relaton-itu 1.14.1 → 1.14.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9c4af93cb7f6ba7f9beff7222f442fb1d0b17ad382308fd492fadbbe7932cd69
-  data.tar.gz: c19c46f3a457914f60cab5992a1443ccf22242ae9ba7c62d141f4de7bdd23fa2
+  metadata.gz: 1bdefbc155c0b9cf2d721888b2512f09913bac76831b421b44444f94dacb4d95
+  data.tar.gz: d5fa79047b6e08c91c8d320c8a95de3a63668a31eac6ae4d958aff3fcb5ce312
 SHA512:
-  metadata.gz: addf42eb6f8631ed0f37fdcfc662edcab0b8877c9316c661b7168e6af64b9e584b23787bf9d5d59372da60f19c7468b378f896279ca6b39d6139c1791f023dc5
-  data.tar.gz: 6789334b80124b7a9bbe33631379a952db3849cd34aa46c2bcbdb922f3f9393e65cb93120cd72b35d3590ab3e0f5b42595c3be815db9ca15042d1bb1e8a42332
+  metadata.gz: 0ae803ee1490114fd1ed254811aa46dc67ac2b9d273f3424018a3e5cf2a34a2a11876f7281184756a9b3ff4b2541bd9aba1eebdf2c39f530f9eacffe4137c0c6
+  data.tar.gz: a18d437f653e9955467dd9f26946cd853c97389835b271029ff721f9c86dfd11bf775bae72cab5ab5ed957467696c3dfbe0debf76670e0499ed6cc03eebcd9ae

data/.github/workflows/rake.yml

@@ -5,6 +5,7 @@ name: rake
 on:
   push:
     branches: [ master, main ]
+    tags: [ v* ]
   pull_request:
 
 jobs:

data/.github/workflows/release.yml

@@ -10,8 +10,9 @@ on:
           Next release version. Possible values: x.y.z, major, minor, patch or pre|rc|etc
         required: true
         default: 'skip'
-  push:
-    tags: [ v* ]
+  repository_dispatch:
+    types: [ do-release ]
+
 
 jobs:
   release:

data/.gitignore

@@ -7,6 +7,7 @@
 /spec/reports/
 /tmp/
 .vscode/
+/vendor/
 .rubocop-https---raw-githubusercontent-com-riboseinc-oss-guides-master-ci-rubocop-yml
 
 # rspec failure tracking

data/.rubocop.yml

@@ -7,6 +7,6 @@ require: rubocop-rails
 inherit_from:
   - https://raw.githubusercontent.com/riboseinc/oss-guides/master/ci/rubocop.yml
 AllCops:
-  TargetRubyVersion: 2.5
+  TargetRubyVersion: 2.7
 Rails:
   Enabled: false

data/Gemfile

@@ -5,7 +5,10 @@ source "https://rubygems.org"
 # Specify your gem's dependencies in relaton_itu.gemspec
 gemspec
 
+gem "equivalent-xml", "~> 0.6"
 gem "pry-byebug"
+gem "rake", "~> 13.0"
+gem "rspec", "~> 3.0"
 gem "ruby-jing"
 gem "simplecov"
 gem "vcr"

data/README.adoc

@@ -69,7 +69,7 @@ item.to_xml bibdata: true
     </bibdata>"
 ----
 
-=== Get code, and year
+=== Get document by code and year
 [source,ruby]
 ----
 RelatonItu::ItuBibliography.get("ITU-T L.163", "2018", {})
@@ -123,7 +123,7 @@ item.link
  #<RelatonBib::TypedUri:0x00007f82d665f0f8 @content=#<Addressable::URI:0xc184 URI:https://www.itu.inthttp//handle.itu.int/11.1002/1000/13786-en?locatt=format:pdf&auth>, @type="obp">]
 ----
 
-=== Create bibliographic item from YAML
+=== Create a bibliographic item from YAML
 [source,ruby]
 ----
 hash = YAML.load_file 'spec/examples/itu_bib_item.yml'
@@ -135,6 +135,25 @@ RelatonItu::ItuBibliographicItem.from_hash hash
 ...
 ----
 
+=== Fetch data
+
+This gem uses the (https://extranet.itu.int/brdocsearch) dataset as a data source. +
+
+The method `RelatonItu::DataFetcher.fetch(output: "data", format: "yaml")` fetches all the documents from the dataset and saves them to the `./data` folder in YAML format.
+Arguments:
+
+- `output` - folder to save documents (default 'data').
+- `format` - the format in which the documents are saved. Possible formats are: `yaml`, `xml`, `bibxxml` (default `yaml`).
+
+[source,ruby]
+----
+RelatonItu::DataFetcher.fetch output: "dir", format: "xml"
+Started at: 2023-05-27 09:21:16 -0400
+Stopped at: 2023-05-27 09:27:45 -0400
+Done in: 390 sec.
+=> nil
+----
+
 == Development
 
 After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake spec` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.

data/lib/relaton_itu/data_fetcher.rb

@@ -0,0 +1,147 @@
+module RelatonItu
+  class DataFetcher
+    def initialize(output, format)
+      @output = output
+      @format = format
+      @ext = format.sub "bibxml", "xml"
+    end
+
+    def files
+      @files ||= []
+    end
+
+    def index
+      @index ||= Relaton::Index.find_or_create :itu, file: "index-v1.yaml"
+    end
+
+    # @return agent [Mechanize]
+    def agent
+      @agent ||= Mechanize.new
+    end
+
+    # @return workers [RelatonBib::WorkersPool]
+    def workers
+      return @workers if @workers
+
+      @workers = RelatonBib::WorkersPool.new 10
+      @workers.tap do |w|
+        w.worker { |row| parse_page(*row) }
+      end
+    end
+
+    # @param url [String]
+    # @param type [String]
+    def parse_page(url, type)
+      doc = agent.get url
+      bib = DataParserR.parse doc, url, type
+      write_file bib
+    rescue => e # rubocop:disable Style/RescueStandardError
+      warn e.message
+      warn e.backtrace
+    end
+
+    def self.fetch(output: "data", format: "yaml")
+      t1 = Time.now
+      puts "Started at: #{t1}"
+      FileUtils.mkdir_p output
+      new(output, format).fetch
+      t2 = Time.now
+      puts "Stopped at: #{t2}"
+      puts "Done in: #{(t2 - t1).round} sec."
+    end
+
+    def fetch
+      fetch_recommendation
+      fetch_question
+      fetch_report
+      fetch_handbook
+      fetch_resolution
+      workers.end
+      workers.result
+      index.save
+    end
+
+    def fetch_recommendation
+      url = "https://extranet.itu.int/brdocsearch/_layouts/15/inplview.aspx?" \
+            "List=%7B0661B581-2413-4E84-BAB2-77E6DB27AF7F%7D&" \
+            "View=%7BC81191DD-48C4-4881-9CB7-FB61C683FE98%7D&" \
+            "ViewCount=123&" \
+            "IsXslView=TRUE&" \
+            "IsCSR=TRUE&" \
+            "ListViewPageUrl=https%3A%2F%2Fextranet.itu.int%2Fbrdocsearch%2FR-REC%2FForms%2Ffolders_inforce.aspx&" \
+            "FolderCTID=0x012001"
+      json_index url, "recommendation"
+    end
+
+    def fetch_question
+      url = "https://extranet.itu.int/brdocsearch/R-QUE/Forms/folders_inforce.aspx"
+      html_index url, "question"
+    end
+
+    def fetch_report
+      url = "https://extranet.itu.int/brdocsearch/_layouts/15/inplview.aspx?" \
+            "List=%7B82E4A13D-C7F3-4844-9E8A-2463C4B7784F%7D&" \
+            "View=%7B94CC1561-E4AC-4317-B402-AA0AADD7F414%7D&" \
+            "ViewCount=407&" \
+            "IsXslView=TRUE&" \
+            "IsCSR=TRUE&" \
+            "ListViewPageUrl=https%3A%2F%2Fextranet.itu.int%2Fbrdocsearch%2FR-REP%2FForms%2FFolders%2520InForce.aspx&" \
+            "FolderCTID=0x012001"
+      json_index url, "technical-report"
+    end
+
+    def fetch_handbook
+      url = "https://extranet.itu.int/brdocsearch/R-HDB/Forms/Folders%20InForce.aspx"
+      html_index url, "handbook"
+    end
+
+    def fetch_resolution
+      url = "https://extranet.itu.int/brdocsearch/R-RES/Forms/Folders%20InForce.aspx"
+      html_index url, "resolution"
+    end
+
+    # #param url [String]
+    # @param type [String]
+    def json_index(url, type) # rubocop:disable Metrics/AbcSize
+      result = agent.post url
+      json = JSON.parse result.body
+      json["Row"].each { |row| workers << [row["serverurl.progid"].sub(/^1/, ""), type] }
+      return unless json["NextHref"]
+
+      nexturl = url.sub(/(Paged|FolderCTID)=.+/, json["NextHref"].match(/(?<=aspx\?).+/).to_s)
+      json_index nexturl, type
+    end
+
+    # #param url [String]
+    # @param type [String]
+    def html_index(url, type)
+      resp = agent.get url
+      result = Nokogiri::HTML resp.body
+      result.xpath("//table//table/tr[position() > 1]").each do |hit|
+        url = hit.at("td/a")[:onclick].match(%r{https://[^']+}).to_s
+        workers << [url, type]
+      end
+    end
+
+    # @param bib [RelatonItu::ItuBibliographicItem]
+    def write_file(bib) # rubocop:disable Metrics/AbcSize
+      id = bib.docidentifier[0].id.gsub(/[\s.]/, "_")
+      file = "#{@output}/#{id}.#{@ext}"
+      if files.include? file
+        warn "File #{file} exists."
+      else
+        files << file
+      end
+      index.add_or_update bib.docidentifier[0].id, file
+      File.write file, content(bib), encoding: "UTF-8"
+    end
+
+    def content(bib)
+      case @format
+      when "yaml" then bib.to_hash.to_yaml
+      when "xml" then bib.to_xml bibdata: true
+      when "bibxml" then bib.to_bibxml
+      end
+    end
+  end
+end

data/lib/relaton_itu/data_parser_r.rb

@@ -0,0 +1,94 @@
+module RelatonItu
+  module DataParserR
+    extend self
+
+    #
+    # Parse ITU-R document.
+    #
+    # @param [Mechanize::Page] doc mechanize page
+    # @param [String] url document url
+    # @param [String] type document type
+    #
+    # @return [RelatonItu::ItuBibliographicItem] bibliographic item
+    #
+    def parse(doc, url, type)
+      RelatonItu::ItuBibliographicItem.new(
+        docid: fetch_docid(doc), title: fetch_title(doc),
+        abstract: fetch_abstract(doc), date: fetch_date(doc), language: ["en"],
+        link: fetch_link(url), script: ["Latn"], docstatus: fetch_status(doc),
+        type: "standard", doctype: type
+      )
+    end
+
+    # @param doc [Mechanize::Page]
+    # @return [Araay<RelatonBib::DocumentIdentifier>]
+    def fetch_docid(doc)
+      # id = doc.at('//h3[.="Number"]/parent::td/following-sibling::td[2]').text # .match(/^[^\s\(]+/).to_s
+      # %r{^(?<id1>[^\s\(\/]+(\/\d+)?)(\/(?<id2>\w+[^\s\(]+))?} =~ id
+      id = doc.at('//div[@id="idDocSetPropertiesWebPart"]/h2').text.match(/^R-\w+-([^-]+(?:-\d{1,3})?)/)[1]
+      [RelatonBib::DocumentIdentifier.new(type: "ITU", id: "ITU-R #{id}", primary: true)]
+      # docid << RelatonBib::DocumentIdentifier.new(type: 'ITU', id: id2) if id2
+      # docid
+    end
+
+    # @param doc [Mechanize::Page]
+    # @return [Araay<RelatonBib::TypedTitleString>]
+    def fetch_title(doc)
+      content = doc.at('//h3[.="Title"]/parent::td/following-sibling::td[2]').text
+      [RelatonBib::TypedTitleString.new(type: "main", content: content, language: "en", script: "Latn")]
+    end
+
+    # @param doc [Mechanize::Page]
+    # @return [Array<RelatonBib::FormattedString>]
+    def fetch_abstract(doc)
+      doc.xpath('//h3[.="Observation"]/parent::td/following-sibling::td[2]').map do |a|
+        c = a.text.strip
+        RelatonBib::FormattedString.new content: c, language: "en", script: "Latn" unless c.empty?
+      end.compact
+    end
+
+    # @param doc [Mechanize::Page]
+    # @return [Araay<RelatonBib::BibliographicDate>]
+    def fetch_date(doc)
+      dates = []
+      date = doc.at('//h3[.="Approval_Date"]/parent::td/following-sibling::td[2]',
+                    '//h3[.="Approval date"]/parent::td/following-sibling::td[2]',
+                    '//h3[.="Approval year"]/parent::td/following-sibling::td[2]')
+      dates << parse_date(date.text, "confirmed") if date
+
+      date = doc.at('//h3[.="Version year"]/parent::td/following-sibling::td[2]')
+      dates << parse_date(date.text, "updated") if date
+      date = doc.at('//div[@id="idDocSetPropertiesWebPart"]/h2').text.match(/(?<=-)(19|20)\d{2}/)
+      dates << parse_date(date.to_s, "published") if date
+      dates
+    end
+
+    # @param date [String]
+    # @param type [String]
+    # @return [RelatonBib::BibliographicDate]
+    def parse_date(date, type)
+      d = case date
+          # when /^\d{4}$/ then date
+          when /(\d{4})(\d{2})/ then "#{$1}-#{$2}"
+          when %r{(\d{1,2})/(\d{1,2})/(\d{4})} then "#{$3}-#{$1}-#{$2}"
+          else date
+          end
+      RelatonBib::BibliographicDate.new(type: type, on: d)
+    end
+
+    # @param url [String]
+    # @return [Array<RelatonBib::TypedUri>]
+    def fetch_link(url)
+      [RelatonBib::TypedUri.new(type: "src", content: url)]
+    end
+
+    # @param doc [Mechanize::Page]
+    # @return [RelatonBib::DocumentStatus, nil]
+    def fetch_status(doc)
+      s = doc.at('//h3[.="Status"]/parent::td/following-sibling::td[2]')
+      return unless s
+
+      RelatonBib::DocumentStatus.new stage: s.text
+    end
+  end
+end

data/lib/relaton_itu/hit.rb

@@ -8,7 +8,7 @@ module RelatonItu
     # Parse page.
     # @return [RelatonItu::ItuBibliographicItem]
     def fetch
-      @fetch ||= Scrapper.parse_page self, hit_collection.gi_imp
+      @fetch ||= Scrapper.parse_page self, imp: hit_collection.gi_imp
     end
   end
 end

data/lib/relaton_itu/hit_collection.rb

@@ -8,6 +8,8 @@ module RelatonItu
   # Page of hit collection.
   class HitCollection < RelatonBib::HitCollection
     DOMAIN = "https://www.itu.int"
+    GH_ITU_R = "https://raw.githubusercontent.com/relaton/relaton-data-itu-r/master/"
+    INDEX_FILE = "index-v1.yaml"
 
     # @return [TrueClass, FalseClass]
     attr_reader :gi_imp
@@ -15,20 +17,22 @@ module RelatonItu
     # @return [Mechanize]
     attr_reader :agent
 
-    # @param ref [String]
-    # @param year [String]
-    def initialize(ref, year = nil) # rubocop:todo Metrics/MethodLength
-      text = ref.sub(/(?<=\.)Imp\s?(?=\d)/, "")
-      super text, year
+    #
+    # @param refid [RelatonItu::Pubid] reference
+    #
+    def initialize(refid) # rubocop:todo Metrics/MethodLength
+      text = refid.to_ref.sub(/(?<=\.)Imp\s?(?=\d)/, "")
+      super text, refid.year
       @agent = Mechanize.new
       agent.user_agent_alias = "Mac Safari"
-      @gi_imp = /\.Imp\d/.match?(ref)
+      @gi_imp = /\.Imp\d/.match?(refid.to_s)
+      @array = []
 
-      case ref
+      case refid.to_ref
       when /^(ITU-T|ITU-R\sRR)/
         request_search
       when /^ITU-R\s/
-        request_document(ref)
+        request_document(refid)
       end
     end
 
@@ -41,17 +45,17 @@ module RelatonItu
       @array = hits JSON.parse(resp.body)
     end
 
-    # @param ref [String] a document ref
-    def request_document(ref) # rubocop:todo Metrics/MethodLength
-      uri = URI::HTTPS.build(
-        host: "raw.githubusercontent.com",
-        path: "/relaton/relaton-data-itu-r/master/data/#{ref.gsub(/[\s.]/, '_')}.yaml",
-      )
+    #
+    # @param refid [RelatonItu::Pubid] a document reference
+    #
+    def request_document(refid) # rubocop:todo Metrics/MethodLength, Metrics/AbcSize
+      index = Relaton::Index.find_or_create :itu, url: "#{GH_ITU_R}index-v1.zip", file: INDEX_FILE
+      row = index.search(refid.to_ref).min_by { |i| i[:id] }
+      return unless row
+
+      uri = URI("#{GH_ITU_R}#{row[:file]}")
       resp = Net::HTTP.get_response(uri)
-      if resp.code == "404"
-        @array = []
-        return
-      end
+      return if resp.code == "404"
 
       hash = YAML.safe_load resp.body
       item_hash = HashConverter.hash_to_bib(hash)

data/lib/relaton_itu/itu_bibliography.rb

@@ -14,134 +14,139 @@ require "date"
 
 module RelatonItu
   # Class methods for search ISO standards.
-  class ItuBibliography
-    class << self
-      # @param text [String]
-      # @return [RelatonItu::HitCollection]
-      def search(text, year = nil)
-        # code = text.sub(/(?<=ITU-T\s\w)\.(\w+\.)(?=\d+)/, ' \1')
-        if text =~ /(ITU-T\s\w)\.(Suppl\.|Annex)\s?(\w?\d+)/
-          correct_ref = "#{$~[1]} #{$~[2]} #{$~[3]}"
-          warn "[relaton-itu] WARNING: Incorrect reference #{text}"
-          warn "[relaton-itu] the reference should be #{correct_ref}"
-        end
-        HitCollection.new text, year
-      rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET,
-             EOFError, Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError,
-             Net::ProtocolError, URI::InvalidURIError => e
-        raise RelatonBib::RequestError, e.message
+  module ItuBibliography
+    extend self
+
+    # @param refid [RelatonItu::Pubid, String] a document reference
+    # @return [RelatonItu::HitCollection]
+    #
+    def search(refid)
+      refid = RelatonItu::Pubid.parse refid if refid.is_a? String
+      if refid.to_ref =~ /(ITU[\s-]T\s\w)\.(Suppl\.|Annex)\s?(\w?\d+)/
+        correct_ref = "#{$~[1]} #{$~[2]} #{$~[3]}"
+        warn "[relaton-itu] WARNING: Incorrect reference #{refid}"
+        warn "[relaton-itu] the reference should be #{correct_ref}"
       end
+      HitCollection.new refid
+    rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET, EOFError, Net::HTTPBadResponse,
+           Net::HTTPHeaderSyntaxError, Net::ProtocolError, URI::InvalidURIError => e
+      raise RelatonBib::RequestError, e.message
+    end
 
-      # @param code [String] the ISO standard Code to look up (e..g "ISO 9000")
-      # @param year [String] the year the standard was published (optional)
-      # @param opts [Hash] options; restricted to :all_parts if all-parts
-      #   reference is required
-      # @return [String] Relaton XML serialisation of reference
-      def get(code, year = nil, opts = {}) # rubocop:disable Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-        if year.nil?
-          /^(?<code1>[^\s]+\s[^\s]+)\s\((?:\d{2}\/)?(?<year1>\d+)\)$/ =~ code
-          unless code1.nil?
-            code = code1
-            year = year1
-          end
-        end
+    # def transform_ref(ref)
+    #   ref.sub(/^ITU[\s-](\w)[\s-](?:REC[\s-])?/, 'ITU-\1')
+    # end
 
-        code += "-1" if opts[:all_parts]
-        ret = itubib_get1(code, year, opts)
-        return nil if ret.nil?
+    # @param code [String] the ISO standard Code to look up (e..g "ISO 9000")
+    # @param year [String] the year the standard was published (optional)
+    # @param opts [Hash] options; restricted to :all_parts if all-parts
+    #   reference is required
+    # @return [String] Relaton XML serialisation of reference
+    def get(code, year = nil, opts = {}) # rubocop:disable Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+      refid = Pubid.parse code
+      refid.year ||= year
+      # if year.nil?
+      #   /^(?<code1>[^\s]+\s[^\s]+)\s\((?:\d{2}\/)?(?<year1>\d+)\)$/ =~ code
+      #   unless code1.nil?
+      #     code = code1
+      #     year = year1
+      #   end
+      # end
 
-        ret = ret.to_most_recent_reference unless year || opts[:keep_year]
-        ret = ret.to_all_parts if opts[:all_parts]
-        ret
-      end
+      ret = itubib_get1(refid)
+      return nil if ret.nil?
 
-      private
+      ret = ret.to_most_recent_reference unless refid.year || opts[:keep_year]
+      ret = ret.to_all_parts if opts[:all_parts]
+      ret
+    end
 
-      def fetch_ref_err(code, year, missed_years) # rubocop:disable Metrics/MethodLength
-        id = year ? "#{code}:#{year}" : code
-        warn "[relaton-itu] WARNING: no match found online for #{id}. " \
-             "The code must be exactly like it is on the standards website."
-        unless missed_years.empty?
-          warn "[relaton-itu] (There was no match for #{year}, though there " \
-               "were matches found for #{missed_years.join(', ')}.)"
-        end
-        if /\d-\d/.match? code
-          warn "[relaton-itu] The provided document part may not exist, or " \
-               "the document may no longer be published in parts."
-        else
-          warn "[relaton-itu] If you wanted to cite all document parts for the reference, " \
-               "use \"#{code} (all parts)\".\nIf the document is not a standard, " \
-               "use its document type abbreviation (TS, TR, PAS, Guide)."
-        end
-        nil
+    private
+
+    def fetch_ref_err(refid, missed_years) # rubocop:disable Metrics/MethodLength
+      warn "[relaton-itu] WARNING: no match found online for #{refid}. " \
+           "The code must be exactly like it is on the standards website."
+      unless missed_years.empty?
+        warn "[relaton-itu] (There was no match for #{refid.year}, though there " \
+             "were matches found for #{missed_years.join(', ')}.)"
       end
+      # if /\d-\d/.match? refid.code
+      #   warn "[relaton-itu] The provided document part may not exist, or " \
+      #        "the document may no longer be published in parts."
+      # else
+      warn "[relaton-itu] If you wanted to cite all document parts for the reference, " \
+           "use \"#{refid} (all parts)\".\nIf the document is not a standard, " \
+           "use its document type abbreviation (TS, TR, PAS, Guide)."
+      # end
+    end
 
-      def search_filter(code, year) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-        %r{
-          ^(?<pref1>ITU)?(?:-(?<type1>\w))?\s?(?<code1>[^\s/]+(?:/\w[.\d]+)?)
-          (?:\s\(?(?<ver1>v\d+)\)?)?
-          (?:\s\((?:(?<month1>\d{2})/)?(?<year1>\d{4})\))?
-          (?:\s-\s(?<buldate1>\d{2}\.\w{1,4}\.\d{4}))?
-          (?:\s(?<corr1>(?:Amd|Cor|Amendment|Corrigendum)\.?\s?\d+))?
-          (?:\s\((?:(?<cormonth1>\d{2})/)?(?<coryear1>\d{4})\))?
-        }x =~ code
-        year ||= year1
-        warn "[relaton-itu] (\"#{code}\") fetching..."
-        result = search(code)
-        code1.sub!(/(?<=\.)Imp(?=\d)/, "") if result.gi_imp
-        if corr1
-          corr1.sub!(/[.\s]+/, " ").sub!("Amendment", "Amd")
-          corr1.sub!("Corrigendum", "Corr")
-        end
-        result.select do |i|
-          next true unless i.hit[:code]
+    def search_filter(refid) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+      # %r{
+      #   ^(?<pref1>ITU)?(?:-(?<type1>\w))?\s?(?<code1>[^\s/]+(?:/\w[.\d]+)?)
+      #   (?:\s\(?(?<ver1>v\d+)\)?)?
+      #   (?:\s\((?:(?<month1>\d{2})/)?(?<year1>\d{4})\))?
+      #   (?:\s-\s(?<buldate1>\d{2}\.\w{1,4}\.\d{4}))?
+      #   (?:\s(?<corr1>(?:Amd|Cor|Amendment|Corrigendum)\.?\s?\d+))?
+      #   (?:\s\((?:(?<cormonth1>\d{2})/)?(?<coryear1>\d{4})\))?
+      # }x =~ code
+      # year ||= year1
+      warn "[relaton-itu] (\"#{refid}\") fetching..."
+      result = search(refid)
+      # refid.code.sub!(/(?<=\.)Imp(?=\d)/, "") if result.gi_imp
+      # if corr1
+      #   corr1.sub!(/[.\s]+/, " ").sub!("Amendment", "Amd")
+      #   corr1.sub!("Corrigendum", "Corr")
+      # end
+      result.select do |i|
+        next true unless i.hit[:code]
 
-          %r{
-            ^(?<pref2>ITU)?(?:-(?<type2>\w))?\s?(?<code2>\S+)
-            (?:\s\(?(?<ver2>v\d+)\)?)?
-            (?:\s\((?:(?<month2>\d{2})/)?(?<year2>\d{4})\))?
-            (?:\s(?<corr2>(?:Amd|Cor)\.\s?\d+))?
-            (?:\s\((?:(?<cormonth2>\d{2})/)?(?<coryear2>\d{4})\))?
-          }x =~ i.hit[:code]
-          /:[^(]+\((?<buldate2>\d{2}\.\w{1,4}\.\d{4})\)/ =~ i.hit[:title]
-          corr2&.sub!(/\.\s?/, " ")
-          pref1 == pref2 && (!type1 || type1 == type2) && code2.include?(code1) &&
-            (!year || year == year2) && (!month1 || month1 == month2) &&
-            corr1 == corr2 && (!coryear1 || coryear1 == coryear2) &&
-            buldate1 == buldate2 && (!cormonth1 || cormonth1 == cormonth2) &&
-            (!ver1 || ver1 == ver2)
-        end
+        pubid = Pubid.parse i.hit[:code]
+        refid === pubid
+        # %r{
+        #   ^(?<pref2>ITU)?(?:-(?<type2>\w))?\s?(?<code2>\S+)
+        #   (?:\s\(?(?<ver2>v\d+)\)?)?
+        #   (?:\s\((?:(?<month2>\d{2})/)?(?<year2>\d{4})\))?
+        #   (?:\s(?<corr2>(?:Amd|Cor)\.\s?\d+))?
+        #   (?:\s\((?:(?<cormonth2>\d{2})/)?(?<coryear2>\d{4})\))?
+        # }x =~ i.hit[:code]
+        # /:[^(]+\((?<buldate2>\d{2}\.\w{1,4}\.\d{4})\)/ =~ i.hit[:title]
+        # corr2&.sub!(/\.\s?/, " ")
+        # pref1 == pref2 && (!type1 || type1 == type2) && code2.include?(code1) &&
+        #   (!year || year == year2) && (!month1 || month1 == month2) &&
+        #   corr1 == corr2 && (!coryear1 || coryear1 == coryear2) &&
+        #   buldate1 == buldate2 && (!cormonth1 || cormonth1 == cormonth2) &&
+        #   (!ver1 || ver1 == ver2)
       end
+    end
 
-      # Sort through the results from Isobib, fetching them three at a time,
-      # and return the first result that matches the code,
-      # matches the year (if provided), and which # has a title (amendments do not).
-      # Only expects the first page of results to be populated.
-      # Does not match corrigenda etc (e.g. ISO 3166-1:2006/Cor 1:2007)
-      # If no match, returns any years which caused mismatch, for error reporting
-      def isobib_results_filter(result, year)
-        missed_years = []
-        result.each do |r|
-          /\((?:\d{2}\/)?(?<pyear>\d{4})\)/ =~ r.hit[:code]
-          if !year || year == pyear
-            ret = r.fetch
-            return { ret: ret } if ret
-          end
-
-          missed_years << pyear
+    # Sort through the results from Isobib, fetching them three at a time,
+    # and return the first result that matches the code,
+    # matches the year (if provided), and which # has a title (amendments do not).
+    # Only expects the first page of results to be populated.
+    # Does not match corrigenda etc (e.g. ISO 3166-1:2006/Cor 1:2007)
+    # If no match, returns any years which caused mismatch, for error reporting
+    def isobib_results_filter(result, refid)
+      missed_years = []
+      result.each do |r|
+        /\((?:\d{2}\/)?(?<pyear>\d{4})\)/ =~ r.hit[:code]
+        if !refid.year || refid.year == pyear
+          ret = r.fetch
+          return { ret: ret } if ret
         end
-        { years: missed_years }
+
+        missed_years << pyear
       end
+      { years: missed_years }
+    end
 
-      def itubib_get1(code, year, _opts)
-        result = search_filter(code, year) || return
-        ret = isobib_results_filter(result, year)
-        if ret[:ret]
-          warn "[relaton-itu] (\"#{code}\") found #{ret[:ret].docidentifier.first&.id}"
-          ret[:ret]
-        else
-          fetch_ref_err(code, year, ret[:years])
-        end
+    def itubib_get1(refid)
+      result = search_filter(refid) || return
+      ret = isobib_results_filter(result, refid)
+      if ret[:ret]
+        warn "[relaton-itu] (\"#{refid}\") found #{ret[:ret].docidentifier.first&.id}"
+        ret[:ret]
+      else
+        fetch_ref_err(refid, ret[:years])
       end
     end
   end

data/lib/relaton_itu/processor.rb

@@ -7,6 +7,7 @@ module RelatonItu
       @prefix = "ITU"
       @defaultprefix = %r{^ITU\s}
       @idtype = "ITU"
+      @datasets = %w[itu-r]
     end
 
     # @param code [String]
@@ -17,6 +18,18 @@ module RelatonItu
       ::RelatonItu::ItuBibliography.get(code, date, opts)
     end
 
+    #
+    # Fetch all the documents from https://extranet.itu.int/brdocsearch/
+    #
+    # @param [String] source source name (itu-r)
+    # @param [Hash] opts
+    # @option opts [String] :output directory to output documents, default is data
+    # @option opts [String] :format output format, default is yaml
+    #
+    def fetch_data(_source, opts)
+      DataFetcher.fetch(**opts)
+    end
+
     # @param xml [String]
     # @return [RelatonItu::ItuBibliographicItem]
     def from_xml(xml)
@@ -34,5 +47,12 @@ module RelatonItu
     def grammar_hash
       @grammar_hash ||= ::RelatonItu.grammar_hash
     end
+
+    #
+    # Remove index file
+    #
+    def remove_index_file
+      Relaton::Index.find_or_create(:itu, url: true, file: HitCollection::INDEX_FILE).remove_file
+    end
   end
 end

data/lib/relaton_itu/pubid.rb

@@ -0,0 +1,142 @@
+module RelatonItu
+  class Pubid
+    class Parser < Parslet::Parser
+      rule(:dash) { str("-") }
+      rule(:dot) { str(".") }
+      rule(:dot?) { dot.maybe }
+      rule(:separator) { match['\s-'] }
+      rule(:space) { match("\s") }
+      rule(:num) { match["0-9"] }
+
+      rule(:prefix) { str("ITU").as(:prefix) }
+      rule(:sector) { separator >> match("[A-Z]").as(:sector) }
+      rule(:type) { separator >> str("REC").as(:type) }
+      rule(:type?) { type.maybe }
+      rule(:code) { separator >> (match["A-Z0-9"].repeat(1) >> match["[:alnum:]/.-"].repeat).as(:code) }
+      rule(:year) { (match["12"] >> num.repeat(3, 3)).as(:year) }
+
+      rule(:month1) { num.repeat(2, 2).as(:month) }
+      rule(:date1) { str(" (") >> (month1 >> str("/")).maybe >> year >> str(")") }
+      rule(:month2) { match["IVX"].repeat(1, 3).as(:month) }
+      rule(:date2) { str(" - ") >> num.repeat(2, 2) >> dot >> month2 >> dot >> year }
+      rule(:date) { date1 | date2 }
+      rule(:date?) { date.maybe }
+
+      rule(:amd) { space >> (str("Amd") | str("Amendment")) >> dot? >> space >> num.repeat(1, 2).as(:amd) }
+      rule(:amd?) { amd.maybe }
+
+      rule(:sup) { space >> str("Suppl") >> dot? >> space >> num.repeat(1, 2).as(:suppl) }
+      rule(:sup?) { sup.maybe }
+
+      rule(:annex) { space >> str("Annex") >> space >> match["[:alnum:]"].repeat(1, 2).as(:annex) }
+      rule(:annex?) { annex.maybe }
+
+      rule(:itu_pubid) { prefix >> sector >> type? >> code >> sup? >> annex? >> date? >> amd? >> any.repeat }
+      root(:itu_pubid)
+    end
+
+    attr_accessor :prefix, :sector, :type, :code, :suppl, :annex, :year, :month, :amd
+
+    #
+    # Create a new ITU publication identifier.
+    #
+    # @param [String] prefix
+    # @param [String] sector
+    # @param [String, nil] type
+    # @param [String] code
+    # @param [String, nil] suppl number
+    # @param [String, nil] year
+    # @param [String, nil] month
+    # @param [String, nil] amd amendment number
+    #
+    def initialize(prefix:, sector:, code:, **args)
+      @prefix = prefix
+      @sector = sector
+      @type = args[:type]
+      @code, year, month = date_from_code code
+      @suppl = args[:suppl]
+      @annex = args[:annex]
+      @year = args[:year] || year
+      @month = roman_to_2digit args[:month] || month
+      @amd = args[:amd]
+    end
+
+    def self.parse(id)
+      id_parts = Parser.new.parse(id).to_h.transform_values(&:to_s)
+      new(**id_parts)
+    rescue Parslet::ParseFailed => e
+      warn "[relaton-itu] WARNING: #{id} is invalid ITU publication identifier"
+      warn e.parse_failure_cause.ascii_tree
+      raise e
+    end
+
+    def to_h(with_type: true) # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+      hash = { prefix: prefix, sector: sector, code: code }
+      hash[:type] = type if type && with_type
+      hash[:suppl] = suppl if suppl
+      hash[:annex] = annex if annex
+      hash[:year] = year if year
+      hash[:month] = month if month
+      hash[:amd] = amd if amd
+      hash
+    end
+
+    def to_ref
+      to_s ref: true
+    end
+
+    def to_s(ref: false) # rubocop:disable Metrics/AbcSize
+      s = "#{prefix}-#{sector}"
+      s << " #{type}" if type && !ref
+      s << " #{code}"
+      s << " Suppl. #{suppl}" if suppl
+      s << " Annex #{annex}" if annex
+      s << date_to_s
+      s << " Amd #{amd}" if amd
+      s
+    end
+
+    def ===(other)
+      hash = to_h with_type: false
+      other_hash = other.to_h with_type: false
+      hash.delete(:month)
+      other_hash.delete(:month)
+      other_hash.delete(:year) unless hash[:year]
+      hash == other_hash
+    end
+
+    private
+
+    def date_from_code(code)
+      /(?<cod>.+?)-(?<date>\d{6})(?:-I|$)/ =~ code
+      return [code, nil, nil] unless cod && date
+
+      [cod, date[0..3], date[4..5]]
+    end
+
+    def roman_to_2digit(num) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
+      return unless num
+
+      roman_nums = { "I" => 1, "V" => 5, "X" => 10 }
+      last = roman_nums[num[-1]]
+      return num unless last
+
+      return roman_nums[num].to_s.rjust(2, "0") if num.size == 1
+
+      num.chars.each_cons(2).reduce(last) do |acc, (a, b)|
+        if roman_nums[a] < roman_nums[b]
+          acc - roman_nums[a]
+        else
+          acc + roman_nums[a]
+        end
+      end.to_s.rjust(2, "0")
+    end
+
+    def date_to_s
+      if month && year then " (#{month}/#{year})"
+      elsif year then " (#{year})"
+      else ""
+      end
+    end
+  end
+end

data/lib/relaton_itu/scrapper.rb

@@ -27,7 +27,7 @@ module RelatonItu
       # Parse page.
       # @param hit [RelatonItu::Hit]
       # @return [Hash]
-      def parse_page(hit, imp = false) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
+      def parse_page(hit, imp: false) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
         doc = get_page hit
         return unless doc.code == "200"
 

data/lib/relaton_itu/version.rb

@@ -1,3 +1,3 @@
 module RelatonItu
-  VERSION = "1.14.1".freeze
+  VERSION = "1.14.3".freeze
 end

data/lib/relaton_itu.rb

@@ -1,7 +1,12 @@
 require "mechanize"
+require "parslet"
+require "digest/md5"
+require "relaton/index"
 require "relaton_itu/version"
+require "relaton_itu/pubid"
 require "relaton_itu/itu_bibliography"
-require "digest/md5"
+require "relaton_itu/data_fetcher"
+require "relaton_itu/data_parser_r"
 
 module RelatonItu
   class Error < StandardError; end

data/relaton-itu.gemspec

@@ -1,5 +1,4 @@
-
-lib = File.expand_path("../lib", __FILE__)
+lib = File.expand_path("lib", __dir__)
 $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 require "relaton_itu/version"
 
@@ -9,10 +8,10 @@ Gem::Specification.new do |spec|
   spec.authors       = ["Ribose Inc."]
   spec.email         = ["open.source@ribose.com"]
 
-  spec.summary       = "RelatonItu: retrieve ITU Standards for bibliographic use "\
-                       "using the BibliographicItem model"
-  spec.description   = "RelatonItu: retrieve ITU Standards for bibliographic use "\
-                       "using the BibliographicItem model"
+  spec.summary       = "RelatonItu: retrieve ITU Standards for bibliographic " \
+                       "use using the BibliographicItem model"
+  spec.description   = "RelatonItu: retrieve ITU Standards for bibliographic " \
+                       "use using the BibliographicItem model"
   spec.homepage      = "https://github.com/metanorma/relaton-itu"
   spec.license       = "MIT"
 
@@ -24,12 +23,10 @@ Gem::Specification.new do |spec|
   spec.bindir        = "exe"
   spec.executables   = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]
-  spec.required_ruby_version = Gem::Requirement.new(">= 2.5.0")
-
-  spec.add_development_dependency "equivalent-xml", "~> 0.6"
-  spec.add_development_dependency "rake", "~> 13.0"
-  spec.add_development_dependency "rspec", "~> 3.0"
+  spec.required_ruby_version = Gem::Requirement.new(">= 2.7.0")
 
   spec.add_dependency "mechanize", "~> 2.8.0"
+  spec.add_dependency "parslet", "~> 2.0.0"
   spec.add_dependency "relaton-bib", "~> 1.14.0"
+  spec.add_dependency "relaton-index", "~> 0.2.0"
 end

metadata

@@ -1,85 +1,71 @@
 --- !ruby/object:Gem::Specification
 name: relaton-itu
 version: !ruby/object:Gem::Version
-  version: 1.14.1
+  version: 1.14.3
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-04-09 00:00:00.000000000 Z
+date: 2023-06-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: equivalent-xml
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.6'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.6'
-- !ruby/object:Gem::Dependency
-  name: rake
+  name: mechanize
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '13.0'
-  type: :development
+        version: 2.8.0
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '13.0'
+        version: 2.8.0
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: parslet
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
-  type: :development
+        version: 2.0.0
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: 2.0.0
 - !ruby/object:Gem::Dependency
-  name: mechanize
+  name: relaton-bib
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.8.0
+        version: 1.14.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.8.0
+        version: 1.14.0
 - !ruby/object:Gem::Dependency
-  name: relaton-bib
+  name: relaton-index
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.14.0
+        version: 0.2.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.14.0
+        version: 0.2.0
 description: 'RelatonItu: retrieve ITU Standards for bibliographic use using the BibliographicItem
   model'
 email:
@@ -106,6 +92,8 @@ files:
 - grammars/relaton-itu-compile.rng
 - grammars/relaton-itu.rng
 - lib/relaton_itu.rb
+- lib/relaton_itu/data_fetcher.rb
+- lib/relaton_itu/data_parser_r.rb
 - lib/relaton_itu/editorial_group.rb
 - lib/relaton_itu/hash_converter.rb
 - lib/relaton_itu/hit.rb
@@ -114,6 +102,7 @@ files:
 - lib/relaton_itu/itu_bibliography.rb
 - lib/relaton_itu/itu_group.rb
 - lib/relaton_itu/processor.rb
+- lib/relaton_itu/pubid.rb
 - lib/relaton_itu/scrapper.rb
 - lib/relaton_itu/structured_identifier.rb
 - lib/relaton_itu/version.rb
@@ -131,14 +120,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 2.5.0
+      version: 2.7.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.6
+rubygems_version: 3.3.26
 signing_key: 
 specification_version: 4
 summary: 'RelatonItu: retrieve ITU Standards for bibliographic use using the BibliographicItem