json-ld 3.1.4 → 3.1.9

data/lib/json/ld/expand.rb

@@ -11,9 +11,9 @@ module JSON::LD
     # The following constant is used to reduce object allocations
     CONTAINER_INDEX_ID_TYPE = Set['@index', '@id', '@type'].freeze
     KEY_ID = %w(@id).freeze
-    KEYS_VALUE_LANGUAGE_TYPE_INDEX_DIRECTION = %w(@value @language @type @index @direction).freeze
+    KEYS_VALUE_LANGUAGE_TYPE_INDEX_DIRECTION = %w(@value @language @type @index @direction @annotation).freeze
     KEYS_SET_LIST_INDEX = %w(@set @list @index).freeze
-    KEYS_INCLUDED_TYPE = %w(@included @type).freeze
+    KEYS_INCLUDED_TYPE_REVERSE = %w(@included @type @reverse).freeze
 
     ##
     # Expand an Array or Object given an active context and performing local context expansion.
@@ -49,7 +49,13 @@ module JSON::LD
             log_depth: log_depth.to_i + 1)
 
           # If the active property is @list or its container mapping is set to @list and v is an array, change it to a list object
-          v = {"@list" => v} if is_list && v.is_a?(Array)
+          if is_list && v.is_a?(Array)
+            # Make sure that no member of v contains an annotation object
+            raise JsonLdError::InvalidAnnotation,
+              "A list element must not contain @annotation." if
+              v.any? {|n| n.is_a?(Hash) && n.key?('@annotation')}
+            v = {"@list" => v}
+          end
 
           case v
           when nil then nil
@@ -81,7 +87,7 @@ module JSON::LD
         log_debug("expand", depth: log_depth.to_i) {"after property_scoped_context: #{context.inspect}"} unless property_scoped_context.nil?
 
         # If element contains the key @context, set active context to the result of the Context Processing algorithm, passing active context and the value of the @context key as local context.
-        if input.has_key?('@context')
+        if input.key?('@context')
           context = context.parse(input.delete('@context'), base: @options[:base])
           log_debug("expand", depth: log_depth.to_i) {"context: #{context.inspect}"}
         end
@@ -142,13 +148,20 @@ module JSON::LD
 
           if output_object['@type'] == '@json' && context.processingMode('json-ld-1.1')
             # Any value of @value is okay if @type: @json
-          elsif !ary.all? {|v| v.is_a?(String) || v.is_a?(Hash) && v.empty?} && output_object.has_key?('@language')
+          elsif !ary.all? {|v| v.is_a?(String) || v.is_a?(Hash) && v.empty?} && output_object.key?('@language')
             # Otherwise, if the value of result's @value member is not a string and result contains the key @language, an invalid language-tagged value error has been detected (only strings can be language-tagged) and processing is aborted.
             raise JsonLdError::InvalidLanguageTaggedValue,
                   "when @language is used, @value must be a string: #{output_object.inspect}"
-          elsif !Array(output_object['@type']).all? {|t|
-                  t.is_a?(String) && RDF::URI(t).valid? && !t.start_with?('_:') ||
-                  t.is_a?(Hash) && t.empty?}
+          elsif output_object['@type'] &&
+            (!Array(output_object['@type']).all? {|t|
+                t.is_a?(String) && RDF::URI(t).valid? && !t.start_with?('_:') ||
+                t.is_a?(Hash) && t.empty?} ||
+             !framing && !output_object['@type'].is_a?(String))
+            # Otherwise, if the result has a @type member and its value is not an IRI, an invalid typed value error has been detected and processing is aborted.
+            raise JsonLdError::InvalidTypedValue,
+                  "value of @type must be an IRI or '@json': #{output_object.inspect}"
+          elsif !framing && !output_object.fetch('@type', '').is_a?(String) &&
+                  RDF::URI(t).valid? && !t.start_with?('_:')
             # Otherwise, if the result has a @type member and its value is not an IRI, an invalid typed value error has been detected and processing is aborted.
             raise JsonLdError::InvalidTypedValue,
                   "value of @type must be an IRI or '@json': #{output_object.inspect}"
@@ -165,6 +178,18 @@ module JSON::LD
 
           # If result contains the key @set, then set result to the key's associated value.
           return output_object['@set'] if output_object.key?('@set')
+        elsif output_object['@annotation']
+          # Otherwise, if result contains the key @annotation,
+          # the array value must all be node objects without an @id property, otherwise, an invalid annotation error has been detected and processing is aborted.
+          raise JsonLdError::InvalidAnnotation,
+            "@annotation must reference node objects without @id." unless
+            output_object['@annotation'].all? {|o| node?(o) && !o.key?('@id')}
+
+          # Additionally, the property must not be used if there is no active property, or the expanded active property is @graph.
+          raise JsonLdError::InvalidAnnotation,
+            "@annotation must not be used on a top-level object." if
+            %w(@graph @included).include?(expanded_active_property || '@graph')
+          
         end
 
         # If result contains only the key @language, set result to null.
@@ -247,10 +272,15 @@ module JSON::LD
 
           # If result has already an expanded property member (other than @type), an colliding keywords error has been detected and processing is aborted.
           raise JsonLdError::CollidingKeywords,
-                "#{expanded_property} already exists in result" if output_object.has_key?(expanded_property) && !KEYS_INCLUDED_TYPE.include?(expanded_property)
+                "#{expanded_property} already exists in result" if output_object.key?(expanded_property) && !KEYS_INCLUDED_TYPE_REVERSE.include?(expanded_property)
 
           expanded_value = case expanded_property
           when '@id'
+            # If expanded active property is `@annotation`, an invalid annotation error has been found and processing is aborted.
+            raise JsonLdError::InvalidAnnotation,
+              "an annotation must not contain a property expanding to @id" if
+              expanded_active_property == '@annotation' && @options[:rdfstar]
+
             # If expanded property is @id and value is not a string, an invalid @id value error has been detected and processing is aborted
             e_id = case value
             when String
@@ -266,12 +296,27 @@ module JSON::LD
                 context.expand_iri(v, as_string: true, base: @options[:base], documentRelative: true)
               end
             when Hash
-              raise JsonLdError::InvalidIdValue,
-                    "value of @id must be a string unless framing: #{value.inspect}" unless framing
-              raise JsonLdError::InvalidTypeValue,
-                    "value of @id must be a an empty object for framing: #{value.inspect}" unless
-                    value.empty?
-              [{}]
+              if framing
+                raise JsonLdError::InvalidTypeValue,
+                      "value of @id must be a an empty object for framing: #{value.inspect}" unless
+                      value.empty?
+                [{}]
+              elsif @options[:rdfstar]
+                # Result must have just a single statement
+                rei_node = expand(value, nil, context, log_depth: log_depth.to_i + 1)
+
+                # Node must not contain @reverse
+                raise JsonLdError::InvalidEmbeddedNode,
+                      "Embedded node with @reverse" if rei_node && rei_node.key?('@reverse')
+                statements = to_enum(:item_to_rdf, rei_node)
+                raise JsonLdError::InvalidEmbeddedNode,
+                      "Embedded node with #{statements.size} statements" unless
+                      statements.count == 1
+                rei_node
+              else
+                raise JsonLdError::InvalidIdValue,
+                      "value of @id must be a string unless framing: #{value.inspect}" unless framing
+              end
             else
               raise JsonLdError::InvalidIdValue,
                     "value of @id must be a string or hash if framing: #{value.inspect}"
@@ -447,6 +492,11 @@ module JSON::LD
             # Spec FIXME: need to be sure that result is an array
             value = as_array(value)
 
+            # Make sure that no member of value contains an annotation object
+            raise JsonLdError::InvalidAnnotation,
+              "A list element must not contain @annotation." if
+              value.any? {|n| n.is_a?(Hash) && n.key?('@annotation')}
+
             value
           when '@set'
             # If expanded property is @set, set expanded value to the result of using this algorithm recursively, passing active context, active property, and value for element.
@@ -465,7 +515,7 @@ module JSON::LD
               log_depth: log_depth.to_i + 1)
 
             # If expanded value contains an @reverse member, i.e., properties that are reversed twice, execute for each of its property and item the following steps:
-            if value.has_key?('@reverse')
+            if value.key?('@reverse')
               #log_debug("@reverse", depth: log_depth.to_i) {"double reverse: #{value.inspect}"}
               value['@reverse'].each do |property, item|
                 # If result does not have a property member, create one and set its value to an empty array.
@@ -504,6 +554,12 @@ module JSON::LD
             nests << key
             # Continue with the next key from element
             next
+          when '@annotation'
+            # Skip unless rdfstar option is set
+            next unless @options[:rdfstar]
+            as_array(expand(value, '@annotation', context,
+              framing: framing,
+              log_depth: log_depth.to_i + 1))
           else
             # Skip unknown keyword
             next

data/lib/json/ld/extensions.rb

@@ -11,10 +11,10 @@ module RDF
   class Statement
     # Validate extended RDF
     def valid_extended?
-      has_subject?    && subject.resource? && subject.valid_extended? &&
-      has_predicate?  && predicate.resource? && predicate.valid_extended? &&
-      has_object?     && object.term? && object.valid_extended? &&
-      (has_graph?      ? (graph_name.resource? && graph_name.valid_extended?) : true)
+      subject?    && subject.resource? && subject.valid_extended? &&
+      predicate?  && predicate.resource? && predicate.valid_extended? &&
+      object?     && object.term? && object.valid_extended? &&
+      (graph?      ? (graph_name.resource? && graph_name.valid_extended?) : true)
     end
   end
 

data/lib/json/ld/flatten.rb

@@ -1,5 +1,7 @@
 # -*- encoding: utf-8 -*-
 # frozen_string_literal: true
+require 'json/canonicalization'
+
 module JSON::LD
   module Flatten
     include Utils
@@ -7,6 +9,10 @@ module JSON::LD
     ##
     # This algorithm creates a JSON object node map holding an indexed representation of the graphs and nodes represented in the passed expanded document. All nodes that are not uniquely identified by an IRI get assigned a (new) blank node identifier. The resulting node map will have a member for every graph in the document whose value is another object with a member for every node represented in the document. The default graph is stored under the @default member, all other graphs are stored under their graph name.
     #
+    # For RDF-star/JSON-LD-star:
+    #   * Values of `@id` can be an object (embedded node); when these are used as keys in a Node Map, they are serialized as canonical JSON, and de-serialized when flattening.
+    #   * The presence of `@annotation` implies an embedded node and the annotation object is removed from the node/value object in which it appears.
+    #
     # @param [Array, Hash] element
     #   Expanded JSON-LD input
     # @param [Hash] graph_map A map of graph name to subjects
@@ -16,12 +22,15 @@ module JSON::LD
     #   Node identifier
     # @param [String] active_property (nil)
     #   Property within current node
+    # @param [Boolean] reverse (false)
+    #   Processing a reverse relationship
     # @param [Array] list (nil)
     #   Used when property value is a list
     def create_node_map(element, graph_map,
                         active_graph: '@default',
                         active_subject: nil,
                         active_property: nil,
+                        reverse: false,
                         list: nil)
       if element.is_a?(Array)
         # If element is an array, process each entry in element recursively by passing item for element, node map, active graph, active subject, active property, and list.
@@ -30,21 +39,45 @@ module JSON::LD
                           active_graph: active_graph,
                           active_subject: active_subject,
                           active_property: active_property,
+                          reverse: false,
                           list: list)
         end
       elsif !element.is_a?(Hash)
         raise "Expected hash or array to create_node_map, got #{element.inspect}"
       else
         graph = (graph_map[active_graph] ||= {})
-        subject_node = graph[active_subject]
+        subject_node = !reverse && graph[active_subject.is_a?(Hash) ? active_subject.to_json_c14n : active_subject]
 
         # Transform BNode types
-        if element.has_key?('@type')
+        if element.key?('@type')
           element['@type'] = Array(element['@type']).map {|t| blank_node?(t) ? namer.get_name(t) : t}
         end
 
         if value?(element)
           element['@type'] = element['@type'].first if element ['@type']
+
+          # For rdfstar, if value contains an `@annotation` member ...
+          # note: active_subject will not be nil, and may be an object itself.
+          if element.key?('@annotation')
+            # rdfstar being true is implicit, as it is checked in expansion
+            as = node_reference?(active_subject) ?
+              active_subject['@id'] :
+              active_subject
+            star_subject = {
+              "@id" => as,
+              active_property => [element]
+            }
+
+            # Note that annotation is an array, make the reified subject the id of each member of that array.
+            annotation = element.delete('@annotation').map do |a|
+              a.merge('@id' => star_subject)
+            end
+
+            # Invoke recursively using annotation.
+            create_node_map(annotation, graph_map,
+                            active_graph: active_graph)
+          end
+
           if list.nil?
             add_value(subject_node, active_property, element, property_is_array: true, allow_duplicate: false)
           else
@@ -64,13 +97,19 @@ module JSON::LD
           end
         else
           # Element is a node object
-          id = element.delete('@id')
-          id = namer.get_name(id) if blank_node?(id)
+          ser_id = id = element.delete('@id')
+          if id.is_a?(Hash)
+            # Index graph using serialized id
+            ser_id = id.to_json_c14n
+          elsif id.nil?
+            ser_id = id = namer.get_name
+          end
 
-          node = graph[id] ||= {'@id' => id}
+          node = graph[ser_id] ||= {'@id' => id}
 
-          if active_subject.is_a?(Hash)
-            # If subject is a hash, then we're processing a reverse-property relationship.
+          if reverse
+            # Note: active_subject is a Hash
+            # We're processing a reverse-property relationship.
             add_value(node, active_property, active_subject, property_is_array: true, allow_duplicate: false)
           elsif active_property
             reference = {'@id' => id}
@@ -81,7 +120,30 @@ module JSON::LD
             end
           end
 
-          if element.has_key?('@type')
+          # For rdfstar, if node contains an `@annotation` member ...
+          # note: active_subject will not be nil, and may be an object itself.
+          # XXX: what if we're reversing an annotation?
+          if element.key?('@annotation')
+            # rdfstar being true is implicit, as it is checked in expansion
+            as = node_reference?(active_subject) ?
+              active_subject['@id'] :
+              active_subject
+            star_subject = reverse ? 
+              {"@id" => node['@id'], active_property => [{'@id' => as}]} :
+              {"@id" => as, active_property => [{'@id' => node['@id']}]}
+
+            # Note that annotation is an array, make the reified subject the id of each member of that array.
+            annotation = element.delete('@annotation').map do |a|
+              a.merge('@id' => star_subject)
+            end
+
+            # Invoke recursively using annotation.
+            create_node_map(annotation, graph_map,
+                            active_graph: active_graph,
+                            active_subject: star_subject)
+          end
+
+          if element.key?('@type')
             add_value(node, '@type', element.delete('@type'), property_is_array: true, allow_duplicate: false)
           end
 
@@ -99,7 +161,8 @@ module JSON::LD
                 create_node_map(value, graph_map,
                                 active_graph: active_graph,
                                 active_subject: referenced_node,
-                                active_property: property)
+                                active_property: property,
+                                reverse: true)
               end
             end
           end
@@ -128,7 +191,72 @@ module JSON::LD
       end
     end
 
+    ##
+    # Create annotations
+    #
+    # Updates a node map from which annotations have been folded into embedded triples to re-extract the annotations.
+    #
+    # Map entries where the key is of the form of a canonicalized JSON object are used to find keys with the `@id` and property components. If found, the original map entry is removed and entries added to an `@annotation` property of the associated value.
+    #
+    # * Keys which are of the form of a canonicalized JSON object are examined in inverse order of length.
+    # * Deserialize the key into a map, and re-serialize the value of `@id`.
+    # * If the map contains an entry with that value (after re-canonicalizing, as appropriate), and the associated antry has a item which matches the non-`@id` item from the map, the node is used to create an `@annotation` entry within that value.
+    #
+    # @param [Hash{String => Hash}] input
+    # @return [Hash{String => Hash}]
+    def create_annotations(node_map)
+      node_map.keys.
+        select {|k| k.start_with?('{')}.
+        sort_by(&:length).
+        reverse.each do |key|
+
+        annotation = node_map[key]
+        # Deserialize key, and re-serialize the `@id` value.
+        emb = annotation['@id'].dup
+        id = emb.delete('@id')
+        property, value = emb.to_a.first
+
+        # If id is a map, set it to the result of canonicalizing that value, otherwise to itself.
+        id = id.to_json_c14n if id.is_a?(Hash)
+
+        next unless node_map.key?(id)
+        # If node map has an entry for id and that entry contains the same property and value from entry:
+        node = node_map[id] 
+
+        next unless node.key?(property)
+
+        node[property].each do |emb_value|
+          next unless emb_value == value.first
+
+          node_map.delete(key)
+          annotation.delete('@id')
+          add_value(emb_value, '@annotation', annotation, property_is_array: true) unless
+            annotation.empty?
+        end
+      end
+    end
+
+    ##
+    # Rename blank nodes recursively within an embedded object
+    #
+    # @param [Object] node
+    # @return [Hash]
+    def rename_bnodes(node)
+      case node
+      when Array
+        node.map {|n| rename_bnodes(n)}
+      when Hash
+        node.inject({}) do |memo, (k, v)|
+          v = namer.get_name(v) if k == '@id' && v.is_a?(String) && blank_node?(v)
+          memo.merge(k => rename_bnodes(v))
+        end
+      else
+        node
+      end
+    end
+
   private
+
     ##
     # Merge nodes from all graphs in the graph_map into a new node map
     #

data/lib/json/ld/format.rb

@@ -57,10 +57,11 @@ module JSON::LD
           lambda: ->(files, **options) do
             out = options[:output] || $stdout
             out.set_encoding(Encoding::UTF_8) if RUBY_PLATFORM == "java"
-            options = options.merge(expandContext: options.delete(:context)) if options.has_key?(:context)
+            options = options.merge(expandContext: options.delete(:context)) if options.key?(:context)
+            options[:base] ||= options[:base_uri]
             if options[:format] == :jsonld
               if files.empty?
-                # If files are empty, either use options[:execute]
+                # If files are empty, either use options[:evaluate] or STDIN
                 input = options[:evaluate] ? StringIO.new(options[:evaluate]) : STDIN
                 input.set_encoding(options.fetch(:encoding, Encoding::UTF_8))
                 JSON::LD::API.expand(input, validate: false, **options) do |expanded|
@@ -93,9 +94,10 @@ module JSON::LD
             raise ArgumentError, "Compacting requires a context" unless options[:context]
             out = options[:output] || $stdout
             out.set_encoding(Encoding::UTF_8) if RUBY_PLATFORM == "java"
+            options[:base] ||= options[:base_uri]
             if options[:format] == :jsonld
               if files.empty?
-                # If files are empty, either use options[:execute]
+                # If files are empty, either use options[:evaluate] or STDIN
                 input = options[:evaluate] ? StringIO.new(options[:evaluate]) : STDIN
                 input.set_encoding(options.fetch(:encoding, Encoding::UTF_8))
                 JSON::LD::API.compact(input, options[:context], **options) do |compacted|
@@ -126,7 +128,7 @@ module JSON::LD
               control: :url2,
               use: :required,
               on: ["--context CONTEXT"],
-              description: "Context to use when compacting.") {|arg| RDF::URI(arg)},
+              description: "Context to use when compacting.") {|arg| RDF::URI(arg).absolute? ? RDF::URI(arg) : StringIO.new(File.read(arg))},
           ]
         },
         flatten: {
@@ -137,9 +139,10 @@ module JSON::LD
           lambda: ->(files, **options) do
             out = options[:output] || $stdout
             out.set_encoding(Encoding::UTF_8) if RUBY_PLATFORM == "java"
+            options[:base] ||= options[:base_uri]
             if options[:format] == :jsonld
               if files.empty?
-                # If files are empty, either use options[:execute]
+                # If files are empty, either use options[:evaluate] or STDIN
                 input = options[:evaluate] ? StringIO.new(options[:evaluate]) : STDIN
                 input.set_encoding(options.fetch(:encoding, Encoding::UTF_8))
                 JSON::LD::API.flatten(input, options[:context], **options) do |flattened|
@@ -162,7 +165,23 @@ module JSON::LD
                 end
               end
             end
-          end
+          end,
+          options: [
+            RDF::CLI::Option.new(
+              symbol: :context,
+              datatype: RDF::URI,
+              control: :url2,
+              use: :required,
+              on: ["--context CONTEXT"],
+              description: "Context to use when compacting.") {|arg| RDF::URI(arg)},
+            RDF::CLI::Option.new(
+              symbol: :createAnnotations,
+              datatype: TrueClass,
+              default: false,
+              control: :checkbox,
+              on: ["--[no-]create-annotations"],
+              description: "Unfold embedded nodes which can be represented using `@annotation`."),
+          ]
         },
         frame: {
           description: "Frame JSON-LD or parsed RDF",
@@ -173,9 +192,10 @@ module JSON::LD
             raise ArgumentError, "Framing requires a frame" unless options[:frame]
             out = options[:output] || $stdout
             out.set_encoding(Encoding::UTF_8) if RUBY_PLATFORM == "java"
+            options[:base] ||= options[:base_uri]
             if options[:format] == :jsonld
               if files.empty?
-                # If files are empty, either use options[:execute]
+                # If files are empty, either use options[:evaluate] or STDIN
                 input = options[:evaluate] ? StringIO.new(options[:evaluate]) : STDIN
                 input.set_encoding(options.fetch(:encoding, Encoding::UTF_8))
                 JSON::LD::API.frame(input, options[:frame], **options) do |framed|
@@ -207,7 +227,7 @@ module JSON::LD
               control: :url2,
               use: :required,
               on: ["--frame FRAME"],
-              description: "Frame to use when serializing.") {|arg| RDF::URI(arg)}
+              description: "Frame to use when serializing.") {|arg| RDF::URI(arg).absolute? ? RDF::URI(arg) : StringIO.new(File.read(arg))}
           ]
         },
       }