relaton-jis 1.14.0

data/grammars/relaton-jis-compile.rng

@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<grammar xmlns="http://relaxng.org/ns/structure/1.0">
+  <include href="basicdoc.rng"/>
+  <include href="relaton-jis.rng"/>
+  <start>
+    <choice>
+      <ref name="bibitem"/>
+      <ref name="bibdata"/>
+    </choice>
+  </start>
+</grammar>

data/grammars/relaton-jis.rng

@@ -0,0 +1,231 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<grammar xmlns="http://relaxng.org/ns/structure/1.0" datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes">
+  <include href="biblio-standoc.rng">
+    <define name="BibDataExtensionType">
+      <optional>
+        <attribute name="schema-version"/>
+      </optional>
+      <ref name="doctype"/>
+      <optional>
+        <ref name="docsubtype"/>
+      </optional>
+      <ref name="editorialgroup"/>
+      <zeroOrMore>
+        <ref name="ics"/>
+      </zeroOrMore>
+      <ref name="structuredidentifier"/>
+      <optional>
+        <ref name="stagename"/>
+      </optional>
+    </define>
+    <define name="bdate">
+      <element name="date">
+        <attribute name="type">
+          <choice>
+            <ref name="BibliographicDateType"/>
+            <text/>
+          </choice>
+        </attribute>
+        <choice>
+          <group>
+            <element name="from">
+              <ref name="ISO8601Date"/>
+            </element>
+            <optional>
+              <element name="to">
+                <ref name="ISO8601Date"/>
+              </element>
+            </optional>
+          </group>
+          <element name="on">
+            <choice>
+              <ref name="ISO8601Date"/>
+              <value>--</value>
+              <value>–</value>
+            </choice>
+          </element>
+        </choice>
+      </element>
+    </define>
+    <define name="organization">
+      <element name="organization">
+        <oneOrMore>
+          <ref name="orgname"/>
+        </oneOrMore>
+        <optional>
+          <ref name="abbreviation"/>
+        </optional>
+        <optional>
+          <ref name="uri"/>
+        </optional>
+        <zeroOrMore>
+          <ref name="org-identifier"/>
+        </zeroOrMore>
+        <zeroOrMore>
+          <ref name="contact"/>
+        </zeroOrMore>
+        <optional>
+          <ref name="technical-committee"/>
+        </optional>
+        <optional>
+          <ref name="subcommittee"/>
+        </optional>
+        <optional>
+          <ref name="workgroup"/>
+        </optional>
+        <optional>
+          <ref name="secretariat"/>
+        </optional>
+      </element>
+    </define>
+    <define name="DocumentType">
+      <choice>
+        <value>british-standard</value>
+        <value>draft-for-development</value>
+        <value>published-document</value>
+        <value>privately-subscribed-standard</value>
+        <value>publicly-available-specification</value>
+        <value>flex-standard</value>
+        <value>international-standard</value>
+        <value>technical-specification</value>
+        <value>technical-report</value>
+        <value>guide</value>
+        <value>international-workshop-agreement</value>
+        <value>industry-technical-agreement</value>
+        <value>standard</value>
+        <value>european-workshop-agreement</value>
+        <value>fast-track-standard</value>
