relaton-iso 1.15.3 → 1.15.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e27efa3183e068be1eb194fb2370fb91323c43435a6bfa28a3984fc5ee8f087d
-  data.tar.gz: a571f80e6fab8d18ead5b7f54666e16624a11b6c89637f3b7936a316ae25ee33
+  metadata.gz: d8f96ec7677b17a7e89ca54cc7230eda13391749c404fcc8d95bca32ddb61d98
+  data.tar.gz: c74831929f8d4d276e1aae6c06990559e51bcc4ba1ea3c627fef33d141463612
 SHA512:
-  metadata.gz: 9225503e6deb080fc61d003df621e46ec5ba2435dd9c201f2771345c468d269a45f89690cb77bda27c8a70e2211d695d53942d3ac192c3b48ae0e9ffabad86ec
-  data.tar.gz: 6cc63106785022a14c4bb48f9e2263df8d6e25716af94905e1b91a79ebea9a3014cdcdf4f2ae1d3172ba3ab34b22c1583bc2f7e84bc95308041e81bbfa3a95c6
+  metadata.gz: 80862ee504de862925e70c188ae40ed13001748f87f9af41aaad00e1e3e2b473779da64e0402ac920ac84af6318e19fe8ec6d73c7d6ff5b752b20f3fd0d2721c
+  data.tar.gz: e896e1986545b722df2393b075ce947d2ab00ca365249498b4a0e738ebcb9669475e72961e2ab90495181ce8ccd5af8df380d697b9433b335e013ab554ddf23a

data/.gitignore

@@ -6,6 +6,7 @@
 /pkg/
 /spec/reports/
 /tmp/
+/vendor/
 
 # rspec failure tracking
 .rspec_status

data/Gemfile

@@ -6,7 +6,10 @@ git_source(:github) { |repo_name| "https://github.com/#{repo_name}" }
 gemspec
 
 gem "byebug"
+gem "equivalent-xml"
 gem "pry-byebug"
+gem "rake", "~> 13.0"
+gem "rspec", "~> 3.0"
 gem "rubocop"
 gem "rubocop-performance"
 gem "rubocop-rails"

data/README.adoc

@@ -49,37 +49,11 @@ item = hit_collection[2].fetch
 ...
 
 item.docidentifier
-=> [#<RelatonIso::DocumentIdentifier:0x00007fa8b7db1f20
-  @id=
-   #<Pubid::Iso::Identifier:0x00007fa8b7db2a60
-    @base=#<Pubid::Iso::Identifier:0x00007fa8b7db2c68 @edition="1", @number="19115"@4, @part="1", @publisher="ISO", @year=2014>,
-    @number="2"@21,
-    @publisher="ISO",
-    @stage=#<Pubid::Iso::Stage:0x00007fa8b7db20d8 @abbr=nil, @harmonized_code=#<Pubid::Iso::HarmonizedStageCode:0x00007fa8b7db2088 @stages=["60.60"]>>,
-    @typed_stage=#<Pubid::Iso::TypedStage:0x00007fa8b7db2920 @type=#<Pubid::Iso::Type:0x00007fa8b7db26c8 @type=:amd>, @typed_stage=nil>,
-    @year=2020>,
-  @language=nil,
-  @primary=true,
-  @scope=nil,
-  @script=nil,
-  @type="ISO">,
- #<RelatonIso::DocumentIdentifier:0x00007fa8b7db1e80
-  @id=
-   #<Pubid::Iso::Identifier:0x00007fa8b7db2a60
-    @base=#<Pubid::Iso::Identifier:0x00007fa8b7db2c68 @edition="1", @number="19115"@4, @part="1", @publisher="ISO", @year=2014>,
-    @number="2"@21,
-    @publisher="ISO",
-    @stage=#<Pubid::Iso::Stage:0x00007fa8b7db20d8 @abbr=nil, @harmonized_code=#<Pubid::Iso::HarmonizedStageCode:0x00007fa8b7db2088 @stages=["60.60"]>>,
-    @typed_stage=#<Pubid::Iso::TypedStage:0x00007fa8b7db2920 @type=#<Pubid::Iso::Type:0x00007fa8b7db26c8 @type=:amd>, @typed_stage=nil>,
-    @year=2020>,
-  @language=nil,
-  @primary=nil,
-  @scope=nil,
-  @script=nil,
-  @type="URN">]
+=> [#<RelatonIso::DocumentIdentifier:0x0000000112a23a88
+...
 
 item.docidentifier.detect { |di| di.type == "URN" }.id
-=> "urn:iso:std:iso:19115:-1:ed-1:amd:2020:v2"
+=> "urn:iso:std:iso:19115:-1:ed-1:stage-60.60:amd:2020:v2"
 ----
 
 === Fetch document by reference and year
@@ -93,9 +67,9 @@ item = RelatonIso::IsoBibliography.get "ISO 19115:2003"
 ...
 
 item = RelatonIso::IsoBibliography.get "ISO 19115", "2003"
-[relaton-iso] ("ISO 19115") fetching from ISO...
-[relaton-iso] ("ISO 19115:2003") Found exact match.
-=> #<RelatonIsoBib::IsoBibliographicItem:0x00007f8c828d3180
+[relaton-iso] ("ISO 19115:2003") Fetching from ISO...
+[relaton-iso] ("ISO 19115:2003") Found ("ISO 19115:2003").
+=> #<RelatonIsoBib::IsoBibliographicItem:0x0000000112c9ca80
 ...
 
 item.docidentifier[0].id
@@ -113,7 +87,7 @@ item = RelatonIso::IsoBibliography.get "ISO 19115"
 ...
 
 item.docidentifier[0].id
-=> "ISO 19115:2003"
+=> "ISO 19115"
 ----
 
 === Fetch a part document
@@ -136,13 +110,13 @@ item.docidentifier[0].id
 ----
 item = RelatonIso::IsoBibliography.get "ISO 19115 (all parts)"
 [relaton-iso] ("ISO 19115") Fetching from ISO...
-[relaton-iso] ("ISO 19115") Found exact match.
+[relaton-iso] ("ISO 19115") Found ("ISO 19115").
 => #<RelatonIsoBib::IsoBibliographicItem:0x00007f8ca216e118
 ...
 
 item = RelatonIso::IsoBibliography.get "ISO 19115", nil, all_parts: true
 [relaton-iso] ("ISO 19115") Fetching from ISO...
-[relaton-iso] ("ISO 19115") Found exact match.
+[relaton-iso] ("ISO 19115") Found ("ISO 19115").
 => #<RelatonIsoBib::IsoBibliographicItem:0x00007f8c830f3d38
 ...
 
@@ -151,12 +125,12 @@ item.docidentifier[0].id
 
 item = RelatonIso::IsoBibliography.get "ISO 19115-1 (all parts)"
 [relaton-iso] ("ISO 19115") Fetching from ISO...
-[relaton-iso] ("ISO 19115") Found exact match.
+[relaton-iso] ("ISO 19115") Found ("ISO 19115").
 => #<RelatonIsoBib::IsoBibliographicItem:0x00007f8c8290e5a0
 
 item = RelatonIso::IsoBibliography.get "ISO 19115-1", nil, all_parts: true
 [relaton-iso] ("ISO 19115") Fetching from ISO...
-[relaton-iso] ("ISO 19115") Found exact match.
+[relaton-iso] ("ISO 19115") Found ("ISO 19115").
 => #<RelatonIsoBib::IsoBibliographicItem:0x00007f8c925355b8
 ...
 
@@ -217,35 +191,31 @@ item.to_xml note: [{ text: "Note", type: "note" }]
 [source,ruby]
 ----
 item.title lang: 'en'
-=> #<RelatonBib::TypedTitleStringCollection:0x00007fa8b7f47c90
+=> #<RelatonBib::TypedTitleStringCollection:0x0000000112783fd0
  @array=
-  [#<RelatonBib::TypedTitleString:0x00007fa8b7be8a40
-    @title=#<RelatonBib::FormattedString:0x00007fa8b7be87c0 @content="Geographic information", @format="text/plain", @language=["en"], @script=["Latn"]>,
+  [#<RelatonBib::TypedTitleString:0x00000001138e2380
+    @title=#<RelatonBib::FormattedString:0x0000000112d496b8 @content="Geographic information", @format="text/plain", @language=["en"], @script=["Latn"]>,
     @type="title-intro">,
-   #<RelatonBib::TypedTitleString:0x00007fa8b7be84a0
-    @title=#<RelatonBib::FormattedString:0x00007fa8b7be83d8 @content="Metadata", @format="text/plain", @language=["en"], @script=["Latn"]>,
+   #<RelatonBib::TypedTitleString:0x00000001138e1f70
+    @title=#<RelatonBib::FormattedString:0x0000000112d495c8 @content="Metadata", @format="text/plain", @language=["en"], @script=["Latn"]>,
     @type="title-main">,
-   #<RelatonBib::TypedTitleString:0x00007fa8b7be8130
+   #<RelatonBib::TypedTitleString:0x00000001138e1d68
     @title=
-     #<RelatonBib::FormattedString:0x00007fa8b7be8018
-      @content="Geographic information – Metadata",
-      @format="text/plain",
-      @language=["en"],
-      @script=["Latn"]>,
+     #<RelatonBib::FormattedString:0x0000000112d49488 @content="Geographic information – Metadata", @format="text/plain", @language=["en"], @script=["Latn"]>,
     @type="main">]>
 
 item.title lang: 'fr'
-=> #<RelatonBib::TypedTitleStringCollection:0x00007fa8b7bcb4b8
+=> #<RelatonBib::TypedTitleStringCollection:0x0000000113067458
  @array=
-  [#<RelatonBib::TypedTitleString:0x00007fa8b7be3ea0
-    @title=#<RelatonBib::FormattedString:0x00007fa8b7be3dd8 @content="Information géographique", @format="text/plain", @language=["fr"], @script=["Latn"]>,
+  [#<RelatonBib::TypedTitleString:0x00000001138e1c28
+    @title=#<RelatonBib::FormattedString:0x0000000112d49438 @content="Information géographique", @format="text/plain", @language=["fr"], @script=["Latn"]>,
     @type="title-intro">,
-   #<RelatonBib::TypedTitleString:0x00007fa8b7be3ba8
-    @title=#<RelatonBib::FormattedString:0x00007fa8b7be3b58 @content="Métadonnées", @format="text/plain", @language=["fr"], @script=["Latn"]>,
+   #<RelatonBib::TypedTitleString:0x00000001138e1b10
+    @title=#<RelatonBib::FormattedString:0x0000000112d49398 @content="Métadonnées", @format="text/plain", @language=["fr"], @script=["Latn"]>,
     @type="title-main">,
-   #<RelatonBib::TypedTitleString:0x00007fa8b7be38d8
+   #<RelatonBib::TypedTitleString:0x00000001138e1908
     @title=
-     #<RelatonBib::FormattedString:0x00007fa8b7be3860
+     #<RelatonBib::FormattedString:0x0000000112d491b8
       @content="Information géographique – Métadonnées",
       @format="text/plain",
       @language=["fr"],
@@ -254,10 +224,9 @@ item.title lang: 'fr'
 
 item = RelatonIso::IsoBibliography.get "ISO 19115:2003"
 [relaton-iso] ("ISO 19115:2003") Fetching from ISO...
-[relaton-iso] ("ISO 19115:2003") Found exact match.
+[relaton-iso] ("ISO 19115:2003") Found ("ISO 19115:2003").
 => #<RelatonIsoBib::IsoBibliographicItem:0x00007fa8870b69e0
 
-item.abstract lang: 'en'
 item.abstract lang: 'en'
 => #<RelatonBib::FormattedString:0x00007fa8870b4f78
  @content=
@@ -274,13 +243,13 @@ Each ISO document has `src` type link and optional `obp`, `rss`, and `pub` link
 [source,ruby]
 ----
 item.link
-=> [#<RelatonBib::TypedUri:0x00007fa8870ac5a8
-  @content=#<Addressable::URI:0xcc6514 URI:https://www.iso.org/standard/26020.html>,
+=> [#<RelatonBib::TypedUri:0x0000000112d66c40
+  @content=#<Addressable::URI:0x93d71c URI:https://www.iso.org/standard/26020.html>,
   @language=nil,
   @script=nil,
   @type="src">,
- #<RelatonBib::TypedUri:0x00007fa8870a7490
-  @content=#<Addressable::URI:0xcc6528 URI:https://www.iso.org/contents/data/standard/02/60/26020.detail.rss>,
+ #<RelatonBib::TypedUri:0x0000000112d66920
+  @content=#<Addressable::URI:0x93d730 URI:https://www.iso.org/contents/data/standard/02/60/26020.detail.rss>,
   @language=nil,
   @script=nil,
   @type="rss">]

data/lib/relaton_iso/config.rb

@@ -0,0 +1,26 @@
+module RelatonIso
+  module Config
+    def configure
+      yield configuration if block_given?
+    end
+
+    def configuration
+      @configuration ||= Configuration.new
+    end
+
+    extend self
+  end
+
+  class Configuration
+    attr_accessor :logger
+
+    def initialize
+      @logger = ::Logger.new $stderr
+      @logger.level = ::Logger::WARN
+      @logger.progname = "relaton-iso"
+      @logger.formatter = proc do |_severity, _datetime, progname, msg|
+        "[#{progname}] #{msg}\n"
+      end
+    end
+  end
+end

data/lib/relaton_iso/hit.rb

@@ -42,9 +42,9 @@ module RelatonIso
     # @return [Pubid::Iso::Identifier]
     def pubid
       @pubid ||= Pubid::Iso::Identifier.parse_from_title(hit[:title])
-    rescue Pubid::Iso::Errors::WrongTypeError => e
-      warn "[relaton-iso] unable to find an identifier in \"#{hit[:title]}\"."
-      warn "[relaton-iso] #{e.message}"
+    rescue Pubid::Iso::Errors::WrongTypeError, Pubid::Iso::Errors::ParseError => e
+      Logger.warn "unable to find an identifier in \"#{hit[:title]}\"."
+      Logger.warn e.message
     end
   end
 end

data/lib/relaton_iso/hit_collection.rb

@@ -6,17 +6,19 @@ require "relaton_iso/hit"
 module RelatonIso
   # Page of hit collection.
   class HitCollection < RelatonBib::HitCollection
-    # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
+    # @return [Boolean] whether the search was performed on GitHub
+    attr_reader :from_gh
 
     # @param text [String] reference to search
     def initialize(text)
       super
-      @array = text.match?(/^ISO[\s\/](?:TC\s184\/SC\s?4|IEC\sDIR\s(?:\d|IEC|JTC))/) ? fetch_github : fetch_iso
+      @from_gh = text.match?(/^ISO[\s\/](?:TC\s184\/SC\s?4|IEC\sDIR\s(?:\d|IEC|JTC))/)
+      @array = from_gh ? fetch_github : fetch_iso
     end
 
     # @param lang [String, NilClass]
     # @return [RelatonIsoBib::IsoBibliographicItem, nil]
-    def to_all_parts(lang = nil) # rubocop:disable Metrics/CyclomaticComplexity
+    def to_all_parts(lang = nil) # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
       # parts = @array.reject { |h| h.hit["docPart"]&.empty? }
       hit = @array.min_by { |h| h.pubid.part.to_i }
       return @array.first&.fetch lang unless hit
@@ -33,7 +35,6 @@ module RelatonIso
       end
       all_parts_item
     end
-    # rubocop:enable Metrics/AbcSize, Metrics/MethodLength
 
     private
 
@@ -63,23 +64,11 @@ module RelatonIso
     # @return [Array<RelatonIso::Hit>]
     #
     def fetch_iso # rubocop:disable Metrics/AbcSize,Metrics/MethodLength
-      # %r{\s(?<num>\d+)(?:-(?<part>[\d-]+))?} =~ text
-      # http = Net::HTTP.new "www.iso.org", 443
-      # http.use_ssl = true
-      # search = ["status=ENT_ACTIVE,ENT_PROGRESS,ENT_INACTIVE,ENT_DELETED"]
-      # search << "docNumber=#{num}"
-      # search << "docPartNo=#{part}" if part
-      # q = search.join "&"
-      # resp = http.get("/cms/render/live/en/sites/isoorg.advancedSearch.do?#{q}",
-      #                 "Accept" => "application/json, text/plain, */*")
       config = Algolia::Search::Config.new(application_id: "JCL49WV5AR", api_key: "dd1b9e1ab383f4d4817d29cd5e96d3f0")
-      client = Algolia::Search::Client.new config, logger: ::Logger.new($stderr)
+      client = Algolia::Search::Client.new config, logger: Config.configuration.logger
       index = client.init_index "all_en"
       resp = index.search text, hitsPerPage: 100, filters: "category:standard"
-      # return [] if resp.body.empty?
 
-      # json = JSON.parse resp.body
-      # json["standards"]
       resp[:hits].map { |h| Hit.new h, self }.sort! do |a, b|
         if a.sort_weight == b.sort_weight && b.hit[:year] = a.hit[:year]
           a.hit[:title] <=> b.hit[:title]
@@ -90,19 +79,5 @@ module RelatonIso
         end
       end
     end
-
-    # @param hit [Hash]
-    # @return [Date]
-    # def parse_date(hit)
-    #   if hit["publicationDate"]
-    #     Date.strptime(hit["publicationDate"], "%Y-%m")
-    #   elsif %r{:(?<year>\d{4})} =~ hit["docRef"]
-    #     Date.strptime(year, "%Y")
-    #   elsif hit["newProjectDate"]
-    #     Date.parse hit["newProjectDate"]
-    #   else
-    #     Date.new 0
-    #   end
-    # end
   end
 end

data/lib/relaton_iso/iso_bibliography.rb

@@ -12,7 +12,7 @@ module RelatonIso
       # @param text [String]
       # @return [RelatonIso::HitCollection]
       def search(text)
-        HitCollection.new text.gsub(/\u2013/, "-")
+        HitCollection.new text.gsub("\u2013", "-")
       rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET,
              EOFError, Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError,
              Net::ProtocolError, OpenSSL::SSL::SSLError, Errno::ETIMEDOUT,
@@ -29,7 +29,7 @@ module RelatonIso
       #
       # @return [RelatonIsoBib::IsoBibliographicItem] Relaton XML serialisation of reference
       def get(ref, year = nil, opts = {}) # rubocop:disable Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity,Metrics/AbcSize
-        code = ref.gsub(/\u2013/, "-")
+        code = ref.gsub("\u2013", "-")
 
         # parse "all parts" request
         code.sub! " (all parts)", ""
@@ -37,43 +37,24 @@ module RelatonIso
 
         query_pubid = Pubid::Iso::Identifier.parse(code)
         query_pubid.year = year if year
+        query_pubid.part = nil if opts[:all_parts]
+        Logger.warn "(\"#{query_pubid}\") Fetching from ISO..."
 
-        resp = isobib_search_filter(query_pubid, opts)
-
-        # Try with ISO/IEC prefix if ISO not found
-        if resp[:hits].empty? && query_pubid.copublisher.nil? &&
-            query_pubid.publisher == "ISO"
-          resp_isoiec = retry_isoiec_prefix(query_pubid, opts)
-          resp = resp_isoiec unless resp_isoiec.nil?
-        end
+        hits, missed_year_ids = isobib_search_filter(query_pubid, opts)
+        tip_ids = look_up_with_any_types_stages(hits, query_pubid, opts)
 
-        # return only first one if not all_parts
-        ret = if !opts[:all_parts] || resp[:hits].size == 1
-                resp[:hits].any? && resp[:hits].first.fetch(opts[:lang])
+        ret = if !opts[:all_parts] || hits.size == 1
+                hits.any? && hits.first.fetch(opts[:lang])
               else
-                resp[:hits].to_all_parts(opts[:lang])
+                hits.to_all_parts(opts[:lang])
               end
 
-        return fetch_ref_err(query_pubid) unless ret
+        return fetch_ref_err(query_pubid, missed_year_ids, tip_ids) unless ret
 
-        # puts "xxxxx #{ret.docidentifier.first.id.inspect}"
         response_docid = ret.docidentifier.first.id.sub(" (all parts)", "")
         response_pubid = Pubid::Iso::Identifier.parse(response_docid)
-        # puts "xxxxx query_pubid(#{query_pubid}) response_pubid(#{response_pubid})"
-
-        if query_pubid.to_s == response_pubid.to_s
-          warn "[relaton-iso] (\"#{query_pubid}\") Found exact match."
-        elsif matches_base?(query_pubid, response_pubid)
-          warn "[relaton-iso] (\"#{query_pubid}\") " \
-               "Found (\"#{response_pubid}\")."
-        elsif matches_base?(query_pubid, response_pubid, any_types_stages: true)
-          warn "[relaton-iso] (\"#{query_pubid}\") TIP: " \
-               "Found with different type/stage, " \
-               "please amend to (\"#{response_pubid}\")."
-        else
-          # when there are all parts
-          warn "[relaton-iso] (\"#{query_pubid}\") Found (\"#{response_pubid}\")."
-        end
+
+        Logger.warn "(\"#{query_pubid}\") Found (\"#{response_pubid}\")."
 
         get_all = (
           (query_pubid.year && opts[:keep_year].nil?) ||
@@ -83,9 +64,9 @@ module RelatonIso
         return ret if get_all
 
         ret.to_most_recent_reference
-
       rescue Pubid::Core::Errors::ParseError
-        warn "[relaton-iso] (\"#{code}\") is not recognized as a standards identifier."
+        Logger.warn "(\"#{code}\") is not recognized as a standards identifier."
+        nil
       end
 
       # @param query_pubid [Pubid::Iso::Identifier]
@@ -109,147 +90,111 @@ module RelatonIso
       # @return [<Type>] <description>
       #
       def matches_base?(query_pubid, pubid, any_types_stages: false) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics?PerceivedComplexity
-        return unless pubid.respond_to?(:publisher)
+        return false unless pubid.respond_to?(:publisher)
 
         query_pubid.publisher == pubid.publisher &&
           query_pubid.number == pubid.number &&
           query_pubid.copublisher == pubid.copublisher &&
-          ((any_types_stages && query_pubid.stage.nil?) || query_pubid.stage == pubid.stage) &&
-          ((any_types_stages && query_pubid.type.nil?) || query_pubid.type == pubid.type)
+          (any_types_stages || query_pubid.stage == pubid.stage) &&
+          (any_types_stages || query_pubid.is_a?(pubid.class))
       end
 
       # @param hit_collection [RelatonIso::HitCollection]
       # @param year [String]
-      # @return [RelatonIso::HitCollection]
-      def filter_hits_by_year(hit_collection, year) # rubocop:disable Metrics/MethodLength,Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
-        missed_years = []
-        return { hits: hit_collection, missed_years: missed_years } if year.nil?
+      # @return [Array<RelatonIso::HitCollection, Array<String>>] hits and missed year IDs
+      def filter_hits_by_year(hit_collection, year)
+        missed_year_ids = Set.new
+        return [hit_collection, missed_year_ids] if year.nil?
 
         # filter by year
         hits = hit_collection.select do |hit|
-          if (hit.pubid.base.nil? && hit.pubid.year.to_s == year.to_s) ||
-              (!hit.pubid.base.nil? && hit.pubid.base.year.to_s == year.to_s) ||
-              (!hit.pubid.base.nil? && hit.pubid.year.to_s == year.to_s)
-            true
-          elsif hit.pubid.year.nil? && hit.hit[:year].to_s == year
-            hit.pubid.year = year
-            true
-          else
-            missed_year = (hit.pubid.year || hit.hit[:year]).to_s
-            if missed_year && !missed_year.empty? && !missed_years.include?(missed_year)
-              missed_years << missed_year
-            end
-            false
-          end
+          hit.pubid.year ||= hit.hit[:year]
+          next true if check_year(year, hit)
+
+          missed_year_ids << hit.pubid.to_s if hit.pubid.year
+          false
         end
 
-        { hits: hits, missed_years: missed_years }
+        [hits, missed_year_ids]
       end
 
       private
 
-      # @param query_pubid [Pubid::Iso::Identifier] PubID with no results
-      def fetch_ref_err(query_pubid) # rubocop:disable Metrics/MethodLength
-        warn "[relaton-iso] (\"#{query_pubid}\") " \
-             "Not found. " \
-             "The identifier must be exactly as shown on the ISO website."
+      def check_year(year, hit) # rubocop:disable Metrics/AbcSize
+        (hit.pubid.base.nil? && hit.pubid.year.to_s == year.to_s) ||
+          (!hit.pubid.base.nil? && hit.pubid.base.year.to_s == year.to_s) ||
+          (!hit.pubid.base.nil? && hit.pubid.year.to_s == year.to_s)
+      end
 
-        if query_pubid.part
-          warn "[relaton-iso] (\"#{query_pubid}\") TIP: " \
-               "If it cannot be found, the document may no longer be published in parts."
-        else
-          warn "[relaton-iso] (\"#{query_pubid}\") TIP: " \
-               "If you wish to cite all document parts for the reference, " \
-               "use (\"#{query_pubid.to_s(format: :ref_undated)} (all parts)\")."
+      # @param pubid [Pubid::Iso::Identifier] PubID with no results
+      def fetch_ref_err(pubid, missed_year_ids, tip_ids) # rubocop:disable Metrics/MethodLength,Metrics/AbcSize
+        Logger.warn "(\"#{pubid}\") Not found."
+
+        if missed_year_ids.any?
+          ids = missed_year_ids.map { |i| "\"#{i}\"" }.join(", ")
+          Logger.warn "(\"#{pubid}\") TIP: No match for edition year " \
+                      "#{pubid.year}, but matches exist for #{ids}."
         end
 
-        unless %w(TS TR PAS Guide).include?(query_pubid.type)
-          warn "[relaton-iso] (\"#{query_pubid}\") TIP: " \
-               "If the document is not an International Standard, use its " \
-               "deliverable type abbreviation (TS, TR, PAS, Guide)."
+        if tip_ids.any?
+          ids = tip_ids.map { |i| "\"#{i}\"" }.join(", ")
+          Logger.warn "(\"#{pubid}\") TIP: Matches exist for #{ids}."
         end
 
-        nil
-      end
+        if pubid.part
+          Logger.warn "(\"#{pubid}\") TIP: If it cannot be found, " \
+                      "the document may no longer be published in parts."
+        else
+          Logger.warn "(\"#{pubid}\") TIP: If you wish to cite " \
+                      "all document parts for the reference, use " \
+                      "(\"#{pubid.to_s(format: :ref_undated)} (all parts)\")."
+        end
 
-      # @param pubid [Pubid::Iso::Identifier]
-      # @param missed_years [Array<String>]
-      def warn_missing_years(pubid, missed_years)
-        warn "[relaton-iso] (\"#{pubid}\") TIP: " \
-             "No match for edition year #{pubid.year}, " \
-             "but matches exist for #{missed_years.uniq.join(', ')}."
+        nil
       end
 
-      # Search for hits using ISO/IEC prefix.
-      #
-      # @param old_pubid [Pubid::Iso::Identifier] reference with ISO prefix
-      # @param opts [Hash]
-      # @return [Array<RelatonIso::Hit>]
-      def retry_isoiec_prefix(old_pubid, opts) # rubocop:disable Metrics/MethodLength
-        return nil unless old_pubid.copublisher.nil? && old_pubid.publisher == "ISO"
-
-        pubid = old_pubid.dup
-        pubid.copublisher = "IEC"
-        warn "[relaton-iso] (\"#{old_pubid}\") Not found, trying with ISO/IEC prefix (\"#{pubid}\")..."
-        resp_isoiec = isobib_search_filter(pubid, opts)
-
-        if resp_isoiec[:hits].empty?
-          warn "[relaton-iso] (\"#{pubid}\") Not found. "
-          return nil
-        end
-
-        warn "[relaton-iso] (\"#{pubid}\") TIP: Found with ISO/IEC prefix, " \
-             "please amend to (\"#{pubid}\")."
+      def look_up_with_any_types_stages(hits, pubid, opts) # rubocop:disable Metrics/MethodLength
+        found_ids = []
+        return found_ids unless !hits.from_gh && hits.empty? && pubid.copublisher.nil?
 
-        resp_isoiec
+        resp, = isobib_search_filter(pubid, opts, any_types_stages: true)
+        resp.map &:pubid
       end
 
+      #
       # Search for hits. If no found then trying missed stages.
       #
       # @param query_pubid [Pubid::Iso::Identifier] reference without correction
       # @param opts [Hash]
-      # @return [Array<RelatonIso::Hit>]
-      def isobib_search_filter(query_pubid, opts) # rubocop:disable Metrics/MethodLength,Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
-        missed_years = []
-        query_pubid.part = nil if opts[:all_parts]
-        warn "[relaton-iso] (\"#{query_pubid}\") Fetching from ISO..."
-
-        # fetch hits collection
+      # @param any_types_stages [Boolean] match with any stages
+      #
+      # @return [Array<RelatonIso::HitCollection, Array<String>>] hits and missed years
+      #
+      def isobib_search_filter(query_pubid, opts, any_types_stages: false) # rubocop:disable Metrics/MethodLength,Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
         query_pubid_without_year = query_pubid.dup
         # remove year for query
         query_pubid_without_year.year = nil
         hit_collection = search(query_pubid_without_year.to_s)
 
         # filter only matching hits
-        res = filter_hits hit_collection, query_pubid, all_parts: opts[:all_parts]
-        return res unless res[:hits].empty?
-
-        missed_years += res[:missed_years]
-
-        # lookup for documents with stages when no match without stage
-        res = filter_hits hit_collection, query_pubid,
-                          all_parts: opts[:all_parts], any_types_stages: true
-        return res unless res[:hits].empty?
-
-        missed_years += res[:missed_years]
-
-        if missed_years.any?
-          warn_missing_years(query_pubid, missed_years)
-        end
-
-        res
+        filter_hits hit_collection, query_pubid, opts[:all_parts], any_types_stages
       end
 
-      # @param hits [RelatonIso::HitCollection]
+      #
+      # Filter hits by query_pubid.
+      #
+      # @param hit_collection [RelatonIso::HitCollection]
       # @param query_pubid [Pubid::Iso::Identifier]
       # @param all_parts [Boolean]
-      # @param any_stages [Boolean]
-      # @return [RelatonIso::HitCollection]
-      def filter_hits(hit_collection, query_pubid, all_parts: false, any_types_stages: false) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
+      # @param any_stypes_tages [Boolean]
+      #
+      # @return [Array<RelatonIso::HitCollection, Array<String>>] hits and missed year IDs
+      #
+      def filter_hits(hit_collection, query_pubid, all_parts, any_stypes_tages) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/PerceivedComplexity
         # filter out
         result = hit_collection.select do |i|
           hit_pubid = i.pubid
-          matches_base?(query_pubid, hit_pubid, any_types_stages: any_types_stages) &&
+          matches_base?(query_pubid, hit_pubid, any_types_stages: any_stypes_tages) &&
             matches_parts?(query_pubid, hit_pubid, all_parts: all_parts) &&
             query_pubid.corrigendums == hit_pubid.corrigendums &&
             query_pubid.amendments == hit_pubid.amendments

data/lib/relaton_iso/logger.rb

@@ -0,0 +1,13 @@
+module RelatonIso
+  module Logger
+    extend self
+
+    def method_missing(method, *args, &block)
+      Config.configuration.logger.send(method, *args, &block)
+    end
+
+    def respond_to_missing?(method_name, include_private = false)
+      Config.configuration.logger.respond_to?(method_name) || super
+    end
+  end
+end

data/lib/relaton_iso/scrapper.rb

@@ -215,7 +215,7 @@ module RelatonIso
         10.times do
           doc = Nokogiri::HTML(resp.body)
           # stop trying if page has a document id
-          return doc if doc.at("//main/div/section/div/div/div/nav/h1")
+          return doc if item_ref doc
 
           resp = Net::HTTP.get_response(uri)
         end
@@ -249,7 +249,7 @@ module RelatonIso
       end
 
       def item_ref(doc)
-        doc.at("//nav[contains(@class, 'heading-condensed')]/h1")&.text
+        doc.at("//main//section/div/div/div//h1")&.text
       end
 
       # Fetch status.

data/lib/relaton_iso/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonIso
-  VERSION = "1.15.3"
+  VERSION = "1.15.5"
 end

data/lib/relaton_iso.rb

@@ -1,6 +1,9 @@
 # frozen_string_literal: true
 
+require "logger"
+require "pubid-iso"
+require "relaton_iso/config"
+require "relaton_iso/logger"
 require "relaton_iso/version"
 require "relaton_iso/iso_bibliography"
-require "pubid-iso"
 require "relaton_iso/document_identifier"

data/relaton_iso.gemspec

@@ -10,10 +10,10 @@ Gem::Specification.new do |spec|
   spec.authors       = ["Ribose Inc."]
   spec.email         = ["open.source@ribose.com"]
 
-  spec.summary       = "RelatonIso: retrieve ISO Standards for bibliographic use "\
-                       "using the IsoBibliographicItem model"
-  spec.description   = "RelatonIso: retrieve ISO Standards for bibliographic use "\
-                       "using the IsoBibliographicItem model"
+  spec.summary       = "RelatonIso: retrieve ISO Standards for bibliographic " \
+                       "use using the IsoBibliographicItem model"
+  spec.description   = "RelatonIso: retrieve ISO Standards for bibliographic " \
+                       "use using the IsoBibliographicItem model"
 
   spec.homepage      = "https://github.com/relaton/relaton-iso"
   spec.license       = "BSD-2-Clause"
@@ -26,11 +26,9 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
   spec.required_ruby_version = Gem::Requirement.new(">= 2.7.0")
 
-  spec.add_development_dependency "equivalent-xml", "~> 0.6"
-  spec.add_development_dependency "rake", "~> 13.0"
-  spec.add_development_dependency "rspec", "~> 3.0"
-
   spec.add_dependency "algolia", "~> 2.3.0"
+  spec.add_dependency "pubid-core", "1.8.5"
   spec.add_dependency "pubid-iso", "~> 0.5.0"
+  spec.add_dependency "relaton-bib", "~> 1.14.12"
   spec.add_dependency "relaton-iso-bib", "~> 1.14.0"
 end

metadata

@@ -1,85 +1,71 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iso
 version: !ruby/object:Gem::Version
-  version: 1.15.3
+  version: 1.15.5
 platform: ruby
 authors:
 - Ribose Inc.
-autorequire:
+autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-05-13 00:00:00.000000000 Z
+date: 2023-08-03 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: algolia
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '13.0'
-  type: :development
+        version: 2.3.0
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '13.0'
+        version: 2.3.0
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: pubid-core
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - '='
       - !ruby/object:Gem::Version
-        version: '3.0'
-  type: :development
+        version: 1.8.5
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - '='
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: 1.8.5
 - !ruby/object:Gem::Dependency
-  name: algolia
+  name: pubid-iso
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.3.0
+        version: 0.5.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.3.0
+        version: 0.5.0
 - !ruby/object:Gem::Dependency
-  name: pubid-iso
+  name: relaton-bib
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 1.14.12
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 1.14.12
 - !ruby/object:Gem::Dependency
   name: relaton-iso-bib
   requirement: !ruby/object:Gem::Requirement
@@ -135,10 +121,12 @@ files:
 - bin/setup
 - bin/thor
 - lib/relaton_iso.rb
+- lib/relaton_iso/config.rb
 - lib/relaton_iso/document_identifier.rb
 - lib/relaton_iso/hit.rb
 - lib/relaton_iso/hit_collection.rb
 - lib/relaton_iso/iso_bibliography.rb
+- lib/relaton_iso/logger.rb
 - lib/relaton_iso/processor.rb
 - lib/relaton_iso/scrapper.rb
 - lib/relaton_iso/version.rb
@@ -147,7 +135,7 @@ homepage: https://github.com/relaton/relaton-iso
 licenses:
 - BSD-2-Clause
 metadata: {}
-post_install_message:
+post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
@@ -162,8 +150,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.9
-signing_key:
+rubygems_version: 3.3.26
+signing_key: 
 specification_version: 4
 summary: 'RelatonIso: retrieve ISO Standards for bibliographic use using the IsoBibliographicItem
   model'