relaton-w3c 1.19.0 → 1.20.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dde70fdc615a88c4343add641e193635a5377af68180d96122b02af747aa50e5
-  data.tar.gz: 363d74d0a170b307006310bbbe2f6e645da40aabf12dbc2a94ebac4887f2452f
+  metadata.gz: c9c777f3bbcbff5de8311b7103f7f64dee70971f5745be10f0415200ec8a3fe9
+  data.tar.gz: c13ba956f36f1d4e18a846aeb5d550aad5dbd0128a63457418032958d9d503ff
 SHA512:
-  metadata.gz: fec1491eaa3108fd8726dfa0af50f54b98d7c28286c936d827088eed8ee7badbe6be430dd4588ad6324936f6cce8824379750a71ac4da9aa8d88788a47fceec4
-  data.tar.gz: e9d4074458f1a01e12ab84619861dda851ce668573d3b3e843a2c7ba8effed42e6c94545b49157626ba922ebfc31b57ff9d5877461626d2578a9ebdf128d314c
+  metadata.gz: efbc142c4590df61f6b625bc1e00a29e95edad3ca97d940ea05f57caec3cf8db9fc5ed2a64ec6d5a4079ef3228fd682bae801b6c659a5afb83279d49c4d0b2e8
+  data.tar.gz: ef68de56840568a3e7d063a5c72876d23ae8d7f02d8cd35169e5a5b619c240edad1622738d38cd5d80642301f892e17a6f40d9fe17e33520bd523440064c9dc0

data/lib/relaton_w3c/data_fetcher.rb

@@ -2,11 +2,12 @@ require "rdf"
 require "linkeddata"
 require "sparql"
 require "mechanize"
-require "relaton_w3c/data_parser"
+require_relative "rdf_archive"
+require_relative "data_parser"
 
 module RelatonW3c
   class DataFetcher
-    attr_reader :data, :group_names
+    attr_reader :data, :group_names, :rdf_archive
 
     #
     # Data fetcher initializer
@@ -20,6 +21,7 @@ module RelatonW3c
       @ext = format.sub(/^bib/, "")
       dir = File.dirname(File.expand_path(__FILE__))
       @group_names = YAML.load_file(File.join(dir, "workgroups.yaml"))
+      @files = Set.new
       @index = DataIndex.create_from_file
       @index1 = Relaton::Index.find_or_create :W3C, file: "index1.yaml"
     end
@@ -31,32 +33,35 @@ module RelatonW3c
     # @param [Strin] output directory to save files, default: "data"
     # @param [Strin] format format of output files (xml, yaml, bibxml), default: yaml
     #
-    def self.fetch(source, output: "data", format: "yaml")
+    def self.fetch(output: "data", format: "yaml")
       t1 = Time.now
       puts "Started at: #{t1}"
       FileUtils.mkdir_p output
-      new(output, format).fetch source
+      new(output, format).fetch
       t2 = Time.now
       puts "Stopped at: #{t2}"
       puts "Done in: #{(t2 - t1).round} sec."
     end
 
+    def rdf_archive
+      @rdf_archive ||= RDFArchive.new
+    end
+
     #
     # Parse documents
     #
     # @param [String] source source name "w3c-tr-archive" or "w3c-rdf"
     #
-    def fetch(source) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
-      each_dataset(source) do |rdf|
-        %i[versioned unversioned].each do |type|
-          send("query_#{type}_docs", rdf).each do |sl|
-            bib = DataParser.parse(rdf, sl, self)
-            add_has_edition_relation(bib) if type == :unversioned
-            save_doc bib
-          rescue StandardError => e
-            link = sl.respond_to?(:link) ? sl.link : sl.version_of
-            Util.error "Error: document #{link} #{e.message}\n#{e.backtrace.join("\n")}"
-          end
+    def fetch # (source) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
+      rdf = rdf_archive.get_data
+      %i[versioned unversioned].each do |type|
+        send("query_#{type}_docs", rdf).each do |sl|
+          bib = DataParser.parse(rdf, sl, self)
+          add_has_edition_relation(bib) if type == :unversioned
+          save_doc bib
+        rescue StandardError => e
+          link = sl.respond_to?(:link) ? sl.link : sl.version_of
+          Util.error "Error: document #{link} #{e.message}\n#{e.backtrace.join("\n")}"
         end
       end
       @index.sort!.save
@@ -132,41 +137,6 @@ module RelatonW3c
       (ids1 & ids2).any?
     end
 
-    #
-    # Yield fetching for each dataset
-    #
-    # @param [String] source source name "w3c-tr-archive" or "w3c-rdf"
-    #
-    # @yield [RDF::Repository] RDF repository
-    #
-    def each_dataset(source, &_block) # rubocop:disable Metrics/MethodLength
-      case source
-      when "w3c-tr-archive"
-        Dir["w3c-tr-archive/*.rdf"].map do |f|
-          @files = []
-          yield RDF::Repository.load(f)
-        end
-      when "w3c-rdf"
-        @files = []
-        rdf = RDF::Repository.load("http://www.w3.org/2002/01/tr-automation/tr.rdf")
-        yield rdf
-        # parse_static_dataset
-      end
-    end
-
-    #
-    # Parse static dataset
-    #
-    # def parse_static_dataset
-    #   Dir[File.expand_path("../../data/*", __dir__)].each do |file|
-    #     xml = File.read file, encoding: "UTF-8"
-    #     save_doc BibXMLParser.parse(xml), warn_duplicate: false
-    #   rescue StandardError => e
-    #     warn "Error: document #{file} #{e.message}"
-    #     warn e.backtrace.join("\n")
-    #   end
-    # end
-
     #
     # Query RDF source for versioned documents
     #
@@ -206,15 +176,9 @@ module RelatonW3c
     #
     # @param [RelatonW3c::W3cBibliographicItem, nil] bib bibliographic item
     #
-    def save_doc(bib, warn_duplicate: true) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+    def save_doc(bib, warn_duplicate: true)
       return unless bib
 
-      c = case @format
-          when "xml" then bib.to_xml(bibdata: true)
-          when "yaml" then bib.to_hash.to_yaml
-          else bib.send("to_#{@format}")
-          end
-      # id = bib.docidentifier.detect(&:primary)&.id || bib.formattedref.content
       file = file_name(bib.docnumber)
       if @files.include?(file)
         Util.warn "File #{file} already exists. Document: #{bib.docnumber}" if warn_duplicate
@@ -223,7 +187,15 @@ module RelatonW3c
         @index.add pubid, file
         @index1.add_or_update pubid.to_hash, file
         @files << file
-        File.write file, c, encoding: "UTF-8"
+      end
+      File.write file, serialize(bib), encoding: "UTF-8"
+    end
+
+    def serialize(bib)
+      case @format
+      when "xml" then bib.to_xml(bibdata: true)
+      when "yaml" then bib.to_hash.to_yaml
+      else bib.send("to_#{@format}")
       end
     end
 

data/lib/relaton_w3c/data_parser.rb

@@ -280,9 +280,9 @@ module RelatonW3c
     end
 
     #
-    # Parse editor drafts relation
+    # Parse editor drafts links
     #
-    # @return [Array<RelatonBib::DocumentRelation>] relation
+    # @return [Array<RelatonBib::TypedUri>] links
     #
     def editor_drafts # rubocop:disable Metrics/MethodLength
       return [] unless @sol.respond_to?(:link)
@@ -311,7 +311,10 @@ module RelatonW3c
         PREFIX doc: <http://www.w3.org/2000/10/swap/pim/doc#>
         PREFIX mat: <http://www.w3.org/2002/05/matrix/vocab#>
         SELECT ?rel
-        WHERE { <#{@sol.link.to_s.strip}> #{predicate} ?rel . }
+        WHERE {
+          <#{@sol.link.to_s.strip}> #{predicate} ?rel .
+          FILTER ( isURI(?rel) )
+        }
       ))
       @rdf.query(sse).order_by(:rel)
     end
@@ -371,7 +374,8 @@ module RelatonW3c
       id = pub_id(url)
       fref = RelatonBib::FormattedRef.new content: id
       docid = RelatonBib::DocumentIdentifier.new(type: "W3C", id: id, primary: true)
-      bib = W3cBibliographicItem.new formattedref: fref, docid: [docid]
+      link = [RelatonBib::TypedUri.new(type: "src", content: url)]
+      bib = W3cBibliographicItem.new formattedref: fref, docid: [docid], link: link
       dsc = RelatonBib::FormattedString.new content: desc if desc
       RelatonBib::DocumentRelation.new(type: type, bibitem: bib, description: dsc)
     end

data/lib/relaton_w3c/processor.rb

@@ -9,7 +9,7 @@ module RelatonW3c
       @prefix = "W3C"
       @defaultprefix = %r{^W3C\s}
       @idtype = "W3C"
-      @datasets = %w[w3c-rdf w3c-tr-archive]
+      @datasets = %w[w3c-rdf]
     end
 
     # @param code [String]
@@ -28,8 +28,8 @@ module RelatonW3c
     # @option opts [String] :output directory to output documents
     # @option opts [String] :format
     #
-    def fetch_data(source, opts)
-      DataFetcher.fetch(source, **opts)
+    def fetch_data(_source, opts)
+      DataFetcher.fetch(**opts)
     end
 
