code cleanup and refactoring. 4store moved from opentox-client.

3 files changed, 141 insertions, 12 deletions

diff --git a/lib/4store.rb b/lib/4store.rb
new file mode 100644
index 0000000..d03a38e
--- /dev/null
+++ b/lib/4store.rb
@@ -0,0 +1,120 @@
+module OpenTox
+  module Backend
+    class FourStore
+
+      @@accept_formats = [ "application/rdf+xml", "text/turtle", "text/plain", "text/uri-list", "text/html", 'application/sparql-results+xml' ]
+      @@content_type_formats = [ "application/rdf+xml", "text/turtle", "text/plain", "application/x-turtle" ]
+
+      def self.list mime_type
+        bad_request_error "'#{mime_type}' is not a supported mime type. Please specify one of #{@@accept_formats.join(", ")} in the Accept Header." unless @@accept_formats.include? mime_type
+        if mime_type =~ /(uri-list|html)/
+          sparql = "SELECT DISTINCT ?g WHERE {GRAPH ?g {?s <#{RDF.type}> <#{klass}>; <#{RDF::DC.modified}> ?o.} } ORDER BY ?o"
+        else 
+          sparql = "CONSTRUCT {?s ?p ?o.} WHERE {?s <#{RDF.type}> <#{klass}>; ?p ?o. }"
+        end
+        query sparql, mime_type
+      end
+
+      def self.get uri, mime_type
+        bad_request_error "'#{mime_type}' is not a supported mime type. Please specify one of #{@@accept_formats.join(", ")} in the Accept Header." unless @@accept_formats.include? mime_type
+        sparql = "CONSTRUCT {?s ?p ?o.} FROM <#{uri}> WHERE { ?s ?p ?o. }"
+        rdf = query sparql, mime_type
+        resource_not_found_error "#{uri} not found." if rdf.empty?
+        rdf
+      end
+
+      def self.post uri, rdf, mime_type
+        bad_request_error "'#{mime_type}' is not a supported content type. Please use one of #{@@content_type_formats.join(", ")}." unless @@content_type_formats.include? mime_type or mime_type == "multipart/form-data"
+        bad_request_error "Request body empty." unless rdf 
+        mime_type = "application/x-turtle" if mime_type == "text/plain" # ntriples is turtle in 4store
+        RestClient.post File.join(four_store_uri,"data")+"/", :data => rdf.gsub(/\\C/,'C'), :graph => uri, "mime-type" => mime_type # remove backslashes in SMILES (4store interprets them as UTF-8 \C even within single quoates)
+        update "INSERT DATA { GRAPH <#{uri}> { <#{uri}> <#{RDF::DC.modified}> \"#{DateTime.now}\" } }"
+      end
+
+      def self.put uri, rdf, mime_type
+        bad_request_error "'#{mime_type}' is not a supported content type. Please use one of #{@@content_type_formats.join(", ")}." unless @@content_type_formats.include? mime_type
+        bad_request_error "Reqest body empty." unless rdf 
+        mime_type = "application/x-turtle" if mime_type == "text/plain"
+        RestClientWrapper.put File.join(four_store_uri,"data",uri), rdf, :content_type => mime_type 
+        update "INSERT DATA { GRAPH <#{uri}> { <#{uri}> <#{RDF::DC.modified}> \"#{DateTime.now}\" } }"
+      end
+
+      def self.delete uri
+        RestClientWrapper.delete data_uri(uri)
+      end
+
+      def self.update sparql
+        RestClient.post(update_uri, :update => sparql )
+      rescue
+        bad_request_error $!.message, update_uri
+      end
+
+      def self.query sparql, mime_type
+        if sparql =~ /SELECT/i
+          # return list unless mime_type
+          case mime_type
+          when 'application/sparql-results+xml' 
+            RestClient.get(sparql_uri, :params => { :query => sparql }, :accept => mime_type).body
+          when 'application/json'
+            RestClient.get(sparql_uri, :params => { :query => sparql }, :accept => mime_type).body
+          when /(uri-list|html)/
+            uri_list = RestClient.get(sparql_uri, :params => { :query => sparql }, :accept => "text/plain").body.gsub(/"|<|>/,'').split("\n").drop(1).join("\n")
+            uri_list = uri_list.to_html if mime_type=~/html/
+            return uri_list
+          else
+            bad_request_error "#{mime_type} is not a supported mime type for SELECT statements."
+          end
+        elsif sparql =~ /CONSTRUCT/i
+          case mime_type
+          when "text/plain", "application/rdf+xml" 
+            RestClient.get(sparql_uri, :params => { :query => sparql }, :accept => mime_type).body
+          when /html|turtle/
+            # TODO: fix and improve
+            nt = RestClient.get(sparql_uri, :params => { :query => sparql }, :accept => "text/plain").body # 4store returns ntriples for turtle
+
+            rdf = RDF::Graph.new
+            RDF::Reader.for(:ntriples).new(nt) do |reader|
+              reader.each_statement { |statement| rdf << statement }
+            end
+            prefixes = {:rdf => "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"}
+            ['OT', 'DC', 'XSD', 'OLO'].each{|p| prefixes[p.downcase.to_sym] = eval("RDF::#{p}.to_s") }
+            # TODO: fails for large datasets?? multi_cell_call
+            turtle = RDF::N3::Writer.for(:turtle).buffer(:prefixes => prefixes)  do |writer|
+              rdf.each{|statement| writer << statement}
+            end
+            regex = Regexp.new '(https?:\/\/[\S]+)([>"])'
+            turtle =  "<html><body>" + turtle.gsub( regex, '<a href="\1">\1</a>\2' ).gsub(/\n/,'<br/>') + "</body></html>" if mime_type =~ /html/ and !turtle.empty?
+            turtle
+          end
+        else
+          # TODO: check if this prevents SPARQL injections
+          bad_request_error "Only SELECT and CONSTRUCT are accepted SPARQL statements."
+        end
+      rescue
+        bad_request_error $!.message, sparql_uri
+      end
+
+      def self.klass
+        RDF::OT[SERVICE.capitalize]
+      end
+
+      def self.four_store_uri
+        # credentials are removed from uri in error.rb
+        $four_store[:uri].sub(%r{//},"//#{$four_store[:user]}:#{$four_store[:password]}@")
+      end
+
+      def self.sparql_uri 
+        File.join(four_store_uri, "sparql") + '/'
+      end
+
+      def self.update_uri 
+        File.join(four_store_uri, "update") + '/'
+      end
+
+      def self.data_uri uri
+        File.join(four_store_uri, "data","?graph=#{uri}")
+      end
+
+    end
+  end
+end
diff --git a/lib/opentox-server.rb b/lib/opentox-server.rb
index 5e3aada..5ff3c22 100644
--- a/lib/opentox-server.rb
+++ b/lib/opentox-server.rb
@@ -2,5 +2,4 @@ require 'opentox-client'
 require 'rack'
 require 'rack/contrib'
 require 'sinatra'
-require File.join(File.dirname(__FILE__),"opentox.rb")
-require File.join(File.dirname(__FILE__),"authorization-helper.rb")
+["4store.rb", "opentox.rb", "authorization-helper.rb"].each {|f| require File.join(File.dirname(__FILE__),f) }
diff --git a/lib/opentox.rb b/lib/opentox.rb
index 894033d..eb35b38 100644
--- a/lib/opentox.rb
+++ b/lib/opentox.rb
@@ -22,11 +22,17 @@ module OpenTox
     configure :development do
       register Sinatra::Reloader
       also_reload "./*.rb"
+      also_reload "./**/*.rb"
       also_reload "../opentox-client/lib/*.rb"
       also_reload File.join(ENV["HOME"],".opentox","config","#{SERVICE}.rb")
     end
 
     before do
+      @uri = uri(request.env['REQUEST_URI'])
+      # fix IE
+      request.env['HTTP_ACCEPT'] += ";text/html" if request.env["HTTP_USER_AGENT"]=~/MSIE/
+      request.env['HTTP_ACCEPT'] = request.params["media"] if request.params["media"]
+
       request.content_type ? response['Content-Type'] = request.content_type : response['Content-Type'] = request.env['HTTP_ACCEPT']
       parse_input if request.request_method =~ /POST|PUT/
       @accept = request.env['HTTP_ACCEPT']
@@ -35,10 +41,6 @@ module OpenTox
       response['Content-Type'] = @accept
     end
 
-    before "/#{SERVICE}/:id" do
-      @uri = uri("/#{SERVICE}/#{params[:id]}")
-    end
-
     helpers do
       def parse_input
         case request.content_type 
@@ -56,7 +58,7 @@ module OpenTox
       end
     end
 
-    # Attention: Error within tasks are catched by Task.create
+    # Attention: Error within tasks are catched by Task.run
     error do
       error = request.env['sinatra.error']
       if error.respond_to? :report
@@ -97,10 +99,18 @@ module OpenTox
     # Default methods, may be overwritten by derived services
     # see http://jcalcote.wordpress.com/2008/10/16/put-or-post-the-rest-of-the-story/
 
-    # Get a list of objects at the server
+    # Get a list of objects at the server or perform a SPARQL query
     get "/#{SERVICE}/?" do
-      $logger.debug "listing all #{SERVICE} objects"
-      FourStore.list @accept
+      if params[:query]
+        case @accept
+        when "text/uri-list" # result URIs are protected by A+A
+          FourStore.query(params[:query], "text/uri-list") 
+        else # prevent searches for protected resources
+          bad_request_error "Accept header '#{@accept}' is disabled for SPARQL queries at service URIs in order to protect private data. Use 'text/uri-list' and repeat the query at the result URIs.", uri("/#{SERVICE}")
+        end
+      else
+        FourStore.list(@accept)
+      end
     end
 
     # Create a new resource
@@ -111,9 +121,9 @@ module OpenTox
       @uri
     end
 
-    # Get resource representation
+    # Get resource representation or perform a SPARQL query
     get "/#{SERVICE}/:id/?" do
-      FourStore.get(@uri, @accept)
+      params[:query] ?  FourStore.query(params[:query], @accept) : FourStore.get(@uri, @accept)
     end
 
     # Modify (i.e. add rdf statments to) a resource