Merge branch 'feature/opentox-client' into development

Conflicts:
	application.rb

10 files changed, 453 insertions, 392 deletions

diff --git a/.gitignore b/.gitignore
index d21a58b..4040c6c 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,6 +1,4 @@
-api_key.rb
-*.sqlite3
-tmp/*
-log/*
-public/*
-data/*
+*.gem
+.bundle
+Gemfile.lock
+pkg/*
diff --git a/Gemfile b/Gemfile
new file mode 100644
index 0000000..36f1a90
--- /dev/null
+++ b/Gemfile
@@ -0,0 +1,4 @@
+source :gemcutter
+gemspec
+gem "opentox-server", :path => "~/opentox-server"
+gem "opentox-client", :path => "~/opentox-client"
diff --git a/Rakefile b/Rakefile
new file mode 100644
index 0000000..2995527
--- /dev/null
+++ b/Rakefile
@@ -0,0 +1 @@
+require "bundler/gem_tasks"
diff --git a/VERSION b/VERSION
new file mode 100644
index 0000000..3eefcb9
--- /dev/null
+++ b/VERSION
@@ -0,0 +1 @@
+1.0.0
diff --git a/application.rb b/application.rb
index d4bd78a..494f050 100644
--- a/application.rb
+++ b/application.rb
@@ -1,412 +1,452 @@
-require 'rubygems'
-gem "opentox-ruby", "~> 4"
-require 'opentox-ruby'
-require 'profiler'
-require 'rjb'
-
-set :lock, true
-
-@@datadir = "data"
-
-@@idfile_path = @@datadir+"/id"
-unless File.exist?(@@idfile_path)
-  id = Dir["./#{@@datadir}/*json"].collect{|f| File.basename(f.sub(/.json/,'')).to_i}.sort.last
-  id = 0 if id.nil?
-  open(@@idfile_path,"w") do |f|
-    f.puts(id)
-  end
-end
-
-helpers do
-  def next_id
-    open(@@idfile_path, "r+") do |f|
-      f.flock(File::LOCK_EX)
-      @id = f.gets.to_i + 1
-      f.rewind
-      f.print @id
-    end
-    return @id
-  end
+require 'roo'
+#require 'profiler'
 
-  def uri(id)
-    url_for "/#{id}", :full
-  end
+module OpenTox
+  class Application < Service
 
-  # subjectid ist stored as memeber variable, not in params
-  def load_dataset(id, params,content_type,input_data)
+    @warnings = []
 
-    @uri = uri id
-    raise "store subject-id in dataset-object, not in params" if params.has_key?(:subjectid) and @subjectid==nil
+    helpers do
 
-    content_type = "application/rdf+xml" if content_type.nil?
-    dataset = OpenTox::Dataset.new(@uri, @subjectid)
-
-    case content_type
-
-    when /yaml/
-      dataset.load_yaml(input_data)
-
-    when /json/
-      dataset.load_json(input_data)
+      def from_csv(csv)
+        from_table CSV.parse(csv)
+      end
 
-    when "text/csv"
-      dataset.load_csv(input_data, @subjectid)
+      def from_spreadsheet spreadsheet
+        extensions = { Excel => ".xls", Excelx => ".xlsx", Openoffice => ".ods" }
+        input = params[:file][:tempfile].path + ".xls"
+        csv_file = params[:file][:tempfile].path + ".csv"
+        File.rename params[:file][:tempfile].path, input # roo needs "correct" extensions
+        spreadsheet.new(input).to_csv csv_file # roo cannot write to strings
+        @body = from_csv File.read(csv_file)
+        @content_type = "text/plain"
+      end
 
-    when /application\/rdf\+xml/
-      dataset.load_rdfxml(input_data, @subjectid)
+=begin
+      def from_sdf(sdf)
+
+        #obconversion = OpenBabel::OBConversion.new
+        #obmol = OpenBabel::OBMol.new
+        #obconversion.set_in_and_out_formats "sdf", "inchi"
+
+        table = []
+
+        properties = []
+        sdf.each_line { |l| properties << l.to_s if l.match(/</) }
+        properties.sort!
+        properties.uniq!
+        properties.collect!{ |p| p.gsub(/<|>/,'').strip.chomp }
+        properties.insert 0, "InChI"
+        table[0] = properties
+
+        rec = 0
+        sdf.split(/\$\$\$\$\r*\n/).each do |s|
+          rec += 1
+          table << []
+          begin
+            # TODO: use compound service
+            compound = OpenTox::Compound.from_sdf sdf
+            #obconversion.read_string obmol, s
+            table.last << obconversion.write_string(obmol).gsub(/\s/,'').chomp 
+          rescue
+            # TODO: Fix, will lead to follow up errors
+            table.last << "Could not convert structure at record #{rec}) have been ignored! \n#{s}"
+          end
+          obmol.get_data.each { |d| table.last[table.first.index(d.get_attribute)] = d.get_value }
+        end
+        from_table table
+      end
+=end
+
+      def from_table table
+
+=begin
+        dataset = OpenTox::Dataset.new @uri
+        puts dataset.uri
+        feature_names = table.shift.collect{|f| f.strip}
+        puts feature_names.inspect
+        dataset.append RDF::OT.Warnings, "Duplicated features in table header." unless feature_names.size == feature_names.uniq.size
+        compound_format = feature_names.shift.strip
+        bad_request_error "#{compound_format} is not a supported compound format. Accepted formats: URI, SMILES, InChI." unless compound_format =~ /URI|URL|SMILES|InChI/i
+        features = []
+        feature_names.each_with_index do |f,i|
+          feature = OpenTox::Feature.new File.join($feature[:uri], SecureRandom.uuid)
+          feature[RDF::DC.title] = f
+          features << feature
+          values = table.collect{|row| row[i+1].strip unless row[i+1].nil?}.uniq.compact # skip compound column
+          if values.size <= 3 # max classes
+            feature.append RDF.type, RDF::OT.NominalFeature
+            feature.append RDF.type, RDF::OT.StringFeature
+            feature[RDF::OT.acceptValue] = values
+          else
+            types = values.collect{|v| feature_type(v)}
+            if types.include?(RDF::OT.NominalFeature)
+              dataset.append RDF::OT.Warnings, "Feature #{f} contains nominal and numeric values."
+            else
+              feature.append RDF.type, RDF::OT.NumericFeature
+            end
+          end
+          feature.put
+        end
+        dataset.features = features
+        compounds = []
+        table.each_with_index do |values,j|
+          c = values.shift
+          puts c
+          puts compound_format
+          values.collect!{|v| v.nil? ? nil : v.strip }
+          #begin
+            case compound_format
+            when /URI|URL/i
+              compound = OpenTox::Compound.new c
+            when /SMILES/i
+              compound = OpenTox::Compound.from_smiles($compound[:uri], c)
+            when /InChI/i
+              compound = OpenTox::Compound.from_inchi($compound[:uri], URI.decode_www_form_component(c))
+            end
+          #rescue
+            #dataset.append RDF::OT.Warnings, "Cannot parse compound \"#{c}\" at position #{j+2}, all entries are ignored."
+            #next
+          #end
+          unless compound_uri.match(/InChI=/)
+            dataset.append RDF::OT.Warnings, "Cannot parse compound \"#{c}\" at position #{j+2}, all entries are ignored."
+            next
+          end
+          compounds << compound
+          unless values.size == features.size
+            dataset.append RDF::OT.Warnings, "Number of values at position #{j+2} (#{values.size}) is different than header size (#{features.size}), all entries are ignored."
+            next
+          end
+
+          dataset << values
 
-    when "chemical/x-mdl-sdfile"
-      dataset.load_sdf(input_data, @subjectid)
+        end
+        dataset.compounds = compounds
+        compounds.duplicates.each do |compound|
+          positions = []
+          compounds.each_with_index{|c,i| positions << i+1 if c.uri == compound.uri}
+          dataset.append RDF::OT.Warnings, "Duplicated compound #{compound.uri} at rows #{positions.join(', ')}. Entries are accepted, assuming that measurements come from independent experiments." 
+        end
+        puts dataset.to_ntriples
+        dataset.to_ntriples
+=end
+
+        @warnings = []
+        ntriples = ["<#{@uri}> <#{RDF.type}> <#{RDF::OT.Dataset}>."]
+        ntriples << ["<#{@uri}> <#{RDF.type}> <#{RDF::OT.OrderedDataset}>."]
+
+        # features
+        feature_names = table.shift.collect{|f| f.strip}
+        @warnings << "Duplicated features in table header." unless feature_names.size == feature_names.uniq.size
+        compound_format = feature_names.shift.strip
+        bad_request_error "#{compound_format} is not a supported compound format. Accepted formats: URI, SMILES, InChI." unless compound_format =~ /URI|URL|SMILES|InChI/i
+        features = []
+        ignored_feature_indices = []
+        feature_names.each_with_index do |f,i|
+          feature = OpenTox::Feature.new File.join($feature[:uri], SecureRandom.uuid)
+          feature[RDF::DC.title] = f
+          features << feature
+          values = table.collect{|row| row[i+1].strip unless row[i+1].nil?}.uniq.compact # skip compound column
+          if values.size <= 3 # max classes
+            feature.append RDF.type, RDF::OT.NominalFeature
+            feature.append RDF.type, RDF::OT.StringFeature
+            feature[RDF::OT.acceptValue] = values
+          else
+            types = values.collect{|v| feature_type(v)}
+            if types.include?(RDF::OT.NominalFeature)
+              @warnings << "Feature #{f} contains nominal and numeric values."
+            else
+              feature.append RDF.type, RDF::OT.NumericFeature
+            end
+          end
+          feature.put
+          ntriples << "<#{feature.uri}> <#{RDF.type}> <#{RDF::OT.Feature}>."
+          ntriples << "<#{feature.uri}> <#{RDF::OLO.index}> #{i} ."
+        end
 