+        <value>expert-commentary</value>
+      </choice>
+    </define>
+    <define name="DocumentSubtype">
+      <choice>
+        <value>specification</value>
+        <value>method-of-test</value>
+        <value>method-of-specifying</value>
+        <value>vocabulary</value>
+        <value>code-of-practice</value>
+      </choice>
+    </define>
+    <define name="structuredidentifier">
+      <element name="structuredidentifier">
+        <optional>
+          <attribute name="type"/>
+        </optional>
+        <group>
+          <ref name="documentnumber"/>
+          <optional>
+            <ref name="tc-documentnumber"/>
+          </optional>
+        </group>
+      </element>
+    </define>
+    <define name="editorialgroup">
+      <element name="editorialgroup">
+        <ref name="ISOProjectGroup"/>
+      </element>
+    </define>
+  </include>
+  <define name="updates_document_type">
+    <element name="updates-document-type">
+      <ref name="DocumentType"/>
+    </element>
+  </define>
+  <define name="ISOProjectGroup">
+    <zeroOrMore>
+      <ref name="agency"/>
+    </zeroOrMore>
+    <oneOrMore>
+      <ref name="technical-committee"/>
+    </oneOrMore>
+    <zeroOrMore>
+      <ref name="subcommittee"/>
+    </zeroOrMore>
+    <zeroOrMore>
+      <ref name="workgroup"/>
+    </zeroOrMore>
+    <optional>
+      <ref name="secretariat"/>
+    </optional>
+  </define>
+  <define name="approvalgroup">
+    <element name="approvalgroup">
+      <ref name="ISOProjectGroup"/>
+    </element>
+  </define>
+  <define name="agency">
+    <element name="agency">
+      <text/>
+    </element>
+  </define>
+  <define name="horizontal">
+    <element name="horizontal">
+      <data type="boolean"/>
+    </element>
+  </define>
+  <define name="documentnumber">
+    <element name="project-number">
+      <optional>
+        <attribute name="part">
+          <data type="int"/>
+        </attribute>
+      </optional>
+      <optional>
+        <attribute name="subpart">
+          <data type="int"/>
+        </attribute>
+      </optional>
+      <optional>
+        <attribute name="amendment">
+          <data type="int"/>
+        </attribute>
+      </optional>
+      <optional>
+        <attribute name="corrigendum">
+          <data type="int"/>
+        </attribute>
+      </optional>
+      <optional>
+        <attribute name="origyr">
+          <ref name="ISO8601Date"/>
+        </attribute>
+      </optional>
+      <text/>
+    </element>
+  </define>
+  <define name="tc-documentnumber">
+    <element name="tc-document-number">
+      <data type="int"/>
+    </element>
+  </define>
+  <define name="subcommittee">
+    <element name="subcommittee">
+      <ref name="IsoWorkgroup"/>
+    </element>
+  </define>
+  <define name="workgroup">
+    <element name="workgroup">
+      <ref name="IsoWorkgroup"/>
+    </element>
+  </define>
+  <define name="secretariat">
+    <element name="secretariat">
+      <text/>
+    </element>
+  </define>
+  <define name="stagename">
+    <element name="stagename">
+      <optional>
+        <attribute name="abbreviation"/>
+      </optional>
+      <text/>
+    </element>
+  </define>
+  <define name="coverimages">
+    <element name="coverimages">
+      <oneOrMore>
+        <ref name="image"/>
+      </oneOrMore>
+    </element>
+  </define>
+</grammar>

data/lib/relaton_jis/bibliographic_item.rb

@@ -0,0 +1,12 @@
+module RelatonJis
+  class BibliographicItem < RelatonIsoBib::IsoBibliographicItem
+    #
+    # Fetch the flavor shcema version
+    #
+    # @return [String] schema version
+    #
+    def ext_schema
+      @ext_schema ||= schema_versions["relaton-model-jis"]
+    end
+  end
+end

data/lib/relaton_jis/bibliography.rb

@@ -0,0 +1,58 @@
+module RelatonJis
+  module Bibliography
+    extend self
+
+    SOURCE = "https://webdesk.jsa.or.jp/books/W11M".freeze
+
+    #
+    # Search JIS by keyword
+    #
+    # @param [String] code JIS documetnt code
+    # @param [String, nil] year JIS document year
+    #
+    # @return [RelatonJis::HitCollection] search result
+    #
+    def search(code, year = nil)
+      agent = Mechanize.new
+      resp = agent.post "#{SOURCE}0010/searchByKeyword", search_type: "JIS", keyword: code
+      disp = JSON.parse resp.body
+      result = agent.get "#{SOURCE}#{disp['disp_screen']}/index"
+      HitCollection.new code, year, result: result.xpath("//div[@class='blockGenaral']")
+    end
+
+    #
+    # Get JIS document by reference
+    #
+    # @param [String] ref JIS document reference
+    # @param [String] year JIS document year
+    # @param [Hash] _opts options
+    #
+    # @return [RelatonJis::BibliographicItem, nil] JIS document
+    #
+    def get(ref, year = nil, _opts = {})
+      warn "[relaton-jis] (\"#{ref}\") fetching..."
+      result = search(ref, year).find
+      if result[:bib]
+        warn "[relaton-jis] (\"#{ref}\") found #{result[:bib].docidentifier[0].id}"
+        return result[:bib]
+      end
+      hint result, ref, year
+    end
+
+    #
+    # Log hint message
+    #
+    # @param [Hash] result search result
+    # @param [String] ref reference to search
+    # @param [String, nil] year year to search
+    #
+    def hint(result, ref, year)
+      warn "[relaton-jis] (\"#{ref}\") not found. The identifier must be " \
+           "exactly as shown on the webdesk.jsa.or.jp website."
+      if result[:missed_years]
+        warn "[relaton-jis] (\"#{ref}\") TIP: No match for edition year #{year}, " \
+             "but matches exist for #{result[:missed_years].uniq.join(', ')}."
+      end
+    end
+  end
+end

data/lib/relaton_jis/hash_converter.rb

@@ -0,0 +1,9 @@
+module RelatonJis
+  class HashConverter < RelatonBib::HashConverter
+    # @param item_hash [Hash]
+    # @return [Relaton3gpp::BibliographicItem]
+    def self.bib_item(item_hash)
+      BibliographicItem.new(**item_hash)
+    end
+  end
+end

data/lib/relaton_jis/hit.rb

@@ -0,0 +1,48 @@
+module RelatonJis
+  class Hit < RelatonBib::Hit
+    #
+    # Create new hit
+    #
+    # @param [Nokogiri::XML::Element] node found node
+    # @param [RelatonJis::HitCollection] collection hit collection
+    #
+    # @return [RelatonJis::Hit] new hit
+    #
+    def self.create(node, collection)
+      a = node.at("./a")
+      hit = { id: a.at("./text()").text.strip, url: a["href"] }
+      new hit, collection
+    end
+
+    #
+    # Check if hit matches reference
+    #
+    # @param [Hash] ref_parts parts of reference
+    # @param [String, nil] year year
+    #
+    # @return [Boolean] true if hit matches reference
+    #
+    def match?(ref_parts, year = nil) # rubocop:disable Metrics/CyclomaticComplexity, Metrics/AbcSize, Metrics/PerceivedComplexity
+      id_parts[:code].include?(ref_parts[:code]) &&
+        (year.nil? || year == id_parts[:year]) &&
+        ((ref_parts[:expl].nil? || !id_parts[:expl].nil?) &&
+         (ref_parts[:expl_num].nil? || ref_parts[:expl_num] == id_parts[:expl_num])) &&
+        ((ref_parts[:amd].nil? || !id_parts[:amd].nil?) &&
+          (ref_parts[:amd_num].nil? || ref_parts[:amd_num] == id_parts[:amd_num]) &&
+          (ref_parts[:amd_year].nil? || ref_parts[:amd_year] == id_parts[:amd_year]))
+    end
+
+    #
+    # Return parts of document id
+    #
+    # @return [Hash] hash with parts of document id
+    #
+    def id_parts
+      @id_parts ||= hit_collection.parse_ref hit[:id]
+    end
+
+    def fetch
+      @fetch ||= Scraper.new(hit[:url]).fetch
+    end
+  end
+end

data/lib/relaton_jis/hit_collection.rb

@@ -0,0 +1,58 @@
+module RelatonJis
+  class HitCollection < RelatonBib::HitCollection
+    #
+    # Initialize hit collection
+    #
+    # @param [String] text reference
+    # @param [String, nil] year year
+    # @param [Nokogiri::XML::NodeSet] result <description>
+    #
+    def initialize(text, year = nil, result:)
+      super text, year
+      @array = result.map { |h| Hit.create h, self }
+    end
+
+    #
+    # Find hit in collection
+    #
+    # @return [Hash] hash with bib ot array of missed years
+    #
+    def find
+      missed_years = []
+      y = year || ref_parts[:year]
+      @array.each do |hit|
+        return { bib: hit.fetch } if hit.match? ref_parts, y
+
+        missed_years << hit.id_parts[:year] if y && hit.match?(ref_parts)
+      end
+      { missed_years: missed_years }
+    end
+
+    #
+    # Return parts of reference
+    #
+    # @return [Hash] hash with parts of reference
+    #
+    def ref_parts
+      @ref_parts ||= parse_ref text
+    end
+
+    #
+    # Parse reference
+    #
+    # @param [String] ref reference
+    #
+    # @return [Hash] hash with parts of reference
+    #
+    def parse_ref(ref)
+      %r{
+        ^(?<code>\w+\s\w\s?\d+)
+        (?::(?<year>\d{4}))?
+        (?:/(?<expl>EXPL(?:ANATION)?)(?:\s(?<expl_num>\d+))?)?
+        (?:/(?<amd>AMDENDMENT)(?:\s(?<amd_num>\d+)(?::(?<amd_year>\d{4}))?)?)?
+      }x =~ ref
+      { code: code, year: year, expl: expl, expl_num: expl_num, amd: amd,
+        amd_num: amd_num, amd_year: amd_year }
+    end
+  end
+end

data/lib/relaton_jis/processor.rb

@@ -0,0 +1,41 @@
+require "relaton/processor"
+
+module RelatonJis
+  class Processor < Relaton::Processor
+    attr_reader :idtype
+
+    def initialize # rubocop:disable Lint/MissingSuper
+      @short = :relaton_jis
+      @prefix = "JIS"
+      @defaultprefix = %r{^JIS\s}
+      @idtype = "JIS"
+    end
+
+    # @param code [String]
+    # @param date [String, NilClass] year
+    # @param opts [Hash]
+    # @return [RelatonJis::BibliographicItem]
+    def get(code, date, opts)
+      ::RelatonJis::Bibliography.get(code, date, opts)
+    end
+
+    # @param xml [String]
+    # @return [RelatonJis::BibliographicItem]
+    def from_xml(xml)
+      ::RelatonJis::XMLParser.from_xml xml
+    end
+
+    # @param hash [Hash]
+    # @return [RelatonJis::BibliographicItem]
+    def hash_to_bib(hash)
+      item_hash = ::RelatonJis::HashConverter.hash_to_bib(hash)
+      ::RelatonJis::BibliographicItem.new(**item_hash)
+    end
+
+    # Returns hash of XML grammar
+    # @return [String]
+    def grammar_hash
+      @grammar_hash ||= ::RelatonJis.grammar_hash
+    end
+  end
+end

data/lib/relaton_jis/scraper.rb

@@ -0,0 +1,132 @@
+module RelatonJis
+  class Scraper
+    ATTRS = %i[
+      fetched title link abstract docid docnumber date type language script
+      docstatus doctype ics contributor editorialgroup structuredidentifier
+    ].freeze
+
+    LANGS = { "和文" => { lang: "ja", script: "Jpan" },
+              "英訳" => { lang: "en", script: "Latn" } }.freeze
+
+    DATETYPES = { "発行年月日" => "issued", "確認年月日" => "confirmed" }.freeze
+    STATUSES = { "有効" => "valid", "廃止" => "withdrawn" }.freeze
+
+    def initialize(url)
+      @url = url
+      @agent = Mechanize.new
+    end
+
+    def fetch
+      @doc = @agent.get(@url).at "//div[@id='main']/section"
+      attrs = ATTRS.each_with_object({}) do |attr, hash|
+        hash[attr] = send "fetch_#{attr}"
+      end
+      BibliographicItem.new(**attrs)
+    end
+
+    def fetch_fetched
+      Date.today.to_s
+    end
+
+    def fetch_title
+      { "ja" => "Jpan", "en" => "Lant" }.map.with_index do |(lang, script), i|
+        content = @doc.at("./h2/text()[#{i + 2}]").text.strip
+        RelatonBib::TypedTitleString.new content: content, language: lang, script: script
+      end
+    end
+
+    def fetch_link
+      src = RelatonBib::TypedUri.new content: @url, type: "src"
+      uri = URI @url
+      domain = "#{uri.scheme}://#{uri.host}"
+      @doc.xpath("./table/tr[th[.='プレビュー']]/td/a").reduce([src]) do |mem, node|
+        href = "#{domain}#{node[:href]}"
+        mem << RelatonBib::TypedUri.new(content: href, type: "pdf")
+      end
+    end
+
+    def fetch_abstract
+      @doc.xpath("./table/tr[th[.='規格概要']]/td").map do |node|
+        RelatonBib::FormattedString.new content: node.text.strip, language: "ja", script: "Jpan"
+      end
+    end
+
+    def fetch_docid
+      id = @doc.at("./h2/text()[1]").text.strip
+      [RelatonBib::DocumentIdentifier.new(id: id, type: "JIS", primary: true)]
+    end
+
+    def fetch_docnumber
+      match = @doc.at("./h2/text()[1]").text.strip.match(/^\w+\s(\w)\s?(\d+)/)
+      "#{match[1]}#{match[2]}"
+    end
+
+    def fetch_date
+      DATETYPES.each_with_object([]) do |(key, type), a|
+        node = @doc.at("./div/div/div/p/text()[contains(.,'#{key}')]")
+        next unless node
+
+        on = node.text.match(/\d{4}-\d{2}-\d{2}/).to_s
+        a << RelatonBib::BibliographicDate.new(type: type, on: on)
+      end
+    end
+
+    def fetch_type
+      "standard"
+    end
+
+    def fetch_language
+      langs_scripts.map { |l| l[:lang] }
+    end
+
+    def fetch_script
+      langs_scripts.map { |l| l[:script] }
+    end
+
+    def langs_scripts
+      @langs_scripts ||= LANGS.each_with_object([]) do |(key, lang), a|
+        l = @doc.at("./div/div/div[@class='blockContentFile']/div/div/p[1]/span[contains(.,'#{key}')]/following-sibling::span")
+        next if l.nil? || l.text.strip == "-"
+
+        a << lang
+      end
+    end
+
+    def fetch_docstatus
+      st = @doc.at("./div/div/div/p/text()[contains(.,'状態')]/following-sibling::span")
+      return unless st
+
+      RelatonBib::DocumentStatus.new(stage: STATUSES[st.text.strip])
+    end
+
+    def fetch_doctype
+      "standard"
+    end
+
+    def fetch_ics
+      @doc.xpath("./table/tr[th[.='ICS']]/td").map do |node|
+        RelatonIsoBib::Ics.new node.text.strip
+      end
+    end
+
+    def fetch_contributor
+      @doc.xpath("./table/tr[th[.='原案作成団体']]/td").map do |node|
+        name = RelatonBib::LocalizedString.new node.text.strip, "ja", "Jpan"
+        org = RelatonBib::Organization.new name: name
+        RelatonBib::ContributionInfo.new entity: org, role: [type: "author"]
+      end
+    end
+
+    def fetch_editorialgroup
+      node = @doc.at("./table/tr[th[.='原案作成団体']]/td")
+      return unless node
+
+      tc = RelatonBib::WorkGroup.new name: node.text.strip
+      RelatonIsoBib::EditorialGroup.new technical_committee: [tc]
+    end
+
+    def fetch_structuredidentifier
+      RelatonIsoBib::StructuredIdentifier.new project_number: fetch_docnumber, type: "JIS"
+    end
+  end
+end

data/lib/relaton_jis/version.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module RelatonJis
+  VERSION = "1.14.0"
+end

data/lib/relaton_jis/xml_parser.rb

@@ -0,0 +1,9 @@
+module RelatonJis
+  class XMLParser < RelatonBib::XMLParser
+    # @param item_hash [Hash]
+    # @return [RelatonSgpp::BibliographicItem]
+    def self.bib_item(item_hash)
+      BibliographicItem.new(**item_hash)
+    end
+  end
+end

data/lib/relaton_jis.rb

@@ -0,0 +1,25 @@
+# frozen_string_literal: true
+
+require "mechanize"
+require "relaton_iso_bib"
+require_relative "relaton_jis/version"
+require_relative "relaton_jis/bibliographic_item"
+require_relative "relaton_jis/xml_parser"
+require_relative "relaton_jis/hash_converter"
+require_relative "relaton_jis/scraper"
+require_relative "relaton_jis/bibliography"
+require_relative "relaton_jis/hit_collection"
+require_relative "relaton_jis/hit"
+
+module RelatonJis
+  class Error < StandardError; end
+
+  # Returns hash of XML reammar
+  # @return [String]
+  def self.grammar_hash
+    gem_path = File.expand_path "..", __dir__
+    grammars_path = File.join gem_path, "grammars", "*"
+    grammars = Dir[grammars_path].sort.map { |gp| File.read gp }.join
+    Digest::MD5.hexdigest grammars
+  end
+end

data/relaton_jis.gemspec

@@ -0,0 +1,42 @@
+# frozen_string_literal: true
+
+require_relative "lib/relaton_jis/version"
+
+Gem::Specification.new do |spec|
+  spec.name = "relaton-jis"
+  spec.version = RelatonJis::VERSION
+  spec.authors = ["Ribose Inc."]
+  spec.email = ["open.source@ribose.com"]
+
+  spec.summary = "RelatonJis: retrieve IETF Standards for bibliographic " \
+                 "use using the BibliographicItem model"
+  spec.description = "RelatonJis: retrieve IETF Standards for bibliographic " \
+                     "use using the BibliographicItem model"
+  spec.homepage = "https://github.com/relaton/relaton-jis"
+  spec.license = "MIT"
+  spec.required_ruby_version = ">= 2.7.0"
+
+  # spec.metadata["allowed_push_host"] = "TODO: Set to your gem server 'https://example.com'"
+
+  spec.metadata["homepage_uri"] = spec.homepage
+  # spec.metadata["source_code_uri"] = "TODO: Put your gem's public repo URL here."
+  # spec.metadata["changelog_uri"] = "TODO: Put your gem's CHANGELOG.md URL here."
+
+  # Specify which files should be added to the gem when it is released.
+  # The `git ls-files -z` loads the files in the RubyGem that have been added into git.
+  spec.files = Dir.chdir(__dir__) do
+    `git ls-files -z`.split("\x0").reject do |f|
+      (f == __FILE__) || f.match(%r{\A(?:(?:bin|test|spec|features)/|\.(?:git|circleci)|appveyor)})
+    end
+  end
+  spec.bindir = "exe"
+  spec.executables = spec.files.grep(%r{\Aexe/}) { |f| File.basename(f) }
+  spec.require_paths = ["lib"]
+
+  # Uncomment to register a new dependency of your gem
+  spec.add_dependency "mechanize", "~> 2.8.0"
+  spec.add_dependency "relaton-iso-bib", "~> 1.14.0"
+
+  # For more information and examples about making a new gem, check out our
+  # guide at: https://bundler.io/guides/creating_gem.html
+end