relaton-ietf 1.9.4 → 1.9.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 639209bd0a93d43f7d9d671d709a40e635ef7978ac33b502afde91b64e03c857
-  data.tar.gz: d33475e9becda05d7eadc6b9e48e3df700aa4620dc6081811708faddd84b650f
+  metadata.gz: 9c56a40e6b16ec11c29b7b10ade3c09b071f4eddc84402d8d22909af910af8fa
+  data.tar.gz: 619f416b504b40dc9515cc2ca05c2bd3abffc2fc800eb473adfd2366bb63e068
 SHA512:
-  metadata.gz: 87f20d0e95cc16aa9462b78cf67eef1ca40e0124bda85ec2a480de960b44d33ef333083ce25fdd2fff6e109b071fdbf2cf27226c210eef65331b1786daa59f85
-  data.tar.gz: 3865277ac823c1fe10731a813ca5085d9e3ced27bb5af6e1bb53cb9a7c952c76e47f249cbb6cc013293c5ba2f46410f309689be307e55e1fc1aff4d508219d88
+  metadata.gz: 27fe700492c5e5d7b4f437a281912c2719cc18ec9656c85db6289a978847043f5e52ba011d2e1801304744166ce9091ba24ce5ec00005d5edde09d4d10a0cd8c
+  data.tar.gz: 94919d5a441b27673e10315d5f5cd8670515de252b0e925fb431d9698bfd8e5a40e719684ad61019e93471eb0add922743a6abd98abafd2402b18de764ca09a5

data/README.adoc

@@ -149,16 +149,6 @@ Arguments:
 - `output` - folder to save documents (default './data').
 - `format` - format in which the documents are saved. Possimle formats are: `yaml`, `xml`, `bibxml` (default `yaml`).
 
-For `ietf-rfcsubseries` dataset only special XML format is supported:
-
-[sourse.xml]
-----
-<referencegroup anchor="BCP14" target="https://www.rfc-editor.org/info/bcp14">
-  <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.2119.xml" />
-  <xi:include href="https://www.rfc-editor.org/refs/bibxml/reference.RFC.8174.xml" />
-</referencegroup>
-----
-
 [source,ruby]
 ----
 RelatonIetf::DataFetcher.fetch "ietf-internet-drafts"

data/lib/relaton/provider_ietf.rb

@@ -7,7 +7,7 @@ module Relaton
         def from_rfcxml(xml)
           doc = Nokogiri::XML xml
           reference = doc.at "/rfc"
-          RelatonIetf::Scrapper.fetch_rfc reference
+          RelatonIetf::BibXMLParser.fetch_rfc reference
         end
       end
     end

data/lib/relaton_ietf/bibxml_parser.rb

@@ -0,0 +1,31 @@
+module RelatonIetf
+  module BibXMLParser
+    include RelatonBib::BibXMLParser
+    extend BibXMLParser
+
+    FLAVOR = "IETF".freeze
+
+    # @param attrs [Hash]
+    # @return [RelatonIetf::IetfBibliographicItem]
+    def bib_item(**attrs)
+      unless attrs.delete(:is_relation)
+        attrs[:fetched] = Date.today.to_s
+        attrs[:place] = ["Fremont, CA"]
+      end
+      RelatonIetf::IetfBibliographicItem.new(**attrs)
+    end
+
+    # @param [RelatonBib::WorkGroup]
+    # @return [RelatonIetf::Committee]
+    def committee(wgr)
+      Committee.new wgr
+    end
+
+    # @param reference [Nokogiri::XML::Element]
+    # @return [Array<Hash>]
+    def contributors(reference)
+      [{ entity: new_org("Internet Engineering Task Force", "IETF"), role: [type: "publisher"] }] + super
+      # persons(reference) + organizations(reference)
+    end
+  end
+end

data/lib/relaton_ietf/data_fetcher.rb

@@ -17,7 +17,7 @@ module RelatonIetf
     def initialize(source, output, format)
       @source = source
       @output = output
-      @format = source == "ietf-rfcsubseries" ? "rfcxml" : format
+      @format = format
       @ext = @format.sub(/^bib|^rfc/, "")
       @files = []
     end
@@ -99,7 +99,6 @@ module RelatonIetf
       c = case @format
           when "xml" then entry.to_xml(bibdata: true)
           when "yaml" then entry.to_hash.to_yaml
-          when "rfcxml" then entry.to_xml
           else entry.send("to_#{@format}")
           end
       file = file_name entry
@@ -120,9 +119,12 @@ module RelatonIetf
     #
     def file_name(entry)
       id = if entry.respond_to? :docidentifier
-             entry.docidentifier.detect { |i| i.type == "Internet-Draft" }&.id&.downcase
+             entry.docidentifier.detect { |i| i.type == "Internet-Draft" }&.id
            end
-      id ||= entry.docnumber.upcase
+      id ||= entry.docnumber
+      if @source == "ietf-internet-drafts" then id.downcase!
+      else id.upcase!
+      end
       name = id.gsub(/[\s,:\/]/, "_").squeeze("_")
       File.join @output, "#{name}.#{@ext}"
     end

data/lib/relaton_ietf/processor.rb

@@ -6,7 +6,7 @@ module RelatonIetf
     def initialize # rubocop:disable Lint/MissingSuper
       @short = :relaton_ietf
       @prefix = "IETF"
-      @defaultprefix = /^RFC /
+      @defaultprefix = /^(IETF|RFC|BCP)\s/
       @idtype = "IETF"
       @datasets = %w[ietf-rfcsubseries ietf-internet-drafts ietf-rfc-entries]
     end

data/lib/relaton_ietf/rfc_index_entry.rb

@@ -3,8 +3,9 @@ module RelatonIetf
     #
     # Document parser initalization
     #
+    # @param [String] name document type name
     # @param [String] doc_id document id
-    # @param [Array<String>] is_also also document ids
+    # @param [Array<String>] is_also included document ids
     #
     def initialize(name, doc_id, is_also)
       @name = name
@@ -13,15 +14,6 @@ module RelatonIetf
       @is_also = is_also
     end
 
-    #
-    # Document id
-    #
-    # @return [Strinng] document id
-    #
-    def docnumber
-      @doc_id
-    end
-
     #
     # Initialize document parser and run it
     #
@@ -35,26 +27,61 @@ module RelatonIetf
       return unless doc_id && is_also.any?
 
       name = doc.name.split("-").first
-      new(name, doc_id.text, is_also)
+      new(name, doc_id.text, is_also).parse
+    end
+
+    def parse
+      IetfBibliographicItem.new(
+        docnumber: docnumber,
+        type: "standard",
+        docid: parse_docid,
+        language: ["en"],
+        script: ["Latn"],
+        link: parse_link,
+        formattedref: formattedref,
+        relation: parse_relation,
+      )
     end
 
     #
-    # Render document as XML
+    # Document id
     #
-    # @return [String] XML
+    # @return [Strinng] document id
     #
-    def to_xml # rubocop:disable Metrics/MethodLength
-      Nokogiri::XML::Builder.new do |xml|
-        anchor = "#{@name.upcase}#{@shortnum}"
-        url = "https://www.rfc-editor.org/info/#{@name}#{@shortnum}"
-        xml.referencegroup("xmlns:xi" => "http://www.w3.org/2001/XInclude",
-                           anchor: anchor, target: url) do
-          @is_also.each do |did|
-            num = did[-4..-1]
-            xml["xi"].send("include", href: "https://www.rfc-editor.org/refs/bibxml/reference.RFC.#{num}.xml")
-          end
-        end
-      end.doc.root.to_xml
+    def docnumber
+      @doc_id
+    end
+
+    def parse_docid
+      [
+        RelatonBib::DocumentIdentifier.new(type: "IETF", scope: "rfc-anchor", id: anchor),
+        # RelatonBib::DocumentIdentifier.new(type: "rfc-anchor", id: anchor),
+      ]
+    end
+
+    # def pub_id
+    #   "IETF #{@name.upcase} #{@shortnum}"
+    # end
+
+    def anchor
+      "#{@name.upcase}#{@shortnum}"
+    end
+
+    def parse_link
+      [RelatonBib::TypedUri.new(type: "src", content: "https://www.rfc-editor.org/info/#{@name}#{@shortnum}")]
+    end
+
+    def formattedref
+      RelatonBib::FormattedRef.new(
+        content: anchor, language: "en", script: "Latn",
+      )
+    end
+
+    def parse_relation
+      @is_also.map do |ref|
+        bib = IetfBibliography.get ref.sub(/^(RFC)(\d+)/, '\1 \2')
+        { type: "includes", bibitem: bib }
+      end
     end
   end
 end

data/lib/relaton_ietf/scrapper.rb

@@ -1,17 +1,10 @@
 # frozen_string_literal: true
 
-require "net/http"
-require "relaton_bib"
-require "relaton_ietf/ietf_bibliographic_item"
-
 module RelatonIetf
   # Scrapper module
   module Scrapper
-    extend RelatonBib::BibXMLParser
     extend Scrapper
 
-    FLAVOR = "IETF"
-
     GH_URL = "https://raw.githubusercontent.com/relaton/relaton-data-ietf/master/data/reference."
 
     # @param text [String]
@@ -24,76 +17,13 @@ module RelatonIetf
       ref.sub!(/(?<=^(?:RFC|BCP|FYI|STD)\s)(\d+)/, num.rjust(4, "0")) if num
       rfc_item ref, is_relation
     rescue Timeout::Error, Errno::EINVAL, Errno::ECONNRESET, EOFError,
-            Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError,
-            Net::ProtocolError, SocketError
+           Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError,
+           Net::ProtocolError, SocketError
       raise RelatonBib::RequestError, "No document found for #{ref} reference"
     end
 
-    # @param reference [Nokogiri::XML::Element, nil]
-    # @param is_relation [Boolean] don't add fetched date for relation
-    # @param url [String, NilClass]
-    # @param ver [String, NilClass] Internet Draft version
-    # @return [RelatonBib::tfBibliographicItem]
-    # def fetch_rfc(reference, is_relation: false, url: nil, ver: nil) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
-    #   return unless reference
-
-    #   hash = {
-    #     id: reference[:anchor],
-    #     type: "standard",
-    #     docid: docids(reference, ver),
-    #     status: status(reference),
-    #     language: [language(reference)],
-    #     script: ["Latn"],
-    #     link: link(reference, url, ver),
-    #     title: titles(reference),
-    #     formattedref: formattedref(reference),
-    #     abstract: abstracts(reference),
-    #     contributor: contributors(reference),
-    #     relation: relations(reference),
-    #     date: dates(reference),
-    #     series: series(reference),
-    #     keyword: reference.xpath("front/keyword").map(&:text),
-    #     doctype: doctype(reference[:anchor]),
-    #   }
-    #   hash[:fetched] = Date.today.to_s unless is_relation
-    #   bib_item(**hash)
-    # end
-
     private
 
-    # @param anchor [String]
-    # @return [String]
-    # def doctype(anchor)
-    #   anchor&.include?("I-D") ? "internet-draft" : "rfc"
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @param url [String]
-    # @param ver [String, NilClass] Internet Draft version
-    # @return [Array<Hash>]
-    # def link(reference, url, ver)
-    #   l = []
-    #   l << { type: "xml", content: url } if url
-    #   l << { type: "src", content: reference[:target] } if reference[:target]
-    #   if /^I-D/.match? reference[:anchor]
-    #     reference.xpath("format").each do |f|
-    #       c = ver ? f[:target].sub(/(?<=-)\d{2}(?=\.)/, ver) : f[:target]
-    #       l << { type: f[:type], content: c }
-    #     end
-    #   end
-    #   l
-    # end
-
-    # @param attrs [Hash]
-    # @return [RelatonIetf::IetfBibliographicItem]
-    def bib_item(**attrs)
-      unless attrs.delete(:is_relation)
-        attrs[:fetched] = Date.today.to_s
-        attrs[:place] = ["Fremont, CA"]
-      end
-      RelatonIetf::IetfBibliographicItem.new(**attrs)
-    end
-
     # @param ref [String]
     # @param is_relation [Boolen, nil]
     # @return [RelatonIetf::IetfBibliographicItem]
@@ -108,17 +38,9 @@ module RelatonIetf
       # doc = Nokogiri::XML get_page(uri)
       # r = doc.at("/referencegroup", "/reference")
       # fetch_rfc r, is_relation: is_relation, url: uri, ver: ver
-      parse get_page(uri), url: uri, is_relation: is_relation, ver: ver
+      BibXMLParser.parse get_page(uri), url: uri, is_relation: is_relation, ver: ver
     end
 
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Hash]
-    # def relations(reference)
-    #   reference.xpath("reference").map do |ref|
-    #     { type: "includes", bibitem: fetch_rfc(ref, is_relation: true) }
-    #   end
-    # end
-
     # @param uri [String]
     # @return [String] HTTP response body
     def get_page(uri)
@@ -127,251 +49,5 @@ module RelatonIetf
 
       res.body
     end
-
-    # @param [RelatonBib::WorkGroup]
-    # @return [RelatonIetf::Committee]
-    def committee(wgr)
-      Committee.new wgr
-    end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [String]
-    # def language(reference)
-    #   reference[:lang] || "en"
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<Hash>]
-    # def titles(reference)
-    #   reference.xpath("./front/title").map do |title|
-    #     { content: title.text, language: language(reference), script: "Latn" }
-    #   end
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [RelatonBib::FormattedRef, nil]
-    # def formattedref(reference)
-    #   return if reference.at "./front/title"
-
-    #   cont = (reference[:anchor] || reference[:docName] || reference[:number])
-    #   if cont
-    #     RelatonBib::FormattedRef.new(
-    #       content: cont, language: language(reference), script: "Latn",
-    #     )
-    #   end
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<RelatonBib::FormattedString>]
-    # def abstracts(ref)
-    #   ref.xpath("./front/abstract").map do |a|
-    #     RelatonBib::FormattedString.new(
-    #       content: a.text.gsub(/\\n\\t{2,4}/, " ").strip,
-    #       language: language(ref), script: "Latn"
-    #     )
-    #   end
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<Hash>]
-    def contributors(reference)
-      [{ entity: new_org("Internet Engineering Task Force", "IETF"), role: [type: "publisher"] }] + super
-      # persons(reference) + organizations(reference)
-    end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<Hash{Symbol=>RelatonBib::Person,Symbol=>Array<String>}>]
-    # def persons(reference)
-    #   reference.xpath("./front/author[@surname]|./front/author[@fullname]")
-    #     .map do |author|
-    #     entity = RelatonBib::Person.new(
-    #       name: full_name(author, reference),
-    #       affiliation: affiliation(author),
-    #       contact: contacts(author.at("./address")),
-    #     )
-    #     { entity: entity, role: [contributor_role(author)] }
-    #   end
-    # end
-
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<Hash{Symbol=>RelatonBib::Organization,
-    #   Symbol=>Array<String>}>]
-    # def organizations(reference)
-    #   publisher = { entity: new_org, role: [type: "publisher"] }
-    #   orgs = reference.xpath("./seriesinfo").reduce([publisher]) do |mem, si|
-    #     next mem unless si[:stream]
-
-    #     mem << { entity: new_org(si[:stream], nil), role: [type: "author"] }
-    #   end
-    #   orgs + reference.xpath(
-    #     "front/author[not(@surname)][not(@fullname)]/organization",
-    #   ).map do |org|
-    #     { entity: new_org(org.text, nil), role: [type: "author"] }
-    #   end
-    # end
-
-    # @param author [Nokogiri::XML::Element]
-    # @param ref [Nokogiri::XML::Element]
-    # @return [RelatonBib::FullName]
-    # def full_name(author, ref)
-    #   lang = language ref
-    #   RelatonBib::FullName.new(
-    #     completename: localized_string(author[:fullname], lang),
-    #     initial: [localized_string(author[:initials], lang)].compact,
-    #     surname: localized_string(author[:surname], lang),
-    #   )
-    # end
-
-    # @param content [String]
-    # @param lang [String]
-    # @return [RelatonBib::LocalizedString]
-    # def localized_string(content, lang)
-    #   return unless content
-
-    #   RelatonBib::LocalizedString.new(content, lang)
-    # end
-
-    # @param postal [Nokogiri::XML::Element]
-    # @return [Array<RelatonBib::Address, RelatonBib::Phone>]
-    # def contacts(addr)
-    #   contacts = []
-    #   return contacts unless addr
-
-    #   postal = addr.at("./postal")
-    #   contacts << address(postal) if postal
-    #   add_contact(contacts, "phone", addr.at("./phone"))
-    #   add_contact(contacts, "email", addr.at("./email"))
-    #   add_contact(contacts, "uri", addr.at("./uri"))
-    #   contacts
-    # end
-
-    # @param postal [Nokogiri::XML::Element]
-    # @rerurn [RelatonBib::Address]
-    # def address(postal) # rubocop:disable Metrics/CyclomaticComplexity
-    #   RelatonBib::Address.new(
-    #     street: [(postal.at("./postalLine") || postal.at("./street"))&.text],
-    #     city: postal.at("./city")&.text,
-    #     postcode: postal.at("./code")&.text,
-    #     country: postal.at("./country")&.text,
-    #     state: postal.at("./region")&.text,
-    #   )
-    # end
-
-    # @param type [String] allowed "phone", "email" or "uri"
-    # @param value [String]
-    # def add_contact(contacts, type, value)
-    #   return unless value
-
-    #   contacts << RelatonBib::Contact.new(type: type, value: value.text)
-    # end
-
-    # @param author [Nokogiri::XML::Element]
-    # @return [RelatonBib::Affiliation]
-    # def affiliation(author)
-    #   organization = author.at("./organization")
-    #   org = if organization.nil? || organization&.text&.empty?
-    #           new_org
-    #         else
-    #           new_org organization.text, organization[:abbrev]
-    #         end
-    #   RelatonBib::Affiliation.new organization: org
-    # end
-
-    # @param name [String]
-    # @param abbr [String]
-    # @return [RelatonBib::Organization]
-    # def new_org(name = "Internet Engineering Task Force", abbr = "IETF")
-    #   RelatonBib::Organization.new name: name, abbreviation: abbr
-    # end
-
-    # @param author [Nokogiri::XML::Document]
-    # @return [Hash]
-    # def contributor_role(author)
-    #   { type: author[:role] || "author" }
-    # end
-
-    # def month(mon)
-    #   return 1 if !mon || mon.empty?
-    #   return mon if /^\d+$/.match? mon
-
-    #   Date::MONTHNAMES.index(mon)
-    # end
-
-    #
-    # Extract date from reference.
-    #
-    # @param reference [Nokogiri::XML::Element]
-    # @return [Array<RelatonBib::BibliographicDate>] published data.
-    #
-    # def dates(reference)
-    #   return unless (date = reference.at "./front/date")
-
-    #   d = [date[:year], month(date[:month]),
-    #        (date[:day] || 1)].compact.join "-"
-    #   date = Time.parse(d).strftime "%Y-%m-%d"
-    #   [RelatonBib::BibliographicDate.new(type: "published", on: date)]
-    # end
-
-    #
-    # Extract document identifiers from reference
-    #
-    # @param reference [Nokogiri::XML::Element]
-    # @param ver [String, NilClass] Internet Draft version
-    #
-    # @return [Array<RelatonBib::DocumentIdentifier>]
-    #
-    # def docids(reference, ver) # rubocop:disable Metrics/MethodLength,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity,Metrics/AbcSize
-    #   id = (reference[:anchor] || reference[:docName] || reference[:number])
-    #   ret = []
-    #   if id
-    #     ret << RelatonBib::DocumentIdentifier.new(
-    #       type: "IETF", id: id.sub(/^(RFC)/, "\\1 "),
-    #     )
-    #   end
-    #   if (id = reference[:anchor])
-    #     ret << RelatonBib::DocumentIdentifier.new(type: "rfc-anchor", id: id)
-    #   end
-    #   names = ["DOI", "Internet-Draft"]
-    #   ret + reference.xpath("./seriesInfo").map do |si|
-    #     next unless names.include? si[:name]
-
-    #     id = si[:value]
-    #     id.sub!(/(?<=-)\d{2}$/, ver) if ver && si[:name] == "Internet-Draft"
-    #     RelatonBib::DocumentIdentifier.new(id: id, type: si[:name])
-    #   end.compact
-    # end
-
-    #
-    # Extract series form reference
-    # @param reference [Nokogiri::XML::Element]
-    #
-    # @return [Array<RelatonBib::Series>]
-    #
-    # def series(reference)
-    #   reference.xpath("./seriesInfo").map do |si|
-    #     next if si[:name] == "DOI" || si[:stream] || si[:status]
-
-    #     RelatonBib::Series.new(
-    #       title: RelatonBib::TypedTitleString.new(
-    #         content: si[:name], language: language(reference), script: "Latn",
-    #       ),
-    #       number: si[:value],
-    #       type: "main",
-    #     )
-    #   end.compact
-    # end
-
-    #
-    # extract status
-    # @param reference [Nokogiri::XML::Element]
-    #
-    # @return [RelatonBib::DocumentStatus]
-    #
-    # def status(reference)
-    #   st = reference.at("./seriesinfo[@status]")
-    #   return unless st
-
-    #   RelatonBib::DocumentStatus.new(stage: st[:status])
-    # end
   end
 end

data/lib/relaton_ietf/version.rb

@@ -1,3 +1,3 @@
 module RelatonIetf
-  VERSION = "1.9.4".freeze
+  VERSION = "1.9.8".freeze
 end

data/lib/relaton_ietf.rb

@@ -1,6 +1,9 @@
 # frozen_string_literal: true
 
+require "net/http"
+require "relaton_bib"
 require "relaton_ietf/version"
+require "relaton_ietf/bibxml_parser"
 require "relaton_ietf/ietf_bibliography"
 require "relaton_ietf/ietf_bibliographic_item"
 require "relaton_ietf/xml_parser"

data/relaton_ietf.gemspec

@@ -38,6 +38,6 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "vcr"
   spec.add_development_dependency "webmock"
 
-  spec.add_dependency "relaton-bib", ">= 1.9.8"
+  spec.add_dependency "relaton-bib", ">= 1.9.12"
   spec.add_dependency "zlib", "~> 1.1.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-ietf
 version: !ruby/object:Gem::Version
-  version: 1.9.4
+  version: 1.9.8
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2021-12-23 00:00:00.000000000 Z
+date: 2021-12-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: equivalent-xml
@@ -128,14 +128,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.9.8
+        version: 1.9.12
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.9.8
+        version: 1.9.12
 - !ruby/object:Gem::Dependency
   name: zlib
   requirement: !ruby/object:Gem::Requirement
@@ -177,6 +177,7 @@ files:
 - grammars/reqt.rng
 - lib/relaton/provider_ietf.rb
 - lib/relaton_ietf.rb
+- lib/relaton_ietf/bibxml_parser.rb
 - lib/relaton_ietf/committee.rb
 - lib/relaton_ietf/data_fetcher.rb
 - lib/relaton_ietf/hash_converter.rb