data","hamster_carcinogenicity.csv")) => :split_validation,
-  #File.new("data/EPAFHM.medi.csv") => :split_validation
-  }
+  File.new(File.join(test_path,"data","EPAFHM.medi.csv")) => :split_validation,
+}
   
 unless defined?($short_tests)
-  #HAMSTER_CV_FEATURE_TYPES = ["bbrc", "last"]
   FILES.merge!({
     File.new(File.join(test_path,"data","hamster_carcinogenicity.csv")) => :crossvalidation,  
-  #  File.new("data/EPAFHM.csv") => :crossvalidation,
-  #  File.new("data/hamster_carcinogenicity.csv") => :bootstrap_validation
+    File.new("data/EPAFHM.medi.csv") => :crossvalidation,
+   # File.new("data/hamster_carcinogenicity.csv") => :bootstrap_validation
     })
 end
 
+FEAT_GEN = {}
+FILES.each do |f,t|
+  if f.path=~/hamster/
+    FEAT_GEN[f] = [ File.join($algorithm[:uri],"fminer/bbrc") ] #FEAT_GEN[f] << File.join($algorithm[:uri],"fminer/last")
+  elsif f.path=~/EPAFHM/
+    FEAT_GEN[f] = [ File.join($algorithm[:uri],"descriptor","physchem") ]
+  else
+    raise "please define feature generation uri for dataset: #{f.path}"
+  end
+end
+
 class ValidationTest < MiniTest::Test
   i_suck_and_my_tests_are_order_dependent!
 
@@ -49,10 +58,17 @@ class ValidationTest < MiniTest::Test
     puts "login and upload datasets"
     OpenTox::RestClientWrapper.subjectid ? puts("logged in: "+OpenTox::RestClientWrapper.subjectid.to_s) : puts("AA disabled")
     FILES.each do |file,type|
-      DATA << { :type => type,
-        :data => ValidationTestUtil.upload_dataset(file),
-        :feat => ValidationTestUtil.prediction_feature_for_file(file),
-        :info => file.path, :delete => true} 
+      data = { :type => type,
+          :data => ValidationTestUtil.upload_dataset(file),
+          :feat => ValidationTestUtil.prediction_feature_for_file(file),
+          :split_ratio => (file.path=~/EPAFHM/ ? 0.98 : 0.9),#only used for split_validation
+          :info => file.path, :delete => true} 
+      FEAT_GEN[file].each do |feat_gen|
+        data[:alg_params] = "feature_generation_uri="+feat_gen
++       data[:alg_params] << ";backbone=false;min_chisq_significance=0.0" if feat_gen=~/fminer/ and data[:info] =~ /mini/
+        data[:alg_params] << ";descriptors="+[ "Openbabel.atoms", "Openbabel.bonds", "Openbabel.dbonds", "Openbabel.HBA1", "Openbabel.HBA2", "Openbabel.HBD", "Openbabel.MP", "Openbabel.MR", "Openbabel.MW", "Openbabel.nF", "Openbabel.sbonds", "Openbabel.tbonds", "Openbabel.TPSA"].join(",") if feat_gen=~/physchem/
+        DATA << data
+      end
     end
   end
   
@@ -88,7 +104,7 @@ class ValidationTest < MiniTest::Test
         p = { 
           :dataset_uri => data[:data],
           :algorithm_uri => File.join($algorithm[:uri],"lazar"),
-          :algorithm_params => "feature_generation_uri="+File.join($algorithm[:uri],"fminer/bbrc"),
+          :algorithm_params => data[:alg_params],
           :prediction_feature => data[:feat],
           :random_seed => 2}
         t = OpenTox::SubTask.new(nil,0,1)
@@ -130,9 +146,9 @@ class ValidationTest < MiniTest::Test
         p = { 
           :dataset_uri => data[:data],
           :algorithm_uri => File.join($algorithm[:uri],"lazar"),
-          :algorithm_params => "feature_generation_uri="+File.join($algorithm[:uri],"fminer/bbrc"),
+          :algorithm_params => data[:alg_params],
           :prediction_feature => data[:feat],
-          :split_ratio => 0.9,
+          :split_ratio => data[:split_ratio],
           :random_seed => 2}
         t = OpenTox::SubTask.new(nil,0,1)
         def t.progress(pct)
@@ -161,7 +177,7 @@ class ValidationTest < MiniTest::Test
         train_compounds = OpenTox::Dataset.find(v.metadata[RDF::OT.trainingDataset.to_s]).compounds
         test_compounds = OpenTox::Dataset.find(v.metadata[RDF::OT.testDataset.to_s]).compounds
         orig_compounds = OpenTox::Dataset.find(data[:data]).compounds
-        assert_equal((orig_compounds.size*0.9).round,train_compounds.size)
+        assert_equal((orig_compounds.size*data[:split_ratio]).round,train_compounds.size)
         assert_equal(orig_compounds.size,(train_compounds+test_compounds).size)
         assert_equal(orig_compounds.uniq.size,(train_compounds+test_compounds).uniq.size)
         
@@ -186,7 +202,7 @@ class ValidationTest < MiniTest::Test
           :training_dataset_uri => data[:train_data],
           :test_dataset_uri => data[:test_data],
           :algorithm_uri => File.join($algorithm[:uri],"lazar"),
-          :algorithm_params => "feature_generation_uri="+File.join($algorithm[:uri],"fminer/bbrc"),
+          :algorithm_params => data[:alg_params],
           :prediction_feature => data[:feat]}
         t = OpenTox::SubTask.new(nil,0,1)
         def t.progress(pct)
@@ -273,56 +289,52 @@ class ValidationTest < MiniTest::Test
     @@cv_identifiers = []
     DATA.each do |data|
       if data[:type]==:crossvalidation
-        HAMSTER_CV_FEATURE_TYPES.each do |fminer|
-          next unless (fminer==HAMSTER_CV_FEATURE_TYPES[0] or data[:info].to_s =~ /hamster_carcinogenicity.csv/)
-          puts "test_crossvalidation "+data[:info].to_s+" "+fminer
-          p = { 
-            :dataset_uri => data[:data],
-            :algorithm_uri => File.join($algorithm[:uri],"lazar"),
-            :algorithm_params => "feature_generation_uri="+File.join($algorithm[:uri],"fminer/"+fminer)+
-             (data[:info] =~ /mini/ ? ";backbone=false;min_chisq_significance=0.0" : ""),
-            :prediction_feature => data[:feat],
-            :num_folds => 10 }
-            #:num_folds => 2 }
-          t = OpenTox::SubTask.new(nil,0,1)
-          def t.progress(pct)
-            if !defined?@last_msg or @last_msg+10<Time.new
-              puts "waiting for crossvalidation: "+pct.to_s
-              @last_msg=Time.new
-            end
-          end
-          def t.waiting_for(task_uri); end
-          cv = OpenTox::Crossvalidation.create(p, t)
-          assert cv.uri.uri?
-          if $aa[:uri]
-            assert_unauthorized do
-              OpenTox::Crossvalidation.find(cv.uri)
-            end
+        puts "test_crossvalidation "+data[:info].to_s+" "+data[:alg_params]
+        p = { 
+          :dataset_uri => data[:data],
+          :algorithm_uri => File.join($algorithm[:uri],"lazar"),
+          :algorithm_params => data[:alg_params],
+          :prediction_feature => data[:feat],
+          :num_folds => 10 }
+          #:num_folds => 2 }
+        t = OpenTox::SubTask.new(nil,0,1)
+        def t.progress(pct)
+          if !defined?@last_msg or @last_msg+10<Time.new
+            puts "waiting for crossvalidation: "+pct.to_s
+            @last_msg=Time.new
           end
-          cv = OpenTox::Crossvalidation.find(cv.uri)
-          assert_valid_date cv
-          assert cv.uri.uri?
-          stats_val = cv.statistics
-          assert_kind_of OpenTox::Validation,stats_val
-          assert_prob_correct(stats_val)
-          
-          algorithm = cv.metadata[RDF::OT.algorithm.to_s]
-          assert algorithm.uri?
-          cv_list = OpenTox::Crossvalidation.list( {:algorithm => algorithm} )
-          assert cv_list.include?(cv.uri)
-          cv_list.each do |cv_uri|
-            #begin catch not authorized somehow
-              alg = OpenTox::Crossvalidation.find(cv_uri).metadata[RDF::OT.algorithm.to_s]
-              assert alg==algorithm,"wrong algorithm for filtered crossvalidation, should be: '"+algorithm.to_s+"', is: '"+alg.to_s+"'"
-            #rescue 
-            #end
+        end
+        def t.waiting_for(task_uri); end
+        cv = OpenTox::Crossvalidation.create(p, t)
+        assert cv.uri.uri?
+        if $aa[:uri]
+          assert_unauthorized do
+            OpenTox::Crossvalidation.find(cv.uri)
           end
-          puts cv.uri unless defined?(DELETE) and DELETE
-          
-          @@cvs << cv
-          @@cv_datasets << data
-          @@cv_identifiers << fminer
         end
+        cv = OpenTox::Crossvalidation.find(cv.uri)
+        assert_valid_date cv
+        assert cv.uri.uri?
+        stats_val = cv.statistics
+        assert_kind_of OpenTox::Validation,stats_val
+        assert_prob_correct(stats_val)
+        
+        algorithm = cv.metadata[RDF::OT.algorithm.to_s]
+        assert algorithm.uri?
+        cv_list = OpenTox::Crossvalidation.list( {:algorithm => algorithm} )
+        assert cv_list.include?(cv.uri)
+        cv_list.each do |cv_uri|
+          #begin catch not authorized somehow
+            alg = OpenTox::Crossvalidation.find(cv_uri).metadata[RDF::OT.algorithm.to_s]
+            assert alg==algorithm,"wrong algorithm for filtered crossvalidation, should be: '"+algorithm.to_s+"', is: '"+alg.to_s+"'"
+          #rescue 
+          #end
+        end
+        puts cv.uri unless defined?(DELETE) and DELETE
+        
+        @@cvs << cv
+        @@cv_datasets << data
+        @@cv_identifiers << data[:alg_params]
       end
     end
   end
diff --git a/test/validation_util.rb b/test/validation_util.rb
index d8373f4..0e82aec 100755
--- a/test/validation_util.rb
+++ b/test/validation_util.rb
@@ -35,6 +35,7 @@ class ValidationTestUtil
   end
   
   def self.prediction_feature_for_file(file)
+    raise "no prediction feature available for #{file.path}" unless @@prediction_features[file.path.to_s]
     @@prediction_features[file.path.to_s]
   end