     # @param xml [String]

data/lib/relaton_w3c/rdf_archive.rb

@@ -0,0 +1,67 @@
+module RelatonW3c
+  class RDFArchive
+    def initialize(file = "archive.rdf")
+      @file = file
+    end
+
+    #
+    # Get RDF data from the updated archive file.
+    #
+    # @return [RDF::Repository]
+    #
+    def get_data
+      if !File.exist?(@file) || File.mtime(@file) < Time.now - 86_400
+        get_archive
+        update_archive
+      end
+      RDF::Repository.load(@file)
+    end
+
+    private
+
+    def update_archive
+      # Load the older RDF/XML file
+      older = Nokogiri::XML File.read(@file, encoding: "UTF-8")
+
+      # Load the newer RDF/XML file
+      url = "http://www.w3.org/2002/01/tr-automation/tr.rdf"
+      newer = Nokogiri::XML OpenURI.open_uri(url).read
+
+      # Create a hash to store rdf:about attributes from the newer file
+      newer_elements = {}
+      newer.root.element_children.each do |element|
+        rdf_about = element.attribute('about')&.value
+        newer_elements[rdf_about.sub(/^http\s:/, "")] = element if rdf_about
+      end
+
+      # Replace elements in the older document
+      older.root.element_children.each do |element|
+        rdf_about = element.attribute('about')&.value
+        if rdf_about && newer_elements[url = rdf_about.sub(/^http\s:/, "")]
+          element.replace(newer_elements[url])
+          newer_elements.delete(url)
+        end
+      end
+
+      # Add remaining new elements to the older document
+      newer_elements.each_value do |element|
+        older.root.add_child(element)
+      end
+
+      # Add new namespaces from the newer document to the older document
+      newer.root.namespace_definitions.each do |ns|
+        unless older.root.namespace_definitions.any? { |old_ns| old_ns.href == ns.href }
+          older.root.add_namespace_definition(ns.prefix, ns.href)
+        end
+      end
+      File.write @file, older.to_xml, encoding: "UTF-8"
+    end
+
+    def get_archive
+      unless File.exist? @file
+        url = "https://raw.githubusercontent.com/relaton/relaton-data-w3c/refs/heads/main/archive.rdf"
+        File.write @file, OpenURI.open_uri(url).read, encoding: "UTF-8"
+      end
+    end
+  end
+end

data/lib/relaton_w3c/version.rb

@@ -1,3 +1,3 @@
 module RelatonW3c
-  VERSION = "1.19.0".freeze
+  VERSION = "1.20.0".freeze
 end

data/lib/relaton_w3c/workgroups.yaml

@@ -506,3 +506,16 @@
 'www.w3.org/groups/wg/rch':
   name: RDF Dataset Canonicalization and Hash Working Group
   abbrev: RCH WG
+'www.w3.org/groups/wg/browser-tools-testing':
+  name: Browser Testing and Tools Working Group
+'www.w3.org/groups/wg/did':
+  name: Decentralized Identifier Working Group
+  abbrev: DID WG
+'www.w3.org/Member/Board':
+  name: Board of Directors
+'www.w3.org/groups/wg/webtransport':
+  name: WebTransport Working Group
+'www.w3.org/groups/wg/fedid':
+  name: Federated Identity Working Group
+'www.w3.org/groups/wg/gpu':
+  name: GPU for the Web Working Group

data/relaton_w3c.gemspec

@@ -35,7 +35,7 @@ Gem::Specification.new do |spec|
   spec.add_dependency "mechanize", "~> 2.10"
   spec.add_dependency "rdf", "~> 3.2"
   spec.add_dependency "rdf-normalize", "~> 0.6"
-  spec.add_dependency "relaton-bib", "~> 1.19.0"
+  spec.add_dependency "relaton-bib", "~> 1.20.0"
   spec.add_dependency "relaton-index", "~> 0.2.8"
   spec.add_dependency "rubyzip", "~> 2.3"
   spec.add_dependency "shex", "~> 0.7"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-w3c
 version: !ruby/object:Gem::Version
-  version: 1.19.0
+  version: 1.20.0
 platform: ruby
 authors:
 - Ribose Inc.
-autorequire: 
+autorequire:
 bindir: exe
 cert_chain: []
-date: 2024-07-04 00:00:00.000000000 Z
+date: 2024-12-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: linkeddata
@@ -72,14 +72,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.19.0
+        version: 1.20.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.19.0
+        version: 1.20.0
 - !ruby/object:Gem::Dependency
   name: relaton-index
   requirement: !ruby/object:Gem::Requirement
@@ -187,6 +187,7 @@ files:
 - lib/relaton_w3c/hit_collection.rb
 - lib/relaton_w3c/processor.rb
 - lib/relaton_w3c/pubid.rb
+- lib/relaton_w3c/rdf_archive.rb
 - lib/relaton_w3c/util.rb
 - lib/relaton_w3c/version.rb
 - lib/relaton_w3c/w3c_bibliographic_item.rb
@@ -199,7 +200,7 @@ licenses:
 - BSD-2-Clause
 metadata:
   homepage_uri: https://github.com/relaton/relaton-wc3
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -215,7 +216,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubygems_version: 3.3.27
-signing_key: 
+signing_key:
 specification_version: 4
 summary: 'RelatonIso: retrieve W3C Standards for bibliographic using the IsoBibliographicItem
   model'