relaton-itu 0.2.2 → 0.2.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 06af239fca985e01b1bf84c15ef3353adba66888
-  data.tar.gz: 07f56fe72469d205e2da95cf4e6c17c5d83742ad
+  metadata.gz: 0e53b2a711cf2600a246b1ef35dff524624ae266
+  data.tar.gz: 5b0234760082b97040111b7fe18814977e28fd6b
 SHA512:
-  metadata.gz: a1b4e86236bc6ad52ab08366e03ac045f846ba56bd5b962852b95b9152bfcd21e1b23065388abb5635278ff37d1fc326d3cfe4c2037657a8af0d9e342ada6d01
-  data.tar.gz: 907809fe72f0293cfa0f0f9194bee933b61f1ced2c75010f6acaa92d29b5f3b3851b5a16d955d6e1b5caf3b5478e7e736b4d18897356b082089b83d48481e896
+  metadata.gz: a3dbf304c78838b895c6da7e869b70bbef5b60f00f12498b0a5101761d052828ebf8e45215e33d728d071cfbf9b536857826270ea09e71ca4e6bef3c9070cd9b
+  data.tar.gz: fde81735c358c19beb765da48726306c7ff020f9f02d66bf1204b9d687b73bf6ff707d6c70727738afa56200312623e36fab33e21310eec49b3a280dcd6c4411

data/.travis.yml

@@ -6,6 +6,7 @@ os:
 - linux
 - osx
 rvm:
+- 2.6
 - 2.5
 - 2.4
 - ruby-head

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    relaton-itu (0.2.2)
+    relaton-itu (0.2.3)
       relaton-iso-bib (~> 0.2.0)
 
 GEM
@@ -24,9 +24,9 @@ GEM
     isoics (0.1.7)
     json (2.2.0)
     method_source (0.9.2)
-    mini_portile2 (2.3.0)
-    nokogiri (1.8.5)
-      mini_portile2 (~> 2.3.0)
+    mini_portile2 (2.4.0)
+    nokogiri (1.10.3)
+      mini_portile2 (~> 2.4.0)
     pry (0.12.2)
       coderay (~> 1.1.0)
       method_source (~> 0.9.0)
@@ -35,12 +35,11 @@ GEM
       pry (~> 0.10)
     public_suffix (3.1.1)
     rake (10.5.0)
-    relaton-bib (0.2.3)
+    relaton-bib (0.2.5)
       addressable
-      nokogiri (~> 1.8.4)
-    relaton-iso-bib (0.2.3)
+      nokogiri (~> 1.10)
+    relaton-iso-bib (0.2.4)
       isoics (~> 0.1.6)
-      nokogiri (~> 1.8.4)
       relaton-bib (~> 0.2.0)
       ruby_deep_clone (~> 0.8.0)
     rspec (3.8.0)

data/appveyor.yml

@@ -7,6 +7,7 @@ cache:
 
 environment:
   matrix:
+    - RUBY_VERSION: 26
     - RUBY_VERSION: 25
     - RUBY_VERSION: 24
     - RUBY_VERSION: _trunk

data/lib/relaton_itu/hit_collection.rb

@@ -28,6 +28,7 @@ module RelatonItu
       #   from = Date.strptime year, "%Y"
       #   to   = from.next_year.prev_day
       # end
+      group = %r{(OB|Operational Bulletin) No} =~ text ? "Publications" : "Recommendations"
       url = "#{DOMAIN}/net4/ITU-T/search/GlobalSearch/Search"
       params = {
         "Input" => ref_nbr,
@@ -36,7 +37,7 @@ module RelatonItu
         "SortBy" => "RELEVANCE",
         "ExactPhrase" => false,
         "CollectionName" => "General",
-        "CollectionGroup" => "Recommendations",
+        "CollectionGroup" => group,
         "Sector" => "t",
         "Criterias" => [{
           "Name" => "Search in",
@@ -82,7 +83,8 @@ module RelatonItu
         code  = h["Media"]["Name"]
         title = h["Title"]
         url   = h["Redirection"]
-        Hit.new({ code: code, title: title, url: url }, self)
+        type  = group.downcase[0...-1]
+        Hit.new({ code: code, title: title, url: url, type: type }, self)
       end
       concat hits
       @fetched = false

data/lib/relaton_itu/itu_bibliography.rb

@@ -73,12 +73,13 @@ module RelatonItu
       end
 
       def search_filter(code)
-        docidrx = %r{^ITU-T\s[^\s]+}
+        docidrx = %r{\w+.\d+} # %r{^ITU-T\s[^\s]+}
+        c = code.match(docidrx).to_s
         warn "fetching #{code}..."
         result = search(code)
         result.select do |i|
           i.hit[:code] &&
-            i.hit[:code].match(docidrx).to_s == code
+            i.hit[:code].match(docidrx).to_s == c
         end
       end
 

data/lib/relaton_itu/scrapper.rb

@@ -13,6 +13,7 @@ module RelatonItu
   # rubocop:disable Metrics/ModuleLength
   module Scrapper
     DOMAIN = "https://www.itu.int"
+    ROMAN_MONTHS = %w[I II III IV V VI VII VIII IX X XI XII].freeze
 
     TYPES = {
       "ISO" => "international-standard",
@@ -50,7 +51,7 @@ module RelatonItu
         doc = get_page hit_data[:url]
 
         # Fetch edition.
-        edition = doc.at("//table/tr/td/span[contains(@id, 'Label8')]/b").text
+        edition = doc.at("//table/tr/td/span[contains(@id, 'Label8')]/b")&.text
 
         ItuBibliographicItem.new(
           fetched: Date.today.to_s,
@@ -59,7 +60,7 @@ module RelatonItu
           language: ["en"],
           script: ["Latn"],
           titles: fetch_titles(hit_data),
-          type: fetch_type(doc),
+          type: hit_data[:type],
           docstatus: fetch_status(doc),
           ics: [], # fetch_ics(doc),
           dates: fetch_dates(doc),
@@ -138,34 +139,26 @@ module RelatonItu
 
       # Fetch status.
       # @param doc [Nokogiri::HTML::Document]
-      # @param status [String]
-      # @return [Hash]
+      # @return [RelatonBib::DocumentStatus, NilClass]
       def fetch_status(doc)
-        s = doc.at("//table/tr/td/span[contains(@id, 'Label7')]").text
-        if s == "In force"
-          status   = "Published"
-          # stage    = "60"
-          # substage = "60"
-        else
-          status   = "Withdrawal"
-          # stage    = "95"
-          # substage = "99"
-        end
+        s = doc.at("//table/tr/td/span[contains(@id, 'Label7')]")
+        return unless s
+
+        status = s.text == "In force" ? "Published" : "Withdrawal"
         RelatonBib::DocumentStatus.new(stage: status)
       end
 
       # Fetch workgroup.
       # @param doc [Nokogiri::HTML::Document]
-      # @return [RelatonItu::EditorialGroup]
+      # @return [RelatonItu::EditorialGroup, NilClass]
       def fetch_workgroup(doc)
-        wg = doc.at('//table/tr/td/span[contains(@id, "Label8")]/a').text
+        wg = doc.at('//table/tr/td/span[contains(@id, "Label8")]/a')
+        return unless wg
+
+        workgroup = wg.text
         EditorialGroup.new(
-          bureau: wg.match(/(?<=-)./).to_s,
-          group: itugroup(wg),
-          # name: "International Telecommunication Union",
-          # abbreviation: "ITU",
-          # url: "www.itu.int",
-          # technical_committee: tc,
+          bureau: workgroup.match(/(?<=-)./).to_s,
+          group: itugroup(workgroup),
         )
       end
 
@@ -209,9 +202,9 @@ module RelatonItu
       # Fetch type.
       # @param doc [Nokogiri::HTML::Document]
       # @return [String]
-      def fetch_type(_doc)
-        "recommendation"
-      end
+      # def fetch_type(_doc)
+      #   "recommendation"
+      # end
 
       # Fetch titles.
       # @param hit_data [Hash]
@@ -248,13 +241,33 @@ module RelatonItu
       # @return [Array<Hash>]
       def fetch_dates(doc)
         dates = []
-        publish_date = doc.at("//table/tr/td/span[contains(@id, 'Label5')]").text
+        pdate = doc.at("//table/tr/td/span[contains(@id, 'Label5')]")
+        publish_date = pdate&.text || ob_date(doc)
         unless publish_date.empty?
           dates << { type: "published", on: publish_date }
         end
         dates
       end
 
+      # Scrape Operational Bulletin date.
+      # @param doc [Nokogiri::HTML::Document]
+      # @return [String]
+      def ob_date(doc)
+        pdate = doc.at('//table/tbody/tr/td[contains(text(), "Year:")]')
+        return unless pdate
+
+        roman_to_arabic pdate.text.match(%r{(?<=Year: )\d{2}.\w+.\d{4}}).to_s
+      end
+
+      # Convert roman month number in string date to arabic number
+      # @param date [String]
+      # @return [String]
+      def roman_to_arabic(date)
+        %r{(?<rmonth>[IVX]+)} =~ date
+        month = ROMAN_MONTHS.index(rmonth) + 1
+        Date.parse(date.sub(%r{[IVX]+}, month.to_s)).to_s
+      end
+
       # Fetch contributors
       # @param doc [Nokogiri::HTML::Document]
       # @return [Array<Hash>]
@@ -300,7 +313,8 @@ module RelatonItu
           name = "International Telecommunication Union"
           url = "www.itu.int"
         end
-        from = doc.at("//table/tr/td/span[contains(@id, 'Label5')]").text
+        fdate = doc.at("//table/tr/td/span[contains(@id, 'Label5')]")
+        from = fdate&.text || ob_date(doc)
         { owner: { name: name, abbreviation: abbreviation, url: url }, from: from }
       end
     end

data/lib/relaton_itu/version.rb

@@ -1,3 +1,3 @@
 module RelatonItu
-  VERSION = "0.2.2".freeze
+  VERSION = "0.2.3".freeze
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-itu
 version: !ruby/object:Gem::Version
-  version: 0.2.2
+  version: 0.2.3
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-07-10 00:00:00.000000000 Z
+date: 2019-07-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler