lib/json/ld/writer.rb in json-ld-1.1.7 vs lib/json/ld/writer.rb in json-ld-1.1.8

@@ -1,5 +1,6 @@
+require 'json/ld/streaming_writer'
 module JSON::LD
   ##
   # A JSON-LD parser in Ruby.
   #
   # Note that the natural interface is to write a whole graph at a time.
@@ -8,11 +9,11 @@
   #
   # @example Obtaining a JSON-LD writer class
   #   RDF::Writer.for(:jsonld)         #=> JSON::LD::Writer
   #   RDF::Writer.for("etc/test.json")
   #   RDF::Writer.for(:file_name      => "etc/test.json")
-  #   RDF::Writer.for(:file_extension => "json")
+  #   RDF::Writer.for(file_extension: "json")
   #   RDF::Writer.for(:content_type   => "application/turtle")
   #
   # @example Serializing RDF graph into an JSON-LD file
   #   JSON::LD::Writer.open("etc/test.json") do |writer|
   #     writer << graph
@@ -34,13 +35,13 @@
   #
   # The writer will add prefix definitions, and use them for creating @context definitions, and minting CURIEs
   #
   # @example Creating @@context prefix definitions in output
   #   JSON::LD::Writer.buffer(
-  #     :prefixes => {
+  #     prefixes: {
   #       nil => "http://example.com/ns#",
-  #       :foaf => "http://xmlns.com/foaf/0.1/"}
+  #       foaf: "http://xmlns.com/foaf/0.1/"}
   #   ) do |writer|
   #     graph.each_statement do |statement|
   #       writer << statement
   #     end
   #   end
@@ -49,10 +50,11 @@
   #
   # @see http://json-ld.org/spec/ED/20110507/
   # @see http://json-ld.org/spec/ED/20110507/#the-normalization-algorithm
   # @author [Gregg Kellogg](http://greggkellogg.net/)
   class Writer < RDF::Writer
+    include StreamingWriter
     include Utils
     format Format
 
     # @!attribute [r] graph
     # @return [RDF::Graph] Graph of statements serialized
@@ -87,20 +89,23 @@
     #   context to use when serializing. Constructed context for native serialization.
     # @option options [IO, Array, Hash, String, Context]     :frame     ({})
     #   frame to use when serializing.
     # @option options [Boolean]  :unique_bnodes   (false)
     #   Use unique bnode identifiers, defaults to using the identifier which the node was originall initialized with (if any).
+    # @option options [Boolean] :stream (false)
+    #   Do not attempt to optimize graph presentation, suitable for streaming large graphs.
     # @yield  [writer] `self`
     # @yieldparam  [RDF::Writer] writer
     # @yieldreturn [void]
     # @yield  [writer]
     # @yieldparam [RDF::Writer] writer
     def initialize(output = $stdout, options = {}, &block)
       options[:base_uri] ||= options[:base] if options.has_key?(:base)
       options[:base] ||= options[:base_uri] if options.has_key?(:base_uri)
       super do
         @repo = RDF::Repository.new
+        @debug = @options[:debug]
 
         if block_given?
           case block.arity
             when 0 then instance_eval(&block)
             else block.call(self)
@@ -122,11 +127,17 @@
     ##
     # Adds a statement to be serialized
     # @param  [RDF::Statement] statement
     # @return [void]
     def write_statement(statement)
-      @repo.insert(statement)
+      case
+      when @options[:stream]
+        stream_statement(statement)
+      else
+        # Add to repo and output in epilogue
+        @repo.insert(statement)
+      end
     end
 
     ##
     # Addes a triple to be serialized
     # @param  [RDF::Resource] subject
@@ -134,22 +145,34 @@
     # @param  [RDF::Value]    object
     # @return [void]
     # @raise  [NotImplementedError] unless implemented in subclass
     # @abstract
     def write_triple(subject, predicate, object)
-      @repo.insert(Statement.new(subject, predicate, object))
+      write_statement(Statement.new(subject, predicate, object))
     end
 
     ##
+    # Necessary for streaming
+    # @return [void] `self`
+    def write_prologue
+      case
+      when @options[:stream]
+        stream_prologue
+      else
+        super
+      end
+    end
+
+    ##
     # Outputs the Serialized JSON-LD representation of all stored statements.
     #
     # If provided a context or prefixes, we'll create a context
     # and use it to compact the output. Otherwise, we return un-compacted JSON-LD
     #
     # @return [void]
     # @see    #write_triple
     def write_epilogue
-      @debug = @options[:debug]
+      return stream_epilogue if @options[:stream]
 
       debug("writer") { "serialize #{@repo.count} statements, #{@options.inspect}"}
       result = API.fromRdf(@repo, @options)
 
       # If we were provided a context, or prefixes, use them to compact the output