relaton-ecma 2.0.0.pre.alpha.2 → 2.0.0.pre.alpha.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: efa9481085cffe170d5ec5a11c1de88313f91b4f9026ccf9ada0689d730172e0
-  data.tar.gz: c65688b86e6a80648f4a1906533a81ab098782581059081d45bffa576ab54531
+  metadata.gz: ada2f96de510e80052f8d09ccb04ffa43718c7582eea853a10f1bcf30b39754f
+  data.tar.gz: 7400f8754bbaf45d9718c1855ea1a5b0363fe8c8080dbfef18177e6c4cddd5ae
 SHA512:
-  metadata.gz: ae251ee4a4c0deb57e7a0d7b6b662a3f8fe06db40041b6da343c56e256c7018fcf40426d64e0558ac06df96292c67a340e7b8ba3dd7b3883b00b6004f70dbb4c
-  data.tar.gz: ce46e548e784c0ec824dfff480176be6639458ea2bb42479f61cbbcf2b484de2d5a7aa6470524202a0ad41d592595d87ecec905275fc0cc7b6d24b386cc4c8f2
+  metadata.gz: a685d9c330a66bd2907eccd9fba5bdb0fa439994e65a931652c045fc5f0ad8183dd7a9ff7afca0f56ed80523562d18e15527c0f35a96617c19c41952a1a5fba2
+  data.tar.gz: b7b9a2fbdb69b84bc24f9af125733b8ee43c3631bdf1dde3407170d1051562dd2ee3de957d5fc07d6ca15a1b53c6548644c23032556ee5d6116be1c1b6a0a4f1

data/README.adoc

@@ -108,7 +108,7 @@ Relaton::Ecma::Bibliography.get '1111'
 [source,ruby]
 ----
 item.to_xml
-=> "<bibitem id="ECMA6" type="standard" schema-version="v1.4.1">
+=> "<bibitem id="ECMA6" type="standard" schema-version="v1.5.6">
       <fetched>2025-12-29</fetched>
       <title language="en" script="Latn">7-bit coded character set</title>
       <uri type="src">https://ecma-international.org/publications-and-standards/standards/ecma-6/</uri>
@@ -122,7 +122,7 @@ With `bibdata: true` option XML output wrapped with `bibdata` element and `ext`
 [source,ruby]
 ----
 item.to_xml bibdata: true
-=> "<bibdata type="standard" schema-version="v1.4.1">
+=> "<bibdata type="standard" schema-version="v1.5.6">
       <fetched>2025-12-29</fetched>
       <title language="en" script="Latn">7-bit coded character set</title>
       <uri type="src">https://ecma-international.org/publications-and-standards/standards/ecma-6/</uri>

data/grammars/basicdoc.rng

@@ -187,6 +187,15 @@ Applicable to modify and delete</a:documentation>
         <a:documentation>Optional caption of this block</a:documentation>
       </attribute>
     </optional>
+    <optional>
+      <attribute name="position">
+        <a:documentation>For an "add" change, whether the change is added before or after the location</a:documentation>
+        <choice>
+          <value>before</value>
+          <value>after</value>
+        </choice>
+      </attribute>
+    </optional>
     <optional>
       <element name="location">
         <a:documentation>The location(s) in the original document which have undergone the change described in this block</a:documentation>
@@ -208,11 +217,15 @@ Applicable to modify and delete</a:documentation>
     </zeroOrMore>
     <optional>
       <element name="newcontent">
-        <a:documentation>New content to be added to the document; applicable to add and modify</a:documentation>
+        <a:documentation>New content to be added to the document; applicable to add and modify.
+Can be blocks and/or sections</a:documentation>
         <ref name="OptionalId"/>
         <zeroOrMore>
           <ref name="BasicBlock"/>
         </zeroOrMore>
+        <zeroOrMore>
+          <ref name="section"/>
+        </zeroOrMore>
       </element>
     </optional>
     <zeroOrMore>

data/grammars/biblio.rng

@@ -1142,11 +1142,11 @@ NOTE: This should preferably be encoded as  a URI or short identifier, rather th
         <a:documentation>Information about how long the current description of the bibliographic item is valid for</a:documentation>
       </ref>
     </optional>
-    <optional>
+    <zeroOrMore>
       <ref name="depiction">
         <a:documentation>Depiction of the bibliographic item, typically an image</a:documentation>
       </ref>
-    </optional>
+    </zeroOrMore>
   </define>
   <define name="ReducedBibliographicItem">
     <a:documentation>Reduced description of a bibliographic resource, without mandatory title and docidentifier, used for document relations
@@ -1939,10 +1939,10 @@ Detailed in https://www.relaton.org/model/relations/</a:documentation>
       <value>hasAnnotation</value>
       <value>draftOf</value>
       <value>hasDraft</value>
-      <value>preliminaryDraftOf</value>
-      <value>hasPreliminaryDraft</value>
-      <value>revisionDraftOf</value>
-      <value>hasRevisionDraft</value>
+      <value>predecessorDraftOf</value>
+      <value>hasPredecessorDraft</value>
+      <value>successorDraftOf</value>
+      <value>hasSuccessorDraft</value>
       <value>editionOf</value>
       <value>hasEdition</value>
       <value>updates</value>
@@ -2063,13 +2063,13 @@ provided that it is not the entire bibliographic item that is so related</a:docu
           <ref name="LocalizedString"/>
         </element>
       </optional>
-      <oneOrMore>
+      <zeroOrMore>
         <element name="taxon">
           <a:documentation>The keywords as a hierarchical taxonomy. For example, the sequence of `taxon` elements
 `pump`, `centrifugal pump`, `line shaft pump` represents a taxonomic classification</a:documentation>
           <ref name="LocalizedString"/>
         </element>
-      </oneOrMore>
+      </zeroOrMore>
       <zeroOrMore>
         <ref name="vocabid">
           <a:documentation>Identifiers for the keyword as a controlled vocabulary</a:documentation>

data/lib/relaton/ecma/bibliography.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal:true
 
+require "mechanize"
+
 module Relaton
   module Ecma
     # IETF bibliography module

data/lib/relaton/ecma/data_fetcher.rb

@@ -4,6 +4,11 @@ require "English"
 require "mechanize"
 require "relaton/core"
 require_relative "../ecma"
+require_relative "parser_common"
+require_relative "page_fetcher"
+require_relative "standard_parser"
+require_relative "memento_parser"
+require_relative "edition_parser"
 require_relative "data_parser"
 
 module Relaton
@@ -16,6 +21,10 @@ module Relaton
         @index ||= Relaton::Index.find_or_create :ecma, file: "#{INDEXFILE}.yaml"
       end
 
+      def log_error(msg)
+        Util.error msg
+      end
+
       def agent
         @agent ||= Mechanize.new.tap { |a| a.user_agent_alias = Mechanize::AGENT_ALIASES.keys.sample }
       end
@@ -62,7 +71,7 @@ module Relaton
 
       # @param hit [Nokogiri::XML::Element]
       def parse_page(hit) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
-        DataParser.new(hit).parse.each { |item| write_file item }
+        DataParser.new(hit, @errors).parse.each { |item| write_file item }
       end
 
       # @param type [String]
@@ -87,6 +96,7 @@ module Relaton
       def fetch(_ = nil)
         SOURCES.each { |source| html_index source }
         index.save
+        report_errors
       end
     end
   end

data/lib/relaton/ecma/data_parser.rb

@@ -1,235 +1,49 @@
 module Relaton
   module Ecma
     class DataParser
-      MATTRS = %i[docidentifier title date source ext].freeze
-      ATTRS = MATTRS + %i[abstract relation edition ext].freeze
+      include ParserCommon
 
       #
       # Initialize parser
       #
       # @param [Nokogiri::XML::Element] hit document hit
+      # @param [Hash] errors error tracking hash
       #
-      def initialize(hit)
+      def initialize(hit, errors = {})
         @hit = hit
-        @bib = {
-          type: "standard", language: ["en"], script: ["Latn"], place: [Bib::Place.new(city: "Geneva")]
-        }
-        @agent = Mechanize.new
+        @errors = errors
       end
 
-      def parse # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
+      # @return [Array<Relaton::Ecma::ItemData>]
+      def parse
         if @hit[:href]
-          @agent.user_agent_alias = Mechanize::AGENT_ALIASES.keys[rand(21)]
-          @doc = get_page @hit[:href]
-          ATTRS.each { |a| @bib[a] = send "fetch_#{a}" }
+          parse_standard
         else
-          MATTRS.each { |a| @bib[a] = send "fetch_mem_#{a}" }
+          parse_memento
         end
-        @bib[:contributor] = contributor
-        items = [ItemData.new(**@bib)]
-        items + parse_editions
       end
 
-      #
-      # Get page with retries
-      #
-      # @param [String] url url to fetch
-      #
-      # @return [Mechanize::Page] document
-      #
-      def get_page(url)
-        3.times do |n|
-          sleep n
-          doc = @agent.get url
-          return doc
-        rescue StandardError => e
-          Util.error e.message
-        end
-      end
-
-      #
-      # Parse editions
-      #
-      # @param [Mechanize::Page] doc document
-      # @param [Hash] bib bibliographic item the last edition
-      #
-      # @return [Array<Relaton::Ecma::ItemData>] editions
-      #
-      def parse_editions # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity
-        return [] unless @doc
-
-        docid = @bib[:docid]
-        @doc.xpath('//div[@id="main"]/div[1]/div/main/article/div/div/standard/div/ul/li').map do |hit|
-          id, ed, @bib[:date], vol = edition_id_parts hit.at("./span", "./a").text
-          @bib[:source] = edition_source(hit) + edition_translation_source(ed)
-          next if ed.nil? || ed.empty?
-
-          @bib[:docidentifier] = id.nil? || id.empty? ? docid : fetch_docidentifier(id)
-          @bib[:edition] = Bib::Edition.new(content: ed)
-          @bib[:extent] = create_extent(vol)
-          ItemData.new(**@bib)
-        end.compact
-      end
-
-      def create_extent(vol)
-        return unless vol && !vol.empty?
-
-        locality = Bib::Locality.new(type: "volume", reference_from: vol)
-        [Bib::Extent.new(locality: [locality])]
-      end
-
-      def edition_source(hit)
-        { "src" => hit.at("./a"), "pdf" => hit.at("./span/a") }.map do |type, a|
-          Bib::Uri.new(type: type, content: a[:href]) if a
-        end.compact
-      end
-
-      #
-      # Parse edition and date
-      #
-      # @param [String] text identifier text
-      #
-      # @return [Array<String,nil,Array<Relaton::Bib::Date>>] edition and date
-      #
-      def edition_id_parts(text) # rubocop:disable Metrics/MethodLength
-        %r{^
-          (?<id>\w+(?:[\d-]+|\sTR/\d+)),?\s
-          (?:Volume\s(?<vol>[\d.]+),?\s)?
-          (?<ed>[\d.]+)(?:st|nd|rd|th)?\sedition
-          (?:[,.]\s(?<dt>\w+\s\d+))?
-        }x =~ text
-        date = [dt].compact.map do |d|
-          on = Date.strptime(d, "%B %Y").strftime("%Y-%m")
-          Bib::Date.new(type: "published", at: on)
-        end
-        [id, ed, date, vol]
-      end
-
-      # @return [Array<Relaton::Bib::Docidentifier>]
-      def fetch_docidentifier(id = nil)
-        id ||= @hit.text
-        [Bib::Docidentifier.new(type: "ECMA", content: id, primary: true)]
-      end
-
-      # @return [Array<Relaton::Bib::Uri>]
-      def fetch_source # rubocop:disable Metrics/AbcSize
-        source = []
-        source << Bib::Uri.new(type: "src", content: @hit[:href]) if @hit[:href]
-        ref = @doc.at('//div[@class="ecma-item-content-wrapper"]/span/a',
-                      '//div[@class="ecma-item-content-wrapper"]/a')
-        source << Bib::Uri.new(type: "pdf", content: ref[:href]) if ref
-        source + edition_translation_source(@bib[:edition]&.content)
-      end
-
-      def fetch_mem_source
-        @hit.xpath("./div/section/div/p/a").map do |a|
-          Bib::Uri.new(type: "pdf", content: a[:href])
-        end
-      end
-
-      def edition_translation_source(edition)
-        translation_source.select { |s| s[:ed] == edition }.map { |s| s[:source] }
-      end
-
-      def translation_source
-        return [] unless @doc
-
-        @translation_source ||= @doc.xpath("//h2[.='Translations']/following-sibling::ul/li").map do |l|
-          a = l.at("span/a")
-          id = l.at("span").text
-          %r{\w+[\d-]+,\s(?<lang>\w+)\sversion,\s(?<ed>[\d.]+)(?:st|nd|rd|th)\sedition} =~ id
-          case lang
-          when "Japanese"
-            { ed: ed, source: Bib::Uri.new(type: "pdf", language: "ja", script: "Jpan", content: a[:href]) }
-          end
-        end.compact
-      end
-
-      # @return [Array<Relaton::Bib::Title>]
-      def fetch_title
-        @doc.xpath('//p[@class="ecma-item-short-description"]').map do |t|
-          Bib::Title.new(content: t.text.strip, language: "en", script: "Latn")
-        end
-      end
-
-      # @return [Array<Relaton::Bib::LocalizedMarkedUpString>]
-      def fetch_abstract
-        content = @doc.xpath('//div[@class="ecma-item-content"]/p').map do |a|
-          a.text.strip.squeeze(" ").gsub("\r\n", "")
-        end.join "\n"
-        return [] if content.empty?
-
-        [Bib::LocalizedMarkedUpString.new(content: content, language: "en", script: "Latn")]
-      end
-
-      # @return [Array<Relaton::Bib::Date>]
-      def fetch_date
-        @doc.xpath('//p[@class="ecma-item-edition"]').map do |d|
-          date = d.text.split(", ").last
-          Bib::Date.new type: "published", at: date
-        end
-      end
-
-      # @return [Array<Relaton::Bib::Relation>]
-      def fetch_relation # rubocop:disable Metrics/AbcSize, Metrics/MethodLength, Metrics/CyclomaticComplexity
-        @doc.xpath("//ul[@class='ecma-item-archives']/li").filter_map do |rel|
-          ref, ed, date, vol = edition_id_parts rel.at("span").text
-          next if ed.nil? || ed.empty?
+      private
 
-          docid = Bib::Docidentifier.new(type: "ECMA", content: ref, primary: true)
-          source = rel.xpath("span/a").map { |l| Bib::Uri.new type: "pdf", content: l[:href] }
-          edition = Bib::Edition.new content: ed
-          extent = create_extent(vol)
-          bibitem = ItemData.new(
-            docidentifier: [docid], formattedref: ref, date: date, edition: edition,
-            source: source, extent: extent
-          )
-          Bib::Relation.new(type: "updates", bibitem: bibitem)
-        end
-      end
-
-      #
-      # @return [Relaton::Bib::Edition, nil]
-      #
-      def fetch_edition
-        cnt = @doc.at('//p[@class="ecma-item-edition"]')&.text&.match(/^\d+(?=(?:st|nd|th|rd))/)&.to_s
-        Bib::Edition.new(content: cnt) if cnt && !cnt.empty?
-      end
-
-      def contributor
-        orgname = Bib::TypedLocalizedString.new(content: "Ecma International", language: "en", script: "Latn")
-        org = Bib::Organization.new name: [orgname]
-        role = Bib::Contributor::Role.new type: "publisher"
-        [Bib::Contributor.new(organization: org, role: [role])]
-      end
-
-      # @return [Array<Relaton::Bib::Docidentifier>]
-      def fetch_mem_docidentifier
-        code = "ECMA MEM/#{@hit.at('div[1]//p').text}"
-        fetch_docidentifier code
-      end
-
-      def fetch_mem_date
-        date = @hit.at("div[2]//p").text
-        on = Date.strptime(date, "%B %Y").strftime "%Y-%m"
-        [Bib::Date.new(type: "published", at: on)]
-      end
-
-      def fetch_mem_title
-        year = @hit.at("div[1]//p").text
-        content = "\"Memento #{year}\" for year #{year}"
-        [Bib::Title.new(content: content, language: "en", script: "Latn")]
-      end
-
-      def fetch_ext
-        Ext.new(doctype: fetch_doctype, flavor: "ecma")
+      def parse_standard
+        doc = PageFetcher.new.get(@hit[:href])
+        parser = StandardParser.new(hit: @hit, doc: doc, errors: @errors)
+        bib = parser.to_bib_hash
+        bib[:contributor] = contributor
+        items = [ItemData.new(**bib)]
+        edition_parser = EditionParser.new(
+          doc: doc, bib: bib, errors: @errors,
+          translation_source: parser.translation_source
+        )
+        items + edition_parser.parse
       end
 
-      def fetch_doctype
-        Bib::Doctype.new content: "document"
+      def parse_memento
+        parser = MementoParser.new(hit: @hit, errors: @errors)
+        bib = parser.to_bib_hash
+        bib[:contributor] = contributor
+        [ItemData.new(**bib)]
       end
-
-      alias_method :fetch_mem_ext, :fetch_ext
     end
   end
 end

data/lib/relaton/ecma/edition_parser.rb

@@ -0,0 +1,80 @@
+module Relaton
+  module Ecma
+    class EditionParser
+      include ParserCommon
+
+      # @param [Mechanize::Page] doc document page
+      # @param [Hash] bib base bibliographic item attributes
+      # @param [Hash] errors error tracking hash
+      # @param [Array] translation_source precomputed translation sources
+      def initialize(doc:, bib:, errors: {}, translation_source: [])
+        @doc = doc
+        @bib = bib
+        @errors = errors
+        @translation_source = translation_source
+      end
+
+      # @return [Array<Relaton::Ecma::ItemData>] editions
+      def parse # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity
+        return [] unless @doc
+
+        docid = @bib[:docidentifier]
+        @doc.xpath('//div[@id="main"]/div[1]/div/main/article/div/div/standard/div/ul/li').map do |hit|
+          bib = @bib.dup
+          id, ed, bib[:date], vol = edition_id_parts hit.at("./span", "./a").text
+          bib[:source] = edition_source(hit) + edition_translation_source(ed)
+          next if ed.nil? || ed.empty?
+
+          bib[:docidentifier] = id.nil? || id.empty? ? docid : fetch_docidentifier(id)
+          @errors[:edition_docidentifier] &&= bib[:docidentifier].empty?
+          bib[:edition] = Bib::Edition.new(content: ed)
+          bib[:extent] = create_extent(vol)
+          @errors[:edition_extent] &&= bib[:extent].nil?
+          ItemData.new(**bib)
+        end.compact
+      end
+
+      #
+      # Parse edition and date
+      #
+      # @param [String] text identifier text
+      #
+      # @return [Array<String,nil,Array<Relaton::Bib::Date>>] edition and date
+      #
+      def edition_id_parts(text) # rubocop:disable Metrics/MethodLength
+        %r{^
+          (?<id>\w+(?:[\d-]+|\sTR/\d+)),?\s
+          (?:Volume\s(?<vol>[\d.]+),?\s)?
+          (?<ed>[\d.]+)(?:st|nd|rd|th)?\sedition
+          (?:[,.]\s(?<dt>\w+\s\d+))?
+        }x =~ text
+        date = [dt].compact.map do |d|
+          on = Date.strptime(d, "%B %Y").strftime("%Y-%m")
+          Bib::Date.new(type: "published", at: on)
+        end
+        [id, ed, date, vol]
+      end
+
+      def edition_source(hit)
+        es = { "src" => hit.at("./a"), "pdf" => hit.at("./span/a") }.map do |type, a|
+          Bib::Uri.new(type: type, content: a[:href]) if a
+        end.compact
+        @errors[:edition_source] &&= es.empty?
+        es
+      end
+
+      def create_extent(vol)
+        return unless vol && !vol.empty?
+
+        locality = Bib::Locality.new(type: "volume", reference_from: vol)
+        [Bib::Extent.new(locality: [locality])]
+      end
+
+      private
+
+      def edition_translation_source(edition)
+        @translation_source.select { |s| s[:ed] == edition }.map { |s| s[:source] }
+      end
+    end
+  end
+end

data/lib/relaton/ecma/memento_parser.rb

@@ -0,0 +1,60 @@
+module Relaton
+  module Ecma
+    class MementoParser
+      include ParserCommon
+
+      ATTRS = %i[docidentifier title date source ext].freeze
+
+      # @param [Nokogiri::XML::Element] hit document hit
+      # @param [Hash] errors error tracking hash
+      def initialize(hit:, errors: {})
+        @hit = hit
+        @errors = errors
+      end
+
+      # @return [Hash] bibliographic item attributes
+      def to_bib_hash
+        bib = default_bib_hash
+        ATTRS.each { |a| bib[a] = send "fetch_#{a}" }
+        bib
+      end
+
+      private
+
+      # @return [Array<Relaton::Bib::Docidentifier>]
+      def fetch_docidentifier
+        code = "ECMA MEM/#{@hit.at('div[1]//p').text}"
+        docid = super(code)
+        @errors[:memento_docidentifier] &&= docid.empty?
+        docid
+      end
+
+      # @return [Array<Relaton::Bib::Title>]
+      def fetch_title
+        year = @hit.at("div[1]//p").text
+        content = "\"Memento #{year}\" for year #{year}"
+        result = [Bib::Title.new(content: content, language: "en", script: "Latn")]
+        @errors[:memento_title] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Date>]
+      def fetch_date
+        date = @hit.at("div[2]//p").text
+        on = Date.strptime(date, "%B %Y").strftime "%Y-%m"
+        result = [Bib::Date.new(type: "published", at: on)]
+        @errors[:memento_date] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Uri>]
+      def fetch_source
+        result = @hit.xpath("./div/section/div/p/a").map do |a|
+          Bib::Uri.new(type: "pdf", content: a[:href])
+        end
+        @errors[:memento_source] &&= result.empty?
+        result
+      end
+    end
+  end
+end

data/lib/relaton/ecma/page_fetcher.rb

@@ -0,0 +1,27 @@
+module Relaton
+  module Ecma
+    class PageFetcher
+      def initialize
+        @agent = Mechanize.new
+        @agent.user_agent_alias = Mechanize::AGENT_ALIASES.keys[rand(21)]
+      end
+
+      #
+      # Get page with retries
+      #
+      # @param [String] url url to fetch
+      #
+      # @return [Mechanize::Page] document
+      #
+      def get(url)
+        3.times do |n|
+          sleep n
+          doc = @agent.get url
+          return doc
+        rescue StandardError => e
+          Util.error e.message
+        end
+      end
+    end
+  end
+end

data/lib/relaton/ecma/parser_common.rb

@@ -0,0 +1,33 @@
+module Relaton
+  module Ecma
+    module ParserCommon
+      def default_bib_hash
+        {
+          type: "standard", language: ["en"], script: ["Latn"], place: [Bib::Place.new(city: "Geneva")]
+        }
+      end
+
+      def contributor
+        orgname = Bib::TypedLocalizedString.new(content: "Ecma International", language: "en", script: "Latn")
+        org = Bib::Organization.new name: [orgname]
+        role = Bib::Contributor::Role.new type: "publisher"
+        [Bib::Contributor.new(organization: org, role: [role])]
+      end
+
+      # @return [Array<Relaton::Bib::Docidentifier>]
+      def fetch_docidentifier(id = nil)
+        return [] if id.nil? || id.empty?
+
+        [Bib::Docidentifier.new(type: "ECMA", content: id, primary: true)]
+      end
+
+      def fetch_ext
+        Ext.new(doctype: fetch_doctype, flavor: "ecma")
+      end
+
+      def fetch_doctype
+        Bib::Doctype.new content: "document"
+      end
+    end
+  end
+end

data/lib/relaton/ecma/standard_parser.rb

@@ -0,0 +1,134 @@
+module Relaton
+  module Ecma
+    class StandardParser
+      include ParserCommon
+
+      ATTRS = %i[docidentifier title date source abstract relation edition ext].freeze
+
+      # @param [Nokogiri::XML::Element] hit document hit
+      # @param [Mechanize::Page] doc fetched document page
+      # @param [Hash] errors error tracking hash
+      def initialize(hit:, doc:, errors: {})
+        @hit = hit
+        @doc = doc
+        @errors = errors
+      end
+
+      # @return [Hash] bibliographic item attributes
+      def to_bib_hash
+        bib = default_bib_hash
+        ATTRS.each { |a| bib[a] = send "fetch_#{a}" }
+        bib
+      end
+
+      # @return [Array] precomputed translation sources
+      def translation_source
+        @translation_source ||= parse_translation_source
+      end
+
+      # @return [Array<Relaton::Bib::Docidentifier>]
+      def fetch_docidentifier
+        result = super(@hit.text)
+        @errors[:standard_docidentifier] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Title>]
+      def fetch_title
+        result = @doc.xpath('//p[@class="ecma-item-short-description"]').map do |t|
+          Bib::Title.new(content: t.text.strip, language: "en", script: "Latn")
+        end
+        @errors[:standard_title] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::LocalizedMarkedUpString>]
+      def fetch_abstract
+        content = @doc.xpath('//div[@class="ecma-item-content"]/p').map do |a|
+          a.text.strip.squeeze(" ").gsub("\r\n", "")
+        end.join "\n"
+        return [] if content.empty?
+
+        result = [Bib::Abstract.new(content: content, language: "en", script: "Latn")]
+        @errors[:standard_abstract] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Date>]
+      def fetch_date
+        result = @doc.xpath('//p[@class="ecma-item-edition"]').map do |d|
+          date = d.text.split(", ").last
+          Bib::Date.new type: "published", at: date
+        end
+        @errors[:standard_date] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Uri>]
+      def fetch_source # rubocop:disable Metrics/AbcSize
+        source = []
+        source << Bib::Uri.new(type: "src", content: @hit[:href]) if @hit[:href]
+        ref = @doc.at('//div[@class="ecma-item-content-wrapper"]/span/a',
+                      '//div[@class="ecma-item-content-wrapper"]/a')
+        source << Bib::Uri.new(type: "pdf", content: ref[:href]) if ref
+        result = source + edition_translation_source(fetch_edition_content)
+        @errors[:standard_source] &&= result.empty?
+        result
+      end
+
+      # @return [Array<Relaton::Bib::Relation>]
+      def fetch_relation # rubocop:disable Metrics/AbcSize, Metrics/MethodLength, Metrics/CyclomaticComplexity
+        edition_parser = EditionParser.new(doc: @doc, bib: {}, errors: @errors)
+        result = @doc.xpath("//ul[@class='ecma-item-archives']/li").filter_map do |rel|
+          ref, ed, date, vol = edition_parser.edition_id_parts rel.at("span").text
+          next if ed.nil? || ed.empty?
+
+          docid = Bib::Docidentifier.new(type: "ECMA", content: ref, primary: true)
+          source = rel.xpath("span/a").map { |l| Bib::Uri.new type: "pdf", content: l[:href] }
+          edition = Bib::Edition.new content: ed
+          extent = edition_parser.create_extent(vol)
+          @errors[:standard_relation_extent] &&= extent.nil?
+          bibitem = ItemData.new(
+            docidentifier: [docid], formattedref: Bib::Formattedref.new(content: ref), date: date, edition: edition,
+            source: source, extent: extent
+          )
+          Bib::Relation.new(type: "updates", bibitem: bibitem)
+        end
+        @errors[:standard_relation] &&= result.empty?
+        result
+      end
+
+      # @return [Relaton::Bib::Edition, nil]
+      def fetch_edition
+        cnt = fetch_edition_content
+        result = Bib::Edition.new(content: cnt) if cnt && !cnt.empty?
+        @errors[:standard_edition] &&= result.nil?
+        result
+      end
+
+      private
+
+      def fetch_edition_content
+        @doc.at('//p[@class="ecma-item-edition"]')&.text&.match(/^\d+(?=(?:st|nd|th|rd))/)&.to_s
+      end
+
+      def edition_translation_source(edition)
+        translation_source.select { |s| s[:ed] == edition }.map { |s| s[:source] }
+      end
+
+      def parse_translation_source
+        return [] unless @doc
+
+        @doc.xpath("//h2[.='Translations']/following-sibling::ul/li").map do |l|
+          a = l.at("span/a")
+          id = l.at("span").text
+          %r{\w+[\d-]+,\s(?<lang>\w+)\sversion,\s(?<ed>[\d.]+)(?:st|nd|rd|th)\sedition} =~ id
+          case lang
+          when "Japanese"
+            { ed: ed, source: Bib::Uri.new(type: "pdf", language: "ja", script: "Jpan", content: a[:href]) }
+          end
+        end.compact
+      end
+    end
+  end
+end

data/lib/relaton/ecma/version.rb

@@ -1,5 +1,5 @@
 module Relaton
   module Ecma
-    VERSION = "2.0.0-alpha.2".freeze
+    VERSION = "2.0.0-alpha.3".freeze
   end
 end

data/relaton_ecma.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |spec| # rubocop:disable Metrics/BlockLength
   spec.require_paths = ["lib"]
 
   spec.add_dependency "mechanize", "~> 2.10"
-  spec.add_dependency "relaton-bib", "~> 2.0.0-alpha.4"
-  spec.add_dependency "relaton-core", "~> 0.0.9"
+  spec.add_dependency "relaton-bib", "~> 2.0.0-alpha.7"
+  spec.add_dependency "relaton-core", "~> 0.0.12"
   spec.add_dependency "relaton-index", "~> 0.2.4"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: relaton-ecma
 version: !ruby/object:Gem::Version
-  version: 2.0.0.pre.alpha.2
+  version: 2.0.0.pre.alpha.3
 platform: ruby
 authors:
 - Ribose Inc.
@@ -29,28 +29,28 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0.pre.alpha.4
+        version: 2.0.0.pre.alpha.7
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0.pre.alpha.4
+        version: 2.0.0.pre.alpha.7
 - !ruby/object:Gem::Dependency
   name: relaton-core
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.0.9
+        version: 0.0.12
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.0.9
+        version: 0.0.12
 - !ruby/object:Gem::Dependency
   name: relaton-index
   requirement: !ruby/object:Gem::Requirement
@@ -96,10 +96,15 @@ files:
 - lib/relaton/ecma/bibliography.rb
 - lib/relaton/ecma/data_fetcher.rb
 - lib/relaton/ecma/data_parser.rb
+- lib/relaton/ecma/edition_parser.rb
 - lib/relaton/ecma/ext.rb
 - lib/relaton/ecma/item.rb
 - lib/relaton/ecma/item_data.rb
+- lib/relaton/ecma/memento_parser.rb
+- lib/relaton/ecma/page_fetcher.rb
+- lib/relaton/ecma/parser_common.rb
 - lib/relaton/ecma/processor.rb
+- lib/relaton/ecma/standard_parser.rb
 - lib/relaton/ecma/util.rb
 - lib/relaton/ecma/version.rb
 - relaton_ecma.gemspec