new branch val_exp

4 files changed, 170 insertions, 13 deletions

diff --git a/.gitignore b/.gitignore
index e3debba..fa8bcaf 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,4 +1,3 @@
-tmp/*
-log/*
-public/*.owl
-*.*~
+.buildpath
+.loadpath
+.project
diff --git a/algorithm_test.rb b/algorithm_test.rb
new file mode 100644
index 0000000..08c7d4c
--- /dev/null
+++ b/algorithm_test.rb
@@ -0,0 +1,134 @@
+require "rubygems"
+require "sinatra"
+before {
+  request.env['HTTP_HOST']="local-ot/algorithm"
+  request.env["REQUEST_URI"]=request.env["PATH_INFO"]
+}
+
+require "opentox-ruby"
+ENV['RACK_ENV'] = 'test'
+require 'application.rb'
+require 'test/unit'
+require 'rack/test'
+LOGGER = Logger.new(STDOUT)
+LOGGER.datetime_format = "%Y-%m-%d %H:%M:%S "
+  
+module Sinatra
+  
+  set :raise_errors, false
+  set :show_exceptions, false
+
+  module UrlForHelper
+    BASE = "http://local-ot/algorithm"
+    def url_for url_fragment, mode=:path_only
+      case mode
+      when :path_only
+        raise "not impl"
+      when :full
+      end
+      "#{BASE}#{url_fragment}"
+    end
+  end
+end
+
+class AlgorithmTest < Test::Unit::TestCase
+  include Rack::Test::Methods
+  
+  def app
+    Sinatra::Application
+  end
+  
+  def test_prediction
+    
+    begin
+      
+      #dataset_uri = "http://apps.ideaconsult.net:8080/ambit2/dataset/603204?pagesize=100&page=0"
+      #test_dataset_uri = "http://apps.ideaconsult.net:8080/ambit2/dataset/603204?pagesize=100&page=1"
+      #feature_dataset_uri = "http://apps.ideaconsult.net:8080/ambit2/dataset/603204?pagesize=200&page=0"
+      #prediction_feature = "http://apps.ideaconsult.net:8080/ambit2/feature/528321"
+      
+      dataset_uri = "http://apps.ideaconsult.net:8080/ambit2/dataset/425254"
+      prediction_feature = "http://apps.ideaconsult.net:8080/ambit2/feature/528321"
+      
+#      dataset_uri = "http://local-ot/dataset/1488"
+#      prediction_feature = "http://local-ot/dataset/1315/feature/Rodent%20carcinogenicity"
+      params = {:dataset_uri=>dataset_uri,
+                :prediction_feature=>prediction_feature,
+                :min_frequency=>7} #multi: 10=>4, 5=>>3000
+      post "/fminer/bbrc",params
+      
+#      params = {:dataset_uri=>dataset_uri,
+#        :prediction_feature=>prediction_feature, :feature_dataset_uri=>feature_dataset_uri}
+#      post "/lazar",params
+      
+      uri = wait_for_task(last_response.body)
+      puts uri
+      
+#      model = uri
+#      puts "model #{model}"
+#      params = {:dataset_uri=>test_dataset_uri}
+#      puts OpenTox::RestClientWrapper.post(model,params)
+      
+      #puts "features: "+OpenTox::Dataset.find(uri).features.size.to_s
+     
+     
+#      params = {:dataset_uri=>"http://local-ot/dataset/1724",
+#        :feature_dataset_uri=>"http://local-ot/dataset/1725"}
+#      #post "/fminer/match",params
+#      #uri = wait_for_task(last_response.body)
+#      #puts uri
+#      
+#      fminer = File.join(CONFIG[:services]["opentox-algorithm"],"fminer/bbrc")
+#      OpenTox::RestClientWrapper.post(fminer,params)
+      
+#      params = {:dataset_uri=>"http://local-ot/dataset/1488",
+#        :prediction_feature=>"http://local-ot/dataset/1315/feature/Rodent%20carcinogenicity",
+#        :min_frequency=>50}
+#      post "/lazar",params
+#      uri = wait_for_task(last_response.body)
+#      puts uri
+      puts "features: "+OpenTox::Dataset.find(uri).features.size.to_s
+      
+   #  fminer = File.join(CONFIG[:services]["opentox-algorithm"],"fminer/bbrc")
+   #  OpenTox::RestClientWrapper.post(fminer,params)
+     
+   rescue => ex
+     rep = OpenTox::ErrorReport.create(ex, "")
+     puts rep.to_yaml
+   end 
+    
+    #get "/lazar",nil,'HTTP_ACCEPT' => "application/rdf+xml"
+    #get "/fminer",nil,'HTTP_ACCEPT' => "application/rdf+xml" 
+    #OpenTox::Algorithm.Generic.find("http://localhost/algorithm/lazar")
+    
+  #  puts last_response.body
+    
+#    webservice = "http://ot.algorithm.de/lazar"
+#    headers = {:dataset_uri=>"http://ot.dataset.de/2", 
+#      :prediction_feature=>"http://localhost/toxmodel/feature%23Hamster%20Carcinogenicity%20(DSSTOX/CPDB)", 
+#      :feature_generation_uri=>"http://ot.algorithm.de/fminer"}
+#    
+#    #puts OpenTox::RestClientWrapper.post(webservice,headers)
+#    post webservice,headers
+#    #puts 
+#    uri = wait_for_task(last_response.body.to_s)
+#    puts uri
+#    puts OpenTox::RestClientWrapper.get(uri,:accept => 'application/rdf+xml')
+#    #get uri
+    
+  end
+  
+      # see test_util.rb
+  def wait_for_task(uri)
+      if uri.task_uri?
+        task = OpenTox::Task.find(uri)
+        task.wait_for_completion
+        raise "task failed: "+uri.to_s if task.error?
+        uri = task.result_uri
+      end
+      return uri
+    end
+  
+  
+  
+end
\ No newline at end of file
diff --git a/fminer.rb b/fminer.rb
index 95ce469..36e1d73 100644
--- a/fminer.rb
+++ b/fminer.rb
@@ -98,16 +98,38 @@ post '/fminer/:method/match?' do
     f_dataset.features.each do |f,m|
       res_dataset.add_feature(f,m)
     end
-    c_dataset.compounds.each do |c|
-      res_dataset.add_compound(c)
-      comp = OpenTox::Compound.new(c)
-      f_dataset.features.each do |f,m|
-        if params[:nr_hits] == "true"
+
+    step_width = 100 / c_dataset.compounds.size.to_f
+    count = 0 
+    
+    if params[:nr_hits] == "true"
+      c_dataset.compounds.each do |c|
+        res_dataset.add_compound(c)
+        comp = OpenTox::Compound.new(c)
+        f_dataset.features.each do |f,m|
           hits = comp.match_hits([m[OT.smarts]])
           res_dataset.add(c,f,hits[m[OT.smarts]]) if hits[m[OT.smarts]]          
-        else
-          res_dataset.add(c,f,1) if comp.match?(m[OT.smarts])
         end
+        count += 1
+        task.progress step_width*count
+      end
+    else
+      LOGGER.debug "match #{c_dataset.compounds.size} compounds with #{f_dataset.features.keys.size} features"
+      
+      obconversion = OpenBabel::OBConversion.new
+      obmol = OpenBabel::OBMol.new
+      obconversion.set_in_format('inchi')
+      smarts_pattern = OpenBabel::OBSmartsPattern.new
+      c_dataset.compounds.each do |c|
+        res_dataset.add_compound(c)
+        comp = OpenTox::Compound.new(c)
+        obconversion.read_string(obmol,comp.inchi)
+        f_dataset.features.each do |f,m|
+          smarts_pattern.init(m[OT.smarts])
+          res_dataset.add(c,f,1) if smarts_pattern.match(obmol)
+        end
+        count += 1
+        task.progress step_width*count if count%100==0
       end
     end
     res_dataset.save @subjectid
@@ -167,7 +189,8 @@ post '/fminer/bbrc/?' do
     fminer.smi = [] # AM LAST: needed for matching the patterns back
 
     # Add data to fminer
-    fminer.add_fminer_data(@@bbrc, params, @value_map)
+    #fminer.add_fminer_data(@@bbrc, params, @value_map)
+    fminer.add_fminer_data(@@bbrc, @value_map)
 
     g_array=fminer.all_activities.values # DV: calculation of global median for effect calculation
     g_median=g_array.to_scale.median
@@ -298,7 +321,8 @@ post '/fminer/last/?' do
     fminer.smi = [] # AM LAST: needed for matching the patterns back
 
     # Add data to fminer
-    fminer.add_fminer_data(@@last, params, @value_map)
+    #fminer.add_fminer_data(@@last, params, @value_map)
+    fminer.add_fminer_data(@@last, @value_map)
 
     raise "No compounds in dataset #{fminer.training_dataset.uri}" if fminer.compounds.size==0
 
diff --git a/libfminer b/libfminer
-Subproject f9e560dc0a7a5d5af439814ab5fa9ce027a025b
+Subproject 4577dcacb352af3fdca925714dc570de9e02582