-    when /multipart\/form-data/ , "application/x-www-form-urlencoded" # file uploads
+        # compounds and values
+        compound_uris = []
+        table.each_with_index do |values,j|
+          values.collect!{|v| v.nil? ? nil : v.strip }
+          compound = values.shift
+          begin
+            case compound_format
+            when /URI|URL/i
+              compound_uri = compound
+            when /SMILES/i
+              compound_uri = OpenTox::Compound.from_smiles($compound[:uri], compound).uri
+            when /InChI/i
+              compound_uri = OpenTox::Compound.from_inchi($compound[:uri], URI.decode_www_form_component(compound)).uri
+            end
+          rescue
+            @warnings << "Cannot parse compound \"#{compound}\" at position #{j+2}, all entries are ignored."
+            next
+          end
+          unless compound_uri.match(/InChI=/)
+            @warnings << "Cannot parse compound \"#{compound}\" at position #{j+2}, all entries are ignored."
+            next
+          end
+          compound_uris << compound_uri
+          unless values.size == features.size
+            @warnings << "Number of values at position #{j+2} (#{values.size}) is different than header size (#{features.size}), all entries are ignored."
+            next
+          end
+          ntriples << "<#{compound_uri}> <#{RDF.type}> <#{RDF::OT.Compound}>."
+          ntriples << "<#{compound_uri}> <#{RDF::OLO.index}> #{j} ."
+
+          values.each_with_index do |v,i|
+            #@warnings << "Empty value for compound #{compound} (row #{j+2}) and feature \"#{feature_names[i]}\" (column #{i+2})." if v.blank?
+            #@warnings << "Empty value in row #{j+2}, column #{i+2} (feature \"#{feature_names[i]}\")." if v.blank?
+
+            data_entry_node = "_:dataentry"+ j.to_s
+            value_node = data_entry_node+ "_value"+ i.to_s
+            ntriples << "<#{@uri}> <#{RDF::OT.dataEntry}> #{data_entry_node} ."
+            ntriples << "#{data_entry_node} <#{RDF.type}> <#{RDF::OT.DataEntry}> ."
+            ntriples << "#{data_entry_node} <#{RDF::OLO.index}> #{j} ."
+            ntriples << "#{data_entry_node} <#{RDF::OT.compound}> <#{compound_uri}> ."
+            ntriples << "#{data_entry_node} <#{RDF::OT.values}> #{value_node} ."
+            ntriples << "#{value_node} <#{RDF::OT.feature}> <#{features[i].uri}> ."
+            ntriples << "#{value_node} <#{RDF::OT.value}> \"#{v}\" ."
+
+          end
 
-      case params[:file][:type]
+        end
+        compound_uris.duplicates.each do |uri|
+          positions = []
+          compound_uris.each_with_index{|c,i| positions << i+1 if c == uri}
+          @warnings << "Duplicated compound #{uri} at rows #{positions.join(', ')}. Entries are accepted, assuming that measurements come from independent experiments." 
+        end
 
-      when "chemical/x-mdl-sdfile"
-        dataset.load_sdf(input_data, @subjectid)
+        ntriples << "<#{@uri}> <#{RDF::OT.Warnings}> \"#{@warnings.join('\n')}\" ."
+        ntriples.join("\n")
+=begin
+=end
+      end
 
-      when /json/
-        dataset.load_json(params[:file][:tempfile].read)
+=begin
+      def to_xlsx
 
-      when /yaml/
-        dataset.load_yaml(params[:file][:tempfile].read)
+        # both simple_xlsx and axlsx create empty documents with OLE2 errors
+        xlsx = @uri.split("/").last+".xlsx"
+        p = Axlsx::Package.new
+        wb = p.workbook
+        wb.add_worksheet(:name => "test") do |sheet|
+          to_table.each { |row| sheet.add_row row; puts row }
+        end
+        p.serialize("test.xlsx")
+
+        p.to_stream
+#```
+        #Tempfile.open(@uri.split("/").last+".xlsx") do |xlsx|
+          SimpleXlsx::Serializer.new(xlsx) do |doc|
+            doc.add_sheet("People") do |sheet|
+              to_table.each { |row| sheet.add_row row }
+            end
+          end
+          send_file xlsx
+        #end
+      end
+=end
 
-      when "application/rdf+xml"
-        dataset.load_rdfxml_file(params[:file][:tempfile], @subjectid)
+      def to_csv
+        csv_string = CSV.generate do |csv|
+          to_table.each { |row| csv << row }
+        end
+      end
 
-      when "text/csv"
-        dataset.load_csv(params[:file][:tempfile].read, @subjectid)
-        dataset.add_metadata({
-        DC.title => File.basename(params[:file][:filename],".csv"),
-        OT.hasSource => File.basename(params[:file][:filename])
-      })
+      def to_table
+=begin
+        table = []
+        dataset = OpenTox::Dataset.new @uri
+        dataset.get
+        table << ["SMILES"] + dataset.features.collect{|f| f.get; f.title}
+        dataset.data_entries.each_with_index do |data_entry,i|
+          table << [dataset.compounds[i]] + data_entry
+        end
+        table
+=end
+        accept = "text/uri-list"
+        table  = []
+        if ordered?
+          sparql = "SELECT DISTINCT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Feature}> . ?s <#{RDF::OLO.index}> ?i} ORDER BY ?i"
+          features = FourStore.query(sparql, accept).split("\n").collect{|uri| OpenTox::Feature.new uri}
+          table << ["SMILES"] + features.collect{ |f| f.get; f[RDF::DC.title] }
+          sparql = "SELECT DISTINCT ?i FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.DataEntry}> . ?s <#{RDF::OLO.index}> ?i} ORDER BY ?i"
+          FourStore.query(sparql, accept).split("\n").each do |data_entry_idx|
+            sparql = "SELECT DISTINCT ?compound FROM <#{@uri}> WHERE {
+              ?data_entry <#{RDF::OLO.index}> #{data_entry_idx} ;
+                <#{RDF::OT.compound}> ?compound. }"
+            compound = OpenTox::Compound.new FourStore.query(sparql, accept).strip
+            sparql = "SELECT ?value FROM <#{@uri}> WHERE {
+              ?data_entry <#{RDF::OLO.index}> #{data_entry_idx} ;
+                <#{RDF::OT.values}> ?v .
+              ?v <#{RDF::OT.feature}> ?f;
+                <#{RDF::OT.value}> ?value .
+              ?f <#{RDF::OLO.index}> ?i.
+
+                } ORDER BY ?i"
+            values = FourStore.query(sparql,accept).split("\n")
+            # Fill up trailing empty cells
+            table << [compound.to_smiles] + values.fill("",values.size,features.size-values.size)
+          end
+        else
+          sparql = "SELECT DISTINCT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Feature}>}"
+          features = FourStore.query(sparql, accept).split("\n").collect{|uri| OpenTox::Feature.new uri}
+          table << ["SMILES"] + features.collect{ |f| f.get; f[RDF::DC.title] }
+          sparql = "SELECT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Compound}>. }"
+          compounds = FourStore.query(sparql, accept).split("\n").collect{|uri| OpenTox::Compound.new uri}
+          compounds.each do |compound|
+            data_entries = []
+            features.each do |feature|
+              sparql = "SELECT ?value FROM <#{@uri}> WHERE {
+                ?data_entry <#{RDF::OT.compound}> <#{compound.uri}>;
+                  <#{RDF::OT.values}> ?v .
+                ?v <#{RDF::OT.feature}> <#{feature.uri}>;
+                  <#{RDF::OT.value}> ?value.
+                  } ORDER BY ?data_entry"
+              FourStore.query(sparql, accept).split("\n").each_with_index do |value,i|
+                data_entries[i] = Array.new(features.size) unless data_entries[i]
+                data_entries[i] << value
+              end
+            end
+            data_entries.each{|data_entry| table << [compound.to_smiles] + data_entry}
+          end
+        end
+        table
+=begin
+=end
+      end
 
-      when /ms-excel/
-        extension =  File.extname(params[:file][:filename])
-        case extension
-        when ".xls"
-          xls = params[:file][:tempfile].path + ".xls"
-          File.rename params[:file][:tempfile].path, xls # roo needs these endings
-          book = Excel.new xls
-        when ".xlsx"
-          xlsx = params[:file][:tempfile].path + ".xlsx"
-          File.rename params[:file][:tempfile].path, xlsx # roo needs these endings
-          book = Excel.new xlsx
+      def feature_type(value)
+        if value.blank?
+          nil
+        elsif value.numeric?
+          RDF::OT.NumericFeature
         else
-          raise "#{params[:file][:filename]} is not a valid Excel input file."
+          RDF::OT.NominalFeature
         end
-        dataset.load_spreadsheet(book, @subjectid)
-        dataset.add_metadata({
-          DC.title => File.basename(params[:file][:filename],extension),
-          OT.hasSource => File.basename(params[:file][:filename])
-        })
+      end
 
-      else
-        raise "MIME type \"#{params[:file][:type]}\" not supported."
+      def ordered?
+        sparql = "SELECT DISTINCT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.OrderedDataset}>}"
+        FourStore.query(sparql, "text/uri-list").split("\n").empty? ? false : true
       end
 
-    else
-      raise "MIME type \"#{content_type}\" not supported."
+      def parse_put
+        task = OpenTox::Task.create $task[:uri], nil, RDF::DC.description => "Dataset upload: #{@uri}" do
+          #Profiler__::start_profile
+          case @content_type
+          when "text/plain", "text/turtle", "application/rdf+xml" # no conversion needed
+          when "text/csv"
+            @body = from_csv @body
+            @content_type = "text/plain"
+          when "application/vnd.ms-excel"
+            from_spreadsheet Excel
+          when "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+            from_spreadsheet Excelx
+          when "application/vnd.oasis.opendocument.spreadsheet"
+            from_spreadsheet Openoffice
+    #      when "chemical/x-mdl-sdfile"
+    #        @body = parse_sdf @body
+    #        @content_type = "text/plain"
+          else
+            bad_request_error "#{@content_type} is not a supported content type."
+          end
+          FourStore.put @uri, @body, @content_type
+          if params[:file]
+            nt = "<#{@uri}> <#{RDF::DC.title}> \"#{params[:file][:filename]}\".\n<#{uri}> <#{RDF::OT.hasSource}> \"#{params[:file][:filename]}\"."
+            FourStore.post(@uri, nt, "text/plain")
+          end
+          #Profiler__::stop_profile
+          #Profiler__::print_profile($stdout)
+          @uri
+        end
+        response['Content-Type'] = "text/uri-list"
+        task.uri
+      end
     end
 
-    dataset.uri = @uri # update uri (also in metdata)
-    dataset.features.keys.each { |f| dataset.features[f][OT.hasSource] = dataset.metadata[OT.hasSource] unless dataset.features[f][OT.hasSource]}
-    File.open("#{@@datadir}/#{@id}.json","w+"){|f| f.puts dataset.to_json}
-  end
-end
-
-before do
-
-  @accept = request.env['HTTP_ACCEPT']
-  @accept = 'application/rdf+xml' if @accept == '*/*' or @accept == '' or @accept.nil?
-  @id = request.path_info.match(/^\/\d+/)
-  unless @id.nil?
-    @id = @id.to_s.sub(/\//,'').to_i
-
-    @uri = uri @id
-    @json_file = "#{@@datadir}/#{@id}.json"
-    raise OpenTox::NotFoundError.new "Dataset #{@id} not found." unless File.exists? @json_file
-
-    extension = File.extname(request.path_info)
-    unless extension.empty?
-     case extension
-     when ".html"
-       @accept = 'text/html'
-     when ".json"
-       @accept = 'application/json'
-     when ".yaml"
-       @accept = 'application/x-yaml'
-     when ".csv"
-       @accept = 'text/csv'
-     when ".rdfxml"
-       @accept = 'application/rdf+xml'
-     when ".xls"
-       @accept = 'application/ms-excel'
-     when ".sdf"
-       @accept = 'chemical/x-mdl-sdfile'
-     else
-       raise OpenTox::NotFoundError.new "File format #{extension} not supported."
-     end
+    before "/#{SERVICE}/:id/:property" do
+      @uri = uri("/#{SERVICE}/#{params[:id]}")
     end
-  end
-
-  # make sure subjectid is not included in params, subjectid is set as member variable
-  params.delete(:subjectid)
-end
-
-## REST API
-
-# Get a list of available datasets
-# @return [text/uri-list] List of available datasets
-get '/?' do
-  uri_list = Dir["./#{@@datadir}/*json"].collect{|f| File.basename(f.sub(/.json/,'')).to_i}.sort.collect{|n| uri n}.join("\n") + "\n"
-  case @accept
-  when /html/
-    response['Content-Type'] = 'text/html'
-    OpenTox.text_to_html uri_list
-  else
-    response['Content-Type'] = 'text/uri-list'
-    uri_list
-  end
-end
 
-# Get a dataset representation
-# @param [Header] Accept one of `application/rdf+xml, application-x-yaml, text/csv, application/ms-excel` (default application/rdf+xml)
-# @return [application/rdf+xml, application-x-yaml, text/csv, application/ms-excel] Dataset representation
-get '/:id' do
-  case @accept
-
-  when /rdf/ # redland sends text/rdf instead of application/rdf+xml
-    file = "#{@@datadir}/#{params[:id]}.rdfxml"
-    unless File.exists? file # lazy rdfxml generation
-      dataset = OpenTox::Dataset.from_json File.read(@json_file)
-      File.open(file,"w+") { |f| f.puts dataset.to_rdfxml }
+    # Create a new resource
+    post "/dataset/?" do
+      @uri = uri("/#{SERVICE}/#{SecureRandom.uuid}")
+      parse_put
     end
-    send_file file, :type => 'application/rdf+xml'
 
-  when /json/
-    send_file @json_file, :type => 'application/json'
-
-  when /yaml/
-    file = "#{@@datadir}/#{params[:id]}.yaml"
-    unless File.exists? file # lazy yaml generation
-      dataset = OpenTox::Dataset.from_json File.read(@json_file)
-      File.open(file,"w+") { |f| f.puts dataset.to_yaml }
+    get "/dataset/:id/?" do
+      #Profiler__::start_profile
+      @accept = "text/html" if @accept == '*/*'
+      case @accept
+      when "application/rdf+xml", "text/turtle", "text/plain", /html/
+        r = FourStore.get(@uri, @accept)
+      else
+        case @accept
+        when "text/csv"
+          r = to_csv
+        #when "application/vnd.ms-excel"
+          #to_spreadsheet Excel
+        #when "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+          #to_xlsx
+        #when "application/vnd.oasis.opendocument.spreadsheet"
+          #to_spreadsheet Openoffice
+        #when "chemical/x-mdl-sdfile"
+        else
+          bad_request_error "'#{@accept}' is not a supported content type."
+        end
+      end
+      #Profiler__::stop_profile
+      #Profiler__::print_profile($stdout)
+      r
     end
-    send_file file, :type => 'application/x-yaml'
-
-  when /html/
-    response['Content-Type'] = 'text/html'
-    OpenTox.text_to_html JSON.pretty_generate(JSON.parse(File.read(@json_file)))
-
-  when "text/csv"
-    response['Content-Type'] = 'text/csv'
-    OpenTox::Dataset.from_json(File.read(@json_file)).to_csv
-
-  when /ms-excel/
-    file = "#{@@datadir}/#{params[:id]}.xls"
-    OpenTox::Dataset.from_json(File.read(@json_file)).to_xls.write(file) unless File.exists? file # lazy xls generation
-    send_file file, :type => 'application/ms-excel'
-
-  when /sdfile/
-    response['Content-Type'] = 'chemical/x-mdl-sdfile'
-    OpenTox::Dataset.from_json(File.read(@json_file)).to_sdf
-
-#  when /uri-list/
-#    response['Content-Type'] = 'text/uri-list'
-#    Yajl::Parser.parse(File.read(@json_file)).to_urilist
-
-  else
-    raise OpenTox::NotFoundError.new "Content-type #{@accept} not supported."
-  end
-end
-
-# Get metadata of the dataset
-# @return [application/rdf+xml] Metadata OWL-DL
-get '/:id/metadata' do
-  metadata = OpenTox::Dataset.from_json(File.read(@json_file)).metadata
-
-  case @accept
-  when /rdf/ # redland sends text/rdf instead of application/rdf+xml
-    response['Content-Type'] = 'application/rdf+xml'
-    serializer = OpenTox::Serializer::Owl.new
-    serializer.add_metadata url_for("/#{params[:id]}",:full), metadata
-    serializer.to_rdfxml
-  when /yaml/
-    response['Content-Type'] = 'application/x-yaml'
-    metadata.to_yaml
-  end
-
-end
-
-# Get a dataset feature
-# @param [Header] Accept one of `application/rdf+xml or application-x-yaml` (default application/rdf+xml)
-# @return [application/rdf+xml,application/x-yaml] Feature metadata
-get %r{/(\d+)/feature/(.*)$} do |id,feature|
-
-  @id = id
-  @uri = uri @id
-  @json_file = "#{@@datadir}/#{@id}.json"
-  feature_uri = url_for("/#{@id}/feature/#{URI.encode(feature)}",:full) # work around  racks internal uri decoding
-  metadata = OpenTox::Dataset.from_json(File.read(@json_file)).features[feature_uri]
-
-  case @accept
-  when /rdf/ # redland sends text/rdf instead of application/rdf+xml
-    response['Content-Type'] = 'application/rdf+xml'
-    serializer = OpenTox::Serializer::Owl.new
-    serializer.add_feature feature_uri, metadata
-    serializer.to_rdfxml
-  when /yaml/
-    response['Content-Type'] = 'application/x-yaml'
-    metadata.to_yaml
-  when /json/
-    response['Content-Type'] = 'application/json'
-    Yajl::Encoder.encode(metadata)
-  end
-
-end
-
-# Get a list of all features
-# @param [Header] Accept one of `application/rdf+xml, application-x-yaml, text/uri-list` (default application/rdf+xml)
-# @return [application/rdf+xml, application-x-yaml, text/uri-list] Feature list
-get '/:id/features' do
-
-  features = OpenTox::Dataset.from_json(File.read(@json_file)).features
-
-  case @accept
-  when /rdf/ # redland sends text/rdf instead of application/rdf+xml
-    response['Content-Type'] = 'application/rdf+xml'
-    serializer = OpenTox::Serializer::Owl.new
-    features.each { |feature,metadata| serializer.add_feature feature, metadata }
-    serializer.to_rdfxml
-  when /yaml/
-    response['Content-Type'] = 'application/x-yaml'
-    features.to_yaml
-  when /json/
-    response['Content-Type'] = 'application/json'
-    Yajl::Encoder.encode(features)
-  when "text/uri-list"
-    response['Content-Type'] = 'text/uri-list'
-    features.keys.join("\n") + "\n"
-  end
-end
-
-# Get a list of all compounds
-# @return [text/uri-list] Feature list
-get '/:id/compounds' do
-  response['Content-Type'] = 'text/uri-list'
-  OpenTox::Dataset.from_json(File.read(@json_file)).compounds.join("\n") + "\n"
-end
 
-# Create a new dataset.
-#
-# Posting without parameters creates and saves an empty dataset (with assigned URI).
-# Posting with parameters creates and saves a new dataset.
-# Data can be submitted either
-# - in the message body with the appropriate Content-type header or
-# - as file uploads with Content-type:multipart/form-data and a specified file type
-# @example
-#   curl -X POST -F "file=@training.csv;type=text/csv" http://webservices.in-silico.ch/dataset
-# @param [Header] Content-type one of `application/x-yaml, application/rdf+xml, multipart/form-data/`
-# @param [BODY] - string with data in selected Content-type
-# @param [optional] file, for file uploads, Content-type should be multipart/form-data, please specify the file type `application/rdf+xml, application-x-yaml, text/csv, application/ms-excel`
-# @return [text/uri-list] Task URI or Dataset URI (empty datasets)
-post '/?' do
-
-  response['Content-Type'] = 'text/uri-list'
-
-  # it could be that the read function works only once!, store in varible
-  input_data = request.env["rack.input"].read
-  @id = next_id
-  @uri = uri @id
-  @json_file = "#{@@datadir}/#{@id}.json"
-  if params.size == 0 and input_data.size==0
-    File.open(@json_file,"w+"){|f| f.puts OpenTox::Dataset.new(@uri).to_json}
-    OpenTox::Authorization.check_policy(@uri, @subjectid) if File.exists? @json_file
-    @uri
-  else
-    task = OpenTox::Task.create("Converting and saving dataset ", @uri) do
-      load_dataset @id, params, request.content_type, input_data
-      OpenTox::Authorization.check_policy(@uri, @subjectid) if File.exists? @json_file
-      @uri
+    # Create or updata a resource
+    put "/dataset/:id/?" do
+      parse_put
     end
-    raise OpenTox::ServiceUnavailableError.newtask.uri+"\n" if task.status == "Cancelled"
-    halt 202,task.uri+"\n"
-  end
-end
-
-# Save a dataset, will overwrite all existing data
-#
-# Data can be submitted either
-# - in the message body with the appropriate Content-type header or
-# - as file uploads with Content-type:multipart/form-data and a specified file type
-# @example
-#   curl -X POST -F "file=@training.csv;type=text/csv" http://webservices.in-silico.ch/dataset/1
-# @param [Header] Content-type one of `application/x-yaml, application/rdf+xml, multipart/form-data/`
-# @param [BODY] - string with data in selected Content-type
-# @param [optional] file, for file uploads, Content-type should be multipart/form-data, please specify the file type `application/rdf+xml, application-x-yaml, text/csv, application/ms-excel`
-# @return [text/uri-list] Task ID
-post '/:id' do
-  response['Content-Type'] = 'text/uri-list'
-  task = OpenTox::Task.create("Converting and saving dataset ", @uri) do
-    FileUtils.rm Dir["#{@@datadir}/#{@id}.*"]
-    load_dataset @id, params, request.content_type, request.env["rack.input"].read
-    @uri
-  end
-  raise OpenTox::ServiceUnavailableError.newtask.uri+"\n" if task.status == "Cancelled"
-  halt 202,task.uri.to_s+"\n"
-end
 
+    # Get metadata of the dataset
+    # @return [application/rdf+xml] Metadata OWL-DL
+    get '/dataset/:id/metadata' do
+      case @accept
+      when "application/rdf+xml", "text/turtle", "text/plain"
+        sparql = "CONSTRUCT {?s ?p ?o.} FROM <#{@uri}> WHERE {<#{@uri}> ?p ?o. }"
+        FourStore.query sparql, @accept
+      else
+        bad_request_error "'#{@accept}' is not a supported content type."
+      end
+    end
 
-# Deletes datasets that have been created by a crossvalidatoin that does not exist anymore
-# (This can happen if a crossvalidation fails unexpectedly)
-delete '/cleanup' do
-  Dir["./#{@@datadir}/*json"].each do |file|
-    dataset = OpenTox::Dataset.from_json File.read(file)
-    if dataset.metadata[DC.creator] && dataset.metadata[DC.creator] =~ /crossvalidation\/[0-9]/
-      begin
-        cv = OpenTox::Crossvalidation.find(dataset.metadata[DC.creator],@subjectid)
-        raise unless cv
-      rescue
-        LOGGER.debug "deleting #{dataset.uri}, crossvalidation missing: #{dataset.metadata[DC.creator]}"
-        begin
-          dataset.delete @subjectid
-        rescue
-        end
+    # Get a list of all features
+    # @param [Header] Accept one of `application/rdf+xml, text/turtle, text/plain, text/uri-list` (default application/rdf+xml)
+    # @return [application/rdf+xml, text/turtle, text/plain, text/uri-list] Feature list 
+    get '/dataset/:id/features' do
+      case @accept
+      when "application/rdf+xml", "text/turtle", "text/plain"
+        sparql = "CONSTRUCT {?s ?p ?o.} FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Feature}>; ?p ?o. }"
+      when "text/uri-list"
+        sparql = "SELECT DISTINCT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Feature}>. }"
+      else
+        bad_request_error "'#{@accept}' is not a supported content type."
       end
+      FourStore.query sparql, @accept
     end
-  end
-  "cleanup done"
-end
 
-# Delete a dataset
-# @return [text/plain] Status message
-delete '/:id' do
-  LOGGER.debug "deleting dataset with id #{@id}"
-  begin
-    FileUtils.rm Dir["#{@@datadir}/#{@id}.*"]
-    if @subjectid and !File.exists? @json_file and @uri
-      begin
-        res = OpenTox::Authorization.delete_policies_from_uri(@uri, @subjectid)
-        LOGGER.debug "Policy deleted for Dataset URI: #{@uri} with result: #{res}"
-      rescue
-        LOGGER.warn "Policy delete error for Dataset URI: #{@uri}"
+    # Get a list of all compounds
+    # @return [text/uri-list] Feature list 
+    get '/dataset/:id/compounds' do
+      case @accept
+      when "application/rdf+xml", "text/turtle", "text/plain"
+        sparql = "CONSTRUCT {?s ?p ?o.} FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Compound}>; ?p ?o. }"
+      when "text/uri-list"
+        sparql = "SELECT DISTINCT ?s FROM <#{@uri}> WHERE {?s <#{RDF.type}> <#{RDF::OT.Compound}>. }"
+      else
+        bad_request_error "'#{@accept}' is not a supported content type."
       end
+      FourStore.query sparql, @accept
     end
-    response['Content-Type'] = 'text/plain'
-    "Dataset #{@id} deleted."
-  rescue
-    raise OpenTox::NotFoundError.new "Dataset #{@id} does not exist."
   end
 end
 
-# Delete all datasets
-# @return [text/plain] Status message
-delete '/?' do
-  FileUtils.rm Dir["#{@@datadir}/*.rdfxml"]
-  FileUtils.rm Dir["#{@@datadir}/*.xls"]
-  FileUtils.rm Dir["#{@@datadir}/*.yaml"]
-  FileUtils.rm Dir["#{@@datadir}/*.json"]
-  response['Content-Type'] = 'text/plain'
-  "All datasets deleted."
-end
diff --git a/config.ru b/config.ru
index a1aab0d..a8d4293 100644
--- a/config.ru
+++ b/config.ru
@@ -1,6 +1,5 @@
-require 'rubygems'
-require 'opentox-ruby'
-require 'config/config_ru'
-run Sinatra::Application
-set :raise_errors, false
-set :show_exceptions, false
\ No newline at end of file
+SERVICE = "dataset"
+require 'bundler'
+Bundler.require
+require './application.rb'
+run OpenTox::Application
diff --git a/data/.gitignore b/data/.gitignore
deleted file mode 100644
index 59e3edd..0000000
--- a/data/.gitignore
+++ /dev/null
@@ -1,3 +0,0 @@
-*.rdfxml
-*.xls
-*.yaml
diff --git a/dataset.gemspec b/dataset.gemspec
new file mode 100644
index 0000000..58b427a
--- /dev/null
+++ b/dataset.gemspec
@@ -0,0 +1,25 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path("../lib", __FILE__)
+
+Gem::Specification.new do |s|
+  s.name        = "feature"
+  s.version     = File.read("./VERSION")
+  s.authors     = ["Christoph Helma"]
+  s.email       = ["helma@in-silico.ch"]
+  s.homepage    = ""
+  s.summary     = %q{OpenTox Dataset Service}
+  s.description = %q{OpenTox Dataset Service}
+  s.license     = 'GPL-3'
+
+  s.rubyforge_project = "dataset"
+
+  s.files         = `git ls-files`.split("\n")
+  s.required_ruby_version = '>= 1.9.2'
+
+  # specify any dependencies here; for example:
+  s.add_runtime_dependency "opentox-server"
+  s.add_runtime_dependency 'roo'
+  #s.add_runtime_dependency 'axlsx'
+  #s.add_runtime_dependency 'simple_xlsx_writer'
+  s.post_install_message = "Please configure your service in ~/.opentox/config/dataset.rb"
+end
diff --git a/public/.gitignore b/public/.gitignore
deleted file mode 100644
index debc7d4..0000000
--- a/public/.gitignore
+++ /dev/null
@@ -1,2 +0,0 @@
-*.rdfxml
-*.xls
diff --git a/public/robots.txt b/public/robots.txt
deleted file mode 100644
index 1f53798..0000000
--- a/public/robots.txt
+++ /dev/null
@@ -1,2 +0,0 @@
-User-agent: *
-Disallow: /