relaton-bsi 1.11.0 → 1.11.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3111f0bbfe4aae50b70dbe3b55e5b3b8111eded6d6e808ced57dc1cec958115b
-  data.tar.gz: 4b78d3666be1b4a5c765d18617112fe11735ab3e6323d26d8e96936e73c0ef65
+  metadata.gz: e6526b98460dccb16324ffdbc6cea2309b8e1eda6b5c06ed8272c4bae289b5fa
+  data.tar.gz: 4d64f8b75a735c5ccfbacf3f03816b62ebd22ea4cb2d564ebbb67b87643fbaa2
 SHA512:
-  metadata.gz: f99e7b8408752f33397540fd8b4094505af023a47090ea737a80d4c723146eb97b0858c7bec74677a05a0824168bdfd4151c0e05985905907e2f85367d6b880b
-  data.tar.gz: c14c76eaca4b46d77af4876d085d694118c19ea7226d40c8d171bd44bdb66f9295e98815e3af29876ce1a4ae4c3042a2715697102721402d49e2b096a422ccf4
+  metadata.gz: b640a7761c3eab1336b54904b31ce8b68477b1736bce1ec2d5b1f4775f8f2c31d9b35e9608e48a922c957e62000c4860f1fecfcb743d6602f2becc5446a7ebab
+  data.tar.gz: 1f80f9e3e3a9a0f613bf239e57b53bb8c9a0e189e841d11d9668a479d959eaa1accbd78a8e6ede2e1a78103a04d2658417eed19ab92712ac693ae1921aafe10d

data/lib/relaton_bsi/bsi_bibliography.rb

@@ -35,7 +35,7 @@ module RelatonBsi
       def get(code, year = nil, opts = {})
         # y = code.split(":")[1]
         year ||= code_parts(code)[:year]
-        ret = bib_get1(code, year, opts)
+        ret = bib_get(code, year, opts)
         return nil if ret.nil?
 
         ret = ret.to_most_recent_reference unless year || opts[:keep_year]
@@ -43,10 +43,27 @@ module RelatonBsi
         ret
       end
 
+      #
+      # Destruct code to its parts.
+      #
+      # @param [String] code document identifier
+      #
+      # @return [MatchData] parts of the code
+      #
+      def code_parts(code)
+        %r{
+          ^(?:BSI\s)?(?<code>(?:[A-Z]+\s)*[^:\s+]+)
+          (?::(?<year>\d{4}))?
+          (?:\+(?<a>[^:\s]+)(?::(?<y>\d{4}))?)?
+          (?:\s(?<rest>.+))?
+        }x.match code
+      end
+
       private
 
       def fetch_ref_err(code, year, missed_years) # rubocop:disable Metrics/MethodLength
-        id = year ? "#{code}:#{year}" : code
+        y = code_parts(code)[:year]
+        id = year && !y ? "#{code}:#{year}" : code
         warn "[relaton-bsi] WARNING: no match found online for #{id}. "\
              "The code must be exactly like it is on the standards website."
         unless missed_years.empty?
@@ -65,29 +82,19 @@ module RelatonBsi
         nil
       end
 
-      def search_filter(code) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-        # %r{^(?:BSI\s)?(?<code1>[^:]+)} =~ code
-        # %r{^(?<code1>[^:]+)} =~ code.sub(/^BSI\s/, "")
-        cp1 = code_parts code
+      #
+      # Search for a BSI standard.
+      #
+      # @param [String] code the BSI standard Code to look up
+      #
+      # @return [RelatonBsi::HitCollection] a collection of hits
+      #
+      def search_filter(code)
+        cp = code_parts code
         warn "[relaton-bsi] (\"#{code}\") fetching..."
-        return [] unless cp1
-
-        result = search(code)
-        result.select do |i|
-          # %r{^(?<code2>[^:]+)} =~ i.hit[:code]
-          cp2 = code_parts i.hit[:code]
-          cp2[:code] == cp1[:code] && (!cp1[:a] || cp2[:a] == cp1[:a]) &&
-            (!cp1[:y] || cp2[:y] == cp1[:y])
-        end
-      end
+        return [] unless cp
 
-      def code_parts(code)
-        %r{
-          ^(?:BSI\s)?(?<code>[^:]+)
-          (?::(?<year>\d{4}))?
-          (?:\+(?<a>[^:]+):)?
-          (?::(?<y>\d{4}))?
-        }x.match code
+        search(code).filter_hits!(cp)
       end
 
       # Sort through the results from Isobib, fetching them three at a time,
@@ -96,10 +103,10 @@ module RelatonBsi
       # Only expects the first page of results to be populated.
       # Does not match corrigenda etc (e.g. ISO 3166-1:2006/Cor 1:2007)
       # If no match, returns any years which caused mismatch, for error reporting
-      def isobib_results_filter(result, year)
+      def results_filter(result, year)
         missed_years = []
         result.each do |r|
-          /:(?<pyear>\d{4})/ =~ r.hit[:code]
+          pyear = code_parts(r.hit[:code])[:year]
           if !year || year == pyear
             ret = r.fetch
             return { ret: ret } if ret
@@ -110,9 +117,9 @@ module RelatonBsi
         { years: missed_years }
       end
 
-      def bib_get1(code, year, _opts)
+      def bib_get(code, year, _opts)
         result = search_filter(code) || return
-        ret = isobib_results_filter(result, year)
+        ret = results_filter(result, year)
         if ret[:ret]
           warn "[relaton-bsi] (\"#{code}\") found #{ret[:ret].docidentifier.first&.id}"
           ret[:ret]

data/lib/relaton_bsi/hit_collection.rb

@@ -8,8 +8,12 @@ module RelatonBsi
   class HitCollection < RelatonBib::HitCollection
     DOMAIN = "https://shop.bsigroup.com"
 
-    # @param ref [String]
-    # @param year [String]
+    #
+    # Initialize a new HitCollection.
+    #
+    # @param ref [String] reference
+    # @param year [String] year
+    #
     def initialize(ref, year = nil)
       super ref, year
       config = Algolia::Search::Config.new(
@@ -19,19 +23,39 @@ module RelatonBsi
       client = Algolia::Search::Client.new config, logger: ::Logger.new($stderr)
       index = client.init_index "shopify_products"
       resp = index.search text # , facetFilters: "product_type:standard"
-      @array = hits resp[:hits]
+      @array = create_hits resp[:hits]
+    end
+
+    #
+    # Filter the search results for a BSI standard.
+    #
+    # @param [MatchData] code_parts parts of document identifier
+    #
+    # @return [self] filtered search results
+    #
+    def filter_hits!(code_parts)
+      hits = filter code_parts
+      hits = filter code_parts, skip_rest: true if hits.empty?
+      hits = filter code_parts, drop_amd: true if hits.empty?
+      @array = hits
+      self
     end
 
     private
 
-    # @param hits [Array<Hash>]
-    # @return [Array<RelatonBsi::Hit>]
-    def hits(hits) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize
+    #
+    # Create hits from search results.
+    #
+    # @param hits [Array<Hash>] search results
+    #
+    # @return [Array<RelatonBsi::Hit>] hits
+    #
+    def create_hits(hits) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize
       hits.each_with_object([]) do |h, obj|
         next unless h[:meta][:global][:publishedDate]
 
         code = h[:meta][:global][:primaryDesignator]
-          .sub(/\s(?:LOOSELEAF|\(A5 LAMINATED\)|-\sTC$)/, "")
+          .sub(/\s?(?:LOOSELEAF|\(A5 LAMINATED\)|-\s?TC$)/, "")
         obj << Hit.new(
           {
             code: code,
@@ -46,5 +70,24 @@ module RelatonBsi
         )
       end.sort_by { |h| h.hit[:date] }.reverse
     end
+
+    #
+    # Select hits that match the document identifier.
+    #
+    # @param [MatchData] code_parts parts of document identifier
+    # @param [Boolean] drop_amd drop amendments and corrigendums
+    # @param [Boolean] skip_rest skip rest suffix of document identifier
+    #
+    def filter(code_parts, drop_amd: false, skip_rest: false) # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
+      @array.select do |i|
+        code = drop_amd ? i.hit[:code].sub(/\+[AC]\d+.*$/, "") : i.hit[:code]
+        cp = BsiBibliography.code_parts code
+        match = cp[:code] == code_parts[:code] && cp[:a] == code_parts[:a] &&
+          (!code_parts[:y] || cp[:y] == code_parts[:y]) &&
+          (skip_rest || cp[:rest] == code_parts[:rest])
+        i.hit[:code] = code if drop_amd && match
+        match
+      end
+    end
   end
 end

data/lib/relaton_bsi/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonBsi
-  VERSION = "1.11.0"
+  VERSION = "1.11.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-bsi
 version: !ruby/object:Gem::Version
-  version: 1.11.0
+  version: 1.11.1
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-04-10 00:00:00.000000000 Z
+date: 2022-04-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: equivalent-xml