relaton-iec 1.7.1 → 1.7.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3989d41422ae3287109099dabba99ecc5c8a5657dd391fa6e069e7116df4fdac
-  data.tar.gz: 222dd6c957c83b1a6c4bdb0a79140bf33f64da6a1725a7ece785a1fcc797bf31
+  metadata.gz: '08ac0e0f5df714485be57354b74b7e172758ffe218467667a63c670ea33ea068'
+  data.tar.gz: e2887cb6b9346fc7d1eb9941fe759dc2d3dd90c9c4826d977e1d9b191ea3ddf8
 SHA512:
-  metadata.gz: 59b511e74a2315143c1ed8aa5a78e231d7b8fb7a09b568676443b39bd7d077e654fba28ff9c160000beb6ba9868684646ab546af4c26605aa2c06bac5849c827
-  data.tar.gz: a9530148a809cb3b898abb86a4b54c9c2c421bdd5a71a0f11ee6000ec93b45731b4bab70c088ee973a1538c42f925d2233f27fa1c71f72abacdb70fba8685584
+  metadata.gz: 77bb3d3b8f2c829147434bec3dbcff6840ef1048c24613c9eca3b4b3c1013f6b58d100b987b36713a200de918b240ba8ee4dd461b8887ba5e234b2b7f56eef13
+  data.tar.gz: 93d012040298e5570b91730a260a50009c79889d905daac3fa274660a740b997c50f1914f243fa7c964207061e4d21b60a48ecc0a6b808a555a8688f6598ab4f

data/README.adoc

@@ -31,7 +31,211 @@ Or install it yourself as:
 
 == Usage
 
-Refer to https://github.com/relaton/relaton-iso/blob/master/README.adoc the usage interface is identical.
+=== Search for a standard using keywords
+
+`RelatonIec::IecBibliography.search(ref, year)` method returns hits collection. Each hit can be used to fetch a document.
+
+- `ref` - reference to search document
+- `year` - filter by year (optional)
+
+[source,ruby]
+----
+require 'relaton_iec'
+=> true
+
+hit_collection = RelatonIec::IecBibliography.search("60050")
+=> <RelatonIec::HitCollection:0x007fe0d7126f28 @ref=60050 @fetched=false>
+
+hit_collection.first
+=> <RelatonIec::Hit:0x007fe1068b4d10 @text="60050" @fetched="false" @fullIdentifier="" @title="IEC 60050-102:2007">
+
+item = hit_collection[2].fetch
+=> #<RelatonIec::IecBibliographicItem:0x007fe1171a06f8
+...
+
+item.docidentifier
+=> [#<RelatonBib::DocumentIdentifier:0x007fe1171bb930 @id="IEC 60050-112:2010", @scope=nil, @type="IEC">,
+ #<RelatonBib::DocumentIdentifier:0x007fe1171bb728 @id="urn:iec:std:iec:60050-112:2010:::en", @scope=nil, @type="URN">]
+
+item.docidentifier.detect { |di| di.type == "URN" }.id
+=> "urn:iec:std:iec:60050-112:2010:::en"
+----
+
+=== Fetch documen by keywords
+
+`RelatonIec::IecBibliography.get(ref, year, opts)` method returns document.
+
+- `ref` - reference to search document
+- `year` - filter by year (optional)
+- `opts` - hash of options (optional). Supported options are `:all_parts` (boolean), `:keep_year` (boolean).
+
+[source,ruby]
+----
+item = RelatonIec::IecBibliography.get("IEC 60050-112:2010")
+[relaton-iec] ("IEC 60050-112") fetching...
+[relaton-iec] ("IEC 60050-112") found IEC 60050-112:2010
+=> #<RelatonIec::IecBibliographicItem:0x007fa64e874f78
+
+item = RelatonIec::IecBibliography.get("IEC 60050-112", "2010", all_parts: true)
+[relaton-iec] ("IEC 60050-112") fetching...
+[relaton-iec] ("IEC 60050-112") found IEC 60050 (all parts)
+=> #<RelatonIec::IecBibliographicItem:0x007fa69e9b3948
+
+item.docidentifier.first
+=> #<RelatonBib::DocumentIdentifier:0x007fa69e9abd10 @id="IEC 60050 (all parts)", @scope=nil, @type="IEC">
+
+item = RelatonIec::IecBibliography.get("IEC 60050-112", "2010")
+[relaton-iec] ("IEC 60050-112") fetching...
+[relaton-iec] ("IEC 60050-112") found IEC 60050-112:2010
+=> #<RelatonIec::IecBibliographicItem:0x007fa69f296da8
+
+item.docidentifier.first
+=> #<RelatonBib::DocumentIdentifier:0x007fa69f2a65f0 @id="IEC 60050-112:2010", @scope=nil, @type="IEC">
+----
+
+=== XML serialization
+
+Possible options:
+
+- *bibdata* - If true then wrapp item with _bibdata_ element and add _ext_ element.
+- *note* - Array of hashes `{ text: "Note", type: "note" }`. These notes will be added to XML.
+
+[source,ruby]
+----
+item.to_xml
+=> "<bibitem id="IEC60050-112-2010" type="standard">
+      <fetched>2021-01-28</fetched>
+      <title type="title-main" format="text/plain" language="en" script="Latn">International Electrotechnical Vocabulary (IEV)</title>
+      <title type="title-part" format="text/plain" language="en" script="Latn">Part 112: Quantities and units</title>
+      <title type="main" format="text/plain" language="en" script="Latn">International Electrotechnical Vocabulary (IEV) - Part 112: Quantities and units</title>
+      <uri type="src">https://webstore.iec.ch/publication/162</uri>
+      <uri type="obp">/preview/info_iec60050-112%7Bed1.0%7Db.pdf</uri>
+      <docidentifier type="IEC">IEC 60050-112:2010</docidentifier>
+      <docidentifier type="URN">urn:iec:std:iec:60050-112:2010:::en</docidentifier>
+      <date type="published">
+        <on>2010-01-27</on>
+      </date>
+      <contributor>
+        <role type="publisher"/>
+        <organization>
+          <name>International Electrotechnical Commission</name>
+          <abbreviation>IEC</abbreviation>
+          <uri>www.iec.ch</uri>
+        </organization>
+      </contributor>
+      <edition>1.0</edition>
+      <language>en</language>
+      <script>Latn</script>
+      <abstract format="text/plain" language="en" script="Latn">
+        IEC 60050-112:2010 gives the general terminology concerning quantities and units, the terminology of SI, terms used in names and definitions of quantities, and some basic concepts in metrology. It cancels and replaces Sections 111-11 and 111-12 of International Standard IEC 60050-111:1996. It has the status of a horizontal standard in accordance with IEC Guide 108.
+      </abstract>
+      <status>
+        <stage>60</stage>
+        <substage>60</substage>
+      </status>
+      <copyright>
+        <from>2010</from>
+        <owner>
+          <organization>
+            <name>International Electrotechnical Commission</name>
+            <abbreviation>IEC</abbreviation>
+            <uri>www.iec.ch</uri>
+          </organization>
+        </owner>
+      </copyright>
+      <place>Geneva</place>
+    </bibitem>"
+
+item.to_xml bibdata: true
+=> "<bibdata type="standard">
+      <fetched>2021-01-28</fetched>
+      <title type="title-main" format="text/plain" language="en" script="Latn">International Electrotechnical Vocabulary (IEV)</title>
+      <title type="title-part" format="text/plain" language="en" script="Latn">Part 112: Quantities and units</title>
+      <title type="main" format="text/plain" language="en" script="Latn">International Electrotechnical Vocabulary (IEV) - Part 112: Quantities and units</title>
+      <uri type="src">https://webstore.iec.ch/publication/162</uri>
+      <uri type="obp">/preview/info_iec60050-112%7Bed1.0%7Db.pdf</uri>
+      <docidentifier type="IEC">IEC 60050-112:2010</docidentifier>
+      <docidentifier type="URN">urn:iec:std:iec:60050-112:2010:::en</docidentifier>
+      ...
+      <ext>
+        <doctype>international-standard</doctype>
+        <editorialgroup>
+          <technical-committee number="1" type="technicalCommittee">TC 1 - Terminology</technical-committee>
+        </editorialgroup>
+        <ics>
+          <code>01.040.01</code>
+          <text>Generalities. Terminology. Standardization. Documentation (Vocabularies)</text>
+        </ics>
+        <ics>
+          <code>01.060</code>
+          <text>Quantities and units</text>
+        </ics>
+        <structuredidentifier type="IEC">
+          <project-number>60050</project-number>
+        </structuredidentifier>
+      </ext>
+    </bibdata>"
+
+item.to_xml note: [{ text: "Note", type: "note" }]
+=> "<bibitem id="IEC60050-112-2010" type="standard">
+      ...
+      <note format="text/plain" type="note">Note</note>
+      ...
+    </bibitem>"
+----
+
+=== Converting reference to URN
+
+URN is document a identifier format. It has fields delimited by colon. If any field is absent then it's place is empty. All values are in lower-case.
+
+URN structure: +
+`urn:sdo_namespace:content_type_namespace:header:project_number:date:type:deliverable:language:relation:adjunct_type:adjunct_number:date[#/=]component_or_related_asset`
+
+* prefix
+- `urn` - value: urn
+- `sdo_namespace` - value: iec
+- `content_type_namespace` - value: std
+* base document information
+- `header` - possible values are: iec, iso, iec-iso, iec-ieee, iec-itu, iec-astm
+- `project_number` - number and partnumber. For example 67654, 60601-1, 61076-7-101
+- `date` - document date (optional). Examples: 2010-03, 2010, 2010-10-11 etc.
+- `type` - documant type (optional). Possible values: ts, tr, pas, guide, is, ser.
+- `deliverable` - (optional) possible values: prv, csv, exv, rlv, cmv
+- `language` - (optional) examples: en, fr, ru, en-fr, en-fr-ru etc.
+* adjunct document information (optional)
+- `relation` - "plus" for consolidations of ajancts with a base document or "/" for ajancts itself
+- `adjunct_type` - possible values: amd, cor, ish
+- `adjunct_number` - adjunct number. Examples: 1, 2, 3, etc.
+- `date` - adjanct date. Example: 2009
+* component or related asset information (optional)
+- `component_or_related_asset` - # or = followed by component id or related asset. Exammples: #fig-1, #sec-1, =forum
+
+For more information see https://github.com/relaton/relaton-iec/issues/22
+
+The method `RelatonIec.code_to_urn(code, lang)` converts document identifier to URN.
+
+* `code` is a document identifier
+* `lang` is a laguage code (optional). Examples: en, fr, en-fr etc.
+
+[source,ruby]
+----
+RelatonIec.code_to_urn "IEC 60050-102:2007/AMD1:2017"
+=> "urn:iec:std:iec:60050-102:2007:::::amd:1:2017"
+
+RelatonIec.code_to_urn "IEC 60034-1:1969+AMD1:1977+AMD2:1979+AMD3:1980 CSV", "en-fr"
+=> "urn:iec:std:iec:60034-1:1969::csv:en-fr:plus:amd:1:1977:plus:amd:2:1979:plus:amd:3:1980"
+----
+
+The method `RelatonIec.urn_to_code(urn)` converts URN to document identifier.
+
+[source,ruby]
+----
+RelatonIec.urn_to_code "urn:iec:std:iec:60050-102:2007:::::amd:1:2017"
+=> ["IEC 60050-102:2007/AMD1:2017", ""]
+
+RelatonIec.urn_to_code "urn:iec:std:iec:60034-1:1969::csv:en-fr:plus:amd:1:1977:plus:amd:2:1979:plus:amd:3:1980"
+=> ["IEC 60034-1:1969+AMD1:1977+AMD2:1979+AMD3:1980 CSV", "en-fr"]
+----
 
 == Development
 

data/bin/rspec

@@ -0,0 +1,29 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+#
+# This file was generated by Bundler.
+#
+# The application 'rspec' is installed as part of a gem, and
+# this file is here to facilitate running it.
+#
+
+require "pathname"
+ENV["BUNDLE_GEMFILE"] ||= File.expand_path("../../Gemfile",
+  Pathname.new(__FILE__).realpath)
+
+bundle_binstub = File.expand_path("../bundle", __FILE__)
+
+if File.file?(bundle_binstub)
+  if File.read(bundle_binstub, 300) =~ /This file was generated by Bundler/
+    load(bundle_binstub)
+  else
+    abort("Your `bin/bundle` was not generated by Bundler, so this binstub cannot run.
+Replace `bin/bundle` by running `bundle binstubs bundler --force`, then run this command again.")
+  end
+end
+
+require "rubygems"
+require "bundler/setup"
+
+load Gem.bin_path("rspec-core", "rspec")

data/lib/relaton_iec.rb

@@ -10,12 +10,76 @@ require "relaton_iec/hash_converter"
 require "digest/md5"
 
 module RelatonIec
-  # 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
+  class << self
+    # Returns hash of XML reammar
+    # @return [String]
+    def 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
+
+    # @param code [String]
+    # @param lang [String]
+    # @return [String, nil]
+    def code_to_urn(code, lang = nil) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
+      rest = code.downcase.sub(%r{
+        (?<head>[^\s]+)\s
+        (?<type>is|ts|tr|pas|srd|guide|tec|wp)?(?(<type>)\s)
+        (?<pnum>[\d-]+)\s?
+        (?<_dd>:)?(?(<_dd>)(?<date>[\d-]+)\s?)
+      }x, "")
+      m = $~
+      return unless m[:head] && m[:pnum]
+
+      deliv = /cmv|csv|exv|prv|rlv|ser/.match(code.downcase).to_s
+      urn = ["urn", "iec", "std", m[:head].split("/").join("-"), m[:pnum], m[:date], m[:type], deliv, lang]
+      (urn + ajunct_to_urn(rest)).join ":"
+    end
+
+    # @param urn [String]
+    # @return [Array<String>, nil] urn & language
+    def urn_to_code(urn) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
+      fields = urn.upcase.split ":"
+      return if fields.size < 5
+
+      head, num, date, type, deliv, lang = fields[3, 8]
+      code = head.gsub("-", "/")
+      code += " #{type}" unless type.nil? || type.empty?
+      code += " #{num}"
+      code += ":#{date}" unless date.nil? || date.empty?
+      code += ajanct_to_code(fields[9..-1])
+      code += " #{deliv}" unless deliv.nil? || deliv.empty?
+      [code, lang&.downcase]
+    end
+
+    private
+
+    # @param fields [Array<String>]
+    # @return [String]
+    def ajanct_to_code(fields)
+      return "" if fields.nil? || fields.empty?
+
+      rel, type, num, date = fields[0..3]
+      code = (rel.empty? ? "/" : "+") + type + num
+      code += ":#{date}" unless date.empty?
+      code + ajanct_to_code(fields[4..-1])
+    end
+
+    # @param rest [String]
+    # @return [Array<String, nil>]
+    def ajunct_to_urn(rest)
+      r = rest.sub(%r{
+        (?<pl>\+|\/)(?(<pl>)(?<adjunct>(amd|cor|ish))(?<adjnum>\d+)\s?)
+        (?<_d2>:)?(?(<_d2>)(?<adjdt>[\d-]+)\s?)
+      }x, "")
+      m = $~ || {}
+      return [] unless m[:adjunct]
+
+      plus = "plus" if m[:pl] == "+"
+      urn = [plus, m[:adjunct], m[:adjnum], m[:adjdt]]
+      urn + ajunct_to_urn(r)
+    end
   end
 end

data/lib/relaton_iec/hit.rb

@@ -8,5 +8,9 @@ module RelatonIec
     def fetch
       @fetch ||= Scrapper.parse_page @hit
     end
+
+    def part
+      @part ||= hit[:code].match(/(?<=-)[\w-]+/)&.to_s
+    end
   end
 end

data/lib/relaton_iec/hit_collection.rb

@@ -6,52 +6,74 @@ require "addressable/uri"
 module RelatonIec
   # Page of hit collection.
   class HitCollection < RelatonBib::HitCollection
+    def_delegators :@array, :detect
+
+    attr_reader :part
+
     DOMAIN = "https://webstore.iec.ch"
 
-    # @param ref_nbr [String]
+    # @param ref [String]
     # @param year [String, nil]
     # @param part [String, nil]
-    def initialize(ref_nbr, year = nil, part = nil)
-      super ref_nbr, year
-      @array = hits ref_nbr, year, part
+    def initialize(ref, year = nil, part = nil)
+      super ref, year
+      @part = part
+      @array = ref ? hits(ref, year) : []
+    end
+
+    # @return [RelatonIec::IecBibliographicItem]
+    def to_all_parts # rubocop:disable Metrics/AbcSize,Metrics/MethodLength,Metrics/CyclomaticComplexity
+      parts = @array.reject { |h| h.part.nil? }
+      hit = parts.min_by &:part
+      return @array.first.fetch lang unless hit
+
+      bibitem = hit.fetch
+      all_parts_item = bibitem.to_all_parts
+      parts.reject { |h| h.hit[:code] == hit.hit[:code] }.each do |hi|
+        isobib = RelatonIec::IecBibliographicItem.new(
+          formattedref: RelatonBib::FormattedRef.new(content: hi.hit[:code])
+        )
+        all_parts_item.relation << RelatonBib::DocumentRelation.new(type: "partOf", bibitem: isobib)
+      end
+      all_parts_item
     end
 
     private
 
     # @param ref [String]
     # @param year [String, nil]
-    # @param part [String, nil]
     # @return [Array<RelatonIec::Hit>]
-    def hits(ref, year, part)
+    def hits(ref, year)
       from, to = nil
       if year
         from = Date.strptime year, "%Y"
         to   = from.next_year.prev_day
       end
-      get_results ref, from, to, part
+      get_results ref, from, to
     end
 
     # @param ref [String]
     # @param from [Date, nil]
     # @param to [Date, nil]
-    # @param part [String, nil]
     # @return [Array<RelatonIec::Hit>]
-    def get_results(ref, from, to, part = nil)
+    def get_results(ref, from, to)
       code = part ? ref.sub(/(?<=-\d)\d+/, "*") : ref
       [nil, "trf", "wr"].reduce([]) do |m, t|
         url = "#{DOMAIN}/searchkey"
         url += "&type=#{t}" if t
         url += "&RefNbr=#{code}&From=#{from}&To=#{to}&start=1"
-        m + results(Addressable::URI.parse(url).normalize, part)
+        m + results(Addressable::URI.parse(url).normalize)
       end
     end
 
     # @param url [String]
-    # @param part [String, nil]
     # @return [Array<RelatonIec::Hit>]
-    def results(uri, part)
+    def results(uri)
       contains = "[contains(.,'Part #{part}:')]" if part
-      Nokogiri::HTML(OpenURI.open_uri(uri)).xpath(
+      resp = OpenURI.open_uri(uri, "User-Agent" => "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_6) "\
+        "AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.111 Safari/537.36")
+      doc = Nokogiri::HTML(resp)
+      doc.xpath(
         "//body/li#{contains}",
         "//ul[contains(@class,'search-results')]/li#{contains}",
         "//ul[contains(@class,'morethesame')]/li#{contains}"

data/lib/relaton_iec/iec_bibliography.rb

@@ -21,7 +21,7 @@ module RelatonIec
       # @param part [String, nil] search for packaged stndard if not nil
       # @return [RelatonIec::HitCollection]
       def search(text, year = nil, part = nil)
-        HitCollection.new text, year, part
+        HitCollection.new text, year&.strip, part
       rescue SocketError, OpenURI::HTTPError, OpenSSL::SSL::SSLError
         raise RelatonBib::RequestError, "Could not access http://www.iec.ch"
       end
@@ -32,28 +32,29 @@ module RelatonIec
       #   reference is required
       # @return [String] Relaton XML serialisation of reference
       def get(code, year = nil, opts = {}) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+        opts[:all_parts] ||= code.match? /\s\(all parts\)/
+        ref = code.sub /\s\(all parts\)/, ""
         if year.nil?
-          /^(?<code1>[^:]+):(?<year1>[^:]+)/ =~ code
+          /^(?<code1>[^:]+):(?<year1>[^:]+)/ =~ ref
           unless code1.nil?
-            code = code1
+            ref = code1
             year = year1
           end
         end
+        return iev if ref.casecmp("IEV").zero?
 
-        return iev if code.casecmp("IEV").zero?
-
-        opts[:all_parts] ||= !(code =~ / \(all parts\)/).nil?
-        code = code.sub(/ \(all parts\)/, "")
-        ret = iecbib_get1(code, year, opts)
+        ret = iecbib_get(ref, year, opts)
         return nil if ret.nil?
 
         ret = ret.to_most_recent_reference unless year || opts[:keep_year]
-        ret = ret.to_all_parts if opts[:all_parts]
         ret
       end
 
       private
 
+      # @param code [String]
+      # @param year [String]
+      # @param missed_years [Array<String>]
       def fetch_ref_err(code, year, missed_years) # rubocop:disable Metrics/MethodLength
         id = year ? "#{code}:#{year}" : code
         warn "[relaton-iec] WARNING: no match found online for #{id}. "\
@@ -77,29 +78,42 @@ module RelatonIec
       # @param hits [Array<RelatonIec::Hit>]
       # @param threads [Integer]
       # @return [Array<RelatonIec::Hit>]
-      def fetch_pages(hits, threads)
-        workers = RelatonBib::WorkersPool.new threads
-        workers.worker { |w| { i: w[:i], hit: w[:hit].fetch } }
-        hits.each_with_index { |hit, i| workers << { i: i, hit: hit } }
-        workers.end
-        workers.result.sort_by { |a| a[:i] }.map { |x| x[:hit] }
-      end
-
-      def isobib_search_filter(code, year) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
-        docidrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+}
-        corrigrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+:[0-9]+/}
-        warn "[relaton-iec] (\"#{code}\") fetching..."
+      # def fetch_pages(hits, threads)
+      #   workers = RelatonBib::WorkersPool.new threads
+      #   workers.worker { |w| { i: w[:i], hit: w[:hit].fetch } }
+      #   hits.each_with_index { |hit, i| workers << { i: i, hit: hit } }
+      #   workers.end
+      #   workers.result.sort_by { |a| a[:i] }.map { |x| x[:hit] }
+      # end
+
+      # @param ref [String]
+      # @param year [String, nil]
+      # @return [RelatonIec::HitCollection]
+      def search_filter(ref, year) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
+        %r{
+          ^(?<code>\S+[^\d]*\s\d+((?:-\w+)+)?)
+          (:(?<year1>\d{4}))?
+          (?<bundle>\+[^\s\/]+)?
+          (\/(?<corr>AMD\s\d+))?
+        }x =~ ref.upcase
+        year ||= year1
+        corr&.sub! " ", ""
+        warn "[relaton-iec] (\"#{ref}\") fetching..."
         result = search(code, year)
-        if result.empty? && /(?<=-)(?<part>\d+)/ =~ code
+        if result.empty? && /(?<=-)(?<part>[\w-]+)/ =~ code
           # try to search packaged standard
           result = search code, year, part
-          ref = code.sub /(?<=-\d)\d+/, ""
-        else ref = code
         end
+        result = search code if result.empty?
+        code&.sub! /((?:-\w+)+)/, ""
         result.select do |i|
-          i.hit[:code] &&
-            i.hit[:code].match(docidrx).to_s.include?(ref) &&
-            corrigrx !~ i.hit[:code]
+          %r{
+            ^(?<code2>\S+[^\d]*\s\d+)((?:-\w+)+)?
+            (:\d{4})?
+            (?<bundle2>\+[^\s\/]+)?
+            (\/(?<corr2>AMD\d+))?
+          }x =~ i.hit[:code]
+          code == code2 && bundle == bundle2 && corr == corr2
         end
       end
 
@@ -144,30 +158,61 @@ module RelatonIec
       # 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) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength
+      def results_filter(result, ref, year, opts) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+        r_code, r_year = code_year ref, result.part
+        r_year ||= year
         missed_years = []
-        result.each_slice(3) do |s| # ISO website only allows 3 connections
-          fetch_pages(s, 3).each_with_index do |r, _i|
-            return { ret: r } if !year
-
-            r.date.select { |d| d.type == "published" }.each do |d|
-              return { ret: r } if year.to_i == d.on(:year)
-
-              missed_years << d.on(:year)
-            end
-          end
-        end
-        { years: missed_years }
+        missed_parts = false
+        # result.each_slice(3) do |s| # ISO website only allows 3 connections
+        ret = if opts[:all_parts]
+                result.to_all_parts
+              else
+                result.detect do |h|
+                  h_code, h_year = code_year h.hit[:code], result.part
+                  missed_parts ||= !opts[:all_parts] && r_code != h_code
+                  missed_years << h_year unless !r_year || h_year == r_year
+                  r_code == h_code && (!year || h_year == r_year)
+                  # fetch_pages(s, 3).each_with_index do |r, _i|
+                  # return { ret: r } if !year
+
+                  # r.date.select { |d| d.type == "published" }.each do |d|
+                  # return { ret: r } if year.to_i == d.on(:year)
+
+                  # missed_years << d.on(:year)
+                  # end
+                  # end
+                end&.fetch
+              end
+        { ret: ret, years: missed_years, missed_parts: missed_parts }
       end
 
-      def iecbib_get1(code, year, _opts)
-        return iev if code.casecmp("IEV").zero?
+      # @param ref [string]
+      # @param part [String, nil]
+      # @return [Array<String, nil>]
+      def code_year(ref, part)
+        %r{
+          ^(?<code>\S+[^\d]*\s\d+((?:-\w+)+)?)
+          (:(?<year>\d{4}))?
+        }x =~ ref
+        code.sub!(/-\d+/, "") if part
+        [code, year]
+      end
 
-        result = isobib_search_filter(code, year) || return
-        ret = isobib_results_filter(result, year)
+      # @param code [String]
+      # @param year [String, nil]
+      # @param opts [Hash]
+      # @return [RelatonIec::IecBibliographicItem, nil]
+      def iecbib_get(code, year, opts) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
+        result = search_filter(code, year) || return
+        ret = results_filter(result, code, year, opts)
         if ret[:ret]
-          warn "[relaton-iec] (\"#{code}\") found "\
-          "#{ret[:ret].docidentifier.first.id}"
+          if ret[:missed_parts]
+            warn "[relaton-iec] WARNING: #{code} found as #{ret[:ret].docidentifier.first.id} "\
+            "but also contain parts. If you wanted to cite all document parts for the reference, use "\
+            "\"#{code} (all parts)\""
+          else
+            warn "[relaton-iec] (\"#{code}\") found #{ret[:ret].docidentifier.first.id}"
+          end
           ret[:ret]
         else
           fetch_ref_err(code, year, ret[:years])

data/lib/relaton_iec/processor.rb

@@ -35,5 +35,11 @@ module RelatonIec
     def grammar_hash
       @grammar_hash ||= ::RelatonIec.grammar_hash
     end
+
+    # @param code [String]
+    # @return [String, nil]
+    def urn_to_code(code)
+      RelatonIec.urn_to_code code
+    end
   end
 end

data/lib/relaton_iec/scrapper.rb

@@ -69,38 +69,13 @@ module RelatonIec
       # @param hit [Hash]
       # @return [Array<RelatonBib::DocumentIdentifier>]
       def fetch_docid(hit)
-        rest = hit[:code].downcase.sub(%r{
-          (?<head>[^\s]+)\s
-          (?<type>is|ts|tr|pas|srd|guide|tec|wp)?(?(<type>)\s)
-          (?<pnum>[\d-]+)\s?
-          (?<_dd>:)?(?(<_dd>)(?<date>[\d-]+)\s?)
-        }x, "")
-        m = $~
-        deliv = /cmv|csv|exv|prv|rlv|ser/.match(hit[:code].downcase).to_s
-        urn = ["urn", "iec", "std", m[:head].split("/").join("-"), m[:pnum],
-               m[:date], m[:type], deliv, "en"]
-        urn += fetch_ajunct(rest)
+        urn = RelatonIec.code_to_urn hit[:code], "en"
         [
           RelatonBib::DocumentIdentifier.new(id: hit[:code], type: "IEC"),
-          RelatonBib::DocumentIdentifier.new(id: urn.join(":"), type: "URN"),
+          RelatonBib::DocumentIdentifier.new(id: urn, type: "URN"),
         ]
       end
 
-      # @param rest [String]
-      # @return [Array<String, nil>]
-      def fetch_ajunct(rest)
-        r = rest.sub(%r{
-          (?<_pl>\+)(?(<_pl>)(?<adjunct>amd)(?<adjnum>\d+)\s?)
-          (?<_d2>:)?(?(<_d2>)(?<adjdt>[\d-]+)\s?)
-        }x, "")
-        m = $~ || {}
-        return [] unless m[:adjunct]
-
-        plus = m[:adjunct] && "plus"
-        urn = [plus, m[:adjunct], m[:adjnum], m[:adjdt]]
-        urn + fetch_ajunct(r)
-      end
-
       # Fetch abstracts.
       # @param doc [Nokigiri::HTML::Document]
       # @return [Array<Array>]
@@ -170,8 +145,10 @@ module RelatonIec
       def fetch_status(doc)
         wip = doc.at('//ROW[STATUS[.="PREPARING"]]')
         if wip
-          statuses = YAML.load_file "lib/relaton_iec/statuses.yml"
+          statuses = YAML.load_file File.join __dir__, "statuses.yml"
           s = wip.at("STAGE").text
+          return unless statuses[s]
+
           stage, substage = statuses[s]["stage"].split "."
         else
           stage    = "60"

data/lib/relaton_iec/statuses.yml

@@ -4,6 +4,9 @@ ACD:
 ACDV:
   status: Approved for CDV
   stage: '30.99'
+ADISSB:
+  status: Preparation of text subcontracted to CO
+  stage: '40.95'
 ADTR:
   status: Approved for DTR
   stage: '40.99'
@@ -13,30 +16,39 @@ ADTS:
 AFDIS:
   status: Approved for FDIS
   stage: '40.99'
+AMW:
+  status: Document under revision
+  stage: '92.20'
+ANW:
+  status: Registration of new project
+  stage: '20.00'
 APUB:
   status: Approved for publication
   stage: '50.99'
+APUBSB:
+  status: Preparation of text subcontracted to CO
+  stage: '50.95'
 BPUB:
   status: Being published
   stage: '60.00'
+BWG:
+  status: Return to drafting phase or redefine project
+  stage: '30.92'
 CAN:
   status: Draft cancelled
   stage: '20.98'
+CCDV:
+  status: Draft circulated as CDV
+  stage: '40.00'
 CD:
   status: Draft circulated as CD
   stage: '30.00'
+CDISH:
+  status: Draft circulated as DISH
+  stage: '50.20'
 CDM:
   status: CD to be discussed at meeting
   stage: '30.20'
-CCDV:
-  status: Draft circulated as CDV
-  stage: '40.00'
-CDVM:
-  status: Rejected CDV to be discussed at a meeting
-  stage: '40.93'
-CFDIS:
-  status: Draft circulated as FDIS
-  stage: '50.20'
 CDPAS:
   status: Draft circulated as DPAS
   stage: '50.20'
@@ -46,12 +58,15 @@ CDTR:
 CDTS:
   status: Draft circulated as DTS
   stage: '50.20'
-DTRM:
-  status: Rejected DTR to be discussed at meeting
-  stage: '50.92'
-DTSM:
-  status: Rejected DTS to be discussed at meeting
-  stage: '50.92'
+CDVM:
+  status: Rejected CDV to be discussed at a meeting
+  stage: '40.93'
+CFDIS:
+  status: Draft circulated as FDIS
+  stage: '50.20'
+DECDISH:
+  status: DISH at editing check
+  stage: '40.99'
 DECFDIS:
   status: FDIS at editing check
   stage: '50.60'
@@ -64,6 +79,21 @@ DEL:
 DELPUB:
   status: Deleted publication
   stage: '90.99'
+DREJ:
+  status: Abandon
+  stage: '30.98'
+DTRM:
+  status: Rejected DTR to be discussed at meeting
+  stage: '50.92'
+DTSM:
+  status: Rejected DTS to be discussed at meeting
+  stage: '50.92'
+MERGED:
+  status: Fragment merged
+  stage: '30.97'
+NADIS:
+  status: Repeat enquiry
+  stage: '40.93'
 NCDV:
   status: CDV rejected
   stage: '40.98'
@@ -84,13 +114,16 @@ PNW:
   stage: '10.00'
 PPUB:
   status: Publication issued
-  stage: '60.60' # ?
+  stage: '60.60'
 PRVC:
   status: Preparation of RVC
   stage: '40.92'
 PRVD:
   status: Preparation of RVD
   stage: '40.92'
+PRVDISH:
+  status: Preparation of RVDISH
+  stage: '40.92' # ?
 PRVDPAS:
   status: Preparation of RVDPAS
   stage: '40.92'
@@ -106,15 +139,34 @@ PRVN:
 PWI:
   status: Preliminary work item
   stage: '00.00'
+RDIS:
+  status: Registration for formal approval
+  stage: '50.00'
+RDISH:
+  status: DISH received and registered
+  stage: '50.00'
 RFDIS:
   status: FDIS received and registered
   stage: '50.00'
 RPUB:
   status: Publication received and registered
   stage: '60.60'
+SPE:
+  stage: SPE # ?
+SPLIT:
+  status: Project Fragmented
+  stage: SPLIT # ?
+SRP:
+  stage: SRP
+SUSPENDED:
+  status: Project Suspended
+  stage: SUSPENDED # ?
 TCDV:
   status: Translation of CDV
   stage: '50.00'
+TDISH:
+  status: Translation of DISH
+  stage: '50.00' # ?
 TDTR:
   status: Translation of DTR
   stage: '50.00'
@@ -129,4 +181,19 @@ TPUB:
   stage: '60.00'
 WPUB:
   status: Publication withdrawn
-  stage: '95.99'
+  stage: '95.99'
+preCD:
+  status: Preparation of CD document
+  stage: preCD
+preCDPAS:
+  status: Preparation of DPAS
+  stage: preCDPAS
+preDISH:
+  status: Preparation of DISH
+  stage: preDISH
+preDTR:
+  status: Preparation of DTR document
+  stage: preDTR
+prePNW:
+  status: Preparation of NP document
+  stage: prePNW

data/lib/relaton_iec/version.rb

@@ -1,3 +1,3 @@
 module RelatonIec
-  VERSION = "1.7.1".freeze
+  VERSION = "1.7.6".freeze
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iec
 version: !ruby/object:Gem::Version
-  version: 1.7.1
+  version: 1.7.6
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2020-11-30 00:00:00.000000000 Z
+date: 2021-02-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: debase
@@ -197,6 +197,7 @@ files:
 - README.adoc
 - Rakefile
 - bin/console
+- bin/rspec
 - bin/setup
 - grammars/basicdoc.rng
 - grammars/biblio.rng