activerdf_net7 1.6.11

data/activerdf-jena/README

@@ -0,0 +1,57 @@
+This is the ActiveRDF adapter to the Jena RDF library.
+
+Features: 
+* supports memory, file and database stores; database
+  stores can be configured both with datasources and raw connection
+  parameters.  Jena currently supports Oracle, MySQL, HSQLDB,
+  PostgreSQL, MS SQL, and Derby.
+* supports reasoners, including Pellet and the built-in Jena reasoners
+* supports Lucene query support in ARQ
+  (since LARQ doesn't allow you to add statements
+  and reindex only those statements, whenever we've add a statement to a
+  triple store, we must rebuild the index at query time)
+
+Requirements: 
+* JRuby is required to natively execute Jena
+* the application can only use pure ruby gems
+
+License:
+included LGPL license (version 2 or later).
+
+
+------------
+
+Installing and running Activerdf with Jena: 
+=============================
+
+Download the newest JRuby binary distribution from http://jruby.codehaus.org/
+
+Unpack it somewhere, e.g. /usr/local/jruby
+
+Then set your environement: 
+
+export JRUBY_HOME=/usr/local/jruby
+export JAVA_HOM=/path/to/java/home
+export PATH=$JRUBY_HOME/bin:$PATH
+
+check if you now have the correct jruby commands in your path: 
+which jruby -> /usr/local/jruby/bin/jruby
+which gem -> /usr/local/jruby/bin/gem
+
+Now you can install rails and activerdf:
+gem install rails --include-dependencies --no-rdoc --no-ri
+gem install activerdf --include-dependencies
+gem install activerdf_jena --include-dependencies
+
+Create an instance of the jena adapter with simple file based persistence, and load some data into it:
+
+this_dir = File.dirname(File.expand_path(__FILE__))
+adapter = ConnectionPool.add_data_source(:type => :jena, 
+:model => "superfunky",
+:file => this_dir + "/jena_persistence")
+adapter.load("file://" + this_dir + "/test_data.rdf", :format => :rdfxml, :into => :default_model )
+
+For more information please see the RDoc, the adapter source code and the unit tests. Go to /path/to/jruby/lib/ruby/gems/1.8/gems/activerdf_jena-0.1/test/ for that.
+
+For more information see 
+http://wiki.activerdf.org/GettingStartedGuide

data/activerdf-jena/Rakefile

@@ -0,0 +1,87 @@
+require 'meta_project'
+require 'rake'
+require 'rake/testtask'
+require 'rake/clean'
+require 'rake/gempackagetask'
+require 'rake/contrib/xforge'
+require '../tools/rakehelp'
+require 'rubygems'
+require 'fileutils'
+include FileUtils
+
+$version  = IO.read('VERSION').strip
+$name     = 'activerdf_jena'
+$project  = MetaProject::Project::XForge::RubyForge.new('activerdf')
+$distdir  = "#$name-#$version"
+
+# setup tests and rdoc files
+setup_tests
+setup_clean ["pkg", "lib/*.bundle", "*.gem", ".config"]
+
+# default task: install
+desc 'test and package gem'
+task :default => :install
+
+# define package task
+setup_gem($name, $version) do |spec|
+  spec.summary = 'ActiveRDF adapter to the Jena RDF store'
+  spec.description = spec.summary
+  spec.author = 'Karsten Huneycutt and Benjamin Heitmann'
+  spec.email = 'benjamin.heitmann@deri.org'
+  spec.homepage = 'http://www.activerdf.org'
+  spec.platform = Gem::Platform::RUBY
+  spec.autorequire = 'active_rdf'
+  spec.add_dependency('gem_plugin', '>= 0.2.1')
+  spec.add_dependency('activerdf', '>= 1.6.4')
+  spec.add_dependency('activerdf_sparql', '>= 1.3.3')
+  spec.autorequire = 'init.rb'
+  spec.files += Dir['ext/**/*.jar']
+end
+
+begin
+  require 'rcov/rcovtask'
+  Rcov::RcovTask.new do |t|
+    t.test_files = FileList["activerdf-*/test/**/*.rb"]
+    t.verbose = true
+  end
+rescue LoadError
+end
+
+# define test_all task
+Rake::TestTask.new do |t|
+  t.name = :test_all
+  t.test_files = FileList["test/**/*.rb", "activerdf-*/test/**/*.rb"]
+end
+
+task :verify_rubyforge do
+  raise "RUBYFORGE_USER environment variable not set!" unless ENV['RUBYFORGE_USER']
+  raise "RUBYFORGE_PASSWORD environment variable not set!" unless ENV['RUBYFORGE_PASSWORD']
+end
+
+desc "release #$name-#$version gem on RubyForge"
+task :release => [ :clean, :verify_rubyforge, :package ] do
+  release_files = FileList["pkg/#$distdir.gem"]
+  Rake::XForge::Release.new($project) do |release|
+    release.user_name     = ENV['RUBYFORGE_USER']
+    release.password      = ENV['RUBYFORGE_PASSWORD']
+    release.files         = release_files.to_a
+    release.release_name  = "#$name #$version"
+    release.package_name  = "activerdf"
+    release.release_notes = ""
+
+    changes = []
+    File.open("CHANGELOG") do |file|
+      current = true
+
+      file.each do |line|
+        line.chomp!
+				if current and line =~ /^==/
+					current = false; next
+				end
+        break if line.empty? and not current
+        changes << line
+      end
+    end
+    release.release_changes = changes.join("\n")
+  end
+end

data/activerdf-jena/TODO

@@ -0,0 +1,18 @@
+administrative tasks:
+=========================
+* add API documentation to code
+* update CHANGELOG before release
+* describe which features work right now, and which should not be used
+* check what problems eyals rake file is causing for karsten
+
+
+before initial release: 
+=======================
+* test persistance to derby and maybe mysql and postgres
+
+after release:
+==============
+* get pellet to working (generally and specially with queries)
+* test reasoning
+* quads
+

data/activerdf-jena/VERSION

@@ -0,0 +1 @@
+0.2

data/activerdf-jena/lib/activerdf_jena/init.rb

@@ -0,0 +1,26 @@
+#
+# Author:  Karsten Huneycutt
+# Copyright 2007 Valkeir Corporation
+# License:  LGPL
+#
+# add the directory in which this file is located to the ruby loadpath
+file =
+if File.symlink?(__FILE__)
+  File.readlink(__FILE__)
+else
+  __FILE__
+end
+$: << File.dirname(File.expand_path(file))
+
+java_dir = File.expand_path(File.join(File.dirname(File.expand_path(file)), "..", "..", "ext"))
+
+Dir.foreach(java_dir) do |jar|
+  $CLASSPATH << File.join(java_dir, jar) if jar =~ /.jar$/
+end
+
+require 'jena'
+require 'ng4j'
+require 'pellet'
+require 'lucene'
+require 'jena_adapter'
+require 'ng4j_adapter'

data/activerdf-jena/lib/activerdf_jena/jena.rb

@@ -0,0 +1,59 @@
+#
+# Author:  Karsten Huneycutt
+# Copyright 2007 Valkeir Corporation
+# License:  LGPL
+#
+require 'java'
+
+module Jena
+
+  module Ontology
+    include_package('com.hp.hpl.jena.ontology')
+  end
+
+  module Model
+    include_package('com.hp.hpl.jena.rdf.model')
+  end
+
+  module DB
+    include_package('com.hp.hpl.jena.db')
+    
+    # this maps downcased Jena database types into drivers
+    DRIVER_MAP = {
+      'oracle' => 'oracle.jdbc.Driver',
+      'mysql' => 'com.mysql.jdbc.Driver',
+      'derby' => 'org.apache.derby.jdbc.EmbeddedDriver',
+      'postgresql' => 'org.postgresql.Driver',
+      'hsql' => 'org.hsqldb.jdbcDriver',
+      'mssql' => 'com.microsoft.sqlserver.jdbc.SQLServerDriver'
+    }
+    
+    DRIVER_MAP.each do |name, driver| 
+      av = "#{name}_available"
+      (class << self ; self ; end).send(:bool_accessor, av.to_sym)
+      begin
+        java.lang.Class.forName driver
+        Jena::DB.send("#{av}=", true)
+      rescue
+        Jena::DB.send("#{av}=", false)
+      end
+    end
+  end
+  
+  module Query
+    include_package('com.hp.hpl.jena.query')
+  end
+
+  module Reasoner
+    include_package('com.hp.hpl.jena.reasoner')
+  end
+
+  module Datatypes
+    include_package('com.hp.hpl.jena.datatypes')
+  end
+
+  module Graph
+    include_package('com.hp.hpl.jena.graph')
+  end
+
+end

data/activerdf-jena/lib/activerdf_jena/jena_adapter.rb

@@ -0,0 +1,515 @@
+#
+# Author:  Karsten Huneycutt
+# Copyright 2007 Valkeir Corporation
+# License:  LGPL
+#
+
+class JenaAdapter < ActiveRdfAdapter
+
+  class JenaAdapterConfigurationError < StandardError
+  end
+
+  class DataSourceDBConnection < Jena::DB::DBConnection
+
+    attr_accessor :datasource, :connection
+
+    def initialize(datasource, type)
+      if datasource.kind_of? javax.sql.DataSource
+        self.datasource = datasource
+      else
+        self.datasource = javax.naming.InitialContext.new.lookup(datasource)
+      end
+      self.setDatabaseType(type)
+    end
+
+    def getConnection
+      if !self.connection || !valid_connection?(self.connection)
+        self.connection = self.datasource.getConnection
+      end
+      self.connection
+    end
+
+    def close
+      self.datasource = nil
+    end
+    
+    def valid_connection?(cnxn)
+      true
+    end
+
+  end
+
+  ConnectionPool.register_adapter(:jena, self)
+
+  bool_accessor :keyword_search, :reasoning
+  bool_accessor :lucene_index_behind
+  attr_accessor :ontology_type, :model_name, :reasoner, :connection
+  attr_accessor :model_maker, :base_model, :model, :lucene_index
+  attr_accessor :root_directory
+
+  # :database 
+  #   either use :url, :type, :username, AND :password (for a
+  #   regular connection) OR :datasource AND :type (for a container 
+  #   connection), default to memory data store
+  #   example for a derby connection:
+  #     :database => {:url => "jdbc:derby:superfunky;create=true", :type => "Derby", :username => "", :password => ""}
+  # :file
+  #   database wins over this, this wins over memory store.  parameter is
+  #   a string or file indicating the root directory for all files.
+  # :model
+  #   name of model to use, default is jena's default
+  # :ontology
+  #   set to language type if this needs to be viewed as an ontology, 
+  #   default nil, available :owl, :owl_dl, :owl_lite, :rdfs
+  #   pellet only supports owl reasoning. 
+  # :reasoner 
+  #   set to reasoner to use -- default nil (none).  options:  :pellet, 
+  #   :transitive, :rdfs, :rdfs_simple, :owl_micro, :owl_mini, :owl, 
+  #   :generic_rule
+  # :lucene
+  #   set to true to enable true lucene indexing of this store, default false
+  def initialize(params = {})
+    dbparams = params[:database]
+    self.ontology_type = params[:ontology]
+    self.reasoner = params[:reasoner]
+    self.keyword_search = params[:lucene]
+    
+    # if the model name is not provided and file persistence is used, then jena just
+    # creates random files in the tmp dir. not good, as we need to know the model name
+    # to have persistence
+    if params[:model]
+      self.model_name = params[:model]
+    else
+      self.model_name = "default"
+    end
+    
+    if params[:file]
+      if params[:file].respond_to? :path
+        self.root_directory = File.expand_path(params[:file].path)
+      else
+        self.root_directory = params[:file]
+      end
+    end
+
+    # do some sanity checking
+    if self.keyword_search? && !LuceneARQ.lucene_available?
+      raise JenaAdapterConfigurationError, "Lucene requested but is not available"
+    end
+    
+    if self.reasoner == :pellet && !Pellet.pellet_available?
+      raise JenaAdapterConfigurationError, "Pellet requested but not available"
+    end
+
+    if self.reasoner && !self.ontology_type
+      raise JenaAdapterConfigurationError, "Ontology model needed for reasoner"
+    end
+
+    if dbparams
+      if dbparams[:datasource]
+        self.connection = DataSourceDBConnection.new(dbparams[:datasource],
+                                                     dbparams[:type])
+      else
+        begin
+          if !Jena::DB.send("#{dbparams[:type].downcase}_available?")
+            raise JenaAdapterConfigurationError, "database type #{dbparams[:type]} not available"
+          end
+        rescue NameError
+          raise JenaAdapterConfigurationError, "database type #{dbparams[:type]} not recognized"
+        end
+
+        self.connection = Jena::DB::DBConnection.new(dbparams[:url], 
+                                                     dbparams[:username],
+                                                     dbparams[:password],
+                                                     dbparams[:type])
+      end
+
+      self.model_maker = Jena::Model::ModelFactory.createModelRDBMaker(connection)
+      
+    elsif self.root_directory
+      self.model_maker = Jena::Model::ModelFactory.createFileModelMaker(self.root_directory)
+    else
+      self.model_maker = Jena::Model::ModelFactory.createMemModelMaker
+    end
+    
+
+    self.base_model = self.model_maker.openModel(model_name)
+    
+    if self.ontology_type
+      rf = map_reasoner_factory(self.reasoner)
+      onturi = map_ontology_type(self.ontology_type)
+
+      spec = 
+        Jena::Ontology::OntModelSpec.new(self.model_maker,
+                                         Jena::Ontology::OntDocumentManager.new,
+                                         rf, onturi)
+
+      self.model = Jena::Model::ModelFactory.
+        createOntologyModel(spec, self.base_model)
+      self.reasoning = true
+    else
+      self.model = self.base_model
+      self.reasoning = false
+    end
+    
+    self.reads = true
+    self.writes = true
+    
+    self
+  end
+
+
+  def size
+    self.model.size
+  end
+
+  def dump
+    it = self.model.listStatements
+    res = ""
+    while it.hasNext
+      res += it.nextStatement.asTriple.toString
+      res += " . \n"
+    end
+    res
+  end
+
+  def close
+    ConnectionPool.remove_data_source(self)
+    self.model.close
+    self.connection.close unless self.connection.nil?
+  end
+
+  def clear
+    self.model.removeAll
+    self.model.prepare if self.model.respond_to? :prepare
+    self.model.rebind if self.model.respond_to? :rebind
+  end
+
+
+  def delete(subject, predicate, object, context = nil)
+    self.lucene_index_behind = true
+    mod = get_model_for_context(context)
+    s = (is_wildcard?(subject) ? nil : build_subject(subject, mod))
+    p = (is_wildcard?(predicate) ? nil : build_predicate(predicate, mod))
+    o = (is_wildcard?(object) ? nil : build_object(object, mod))
+    mod.removeAll(s, p, o)
+    mod.prepare if mod.respond_to? :prepare
+    mod.rebind if mod.respond_to? :rebind
+  end
+
+  def add(subject, predicate, object, context = nil)
+    self.lucene_index_behind = true
+    mod = get_model_for_context(context)
+    mod.add(build_statement(subject, predicate, object))
+    mod.prepare if mod.respond_to? :prepare
+    mod.rebind if mod.respond_to? :rebind
+  end
+
+  def flush
+    # no-op
+  end
+
+
+  # :format
+  #   format -- :ntriples, :n3, or :rdfxml, default :rdfxml
+  # :into
+  #   either the name of a model, :default_model for the main model, or 
+  #   :submodel to load into an anonymous memory model, default is :submodel
+  #   if this is an ontology, :default_model if it's not.
+  # :rebind 
+  #   rebind with the inferencer, default true; no effect if no inferencer
+  def load(uri, params = {})
+    into = params[:into] ? params[:into] : 
+      (self.ontology_type ? :submodel : :default_model)
+    format = params[:format] ? params[:format] : :rdfxml
+    rebind = params[:rebind] ? params[:rebind] : true
+    
+    jena_format =
+      case format
+      when :rdfxml
+        "RDF/XML"
+      when :ntriples
+        "N-TRIPLE"
+      when :n3
+        "N3"
+      end
+
+    case into
+    when :default_model
+      self.model.read(uri, jena_format)
+      
+    when :submodel
+      self.model.addSubModel(Jena::Model::ModelFactory.createDefaultModel.read(uri, jena_format))
+      
+    else
+      self.model.addSubModel(self.model_maker.createModel(into).read(uri, jena_format))
+    end
+    
+    if rebind && self.reasoner && self.model.respond_to?(:rebind)
+      self.model.rebind
+    end
+
+    self.lucene_index_behind = true
+    
+  end  
+
+  # this method gets called by the ActiveRDF query engine
+  def query(query, params = {})
+
+    if self.keyword_search? && query.keyword?
+            
+      # duplicate the query
+      query_with_keywords = query.dup
+      
+      # now duplicate the where stuff so we can fiddle with it...
+      # this is GROSS -- fix this if Query ever sprouts a proper
+      # deep copy or a where_clauses setter
+      query_with_keywords.instance_variable_set("@where_clauses", query.where_clauses.dup)
+
+      # now, for each of the keyword clauses, set up the search
+      query.keywords.each do |var, keyword|
+        # use this if activerdf expects the subject to come back and not the
+        # literal and using indexbuilderstring
+        #query.where("lucene_literal_#{var}".to_sym, LuceneARQ::KEYWORD_PREDICATE, keyword)
+        #query.where(var, "lucene_property_#{var}".to_sym, "lucene_literal_#{var}".to_sym)
+
+        # use this if activerdf expects the literal to come back, not the 
+        # subject, or if using indexbuildersubject (which makes the subject
+        # come back instead of the literal
+        query_with_keywords.where(var, RDFS::Resource.new(LuceneARQ::KEYWORD_PREDICATE), keyword)
+
+      end
+
+    else
+      query_with_keywords = query
+    end
+
+    # jena knows about lucene, so use the query object that has the keyword
+    # search requests expanded.
+    jena_results = query_jena(query_with_keywords)
+
+    # use the conjunctive query facility in pellet to get additional 
+    # answers, if we're using pellet and we don't have a pure keyword
+    # query
+    if self.reasoner == :pellet && query.where_clauses.size > 0
+      # pellet doesn't know about lucene, so we use the original query
+      # object
+      pellet_results = query_pellet(query)
+      results = (jena_results + pellet_results).uniq!
+    else
+      results = jena_results
+    end
+
+    if query.ask?
+      return [[true]] if results.size > 0
+      return [[false]]
+    end
+    
+    if query.count?
+      return results.size
+    end
+
+    results
+    
+  end
+
+  # ==========================================================================
+  # put private methods here to seperate api methods from the 
+  # inner workings of the adapter
+  private
+
+  def map_ontology_type(type)
+    case type
+    when :rdfs
+      'http://www.w3.org/2000/01/rdf-schema#'
+    when :owl
+      'http://www.w3.org/2002/07/owl#'
+    when :owl_dl
+      'http://www.w3.org/TR/owl-features/#term_OWLDL'
+    when :owl_lite
+      'http://www.w3.org/TR/owl-features/#term_OWLLite'
+    else
+      type
+    end
+  end
+
+
+  def map_reasoner_factory(type)
+    case type
+    when :pellet
+      Pellet.reasoner_factory
+      
+    when :transitive
+      com.hp.hpl.jena.reasoner.transitiveReasoner.TransitiveReasonerFactory.theInstance
+      
+    when :rdfs
+      com.hp.hpl.jena.reasoner.rulesys.RDFSFBRuleReasonerFactory.theInstance
+      
+    when :rdfs_simple
+      com.hp.hpl.jena.reasoner.rulesys.RDFSRuleReasonerFactory.theInstance
+      
+    when :owl_micro
+      com.hp.hpl.jena.reasoner.rulesys.OWLMicroReasonerFactory.theInstance
+      
+    when :owl_mini
+      com.hp.hpl.jena.reasoner.rulesys.OWLMiniReasonerFactory.theInstance
+      
+    when :owl 
+      com.hp.hpl.jena.reasoner.rulesys.OWLFBRuleReasonerFactory.theInstance
+      
+    when :generic_rule
+      com.hp.hpl.jena.reasoner.rulesys.GenericRuleReasonerFactory.theInstance
+      
+    else
+      type
+    end
+  end
+
+  def appropriate_model(submodel)
+    (submodel && (submodel != self.model))? submodel : self.model
+  end
+
+  def build_object(object, submodel = nil)
+    mod = appropriate_model(submodel)
+    if object.respond_to? :uri
+      o = mod.getResource(object.uri)
+    else
+      #xlate to literal
+      if !object.kind_of? Literal
+        objlit = Literal.new object
+      else
+        objlit = object
+      end
+      
+      if objlit.type
+        type = Jena::Datatypes::TypeMapper.getInstance.getTypeByName(objlit.type.uri)
+        o = mod.createTypedLiteral(objlit.value, type)
+      elsif objlit.language
+        o = mod.createLiteral(objlit.value, objlit.language)
+      else
+        o = mod.createTypedLiteral(objlit.value, nil)
+      end
+    end    
+    return o
+  end
+
+  def build_subject(subject, submodel = nil) 
+    # ensure it exists in the parent model
+    self.model.getResource(subject.uri) if submodel
+    appropriate_model(submodel).getResource(subject.uri)
+  end
+
+  def build_predicate(predicate, submodel = nil)
+    appropriate_model(submodel).getProperty(predicate.uri)
+  end
+
+  def build_statement(subject, predicate, object, submodel = nil)
+    s = build_subject(subject, submodel)
+    p = build_predicate(predicate, submodel)
+    o = build_object(object, submodel)
+    mod = submodel ? submodel : self.model
+    mod.createStatement(s, p, o)
+  end
+
+
+  def is_wildcard?(thing)
+    (thing == nil) || thing.kind_of?(Symbol)
+  end
+
+  def get_model_for_context(context)
+    if (context == nil || context == self.model_name)
+      self.model
+    else
+      subm = self.model_maker.openModel(context)
+      self.model.addSubModel(subm)
+      subm
+    end
+  end
+
+  def query_jena(query)
+    query_sparql = translate(query)    
+    
+    qexec = Jena::Query::QueryExecutionFactory.create(query_sparql, self.model)
+
+    # PROBABLY A VERY EXPENSIVE OPERATION (rebuilds lucene index if ANYTHING
+    # changed -- this seems to be the only way, since you have to close
+    # the index after you build it...
+    if query.keyword? && self.keyword_search?
+      LuceneARQ::LARQ.setDefaultIndex(qexec.getContext, retrieve_lucene_index)
+    end
+
+    begin 
+      results = perform_query(query, qexec)
+    ensure
+      qexec.close
+    end
+
+    results
+  end
+
+  def query_pellet(query)
+    query_sparql = translate(query)
+    jena_query = Jena::Query::QueryFactory.create(query_sparql)
+    
+    # bail if not a select
+    return [] if !jena_query.isSelectType
+
+    qexec = Pellet::Query::PelletQueryExecution.new(jena_query, self.model)
+    
+    begin
+      results = perform_query(query, qexec)
+    ensure
+      qexec.close
+    end
+    
+    results
+  end
+
+  def perform_query(query, qexec)
+    results = qexec.execSelect
+    arr_results = []
+    
+    while results.hasNext
+      row = results.nextSolution
+      res_row = []
+      query.select_clauses.each do |kw|
+        thing = row.get(kw.to_s)
+        if thing.kind_of? Jena::Model::Resource
+          if thing.isAnon
+            res_row << BNode.new(thing.getId.to_s)
+          else
+            res_row << RDFS::Resource.new(thing.to_s)
+          end
+        elsif thing.kind_of? Jena::Model::Literal
+          if thing.getLanguage == "" and thing.getDatatypeURI.nil?
+            # plain literal
+            res_row << thing.getString
+          elsif thing.getLanguage == ""
+            # datatyped literal
+            res_row << Literal.new(thing.getValue, RDFS::Resource.new(thing.getDatatypeURI))
+          elsif thing.getDatatypeURI.nil?
+            # language tagged literal 
+            res_row << Literal.new(thing.getLexicalForm, "@" + thing.getLanguage)
+          else
+            raise ActiveRdfError, "Jena Sparql returned a strange literal"
+          end
+        else
+          raise ActiveRdfError, "Returned thing other than resource or literal"
+        end
+      end
+      arr_results << res_row
+    end
+    arr_results
+  end
+
+  def retrieve_lucene_index
+    if self.lucene_index_behind?
+      builder = LuceneARQ::IndexBuilderSubject.new
+      builder.indexStatements(self.model.listStatements)
+      builder.closeForWriting
+      self.lucene_index = builder.getIndex
+      self.lucene_index_behind = false
+    end
+    self.lucene_index
+  end
+
+end