relaton-jis 2.0.0.pre.alpha.1 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: eb4b54d49f314da1369dec1e4446b108ddd68b92555c88c8951233fd7e6b31dc
-  data.tar.gz: 2660c24bf8a0916a28010eec9b434953233076bb69775c9852473780f2317373
+  metadata.gz: ac14dd89a5e207e3de88b287d2fa8a7a4b18fee184eca59e6083fc9733138c1a
+  data.tar.gz: 6c38de3d1235a9ab8cb1a78c2aa3c8d82930049f02733923650b7c1ee3a73d3c
 SHA512:
-  metadata.gz: aa98951e0c115564f41f81a2de3968617a2699559f4e8a38cc0eaebd6bfeb2feeee16222d01875db30dbcde676a2220671ccdd2709ab4043559874a11470ac9b
-  data.tar.gz: 19c4b2f8e43fb87d8f58ff759d3af106b8b8ca3f5730dbffcfd8a089b1f2d40690a22862df507401313a3e1bd4c3c46941e2aa37ff32005b398268a4f976d5ed
+  metadata.gz: a24537ee87d90330d588c1e7018b211af83973c94c90dad55cc28825f625d8aa20fd6a6ddc086771e7b05076f3f2310d9ac22bd6c1fdfe4343fae631da10bd6e
+  data.tar.gz: 7fecb2b10ff53ae3bfefcfa75bfe91d059c641a7eaef69ebf73d6530801287625904876b493e5e3b2099d58ecd36730a3bc72d1c67190c132f879cb144f74f03

data/grammars/basicdoc.rng

@@ -187,6 +187,15 @@ Applicable to modify and delete</a:documentation>
         <a:documentation>Optional caption of this block</a:documentation>
       </attribute>
     </optional>
+    <optional>
+      <attribute name="position">
+        <a:documentation>For an "add" change, whether the change is added before or after the location</a:documentation>
+        <choice>
+          <value>before</value>
+          <value>after</value>
+        </choice>
+      </attribute>
+    </optional>
     <optional>
       <element name="location">
         <a:documentation>The location(s) in the original document which have undergone the change described in this block</a:documentation>
@@ -208,11 +217,15 @@ Applicable to modify and delete</a:documentation>
     </zeroOrMore>
     <optional>
       <element name="newcontent">
-        <a:documentation>New content to be added to the document; applicable to add and modify</a:documentation>
+        <a:documentation>New content to be added to the document; applicable to add and modify.
+Can be blocks and/or sections</a:documentation>
         <ref name="OptionalId"/>
         <zeroOrMore>
           <ref name="BasicBlock"/>
         </zeroOrMore>
+        <zeroOrMore>
+          <ref name="section"/>
+        </zeroOrMore>
       </element>
     </optional>
     <zeroOrMore>

data/grammars/biblio.rng

@@ -1142,11 +1142,11 @@ NOTE: This should preferably be encoded as  a URI or short identifier, rather th
         <a:documentation>Information about how long the current description of the bibliographic item is valid for</a:documentation>
       </ref>
     </optional>
-    <optional>
+    <zeroOrMore>
       <ref name="depiction">
         <a:documentation>Depiction of the bibliographic item, typically an image</a:documentation>
       </ref>
-    </optional>
+    </zeroOrMore>
   </define>
   <define name="ReducedBibliographicItem">
     <a:documentation>Reduced description of a bibliographic resource, without mandatory title and docidentifier, used for document relations
@@ -1939,10 +1939,10 @@ Detailed in https://www.relaton.org/model/relations/</a:documentation>
       <value>hasAnnotation</value>
       <value>draftOf</value>
       <value>hasDraft</value>
-      <value>preliminaryDraftOf</value>
-      <value>hasPreliminaryDraft</value>
-      <value>revisionDraftOf</value>
-      <value>hasRevisionDraft</value>
+      <value>predecessorDraftOf</value>
+      <value>hasPredecessorDraft</value>
+      <value>successorDraftOf</value>
+      <value>hasSuccessorDraft</value>
       <value>editionOf</value>
       <value>hasEdition</value>
       <value>updates</value>
@@ -2063,13 +2063,13 @@ provided that it is not the entire bibliographic item that is so related</a:docu
           <ref name="LocalizedString"/>
         </element>
       </optional>
-      <oneOrMore>
+      <zeroOrMore>
         <element name="taxon">
           <a:documentation>The keywords as a hierarchical taxonomy. For example, the sequence of `taxon` elements
 `pump`, `centrifugal pump`, `line shaft pump` represents a taxonomic classification</a:documentation>
           <ref name="LocalizedString"/>
         </element>
-      </oneOrMore>
+      </zeroOrMore>
       <zeroOrMore>
         <ref name="vocabid">
           <a:documentation>Identifiers for the keyword as a controlled vocabulary</a:documentation>

data/lib/relaton/jis/data_fetcher.rb

@@ -15,10 +15,6 @@ module Relaton
         @mutex = Mutex.new
       end
 
-      def gh_issue_channel
-        ["relaton/relaton-jis", "Error fetching JIS documents"]
-      end
-
       def log_error(msg)
         Util.error msg
       end
@@ -52,7 +48,7 @@ module Relaton
       def fetch_doc(url) # rubocop:disable Metrics/MethodLength
         attempts = 0
         begin
-          bib = Scraper.new(url).fetch
+          bib = Scraper.new(url, @errors).fetch
         rescue StandardError => e
           attempts += 1
           if attempts < 5
@@ -73,6 +69,7 @@ module Relaton
         resp = agent.get "#{URL}W11M0070/index"
         parse_page resp
         index.save
+        report_errors
       end
 
       def initial_post

data/lib/relaton/jis/hit_collection.rb

@@ -74,7 +74,7 @@ module Relaton
           content: hit.hit[:id], type: "JIS", primary: true,
         )
         bibitem = Bib::ItemData.new(
-          formattedref: hit.hit[:id], docidentifier: [docid],
+          formattedref: Bib::Formattedref.new(content: hit.hit[:id]), docidentifier: [docid],
         )
         Bib::Relation.new(type: "instanceOf", bibitem: bibitem)
       end

data/lib/relaton/jis/scraper.rb

@@ -15,9 +15,10 @@ module Relaton
       DATETYPES = { "発行年月日" => "issued", "確認年月日" => "confirmed" }.freeze
       STATUSES = { "有効" => "valid", "廃止" => "withdrawn" }.freeze
 
-      def initialize(url)
+      def initialize(url, errors = {})
         @url = url
         @agent = Mechanize.new
+        @errors = errors
       end
 
       def fetch # rubocop:disable Metrics/MethodLength
@@ -25,18 +26,18 @@ module Relaton
         contributors = fetch_contributor
         eg_contributor = fetch_editorialgroup_contributor
         contributors << eg_contributor if eg_contributor
-        attrs = ATTRS.each_with_object({}) do |attr, hash|
-          hash[attr] = send "fetch_#{attr}"
-        end
+        attrs = ATTRS.to_h { |attr| [attr, send("fetch_#{attr}")] }
         attrs[:contributor] = contributors
         Bib::ItemData.new(**attrs)
       end
 
       def fetch_title
-        { "ja" => "Jpan", "en" => "Latn" }.map.with_index do |(lang, script), i|
+        result = { "ja" => "Jpan", "en" => "Latn" }.map.with_index do |(lang, script), i|
           content = @doc.at("./h2/text()[#{i + 2}]").text.strip
           Bib::Title.new content: content, language: lang, script: script
         end
+        @errors[:title] &&= result.empty?
+        result
       end
 
       def fetch_source # rubocop:disable Metrics/MethodLength
@@ -44,45 +45,60 @@ module Relaton
         uri = URI @url
         domain = "#{uri.scheme}://#{uri.host}"
         xpath = "./dl/dt[.='プレビュー']/following-sibling::dd[1]/a"
-        @doc.xpath(xpath).reduce([src]) do |mem, node|
+        result = @doc.xpath(xpath).reduce([src]) do |mem, node|
           href = "#{domain}#{node[:href]}"
           mem << Bib::Uri.new(content: href, type: "pdf")
         end
+        @errors[:source] &&= result.empty?
+        result
       end
 
       def fetch_abstract
-        @doc.xpath("//div[@id='honbun']").map do |node|
-          Bib::LocalizedMarkedUpString.new(
+        result = @doc.xpath("//div[@id='honbun']").map do |node|
+          Bib::Abstract.new(
             content: node.text.strip,
             language: "ja", script: "Jpan"
           )
         end
+        @errors[:abstract] &&= result.empty?
+        result
       end
 
       def fetch_docidentifier
         docid = document_id
+        @errors[:docidentifier] &&= docid.nil? || docid.empty?
+        return [] if docid.nil? || docid.empty?
+
         [Docidentifier.new(
           content: docid, type: "JIS", primary: true,
         )]
       end
 
       def fetch_docnumber
-        match = document_id.match(/^\w+\s(\w)\s?(\d+)/)
+        docid = document_id
+        match = docid&.match(/^\w+\s(\w)\s?(\d+)/)
+        @errors[:docnumber] &&= match.nil?
+        return unless match
+
         "#{match[1]}#{match[2]}"
       end
 
       def document_id
-        @document_id ||= @doc.at("./h2/text()[1]").text.strip
+        @document_id ||= @doc.at("./h2/text()[1]")&.text&.strip
       end
 
       def fetch_date
-        DATETYPES.each_with_object([]) do |(key, type), a|
+        result = DATETYPES.each_with_object([]) do |(key, type), a|
           node = @doc.at("./div/div/div/p/text()[contains(.,'#{key}')]")
           next unless node
 
           at = node.text.match(/\d{4}-\d{2}-\d{2}/).to_s
+          next if at.empty?
+
           a << Bib::Date.new(type: type, at: at)
         end
+        @errors[:date] &&= result.empty?
+        result
       end
 
       def fetch_type
@@ -97,15 +113,19 @@ module Relaton
         langs_scripts.map { |l| l[:script] }
       end
 
-      def langs_scripts
-        @langs_scripts ||= LANGS.each_with_object([]) do |(key, lang), a|
-          l = @doc.at(
-            "./div/div/div[@class='blockContentFile']/div/div/p[1]" \
-            "/span[contains(.,'#{key}')]/following-sibling::span",
-          )
-          next if l.nil? || l.text.strip == "-"
+      def langs_scripts # rubocop:disable Metrics/MethodLength
+        @langs_scripts ||= begin
+          result = LANGS.each_with_object([]) do |(key, lang), a|
+            l = @doc.at(
+              "./div/div/div[@class='blockContentFile']/div/div/p[1]" \
+              "/span[contains(.,'#{key}')]/following-sibling::span",
+            )
+            next if l.nil? || l.text.strip == "-"
 
-          a << lang
+            a << lang
+          end
+          @errors[:language] &&= result.empty?
+          result
         end
       end
 
@@ -113,24 +133,30 @@ module Relaton
         xpath = "./div/div/div/p/text()[contains(.,'状態')]" \
                 "/following-sibling::span"
         st = @doc.at(xpath)
-        return unless st
+        status_val = STATUSES[st&.text&.strip]
+        @errors[:status] &&= status_val.nil?
+        return unless status_val
 
-        stage = Bib::Status::Stage.new(content: STATUSES[st.text.strip])
+        stage = Bib::Status::Stage.new(content: status_val)
         Bib::Status.new(stage: stage)
       end
 
-      def fetch_doctype
+      def fetch_doctype # rubocop:disable Metrics/CyclomaticComplexity
         type = case document_id
                when /JIS\s[A-Z]\s[\w-]+:\d{4}\/AMENDMENT/ then "amendment"
                when /JIS\s[A-Z]\s[\w-]+/ then "japanese-industrial-standard"
                when /TR[\s\/][\w-]+/ then "technical-report"
                when /TS[\s\/][\w-]+/ then "technical-specification"
                end
+        @errors[:doctype] &&= type.nil?
+        return unless type
+
         Doctype.new content: type
       end
 
       def fetch_ics
         td = @doc.at("./table/tr[th[.='ICS']]/td")
+        @errors[:ics] &&= td.nil?
         return [] unless td
 
         td.text.strip.split.map { |code| Bib::ICS.new code: code }
@@ -141,10 +167,12 @@ module Relaton
           "一般財団法人　日本規格協会", "authorizer"
         )
         xpath = "./table/tr[th[.='原案作成団体']]/td"
-        @doc.xpath(xpath).reduce([authorizer]) do |a, node|
+        result = @doc.xpath(xpath).reduce([authorizer]) do |a, node|
           a << create_contrib(node.text.strip, "author")
           a << create_contrib(node.text.strip, "publisher")
         end
+        @errors[:contributor] &&= result.empty?
+        result
       end
 
       def create_contrib(name, role)
@@ -167,6 +195,7 @@ module Relaton
 
       def fetch_editorialgroup_contributor # rubocop:disable Metrics/MethodLength
         node = @doc.at("./table/tr[th[.='原案作成団体']]/td")
+        @errors[:editorialgroup] &&= node.nil?
         return unless node
 
         subdivision = Bib::Subdivision.new(

data/lib/relaton/jis/version.rb

@@ -2,6 +2,6 @@
 
 module Relaton
   module Jis
-    VERSION = "2.0.0-alpha.1"
+    VERSION = "2.0.0"
   end
 end

data/relaton_jis.gemspec

@@ -35,8 +35,9 @@ Gem::Specification.new do |spec|
 
   # Uncomment to register a new dependency of your gem
   spec.add_dependency "mechanize", "~> 2.10"
+  spec.add_dependency "relaton-core", "~> 0.0.13"
   spec.add_dependency "relaton-index", "~> 0.2.0"
-  spec.add_dependency "relaton-iso", "~> 2.0.0-alpha.2"
+  spec.add_dependency "relaton-iso", "~> 2.0.0"
 
   # For more information and examples about making a new gem, check out our
   # guide at: https://bundler.io/guides/creating_gem.html

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: relaton-jis
 version: !ruby/object:Gem::Version
-  version: 2.0.0.pre.alpha.1
+  version: 2.0.0
 platform: ruby
 authors:
 - Ribose Inc.
@@ -23,6 +23,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2.10'
+- !ruby/object:Gem::Dependency
+  name: relaton-core
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.13
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.13
 - !ruby/object:Gem::Dependency
   name: relaton-index
   requirement: !ruby/object:Gem::Requirement
@@ -43,14 +57,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0.pre.alpha.2
+        version: 2.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0.pre.alpha.2
+        version: 2.0.0
 description: 'Relaton::Jis: retrieve IETF Standards for bibliographic use using the
   BibliographicItem model'
 email: