relaton-w3c 1.1.0 → 1.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5756c84e1bdc95815fce77b2a058687f220340c1ff154b5ee40d1d07934a71c0
-  data.tar.gz: 5148b2d1fc04c37f310c3dde66847000941d7056940a9d7d0c1dd2dadea59a39
+  metadata.gz: 42a40987b4156671fee680b2ef9eb9a6b42f5a356f62c09d7f56eaf9ae466578
+  data.tar.gz: dcd2de5ae0c8582a2dc3fb29473faee5367c81cf48a394f6cf09b45134c66e0d
 SHA512:
-  metadata.gz: caa8636f7bdb8a257c86acc8755a84fb30fd9f324806940881940346e93d81880a572edde4f5a5041ebd76fa27f00f95c5a3e90ef806ad0c4570e6ec9bece604
-  data.tar.gz: e1490c909e083aabdbcbec63f3c361e847101f9e829bc804e73daf31e9acb3a5cc55fb13ecfa6f3b8911fb6565f4182baeb33aad397cf25f73e56bdf3a96c186
+  metadata.gz: b94a6b93e792f44145b7b70139c673e534addc6989272c623ccfd3a1828b3376755f7193ca9dbcf35049f169a9c0084c3e8111977653ccb62fb5e8d7a2a18c9b
+  data.tar.gz: 272f89a76aeceacd9db2366cd80a0316023c3594a9fc058bed048b883ecea0dd07e63769e14988da0ea6506581ebd0d76539ee4097750bab5ee9ea5b415f6b24

data/.github/workflows/ubuntu.yml

@@ -35,4 +35,5 @@ jobs:
           bundle install --jobs 4 --retry 3
       - name: Run specs
         run: |
+          unset JAVA_TOOL_OPTIONS
           bundle exec rake

data/.rubocop.yml

@@ -5,6 +5,6 @@
 inherit_from:
   - https://raw.githubusercontent.com/riboseinc/oss-guides/master/ci/rubocop.yml
 AllCops:
-  TargetRubyVersion: 2.3
+  TargetRubyVersion: 2.4
 Rails:
-  Enabled: true
+  Enabled: false

data/grammars/biblio.rng

@@ -532,7 +532,7 @@
   </define>
   <define name="LocalityType">
     <data type="string">
-      <param name="pattern">section|clause|part|paragraph|chapter|page|whole|table|annex|figure|note|list|example|volume|issue|time|locality:[a-zA-Z0-9_]+</param>
+      <param name="pattern">section|clause|part|paragraph|chapter|page|whole|table|annex|figure|note|list|example|volume|issue|time|anchor|locality:[a-zA-Z0-9_]+</param>
     </data>
   </define>
   <define name="referenceFrom">

data/lib/relaton_w3c/hit_collection.rb

@@ -45,7 +45,12 @@ module RelatonW3c
       /(?<title>.+)\s(?<date>\d{4}-\d{2}-\d{2})$/ =~ title_date
       title ||= title_date
       result = data.select do |hit|
-        hit["title"] == title && type_date_filter(hit, type, date)
+        (hit["title"].casecmp?(title) ||
+          hit["link"].split("/").last.match?(/-#{title}-/)) &&
+          type_date_filter(hit, type, date)
+      end
+      if result.empty?
+        result = data.select { |h| h["link"].split("/").last.match? /#{title}/ }
       end
       result.map { |h| Hit.new(h, self) }
     end
@@ -54,7 +59,7 @@ module RelatonW3c
     # @param type [String]
     # @param date [String]
     # @return [TrueClass, FalseClass]
-    def type_date_filter(hit, type, date)
+    def type_date_filter(hit, type, date) # rubocop:disable Metrics/AbcSize
       if type && hit["type"] != short_type(type) || date && hit["date"] != date
         history = get_history hit, type, date
         return false unless history.any?

data/lib/relaton_w3c/scrapper.rb

@@ -13,11 +13,12 @@ module RelatonW3c
 
       # @param hit [Hash]
       # @return [RelatonW3c::W3cBibliographicItem]
-      def parse_page(hit)
+      def parse_page(hit) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
         resp = Net::HTTP.get_response URI.parse(hit["link"])
         doc = resp.code == "200" ? Nokogiri::HTML(resp.body) : nil
         W3cBibliographicItem.new(
           type: "standard",
+          docid: fetch_docid(hit),
           fetched: Date.today.to_s,
           language: ["en"],
           script: ["Latn"],
@@ -28,28 +29,38 @@ module RelatonW3c
           doctype: fetch_doctype(hit, doc),
           contributor: fetch_contributor(hit, doc),
           relation: fetch_relation(doc),
-          keyword: hit["keyword"],
+          keyword: hit["keyword"]
         )
       end
 
       private
 
+      # @param hit [Hash]
+      # @return [Array<RelatonBib::DocumentIdentifier>]
+      def fetch_docid(hit)
+        id = hit["link"].split("/").last
+        [RelatonBib::DocumentIdentifier.new(id: id, type: "W3C")]
+      end
+
       # @param hit [Hash]
       # @param doc [Nokogiri::HTML::Document]
       # @return [Array<RelatonBib::TypedTitleString>]
-      def fetch_title(hit, doc)
+      def fetch_title(hit, doc) # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity
         titles = []
         if doc
-          title = doc.at("//h1[@id='title']").text
-          titles << { content: title, type: "main" }
-          subtitle = doc.at("//h2[@id='subtitle']").text
-          titles << { content: subtitle, tipe: "subtitle" }
-        elsif hit["title"]
+          title = doc.at("//*[contains(@id, 'title')]")&.text
+          titles << { content: title, type: "main" } if title && !title.empty?
+          subtitle = doc.at(
+            "//h2[@id='subtitle']|//p[contains(@class, 'subline')]"
+          )&.text
+          titles << { content: subtitle, tipe: "subtitle" } if subtitle
+        end
+        if titles.empty? && hit["title"]
           titles << { content: hit["title"], type: "main" }
         end
         titles.map do |t|
           title = RelatonBib::FormattedString.new(
-            content: t[:content], language: "en", script: "Latn",
+            content: t[:content], language: "en", script: "Latn"
           )
           RelatonBib::TypedTitleString.new(type: t[:type], title: title)
         end
@@ -75,10 +86,27 @@ module RelatonW3c
       # @param doc [Nokogiri::HTML::Document, NilClass]
       # @return [Array<RelatonBib::BibliographicDate>]
       def fetch_date(hit, doc)
-        on = hit["datepub"] || doc && doc.at("//h2/time[@datetime]")[:datetime]
+        on = hit["datepub"] || doc&.at("//h2/time[@datetime]")&.attr(:datetime)
+        on ||= fetch_date1(doc) || fetch_date2(doc)
         [RelatonBib::BibliographicDate.new(type: "published", on: on)] if on
       end
 
+      # @param doc [Nokogiri::HTML::Document, NilClass]
+      # @return [String]
+      def fetch_date1(doc)
+        d = doc&.at("//h2[@property='dc:issued']")&.attr(:content)
+        d&.match(/\d{4}-\d{2}-\d{2}/)&.to_s
+      end
+
+      # @param doc [Nokogiri::HTML::Document, NilClass]
+      # @return [String]
+      def fetch_date2(doc)
+        d = doc&.at("//h2[contains(@id, 'w3c-recommendation')]")
+        return unless d
+
+        Date.parse(d.attr(:id.match(/\d{2}-\w+-\d{4}/).to_s)).to_s
+      end
+
       # @param hit [Hash]
       # @param doc [Nokogiri::HTML::Document, NilClass]
       # @return [String]
@@ -96,17 +124,19 @@ module RelatonW3c
       # @param hit [Hash]
       # @param doc [Nokogiri::HTML::Document, NilClass]
       # @return [Array<RelatonBib::ContributionInfo>]
-      def fetch_contributor(hit, doc)
+      def fetch_contributor(hit, doc) # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity
         if doc
-          editors = find_contribs(doc, "Editors").map do |ed|
-            parse_contrib ed, "editor"
+          editors = find_contribs(doc, "Editors").reduce([]) do |mem, ed|
+            c = parse_contrib ed, "editor"
+            mem << c if c
+            mem
           end
-          contribs = find_contribs(doc, "Authors").reduce(editors) do |mem, athr|
-            ed = mem.detect { |e| e[:id] && e[:id] == athr["data-editor-id"] }
+          contribs = find_contribs(doc, "Authors").reduce(editors) do |mem, ath|
+            ed = mem.detect { |e| e[:id] && e[:id] == ath["data-editor-id"] }
             if ed
               ed[:role] << { type: "author" }
             else
-              mem << parse_contrib(athr, "author")
+              mem << parse_contrib(ath, "author")
             end
             mem
           end
@@ -131,6 +161,8 @@ module RelatonW3c
       # @return [Hash]
       def parse_contrib(element, type)
         p = element.at("a")
+        return unless p
+
         contrib = {
           name: p.text,
           url: p[:href],

data/lib/relaton_w3c/version.rb

@@ -1,3 +1,3 @@
 module RelatonW3c
-  VERSION = "1.1.0".freeze
+  VERSION = "1.3.3".freeze
 end

data/lib/relaton_w3c/w3c_bibliographic_item.rb

@@ -5,35 +5,12 @@ module RelatonW3c
       proposedRecommendation recommendation retired workingDraft
     ].freeze
 
-    attr_reader :doctype
-
     # @param doctype [String]
     def initialize(**args)
       if args[:doctype] && !TYPES.include?(args[:doctype])
         warn "[relaton-w3c] invalid document type: #{args[:doctype]}"
       end
-      @doctype = args.delete :doctype
       super **args
     end
-
-    # @param builder [Nokogiri::XML::Builder, NilClass]
-    # @param opts [Hash]
-    # @option opts [TrueClass, FalseClass, NilClass] bibdata
-    def to_xml(builder = nil, **opts)
-      super builder, **opts do |b|
-        if opts[:bibdata] && doctype
-          b.ext do |e|
-            e.doctype doctype if doctype
-          end
-        end
-      end
-    end
-
-    # @return [Hash]
-    def to_hash
-      hash = super
-      hash["doctype"] = doctype if doctype
-      hash
-    end
   end
 end

data/relaton_w3c.gemspec

@@ -39,5 +39,5 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "vcr"
   spec.add_development_dependency "webmock"
 
-  spec.add_dependency "relaton-bib", ">= 1.1.0"
+  spec.add_dependency "relaton-bib", ">= 1.3.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-w3c
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.3.3
 platform: ruby
 authors:
 - Ribose Inc.
-autorequire: 
+autorequire:
 bindir: exe
 cert_chain: []
-date: 2020-06-03 00:00:00.000000000 Z
+date: 2020-09-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: debase
@@ -114,14 +114,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.1.0
+        version: 1.3.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.1.0
+        version: 1.3.0
 description: 'RelatonIso: retrieve W3C Standards for bibliographic using the IsoBibliographicItem
   model'
 email:
@@ -162,7 +162,7 @@ licenses:
 - BSD-2-Clause
 metadata:
   homepage_uri: https://github.com/relaton/relaton-wc3
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -178,7 +178,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubygems_version: 3.0.6
-signing_key: 
+signing_key:
 specification_version: 4
 summary: 'RelatonIso: retrieve W3C Standards for bibliographic using the IsoBibliographicItem
   model'