relaton-iso 0.5.1 → 0.5.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 5d15a988eb361c24d1cd59bf99f23b66421d19ba
-  data.tar.gz: 7718cbe3d0ec06f333e78b54b41535e3ba75c1b4
+  metadata.gz: 251529f7af2fbfc0a28760e6f7100a3245091229
+  data.tar.gz: d9be71b4df57130f0437b9112072741c37cd9b08
 SHA512:
-  metadata.gz: bc6c77ab77e9cb7d67fe0d6ea6b158efebe1235f5c795af5b55d7e9ad6d3b87af1b81e8d0fbea2ea1d893cb20975271132e41b3898e64ac1cf2cd09439825a5c
-  data.tar.gz: c00ee4222a3028db3eb5a0e50ba98906f3e9d60c52e1595bcb301b32f6ff9a9b813ed74f23dd8a0046d1f74da74e198cec31fb8cea54802e417471239b7deb7e
+  metadata.gz: e4fc36c5c32d790c39f7b34034960fb73da5329f1ef2f63af2a0417dd8663e915a622a9981abc10cdf334f8dbda5db12708cd3f162428ba8a9283984de3b2793
+  data.tar.gz: eaeadce774f686eaedfcc096071e4fe03b25b72f8c76b702ad302eb21e659ddf5d26d49411064aa5c5c75a3fe1e44dbc86bcc3ec049c116c1ee6fa0b37304acf

data/.travis.yml

@@ -10,7 +10,7 @@ rvm:
 - 2.4
 - ruby-head
 before_install: 
-- gem install bundler -v 2.0.1
+- gem install bundler -v "~> 2"
 - bundle update
 matrix:
   allow_failures:

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    relaton-iso (0.5.1)
+    relaton-iso (0.5.2)
       algoliasearch
       relaton-iec (~> 0.3.0)
       relaton-iso-bib (~> 0.2.0)
@@ -44,7 +44,7 @@ GEM
     relaton-bib (0.2.3)
       addressable
       nokogiri (~> 1.8.4)
-    relaton-iec (0.3.0)
+    relaton-iec (0.3.1)
       addressable
       relaton-iso-bib (~> 0.2.0)
     relaton-iso-bib (0.2.3)

data/appveyor.yml

@@ -22,6 +22,7 @@ install:
 build_script:
   - set PATH=C:\Ruby%RUBY_VERSION%\bin;%PATH%
   - set GIT_TERMINAL_PROMPT=0
+  - gem install bundler -v "~> 2"
   - bundle config --local path vendor/bundle
   - bundle update
   - bundle install

data/lib/relaton_iso/iso_bibliography.rb

@@ -31,8 +31,15 @@ module RelatonIso
       #   :keep_year if undated reference should return actual reference with year
       # @return [String] Relaton XML serialisation of reference
       def get(code, year, opts)
+        %r{
+          ^(?<code1>[^\s]+\s[^/]+) # match code
+          /?
+          (?<corr>(Amd|CD Amd|Cor|CD Cor)\s\d+:?(\d{4})?(/Cor \d+:\d{4})?) # match correction
+        }x =~ code
+        code = code1 if code1
+
         if year.nil?
-          /^(?<code1>[^:]+):(?<year1>[^:]+)$/ =~ code
+          /^(?<code1>[^\s]+\s[\d-]+):?(?<year1>\d{4})?/ =~ code
           unless code1.nil?
             code = code1
             year = year1
@@ -41,11 +48,11 @@ module RelatonIso
         code += "-1" if opts[:all_parts]
         return RelatonIec::IecBibliography.get(code, year, opts) if %r[^ISO/IEC DIR].match code
 
-        ret = isobib_get1(code, year, opts)
+        ret = isobib_get1(code, year, corr)
         if ret.nil? && code =~ %r[^ISO\s]
           c = code.gsub "ISO", "ISO/IEC"
           warn "Attempting ISO/IEC retrieval"
-          ret = isobib_get1(c, year, opts)
+          ret = isobib_get1(c, year, corr)
         end
         return nil if ret.nil?
 
@@ -81,20 +88,24 @@ module RelatonIso
         workers.result.sort { |x, y| x[:i] <=> y[:i] }.map { |x| x[:hit] }
       end
 
-      def isobib_search_filter(code)
-        docidrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+}
-        corrigrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+:[0-9]+/}
+      def isobib_search_filter(code, corr)
+        # docidrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+}
+        # corrigrx = %r{^(ISO|IEC)[^0-9]*\s[0-9-]+:[0-9]+/}
         warn "fetching #{code}..."
         result = search(code)
-        result.each do |page|
-          ret = page.select do |i|
+        result.reduce([]) do |ret, page|
+          ret += page.select do |i|
             i.hit["title"] &&
-              i.hit["title"].match(docidrx).to_s == code &&
-              corrigrx !~ i.hit["title"]
+              i.hit["title"] =~ %r{^#{code}} && (
+                corr && %r{^#{code}[d-]*(:\d{4})?/#{corr}} =~ i.hit["title"] ||
+                %r{^#{code}[\d-]*(:\d{4})?/} !~ i.hit["title"] && !corr
+              )
           end
-          return ret unless ret.empty?
+          return ret if ret.size > 9
+
+          ret
         end
-        []
+        # []
       end
 
       # Sort through the results from RelatonIso, fetching them three at a time,
@@ -120,9 +131,9 @@ module RelatonIso
         { years: missed_years }
       end
 
-      def isobib_get1(code, year, _opts)
+      def isobib_get1(code, year, corr)
         # return iev(code) if /^IEC 60050-/.match code
-        result = isobib_search_filter(code) || return
+        result = isobib_search_filter(code, corr) || return
         ret = isobib_results_filter(result, year)
         return ret[:ret] if ret[:ret]
 

data/lib/relaton_iso/scrapper.rb

@@ -70,7 +70,7 @@ module RelatonIso
           type: fetch_type(hit_data["title"]),
           docstatus: fetch_status(doc, hit_data["status"]),
           ics: fetch_ics(doc),
-          dates: fetch_dates(doc),
+          dates: fetch_dates(doc, hit_data["title"]),
           contributors: fetch_contributors(hit_data["title"]),
           editorialgroup: fetch_workgroup(doc),
           abstract: abstract,
@@ -341,19 +341,33 @@ module RelatonIso
         end
       end
 
+      # rubocop:disable Metrics/MethodLength
       # Fetch dates
       # @param doc [Nokogiri::HTML::Document]
       # @return [Array<Hash>]
-      def fetch_dates(doc)
+      def fetch_dates(doc, title)
         dates = []
-        publish_date = doc.xpath("//span[@itemprop='releaseDate']").text
-        unless publish_date.empty?
-          dates << { type: "published", on: publish_date }
+        %r{^[^\s]+\s[\d-]+:(?<ref_date_str>\d{4})} =~ title
+        pub_date_str = doc.xpath("//span[@itemprop='releaseDate']").text
+        if ref_date_str
+          ref_date = Date.strptime ref_date_str, "%Y"
+          if pub_date_str.empty?
+            dates << { type: "published", on: ref_date_str }
+          else
+            pub_date = Date.strptime pub_date_str, "%Y"
+            if pub_date.year > ref_date.year
+              dates << { type: "published", on: ref_date_str }
+              dates << { type: "updated", on: pub_date_str }
+            else
+              dates << { type: "published", on: pub_date_str }
+            end
+          end
+        elsif !pub_date_str.empty?
+          dates << { type: "published", on: pub_date_str }
         end
         dates
       end
 
-      # rubocop:disable Metrics/MethodLength
       def fetch_contributors(title)
         title.sub(/\s.*/, "").split("/").map do |abbrev|
           case abbrev
@@ -408,20 +422,6 @@ module RelatonIso
         { owner: { name: owner_name }, from: from }
       end
     end
-
-    # private
-    #
-    # def next_hits_page(next_page)
-    #   page = @index.search @text, facetFilters: ['category:standard'],
-    #                               page:         next_page
-    #   page.each do |key, value|
-    #     if key == 'hits'
-    #       @docs[key] += value
-    #     else
-    #       @docs[key] = value
-    #     end
-    #   end
-    # end
   end
   # rubocop:enable Metrics/ModuleLength
 end

data/lib/relaton_iso/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonIso
-  VERSION = "0.5.1"
+  VERSION = "0.5.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iso
 version: !ruby/object:Gem::Version
-  version: 0.5.1
+  version: 0.5.2
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-06-27 00:00:00.000000000 Z
+date: 2019-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler