relaton-itu 1.6.pre1 → 1.7.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 58aad32533947d32d65def46311fbd861058969e52e40c57e5980f7f135b7f05
-  data.tar.gz: 16be516de75d4cafce79f5d5c865637a712f0ea47e710681e9a0f2b36084b624
+  metadata.gz: 971645ffb5b0111df7d841c47f6ac1a64904cc85274218f2f6ff622dbba3b576
+  data.tar.gz: 68bc7e0bbb844896c7bf30c6d3e5747a9d6f0e68aa9384f9e761f2881f3be2bc
 SHA512:
-  metadata.gz: '069aec7fac43a05b079b406ae103ed9a5452639ca376bc7dd716ff5a72acd1ce630fe139dc7c53d45015a37ed8a35687c29b7e7daac6e2124c72995c02bb89c2'
-  data.tar.gz: e0945e7d9e5c6a26c54d834a3c15c115e872c3eb62f2d4e2ad1bac7b80760fc431051d6b612cffbd32bbef797b0271aa0e1c2018254897a4736bb7996c3097fc
+  metadata.gz: 83a9d8ee974bfe44cad5715db0199dfcf1eec9709d2a1d778d08cdc21379a69063bb6d6f0ee818b962f825c684d680bb4dda5cbba9b0ca48b24dc8dde4876cdd
+  data.tar.gz: 96a7cf63dd9ff84a115959d7766544e8d8f5f1328fa976737875b7808152939f1168e6a1644e0eb5e4af5cd22f6077b232cf45968e6945bc1a86d2a7721f8e19

data/README.adoc

@@ -87,6 +87,16 @@ RelatonItu::ItuBibliography.get("ITU-T L.163", "2018", {})
 ...
 ----
 
+=== Get amendment
+[source,ruby]
+----
+RelatonItu::ItuBibliography.get "ITU-T G.989.2/Amd 1"
+[relaton-itu] ("ITU-T G.989.2/Amd 1") fetching...
+[relaton-itu] ("ITU-T G.989.2/Amd 1") found ITU-T G.989.2/Amd 1
+=> #<RelatonItu::ItuBibliographicItem:0x007fe5409e5840
+...
+----
+
 === Create bibliographic item form YAML
 [source,ruby]
 ----

data/grammars/isodoc.rng

@@ -24,6 +24,14 @@
     <start>
       <ref name="standard-document"/>
     </start>
+    <define name="doctype">
+      <element name="doctype">
+        <optional>
+          <attribute name="abbreviation"/>
+        </optional>
+        <ref name="DocumentType"/>
+      </element>
+    </define>
     <define name="hyperlink">
       <element name="link">
         <attribute name="target">
@@ -141,6 +149,11 @@
             <data type="boolean"/>
           </attribute>
         </optional>
+        <optional>
+          <attribute name="key">
+            <data type="boolean"/>
+          </attribute>
+        </optional>
         <oneOrMore>
           <ref name="dt"/>
           <ref name="dd"/>
@@ -1164,49 +1177,7 @@
   </define>
   <define name="annex">
     <element name="annex">
-      <optional>
-        <attribute name="id">
-          <data type="ID"/>
-        </attribute>
-      </optional>
-      <optional>
-        <attribute name="language"/>
-      </optional>
-      <optional>
-        <attribute name="script"/>
-      </optional>
-      <optional>
-        <attribute name="inline-header">
-          <data type="boolean"/>
-        </attribute>
-      </optional>
-      <attribute name="obligation">
-        <choice>
-          <value>normative</value>
-          <value>informative</value>
-        </choice>
-      </attribute>
-      <optional>
-        <ref name="section-title"/>
-      </optional>
-      <group>
-        <group>
-          <zeroOrMore>
-            <ref name="BasicBlock"/>
-          </zeroOrMore>
-          <zeroOrMore>
-            <ref name="note"/>
-          </zeroOrMore>
-        </group>
-        <zeroOrMore>
-          <choice>
-            <ref name="annex-subsection"/>
-            <ref name="terms"/>
-            <ref name="definitions"/>
-            <ref name="references"/>
-          </choice>
-        </zeroOrMore>
-      </group>
+      <ref name="Annex-Section"/>
     </element>
   </define>
   <define name="terms">

data/grammars/itu.rng

@@ -41,6 +41,9 @@
         <value>technical-report</value>
         <value>joint-itu-iso-iec</value>
         <value>resolution</value>
+        <value>service-publication</value>
+        <value>handbook</value>
+        <value>question</value>
       </choice>
     </define>
     <define name="editorialgroup">
@@ -97,6 +100,9 @@
       <optional>
         <ref name="meeting"/>
       </optional>
+      <optional>
+        <ref name="meeting-place"/>
+      </optional>
       <optional>
         <ref name="meeting-date"/>
       </optional>
@@ -127,6 +133,14 @@
   </include>
   <define name="meeting">
     <element name="meeting">
+      <optional>
+        <attribute name="acronym"/>
+      </optional>
+      <text/>
+    </element>
+  </define>
+  <define name="meeting-place">
+    <element name="meeting-place">
       <text/>
     </element>
   </define>

data/lib/relaton_itu/hit_collection.rb

@@ -54,7 +54,7 @@ module RelatonItu
               "Selected" => false,
               "Value" => "",
               "Label" => "Name",
-              "Target" => "/name_s",
+              "Target" => "\\/name_s",
               "TypeName" => "CHECKBOX",
               "GetCriteriaType" => 0,
             },
@@ -62,7 +62,7 @@ module RelatonItu
               "Selected" => false,
               "Value" => "",
               "Label" => "Short description",
-              "Target" => "/short_description_s",
+              "Target" => "\\/short_description_s",
               "TypeName" => "CHECKBOX",
               "GetCriteriaType" => 0,
             },
@@ -70,7 +70,7 @@ module RelatonItu
               "Selected" => false,
               "Value" => "",
               "Label" => "File content",
-              "Target" => "/file",
+              "Target" => "\\/file",
               "TypeName" => "CHECKBOX",
               "GetCriteriaType" => 0,
             },
@@ -79,9 +79,8 @@ module RelatonItu
           "Selected" => false,
         }],
         "Topics" => "",
-        "ClientData" => { "ip" => "" },
+        "ClientData" => {},
         "Language" => "en",
-        "IP" => "",
         "SearchType" => "All",
       }
     end

data/lib/relaton_itu/itu_bibliographic_item.rb

@@ -4,7 +4,7 @@ module RelatonItu
       recommendation recommendation-supplement recommendation-amendment
       recommendation-corrigendum recommendation-errata recommendation-annex
       focus-group implementers-guide technical-paper technical-report
-      joint-itu-iso-iec
+      joint-itu-iso-iec resolution service-publication handbook question
     ].freeze
 
     # @params structuredidentifier [RelatonItu::StructuredIdentifier]

data/lib/relaton_itu/itu_bibliography.rb

@@ -30,7 +30,7 @@ module RelatonItu
       # @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 = {})
+      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?
@@ -50,7 +50,7 @@ module RelatonItu
 
       private
 
-      def fetch_ref_err(code, year, missed_years)
+      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."
@@ -67,14 +67,35 @@ module RelatonItu
         nil
       end
 
-      def search_filter(code)
-        docidrx = %r{\w+\.\d+|\w\sSuppl\.\s\d+} # %r{^ITU-T\s[^\s]+}
-        c = code.sub(/Imp\s?/, "").match(docidrx).to_s
+      def search_filter(code, year) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+        %r{
+          ^(?<pref1>ITU)?(-(?<type1>\w))?\s?(?<code1>[^\s\/]+)
+          (\s\(((?<month1>\d{2})\/)?(?<year1>\d{4})\))?
+          (\s-\s(?<buldate1>\d{2}\.\w{1,4}\.\d{4}))?
+          (\/(?<corr1>(Amd|Cor)\s?\d+))?
+          (\s\(((?<cormonth1>\d{2})\/)?(?<coryear1>\d{4})\))?
+        }x =~ code
+        year ||= year1
+        # docidrx = %r{\w+\.\d+|\w\sSuppl\.\s\d+} # %r{^ITU-T\s[^\s]+}
+        # c = code.sub(/Imp\s?/, "").match(docidrx).to_s
         warn "[relaton-itu] (\"#{code}\") fetching..."
         result = search(code)
+        code1.sub! /(?<=\.)Imp(?=\d)/, "" if result.gi_imp
         result.select do |i|
-          i.hit[:code] &&
-            i.hit[:code].match(docidrx).to_s == c
+          %r{
+            ^(?<pref2>ITU)?(-(?<type2>\w))?\s?(?<code2>[\S]+)
+            (\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) && code1 == code2 &&
+            (!year || year == year2) && (!month1 || month1 == month2) &&
+            corr1 == corr2 && (!coryear1 || coryear1 == coryear2) &&
+            buldate1 == buldate2 && (!cormonth1 || cormonth1 == cormonth2)
+          # i.hit[:code] &&
+          #   i.hit[:code].match(docidrx).to_s == c
         end
       end
 
@@ -86,24 +107,19 @@ module RelatonItu
       # If no match, returns any years which caused mismatch, for error reporting
       def isobib_results_filter(result, year)
         missed_years = []
-        # result.each_slice(3) do |s| # ISO website only allows 3 connections
-        #   fetch_pages(s, 3).each do |r|
         result.each do |r|
           return { ret: r.fetch } if !year
 
           /\(\d{2}\/(?<pyear>\d{4})\)/ =~ r.hit[:code]
-          # r.date.select { |d| d.type == "published" }.each do |d|
           return { ret: r.fetch } if year == pyear
 
           missed_years << pyear
-          # end
-          # end
         end
         { years: missed_years }
       end
 
       def itubib_get1(code, year, _opts)
-        result = search_filter(code) || return
+        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}"

data/lib/relaton_itu/scrapper.rb

@@ -44,7 +44,7 @@ module RelatonItu
         ItuBibliographicItem.new(
           fetched: Date.today.to_s,
           type: "standard",
-          docid: fetch_docid(doc),
+          docid: fetch_docid(doc, hit_data[:title]),
           edition: edition,
           language: ["en"],
           script: ["Latn"],
@@ -59,7 +59,7 @@ module RelatonItu
           copyright: fetch_copyright(hit_data[:code], doc),
           link: fetch_link(doc, url),
           relation: fetch_relations(doc),
-          place: ["Geneva"],
+          place: ["Geneva"]
         )
       end
       # rubocop:enable Metrics/AbcSize
@@ -92,7 +92,7 @@ module RelatonItu
         uri = URI url
         resp = Net::HTTP.get_response(uri)
         until resp.code == "200"
-          uri = URI resp["location"] if resp.code =~ /^30/
+          uri = URI resp["location"] if resp.code.match? /^30/
           resp = Net::HTTP.get_response(uri)
         end
         [uri.to_s, Nokogiri::HTML(resp.body)]
@@ -105,18 +105,32 @@ module RelatonItu
 
       # Fetch docid.
       # @param doc [Nokogiri::HTML::Document]
+      # @param title [String]
       # @return [Hash]
-      def fetch_docid(doc)
-        doc.xpath(
+      def fetch_docid(doc, title)
+        docids = doc.xpath(
           "//span[@id='ctl00_content_main_uc_rec_main_info1_rpt_main_ctl00_lbl_rec']",
           "//td[.='Identical standard:']/following-sibling::td",
-          "//div/table[1]/tr[4]/td/strong",
-        ).map do |code|
-          id = code.text.match(%r{^.*?(?= \()|\w\.Imp\s?\d+}).to_s.squeeze(" ")
-          type = id.match(%r{^\w+}).to_s
-          type = "ITU" if type == "G"
-          RelatonBib::DocumentIdentifier.new(type: type, id: id)
-        end
+          "//div/table[1]/tr[4]/td/strong"
+        ).map { |c| createdocid c.text }
+        docids << createdocid(title) unless docids.any?
+        docids
+      end
+
+      def createdocid(text)
+        %r{
+          ^(?<code>((ITU-\w|ISO\/IEC)\s)?[^\(:]+)
+          (\(((?<month>\d{2})\/)?(?<year>\d{4})\))?
+          (:[^\(]+\((?<buldate>\d{2}\.\w{1,4}\.\d{4})\))?
+          (\s(?<corr>(Amd|Cor)\.\s?\d+))?
+          # (\s\(((?<cormonth>\d{2})\/)?(?<coryear>\d{4})\))?
+        }x =~ text.squeeze(" ")
+        corr&.sub! /\.\s?/, " "
+        id = [code.sub(/[[:space:]]$/, ""), corr].compact.join "/"
+        id += " - #{buldate}" if buldate
+        type = id.match(%r{^\w+}).to_s
+        type = "ITU" if type == "G"
+        RelatonBib::DocumentIdentifier.new(type: type, id: id)
       end
 
       # Fetch status.
@@ -142,13 +156,13 @@ module RelatonItu
         group = wg && itugroup(wg.text)
         EditorialGroup.new(
           bureau: code.match(/(?<=-)./).to_s,
-          group: group,
+          group: group
         )
       end
 
       # @param name [String]
       # @return [RelatonItu::ItuGroup]
-      def itugroup(name)
+      def itugroup(name) # rubocop:disable Metrics/MethodLength
         if name.include? "Study Group"
           type = "study-group"
           acronym = "SG"
@@ -190,7 +204,7 @@ module RelatonItu
       # Fetch dates
       # @param doc [Nokogiri::HTML::Document]
       # @return [Array<Hash>]
-      def fetch_dates(doc)
+      def fetch_dates(doc) # rubocop:disable Metrics/CyclomaticComplexity
         dates = []
         date = doc.at("//table/tr/td/span[contains(@id, 'Label5')]",
                       "//p[contains(.,'Approved in')]")
@@ -244,7 +258,7 @@ module RelatonItu
         links = [{ type: "src", content: url }]
         obp_elm = doc.at(
           '//a[@title="Persistent link to download the PDF file"]',
-          "//font[contains(.,'PDF')]/../..",
+          "//font[contains(.,'PDF')]/../.."
         )
         links << typed_link("obp", obp_elm) if obp_elm
         wrd_elm = doc.at("//font[contains(.,'Word')]/../..")

data/lib/relaton_itu/version.rb

@@ -1,3 +1,3 @@
 module RelatonItu
-  VERSION = "1.6.pre1".freeze
+  VERSION = "1.7.3".freeze
 end

data/relaton-itu.gemspec

@@ -37,5 +37,5 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "vcr", "~> 5.0.0"
   spec.add_development_dependency "webmock"
 
-  spec.add_dependency "relaton-bib", "~> 1.6.pre"
+  spec.add_dependency "relaton-bib", "~> 1.7.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-itu
 version: !ruby/object:Gem::Version
-  version: 1.6.pre1
+  version: 1.7.3
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2020-11-12 00:00:00.000000000 Z
+date: 2021-01-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: debase
@@ -156,14 +156,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.6.pre
+        version: 1.7.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.6.pre
+        version: 1.7.0
 description: 'RelatonItu: retrieve ITU Standards for bibliographic use using the BibliographicItem
   model'
 email:
@@ -218,9 +218,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: 2.4.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">"
+  - - ">="
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubygems_version: 3.0.6
 signing_key: