relaton-iso 0.6.4 → 0.6.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 66e86b23d321882dcc79e8f5fb47fbde454b238e
-  data.tar.gz: c9c1e37df8b4054e274ed5a77737a7e909380873
+  metadata.gz: 0ff396da8dad98ab886022064ab47bfe29d85f6a
+  data.tar.gz: b03ac90297c301edac42193e2d37119a24b7348f
 SHA512:
-  metadata.gz: 6f70fc424d1fcbbd786b1389412a170f7e48d589cf030771ea6847824f8cf62b379f34ba22df82b94f7b5663981a166654eed16b08eeab69fc2cc14ee971fa15
-  data.tar.gz: e2fc938d41c9c1d8d59d81388d6b653757c66f4e2eb6158284f540a8394427bb68adc3d0fe342742a98ea11e9485a20ea70000f42566d72003d9773b365d8f47
+  metadata.gz: 40dbc389dbee1e726b882944102eb674289d38b8e73d32f27f29dceeae875f40f3ec272b82a22eab23891f6732235531397e67001575d2e0922317fabbb70d15
+  data.tar.gz: fdacef79ca7cae29db9a2690b99c097c31fd241253176f3babb2bf3d2eb7113dd178fa7c01b0c7da232c86fa2831917af98b2165413cd512f8fb6feb0b9e5af7

data/Gemfile.lock

@@ -1,15 +1,15 @@
 PATH
   remote: .
   specs:
-    relaton-iso (0.6.4)
+    relaton-iso (0.6.5)
       relaton-iec (~> 0.4.0)
       relaton-iso-bib (~> 0.3.0)
 
 GEM
   remote: https://rubygems.org/
   specs:
-    addressable (2.6.0)
-      public_suffix (>= 2.0.2, < 4.0)
+    addressable (2.7.0)
+      public_suffix (>= 2.0.2, < 5.0)
     byebug (11.0.1)
     coderay (1.1.2)
     crack (0.4.3)
@@ -34,15 +34,15 @@ GEM
     pry-byebug (3.7.0)
       byebug (~> 11.0)
       pry (~> 0.10)
-    public_suffix (3.1.1)
+    public_suffix (4.0.1)
     rake (10.5.0)
-    relaton-bib (0.3.5)
+    relaton-bib (0.3.6)
       addressable
-      nokogiri (~> 1.10)
+      nokogiri
     relaton-iec (0.4.3)
       addressable
       relaton-iso-bib (~> 0.3.0)
-    relaton-iso-bib (0.3.4)
+    relaton-iso-bib (0.3.5)
       isoics (~> 0.1.6)
       relaton-bib (~> 0.3.0)
       ruby_deep_clone (~> 0.8.0)

data/lib/relaton_iso.rb

@@ -2,7 +2,11 @@
 
 require "relaton_iso/version"
 require "relaton_iso/iso_bibliography"
-if defined? Relaton
-  require_relative "relaton/processor"
-  Relaton::Registry.instance.register(Relaton::RelatonIso::Processor)
-end
+
+# if defined? Relaton
+#   require "relaton_iso/processor"
+#   # don't register the gem if it's required form relaton's registry
+#   return if caller.detect { |c| c.include? "register_gems" }
+
+#   Relaton::Registry.instance.register(RelatonIso::Processor)
+# end

data/lib/relaton_iso/iso_bibliography.rb

@@ -15,7 +15,7 @@ module RelatonIso
         HitCollection.new text
       rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET, EOFError,
              Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError, Net::ProtocolError,
-             OpenSSL::SSL::SSLError
+             OpenSSL::SSL::SSLError, Errno::ETIMEDOUT
         raise RelatonBib::RequestError, "Could not access http://www.iso.org"
       end
 
@@ -34,12 +34,13 @@ module RelatonIso
         %r{
           ^(?<code1>[^\s]+\s[^/]+) # match code
           /?
-          (?<corr>(Amd|DAmd|CD Amd|Cor|CD Cor)\s\d+:?(\d{4})?(/Cor \d+:\d{4})?) # match correction
+          (?<corr>(Amd|DAmd|(CD|WD|AWI|NP)\sAmd|Cor|CD\sCor|FDAmd)\s\d+ # correction name
+          :?(\d{4})?(/Cor\s\d+:\d{4})?) # match correction year
         }x =~ code
         code = code1 if code1
 
         if year.nil?
-          /^(?<code1>[^\s]+\s[\d-]+):?(?<year1>\d{4})?/ =~ code
+          /^(?<code1>[^\s]+(\s\w+)?\s[\d-]+):?(?<year1>\d{4})?/ =~ code
           unless code1.nil?
             code = code1
             year = year1

data/lib/relaton_iso/processor.rb

@@ -0,0 +1,35 @@
+require "relaton/processor"
+
+module RelatonIso
+  class Processor < Relaton::Processor
+    attr_reader :idtype
+
+    def initialize
+      @short = :relaton_iso
+      @prefix = "ISO"
+      @defaultprefix = %r{^(ISO)[ /]}
+      @idtype = "ISO"
+    end
+
+    # @param code [String]
+    # @param date [String, NilClass] year
+    # @param opts [Hash]
+    # @return [RelatonIsoBib::IsoBibliographicItem]
+    def get(code, date, opts)
+      ::RelatonIso::IsoBibliography.get(code, date, opts)
+    end
+
+    # @param xml [String]
+    # @return [RelatonIsoBib::IsoBibliographicItem]
+    def from_xml(xml)
+      ::RelatonIsoBib::XMLParser.from_xml xml
+    end
+
+    # @param hash [Hash]
+    # @return [RelatonIsoBib::IsoBibliographicItem]
+    def hash_to_bib(hash)
+      item_hash = ::RelatonIsoBib::HashConverter.hash_to_bib(hash)
+      ::RelatonIsoBib::IsoBibliographicItem.new item_hash
+    end
+  end
+end

data/lib/relaton_iso/scrapper.rb

@@ -42,7 +42,7 @@ module RelatonIso
 
         RelatonIsoBib::IsoBibliographicItem.new(
           fetched: Date.today.to_s,
-          docid: fetch_docid(doc),
+          docid: fetch_docid(hit_data["docRef"]),
           docnumber: fetch_docnumber(doc),
           edition: edition,
           language: langs(doc).map { |l| l[:lang] },
@@ -65,46 +65,6 @@ module RelatonIso
 
       private
 
-      # Start algolia search workers.
-      # @param text[String]
-      # @param iso_workers [RelatonBib::WorkersPool]
-      # @reaturn [RelatonBib::WorkersPool]
-      # def start_algolia_search(text, iso_workers)
-      #   index = Algolia::Index.new "all_en"
-      #   algolia_workers = RelatonBib::WorkersPool.new
-      #   algolia_workers.worker do |page|
-      #     algolia_worker(index, text, page, algolia_workers, iso_workers)
-      #   end
-
-      #   # Add first page so algolia worker will start.
-      #   algolia_workers << 0
-      # end
-
-      # Fetch ISO documents.
-      # @param hit [Hash]
-      # @param isiso_workers [RelatonIso::WorkersPool]
-      # def iso_worker(hit, iso_workers)
-      #   print "Parse #{iso_workers.size} of #{iso_workers.nb_hits}  \r"
-      #   parse_page hit
-      # end
-
-      # Fetch hits from algolia search service.
-      # @param index[Algolia::Index]
-      # @param text [String]
-      # @param page [Integer]
-      # @param algolia_workers [RelatonBib::WorkersPool]
-      # @param isiso_workers [RelatonBib::WorkersPool]
-      # def algolia_worker(index, text, page, algolia_workers, iso_workers)
-      #   res = index.search text, facetFilters: ["category:standard"], page: page
-      #   next_page = res["page"] + 1
-      #   algolia_workers << next_page if next_page < res["nbPages"]
-      #   res["hits"].each do |hit|
-      #     iso_workers.nb_hits = res["nbHits"]
-      #     iso_workers << hit
-      #   end
-      #   iso_workers.end unless next_page < res["nbPages"]
-      # end
-
       # Fetch titles and abstracts.
       # @param doc [Nokigiri::HTML::Document]
       # @return [Array<Array>]
@@ -115,10 +75,6 @@ module RelatonIso
         langs(doc).each do |lang|
           # Don't need to get page for en. We already have it.
           d = lang[:path] ? get_page(lang[:path])[0] : doc
-
-          # Check if unavailable for the lang.
-          next if d.css("h5.help-block").any?
-
           titles << fetch_title(d, lang[:lang])
 
           # Fetch abstracts.
@@ -174,7 +130,7 @@ module RelatonIso
         [Nokogiri::HTML(resp.body), url]
       rescue SocketError, Timeout::Error, Errno::EINVAL, Errno::ECONNRESET, EOFError,
              Net::HTTPBadResponse, Net::HTTPHeaderSyntaxError, Net::ProtocolError,
-             OpenSSL::SSL::SSLError
+             OpenSSL::SSL::SSLError, Errno::ETIMEDOUT
         raise RelatonBib::RequestError, "Could not access #{url}"
       end
       # rubocop:enable Metrics/AbcSize, Metrics/MethodLength
@@ -182,11 +138,8 @@ module RelatonIso
       # Fetch docid.
       # @param doc [Nokogiri::HTML::Document]
       # @return [Array<RelatonBib::DocumentIdentifier>]
-      def fetch_docid(doc)
-        item_ref = doc.at("//nav[contains(@class, 'heading-condensed')]/h1")
-        return [] unless item_ref
-
-        [RelatonBib::DocumentIdentifier.new(id: item_ref.text, type: "ISO")]
+      def fetch_docid(doc_ref)
+        [RelatonBib::DocumentIdentifier.new(id: doc_ref, type: "ISO")]
       end
 
       def fetch_docnumber(doc)

data/lib/relaton_iso/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonIso
-  VERSION = "0.6.4"
+  VERSION = "0.6.5"
 end

data/relaton_iso.gemspec

@@ -15,7 +15,7 @@ Gem::Specification.new do |spec|
   spec.description   = "RelatonIso: retrieve ISO Standards for bibliographic use "\
                        "using the IsoBibliographicItem model"
 
-  spec.homepage      = "https://github.com/metanorma/relaton-iso"
+  spec.homepage      = "https://github.com/relaton/relaton-iso"
   spec.license       = "BSD-2-Clause"
 
   spec.files         = `git ls-files -z`.split("\x0").reject do |f|

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iso
 version: !ruby/object:Gem::Version
-  version: 0.6.4
+  version: 0.6.5
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-08-20 00:00:00.000000000 Z
+date: 2019-09-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -227,15 +227,15 @@ files:
 - bin/rspec
 - bin/safe_yaml
 - bin/setup
-- lib/relaton/processor.rb
 - lib/relaton_iso.rb
 - lib/relaton_iso/hit.rb
 - lib/relaton_iso/hit_collection.rb
 - lib/relaton_iso/iso_bibliography.rb
+- lib/relaton_iso/processor.rb
 - lib/relaton_iso/scrapper.rb
 - lib/relaton_iso/version.rb
 - relaton_iso.gemspec
-homepage: https://github.com/metanorma/relaton-iso
+homepage: https://github.com/relaton/relaton-iso
 licenses:
 - BSD-2-Clause
 metadata: {}

data/lib/relaton/processor.rb

@@ -1,22 +0,0 @@
-require "relaton/processor"
-
-module Relaton
-  module RelatonIso
-    class Processor < Relaton::Processor
-      def initialize
-        @short = :relaton_iso
-        @prefix = "ISO"
-        @defaultprefix = %r{^(ISO)[ /]}
-        @idtype = "ISO"
-      end
-
-      def get(code, date, opts)
-        ::RelatonIso::IsoBibliography.get(code, date, opts)
-      end
-
-      def from_xml(xml)
-        RelatonIsoBib::XMLParser.from_xml xml
-      end
-    end
-  end
-end