relaton-bsi 1.8.4 → 1.9.3

data/lib/relaton_bsi/bsi_bibliography.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "mechanize"
+# require "mechanize"
 require "relaton_iso_bib"
 require "relaton_bsi/bsi_bibliographic_item"
 require "relaton_bsi/scrapper"
@@ -20,21 +20,25 @@ module RelatonBsi
         HitCollection.new code, year
       rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET,
              EOFError, Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError,
-             Net::ProtocolError
-        raise RelatonBib::RequestError, "Could not access #{HitCollection::DOMAIN}"
+             Net::ProtocolError, Algolia::AlgoliaUnreachableHostError => e
+        raise RelatonBib::RequestError, e.message
       end
 
+      #
       # @param code [String] the BSI standard Code to look up
       # @param year [String] the year the standard was published (optional)
-      # @param opts [Hash] options; restricted to :all_parts if all-parts reference is required
+      # @param opts [Hash] options
+      # @option opts [Boolean] :all_parts if all-parts reference is required
+      # @option opts [Boolean] :no_year if last published document is required
+      #
       # @return [String] Relaton XML serialisation of reference
-      def get(code, year = nil, opts = {}) # rubocop:disable Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
-        c, y = code.split ':'
+      def get(code, year = nil, opts = {})
+        c, y = code.split ":"
         year ||= y
         ret = bib_get1(c, year, opts)
         return nil if ret.nil?
 
-        # ret = ret.to_most_recent_reference unless year || opts[:keep_year]
+        ret = ret.to_most_recent_reference unless year || opts[:keep_year]
         # ret = ret.to_all_parts if opts[:all_parts]
         ret
       end
@@ -44,18 +48,19 @@ module RelatonBsi
       def fetch_ref_err(code, year, missed_years) # rubocop:disable Metrics/MethodLength
         id = year ? "#{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."
+             "The code must be exactly like it is on the standards website."
         unless missed_years.empty?
-          warn "[relaton-bsi] (There was no match for #{year}, though there were matches "\
-            "found for #{missed_years.join(', ')}.)"
+          warn "[relaton-bsi] (There was no match for #{year}, though there "\
+               "were matches found for #{missed_years.join(', ')}.)"
         end
         # if /\d-\d/.match? code
-        #   warn "[relaton-bsi] The provided document part may not exist, or the document "\
-        #     "may no longer be published in parts."
+        #   warn "[relaton-bsi] The provided document part may not exist, or "\
+        #     "the document may no longer be published in parts."
         # else
-        #   warn "[relaton-bsi] If you wanted to cite all document parts for the reference, "\
-        #     "use \"#{code} (all parts)\".\nIf the document is not a standard, "\
-        #     "use its document type abbreviation (TS, TR, PAS, Guide)."
+        #   warn "[relaton-bsi] If you wanted to cite all document parts for "\
+        #     "the reference, use \"#{code} (all parts)\".\nIf the document "\
+        #     "is not a standard, use its document type abbreviation (TS, TR, "\
+        #     "PAS, Guide)."
         # end
         nil
       end
@@ -68,8 +73,6 @@ module RelatonBsi
 
         result = search(code)
         result.select do |i|
-          next true unless i.hit[:code]
-
           %r{^(?<code2>[^:]+)} =~ i.hit[:code]
           code2.include?(code1)
         end

data/lib/relaton_bsi/hit_collection.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "algolia"
 require "relaton_bsi/hit"
 
 module RelatonBsi
@@ -7,31 +8,40 @@ module RelatonBsi
   class HitCollection < RelatonBib::HitCollection
     DOMAIN = "https://shop.bsigroup.com"
 
-    # @return [Mechanize]
-    attr_reader :agent
 
     # @param ref [String]
     # @param year [String]
     def initialize(ref, year = nil)
       super ref, year
-      @agent = Mechanize.new
-      resp = agent.get "#{DOMAIN}/SearchResults/?q=#{ref}"
-      @array = hits resp
+      # @agent = Mechanize.new
+      # resp = agent.get "#{DOMAIN}/SearchResults/?q=#{ref}"
+      config = Algolia::Search::Config.new(application_id: "575YE157G9", api_key: "a057b4e74099445df2eddb7940828a10")
+      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]
     end
 
     private
 
-    # @param resp [Mechanize::Page]
+    # @param hits [Array<Hash>]
     # @return [Array<RelatonBsi::Hit>]
-    def hits(resp)
-      resp.xpath("//div[@class='resultsInd']").map do |h|
-        ref = h.at('div/h2/a')
-        code = ref.text.strip
-        url = ref[:href]
-        d = h.at("//div/strong[.='Published Date:']/following-sibling::strong").text
-        date = Date.parse(d)
-        Hit.new({ code: code, url: url, date: date }, self)
-      end
+    def hits(hits) # rubocop:disable Metrics/MethodLength, Metrics/AbcSize
+      hits.map do |h|
+        code = h[:meta][:global][:primaryDesignator].sub(/\sLOOSELEAF|\s\(A5 LAMINATED\)/, "")
+        Hit.new(
+          {
+            code: code,
+            title: h[:title],
+            url: h[:handle],
+            date: h[:meta][:global][:publishedDate],
+            publisher: h[:meta][:global][:publisher],
+            status: h[:meta][:global][:status],
+            ics: h[:meta][:global][:icsCodesAlgoliaStringArray],
+            doctype: h[:product_type],
+          }, self
+        )
+      end.sort_by { |h| h.hit[:date] }.reverse
     end
   end
 end