relaton-oasis 1.14.4 → 1.16.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c3258c1d1302d56309c6e63545e12fb6b8b014b8adbabbfaa3d5ca5c3e282fb3
-  data.tar.gz: '09380e04f60fcdd7863098c3ea15cc0ac685fbee7d5055aeaab7cd5b34d284bd'
+  metadata.gz: 3908ef54f79de05bfa9ee81c64f27b03e923f13fdd68df2c3a78b14bdc137bcf
+  data.tar.gz: 8e120eac3b886e8536532747a89b64f3c7fd9e64d098435e3e2135bde6f17acb
 SHA512:
-  metadata.gz: dd97e71ec0a28bb367e98566fe49e3b167cfab89f7e2b045873065ad8270cbb6c6b28331ff934b4136ac02b83a112d595db5d477ba2618ea42af22e5db80029f
-  data.tar.gz: c6c98cdfb711e0464a9bbe51f555625ca53d579190a511055ae4310756a0fc86505c2671c4c029b33a67fbb7fd64eb4083aec124fb488c5f7334073a663fa796
+  metadata.gz: 2f97cd29537cce17fa2f3ce144a7074727a277131a1ea5739480efe6adc3c7dfd6b7b2fc913c029e9ade56694f243311b7b30f7c5df878d34a46af6d80437fb3
+  data.tar.gz: 7e9373716c8c693c1532c7a0b413dd0d114597baba2c4a95be9d76b3796d76b16a38e2c44255f45b53642428ddbf7cb55c2562a6688c0cf63ca87ceb3f82ade8

data/.github/workflows/release.yml

@@ -7,7 +7,8 @@ on:
     inputs:
       next_version:
         description: |
-          Next release version. Possible values: x.y.z, major, minor, patch or pre|rc|etc
+          Next release version. Possible values: x.y.z, major, minor, patch (or pre|rc|etc).
+          Also, you can pass 'skip' to skip 'git tag' and do 'gem push' for the current version
         required: true
         default: 'skip'
   repository_dispatch:

data/Gemfile

@@ -15,5 +15,5 @@ gem "rubocop-performance"
 gem "rubocop-rails"
 gem "ruby-jing"
 gem "simplecov"
-gem "vcr", "~> 5.0.0"
+gem "vcr"
 gem "webmock"

data/README.adoc

@@ -51,7 +51,7 @@ item = RelatonOasis::OasisBibliography.get "OASIS amqp-core"
 [source,ruby]
 ----
 item.to_xml
-=> "<bibitem id="OASISamqp-core" type="standard" schema-version="v1.2.1">
+=> "<bibitem id="OASISamqp-core" type="standard" schema-version="v1.2.4">
       <fetched>2022-12-05</fetched>
       <title type="main" format="text/plain" language="en" script="Latn">Advanced Message Queueing Protocol (AMQP) v1.0</title>
       <docidentifier type="OASIS" primary="true">OASIS amqp-core</docidentifier>
@@ -64,12 +64,12 @@ With argument `bibdata: true` it ouputs XML wrapped by `bibdata` element and add
 [source,ruby]
 ----
 item.to_xml bibdata: true
-=> "<bibdata type="standard" schema-version="v1.2.1">
+=> "<bibdata type="standard" schema-version="v1.2.4">
       <fetched>2022-12-05</fetched>
       <title type="main" format="text/plain" language="en" script="Latn">Advanced Message Queueing Protocol (AMQP) v1.0</title>
       <docidentifier type="OASIS" primary="true">OASIS amqp-core</docidentifier>
       ...
-      <ext schema-version="v1.0.0">
+      <ext schema-version="v1.0.1">
         <doctype>standard</doctype>
         <editorialgroup>
           <technical-committee>OASIS Advanced Message Queuing Protocol (AMQP) TC</technical-committee>
@@ -110,7 +110,7 @@ RelatonOasis::XMLParser.from_xml File.read("spec/fixtures/oasis_bibdata.xml")
 [source,ruby]
 ----
 hash = YAML.load_file "spec/fixtures/oasis_bibdata.yaml"
-=> {"schema-version"=>"v1.2.1",
+=> {"schema-version"=>"v1.2.4",
  "id"=>"OASISamqp-core",
 ...
 

data/grammars/biblio.rng

@@ -1283,7 +1283,7 @@
       <value>mergedInto</value>
       <value>splits</value>
       <value>splitInto</value>
-      <value>instance</value>
+      <value>instanceOf</value>
       <value>hasInstance</value>
       <value>exemplarOf</value>
       <value>hasExemplar</value>

data/grammars/relaton-oasis.rng

@@ -54,6 +54,8 @@
       <value>Security</value>
       <value>SOA</value>
       <value>Web-Services</value>
+      <value>Software-Development</value>
+      <value>Virtualization</value>
     </choice>
   </define>
 </grammar>

data/lib/relaton_oasis/data_fetcher.rb

@@ -63,7 +63,7 @@ module RelatonOasis
     #
     # Save document to file
     #
-    # @param [RelatonOasis::OasisBibliographicItem] doc 
+    # @param [RelatonOasis::OasisBibliographicItem] doc
     #
     def save_doc(doc) # rubocop:disable Metrics/MethodLength
       c = case @format

data/lib/relaton_oasis/data_parser.rb

@@ -10,10 +10,16 @@ module RelatonOasis
     #
     def initialize(node)
       @node = node
-      @title = @node.at("./summary/div/h2").text
-      @text = @node.at(
+    end
+
+    def title
+      @title ||= @node.at("./summary/div/h2").text
+    end
+
+    def text
+      @text ||= @node.at(
         "./div/div/div[contains(@class, 'standard__grid--cite-as')]/p[em or i or a or span]",
-      )&.text
+      )&.text&.strip
     end
 
     #
@@ -27,8 +33,10 @@ module RelatonOasis
         doctype: parse_doctype,
         title: parse_title,
         docid: parse_docid,
+        link: parse_link,
         docnumber: parse_docnumber,
         date: parse_date,
+        contributor: parse_contributor,
         abstract: parse_abstract,
         language: ["en"],
         script: ["Latn"],
@@ -44,7 +52,7 @@ module RelatonOasis
     # @return [Array<RelatonBib::TypedTitleString>] <description>
     #
     def parse_title
-      [RelatonBib::TypedTitleString.new(type: "main", content: @title, language: "en", script: "Latn")]
+      [RelatonBib::TypedTitleString.new(type: "main", content: title, language: "en", script: "Latn")]
     end
 
     #
@@ -60,7 +68,7 @@ module RelatonOasis
       end
     end
 
-    #
+    # #
     # Parse abstract.
     #
     # @return [Array<RelatonBib::FormattedString>] abstract
@@ -87,6 +95,19 @@ module RelatonOasis
       RelatonBib::EditorialGroup.new tc
     end
 
+    def parse_authorizer
+      @node.xpath("./div[@class='standard__details']/a").map do |a|
+        cnt = RelatonBib::Contact.new(type: "uri", value: a[:href])
+        org = RelatonBib::Organization.new name: a.text.strip, contact: [cnt]
+        role = { type: "authorizer", description: ["Committee"] }
+        RelatonBib::ContributionInfo.new entity: org, role: [role]
+      end
+    end
+
+    def link_node
+      @link_node ||= @node.at("./div/div/div[contains(@class, 'standard__grid--cite-as')]/p[strong or span/strong]/a")
+    end
+
     #
     # Parse relation.
     #
@@ -118,6 +139,30 @@ module RelatonOasis
       ).map { |p| p.text.gsub(/^\[{1,2}|\]$/, "").strip }
     end
 
+    def parse_link
+      return [] if parts.size > 1
+
+      links.map do |l|
+        type = l[:href].match(/\.(\w+)$/)&.captures&.first
+        type ||= "src"
+        type.sub!("docx", "doc")
+        type.sub!("html", "src")
+        RelatonBib::TypedUri.new(type: type, content: l[:href])
+      end
+    end
+
+    def parts
+      @parts ||= @node.xpath(
+        "./div/div/div[contains(@class, 'standard__grid--cite-as')]/p[strong or span/strong]",
+      )
+    end
+
+    def links
+      l = @node.xpath("./div/div/div[1]/p[1]/a[@href]")
+      l = @node.xpath("./div/div/div[1]/p[2]/a[@href]") if l.empty?
+      l
+    end
+
     #
     # Parse document number.
     #
@@ -140,7 +185,7 @@ module RelatonOasis
     # @return [String] document identifier
     #
     def parts_to_docid(parts)
-      id = parts[1..-1].each_with_object(parts[0].split("-")) do |part, acc|
+      id = parts[1..].each_with_object(parts[0].split("-")) do |part, acc|
         chunks = part.split "-"
         chunks.each.with_index do |chunk, idx|
           unless chunk.casecmp(acc[idx])&.zero?

data/lib/relaton_oasis/data_parser_utils.rb

@@ -1,6 +1,93 @@
 module RelatonOasis
   # Common methods for document and part parsers.
   module DataParserUtils
+    #
+    # Parse contributor.
+    #
+    # @return [Array<RelatonBib::ContributionInfo>] contributors
+    #
+    def parse_contributor
+      publisher_oasis + parse_authorizer + parse_chairs + parse_editors
+    end
+
+    def publisher_oasis
+      cnt = RelatonBib::Contact.new type: "uri", value: "https://www.oasis-open.org/"
+      entity = RelatonBib::Organization.new name: "OASIS", contact: [cnt]
+      role = [
+        { type: "authorizer", description: ["Standards Development Organization"] },
+        { type: "publisher" },
+      ]
+      [RelatonBib::ContributionInfo.new(entity: entity, role: role)]
+    end
+
+    def parse_editors_from_text
+      return [] unless text
+
+      text.match(/(?<=Edited\sby\s)[^.]+/).to_s.split(/,?\sand\s|,\s/).map do |c|
+        RelatonBib::ContributionInfo.new(role: [type: "editor"], entity: create_person(c))
+      end
+    end
+
+    def page
+      return @page if defined? @page
+
+      if link_node && link_node[:href].match?(/\.html$/)
+        agent = Mechanize.new
+        agent.agent.allowed_error_codes = [404]
+        sleep 1 # to avoid 429 error
+        resp = agent.get link_node[:href]
+        @page = resp if resp.code == "200"
+      end
+    end
+
+    def parse_chairs
+      return [] unless page
+
+      page.xpath(
+        "//p[preceding-sibling::p[starts-with(., 'Chair')]][following-sibling::p[starts-with(., 'Editor')]]",
+      ).map { |p| create_contribution_info(p, "editor", ["Chair"]) }
+    end
+
+    def parse_editors
+      return parse_editors_from_text unless page
+
+      page.xpath(
+        "//p[contains(@class, 'Contributor')][preceding-sibling::p[starts-with(., 'Editor')]]" \
+        "[following-sibling::p[contains(@class, 'Title')]]",
+      ).map { |p| create_contribution_info(p, "editor") }
+    end
+
+    def create_contribution_info(person, type, description = [])
+      name = person.text.match(/^[^(]+/).to_s.strip
+      email, org = person.xpath ".//a[@href]"
+      entity = create_person name, email, org
+      role = { type: type, description: description }
+      RelatonBib::ContributionInfo.new(role: [role], entity: entity)
+    end
+
+    def create_person(name, email = nil, org = nil)
+      forename, surname = name.split
+      fn = RelatonBib::Forename.new(content: forename, language: ["en"], script: ["Latn"])
+      sn = RelatonBib::LocalizedString.new(surname, "en", "Latn")
+      name = RelatonBib::FullName.new(surname: sn, forename: [fn])
+      RelatonBib::Person.new(name: name, contact: contact(email), affiliation: affiliation(org))
+    end
+
+    def contact(email)
+      return [] unless email
+
+      [RelatonBib::Contact.new(type: "email", value: email[:href].split(":")[1])]
+    end
+
+    def affiliation(org)
+      return [] unless org
+
+      cnt = RelatonBib::Contact.new(type: "uri", value: org[:href])
+      org_name = org.text.gsub(/[\r\n]+/, " ")
+      organization = RelatonBib::Organization.new name: org_name, contact: [cnt]
+      [RelatonBib::Affiliation.new(organization: organization)]
+    end
+
     #
     # Parse document identifier specification.
     #
@@ -9,7 +96,7 @@ module RelatonOasis
     # @return [String] document identifier with specification if needed
     #
     def parse_spec(num)
-      id = case @text
+      id = case text
            when /OASIS Project Specification (\d+)/ then "#{num}-PS#{$1}"
            when /Committee Specification (\d+)/ then "#{num}-CS#{$1}"
            else num
@@ -27,11 +114,10 @@ module RelatonOasis
     def parse_part(docid)
       return docid if docid.match?(/(?:Part|Pt)\d+/i)
 
-      id = case @title
-           when /Part\s(\d+)/ then "#{docid}-Pt#{$1}"
-           else docid
-           end
-      parse_errata(id)
+      case title
+      when /Part\s(\d+)/ then "#{docid}-Pt#{$1}"
+      else docid
+      end
     end
 
     #
@@ -42,11 +128,11 @@ module RelatonOasis
     # @return [String] document identifier with errata if needed
     #
     def parse_errata(id)
-      return id if id.match?(/errata\d+/i)
+      return id.sub("errata", "Errata") if id.match?(/errata\d+/i)
 
-      case @title
-      when /Plus\sErrata\s(\d+)/ then "#{id}-plus-errata#{$1}"
-      when /Errata\s(\d+)/ then "#{id}-errata#{$1}"
+      case title
+      when /Plus\sErrata\s(\d+)/ then "#{id}-plus-Errata#{$1}"
+      when /Errata\s(\d+)/ then "#{id}-Errata#{$1}"
       else id
       end
     end
@@ -67,7 +153,7 @@ module RelatonOasis
     # @return [String] document type
     #
     def parse_doctype
-      case @text
+      case text
       when /OASIS Project Specification/, /Committee Specification/
         "specification"
       when /Technical Memorandum/ then "memorandum"

data/lib/relaton_oasis/data_part_parser.rb

@@ -10,13 +10,21 @@ module RelatonOasis
     #
     def initialize(node)
       @node = node
-      @text = node.xpath(
+    end
+
+    def text
+      @text ||= @node.xpath(
         "./strong/following-sibling::text()|./span/strong/../following-sibling::text()",
       ).text.strip
+    end
+
+    def title
+      return @title if @title
+
       t = @node.at("./span[@class='citationTitle' or @class='citeTitle']|./em|./i")
       @title = if t then t.text
                else
-                 @text.match(/(?<content>.+)\s(?:Edited|\d{2}\s\w+\d{4})/)[:content]
+                 text.match(/(?<content>.+)\s(?:Edited|\d{2}\s\w+\d{4})/)[:content]
                end.strip
     end
 
@@ -34,10 +42,10 @@ module RelatonOasis
         link: parse_link,
         docnumber: parse_docnumber,
         date: parse_date,
-        # abstract: parse_abstract,
+        abstract: parse_abstract,
         language: ["en"],
         script: ["Latn"],
-        # editorialgroup: parse_editorialgroup,
+        editorialgroup: parse_editorialgroup,
         relation: parse_relation,
         contributor: parse_contributor,
         technology_area: parse_technology_area,
@@ -50,7 +58,7 @@ module RelatonOasis
     # @return [Array<RelatonBib::TypedTitleString>] title
     #
     def parse_title
-      [RelatonBib::TypedTitleString.new(type: "main", content: @title, language: "en", script: "Latn")]
+      [RelatonBib::TypedTitleString.new(type: "main", content: title, language: "en", script: "Latn")]
     end
 
     #
@@ -61,9 +69,10 @@ module RelatonOasis
     def parse_docnumber
       ref = @node.at("./span[@class='citationLabel']/strong|./strong|b/span")
       num = ref.text.match(/[^\[\]]+/).to_s
+      id = parse_errata(num)
       # some part references need to be added by "Pt" to be distinguishable from root doc
-      num += "-Pt" if %w[CMIS-v1.1 DocBook-5.0 XACML-V3.0 mqtt-v3.1.1 OData-JSON-Format-v4.0].include?(num)
-      parse_spec num
+      id += "-Pt" if %w[CMIS-v1.1 DocBook-5.0 XACML-V3.0 mqtt-v3.1.1 OData-JSON-Format-v4.0].include?(id)
+      parse_spec id
     end
 
     #
@@ -82,29 +91,28 @@ module RelatonOasis
     # @return [Array<RelatonBib::BibliographicDate>] bibliographic dates
     #
     def parse_date
-      /(?<on>\d{1,2}\s\w+\s\d{4})/ =~ @text
+      /(?<on>\d{1,2}\s\w+\s\d{4})/ =~ text
       [RelatonBib::BibliographicDate.new(on: Date.parse(on).to_s, type: "issued")]
     end
 
-    # def parse_abstract
-    #   /\d{2}\s\w+\s\d{4}\.\s(?<content>.+)\sLatest\sversion/ =~ @text
-    #   [RelatonBib::FormattedString.new(content: content, language: "en", script: "Latn")]
-    # end
+    def parse_abstract
+      page.xpath("//p[preceding-sibling::p[starts-with(., 'Abstract')]][1]").map do |p|
+        cnt = p.text.gsub(/[\r\n]+/, " ").strip
+        RelatonBib::FormattedString.new(content: cnt, language: "en", script: "Latn")
+      end
+    end
 
     #
-    # Parse contributor.
+    # Parse technical committee.
     #
-    # @return [Array<RelatonBib::ContributionInfo>] contributors
+    # @return [RelatonBib::EditorialGroup] technical committee
     #
-    def parse_contributor
-      @text.match(/(?<=Edited\sby\s)[^.]+/).to_s.split(/,\s|,?\sand\s/).map do |c|
-        forename, surname = c.split
-        fn = RelatonBib::LocalizedString.new(forename, "en", "Latn")
-        sn = RelatonBib::LocalizedString.new(surname, "en", "Latn")
-        name = RelatonBib::FullName.new(surname: sn, forename: [fn])
-        entity = RelatonBib::Person.new(name: name)
-        RelatonBib::ContributionInfo.new(role: [type: "editor"], entity: entity)
+    def parse_editorialgroup
+      tc = page.xpath("//p[preceding-sibling::p[starts-with(., 'Technical')]][1]//a").map do |a|
+        wg = RelatonBib::WorkGroup.new name: a.text.strip
+        RelatonBib::TechnicalCommittee.new wg
       end
+      RelatonBib::EditorialGroup.new tc
     end
 
     #
@@ -119,6 +127,21 @@ module RelatonOasis
       [RelatonBib::DocumentRelation.new(type: "partOf", bibitem: bib)]
     end
 
+    def parse_authorizer
+      return [] unless page
+
+      page.xpath("//p[preceding-sibling::p[starts-with(., 'Technical')]][1]//a").map do |a|
+        cnt = RelatonBib::Contact.new(type: "uri", value: a[:href])
+        org = RelatonBib::Organization.new name: a.text.gsub(/[\r\n]+/, " ").strip, contact: [cnt]
+        role = { type: "authorizer", description: ["Committee"] }
+        RelatonBib::ContributionInfo.new entity: org, role: [role]
+      end
+    end
+
+    def link_node
+      @link_node = @node.at("./a")
+    end
+
     #
     # Parse technology area.
     #

data/lib/relaton_oasis/oasis_bibliographic_item.rb

@@ -3,7 +3,7 @@ module RelatonOasis
     AREAS = %w[Cloud Content-Technologies Cybersecurity e-Business eGov/Legal
                Emergency-Management Energy Information-Modeling IoT
                Lifecycle-Integration Localization Messaging Privacy/Identity
-               Security SOA Web-Services].freeze
+               Security SOA Web-Services Software-Development Virtualization].freeze
 
     attr_reader :technology_area
 

data/lib/relaton_oasis/oasis_bibliography.rb

@@ -34,7 +34,7 @@ module RelatonOasis
       #   reference is required
       # @return [RelatonOasis::OasisBibliographicItem, nil]
       def get(code, year = nil, _opts = {}) # rubocop:disable Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-        warn "[relaton-oasis] (#{code} fetching..."
+        warn "[relaton-oasis] (#{code}) fetching..."
         bibitem = search code, year
         if bibitem
           docid = bibitem.docidentifier.detect(&:primary).id

data/lib/relaton_oasis/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonOasis
-  VERSION = "1.14.4"
+  VERSION = "1.16.0"
 end

data/relaton_oasis.gemspec

@@ -33,7 +33,7 @@ Gem::Specification.new do |spec|
 
   spec.add_dependency "mechanize", "~> 2.8.0"
   spec.add_dependency "multi_json", "~> 1.15.0"
-  spec.add_dependency "relaton-bib", "~> 1.14.0"
+  spec.add_dependency "relaton-bib", "~> 1.16.0"
   spec.add_dependency "relaton-index", "~> 0.2.0"
 
   # For more information and examples about making a new gem, checkout our

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-oasis
 version: !ruby/object:Gem::Version
-  version: 1.14.4
+  version: 1.16.0
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-05-25 00:00:00.000000000 Z
+date: 2023-09-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mechanize
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.14.0
+        version: 1.16.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.14.0
+        version: 1.16.0
 - !ruby/object:Gem::Dependency
   name: relaton-index
   requirement: !ruby/object:Gem::Requirement