relaton-iso 0.8.0 → 0.8.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a73ac19c4da469921da4bad644b50f148d750476852e1a43c85c2cf0ffe52b87
-  data.tar.gz: b7bf3b134c1dc1d851b136a78d35241572cd9706f4c451791e268932b4a68b70
+  metadata.gz: 23cfcc3b3c94cf0988d6087a1afacb4bcc6065249b4f6785879e8242384bf00d
+  data.tar.gz: f973dc6ba692b89acd1c6b4f9f75b9261ca7aff09557ed58605a7b1b8e499719
 SHA512:
-  metadata.gz: bd884caa8c53e3d4b020335e1adbf2abab31f657069a249f821e4cac21a1d126b69d8d3c281058217b3d3aaa9c5f2f4a6d6b361ca328c8a0872fb5e1cba16565
-  data.tar.gz: 0c92680ecce087a3437176f5d980b59e5fca2416d9f99ac5cdd094b93106c7286cf2f9fded7544c156b54aecf988dc83917c2cf08ddfb8fa05c6f8049bf578d2
+  metadata.gz: 8d88890b35c169076ceb3350eedb036116a4fe30403ae23e6e9300718d5cfd8d6d37f49a6724bdb2c4b459a7782bd429cab660d4a195bcef92ac0dc347021c6f
+  data.tar.gz: e68d82ea105bba5dc0ab989a35eb02062adefdb45302c17065f9b47b3bf3f3b6eb9ea4e3caa43177782235013f984ea6fe574cad8e95fe1a535319d63c8b573e

data/README.adoc

@@ -183,6 +183,36 @@ RelatonIso::IsoBibliography.get('ISO 19115', "2014", {all_parts: true}).title
   @type="main">]
 ----
 
+=== Get specific language
+
+[source,ruby]
+----
+item = RelatonIso::IsoBibliography.get 'ISO 19115', nil, {lang: "en"}
+item.to_xml
+=> <bibitem id="ISO19115-1-2014" type="standard">
+  <fetched>2020-01-22</fetched>
+  <title type="title-intro" format="text/plain" language="en" script="Latn">Geographic information</title>
+  <title type="title-main" format="text/plain" language="en" script="Latn">Metadata</title>
+  <title type="main" format="text/plain" language="en" script="Latn">Geographic information – Metadata</title>
+  <uri type=\"src\">https://www.iso.org/standard/53798.html</uri>
+  ...
+</bibitem>
+
+item = RelatonIso::IsoBibliography.get 'ISO 19115', nil, {lang: "fr"}
+item.to_xml
+=> <bibitem id="ISO19115-1-2014" type="standard">
+  <fetched>2020-01-22</fetched>
+  <title type="title-intro" format="text/plain" language="en" script="Latn">Geographic information</title>
+  <title type="title-main" format="text/plain" language="en" script="Latn">Metadata</title>
+  <title type="main" format="text/plain" language="en" script="Latn">Geographic information – Metadata</title>
+  <title type="title-intro" format="text/plain" language="fr" script="Latn">Information géographique</title>
+  <title type="title-main" format="text/plain" language="fr" script="Latn">Métadonnées</title>
+  <title type="main" format="text/plain" language="fr" script="Latn">Information géographique – Métadonnées</title>
+  <uri type="src">https://www.iso.org/standard/53798.html</uri>
+  ...
+</bibitem>
+----
+
 == Development
 
 After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake spec` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.

data/lib/relaton_iso/hit.rb

@@ -7,9 +7,10 @@ module RelatonIso
     attr_reader :hit_collection
 
     # Parse page.
+    # @param lang [String, NilClass]
     # @return [RelatonIso::IsoBibliographicItem]
-    def fetch
-      @fetch ||= Scrapper.parse_page @hit
+    def fetch(lang = nil)
+      @fetch ||= Scrapper.parse_page @hit, lang
     end
 
     # @param builder [Nokogiri::XML::Builder]

data/lib/relaton_iso/hit_collection.rb

@@ -10,25 +10,23 @@ module RelatonIso
 
     def_delegators :@array, :<<, :[], :first, :empty?, :any?, :size
 
-    # @return [TrueClass, FalseClass]
-    # attr_reader :fetched
-
-    # @return [RelatonIso::HitPages]
-    # attr_reader :hit_pages
-
-    # @return [String]
+    # @return [String, NilClass]
     attr_reader :text
 
-    # @param hits [Array<Hash>]
+    # @param text [String] reference to search
     def initialize(text)
       @array = []
       @text = text
-      %r{\s(?<num>\d+)(-(?<part>\d+))?} =~ text
+      %r{\s(?<num>\d+)(-(?<part>[\d-]+))?} =~ text
       http = Net::HTTP.new "www.iso.org", 443
       http.use_ssl = true
       search = ["status=ENT_ACTIVE,ENT_PROGRESS,ENT_INACTIVE,ENT_DELETED"]
       search << "docNumber=#{num}"
       search << "docPartNo=#{part}" if part
+      # if year
+      #   search << "stageDateStart=#{Date.new(year.to_i).strftime("%Y-%m-%d")}"
+      #   search << "stageDateEnd=#{Date.new(year.to_i, 12, 31).strftime("%Y-%m-%d")}"
+      # end
       q = search.join "&"
       resp = http.get("/cms/render/live/en/sites/isoorg.advancedSearch.do?#{q}",
                       "Accept" => "application/json, text/plain, */*")
@@ -55,27 +53,14 @@ module RelatonIso
       self
     end
 
-    # @return [RelatonIso::HitCollection]
-    # def fetch
-    #   return self if @fetched
-
-    #   workers = RelatonBib::WorkersPool.new 4
-    #   workers.worker(&:fetch)
-    #   @array.each do |hit|
-    #     workers << hit
-    #   end
-    #   workers.end
-    #   workers.result
-    #   @fetched = true
-    #   self
-    # end
-
-    def to_all_parts
+    # @param lang [String, NilClass]
+    # @return [RelatonIsoBib::IsoBibliographicItem]
+    def to_all_parts(lang = nil)
       parts = @array.select { |h| !h.hit["docPart"].empty? }
       hit = parts.min_by { |h| h.hit["docPart"].to_i }
-      return @array.first.fetch unless hit
+      return @array.first.fetch lang unless hit
 
-      bibitem = hit.fetch
+      bibitem = hit.fetch lang
       bibitem.to_all_parts
       parts.reject { |h| h.hit["docRef"] == hit.hit["docRef"] }.each do |hi|
         isobib = RelatonIsoBib::IsoBibliographicItem.new(

data/lib/relaton_iso/iso_bibliography.rb

@@ -19,24 +19,19 @@ module RelatonIso
         raise RelatonBib::RequestError, "Could not access http://www.iso.org"
       end
 
-      # @param text [String]
-      # @return [Array<RelatonIso::IsoBibliographicItem>]
-      # def search_and_fetch(text)
-      #   Scrapper.get(text)
-      # end
-
       # @param ref [String] the ISO standard Code to look up (e..g "ISO 9000")
-      # @param year [String] the year the standard was published (optional)
-      # @param opts [Hash] options; restricted to :all_parts if all-parts reference is required,
-      #   :keep_year if undated reference should return actual reference with year
+      # @param year [String, NilClass] the year the standard was published
+      # @param opts [Hash] options; restricted to :all_parts if all-parts
+      #   reference is required, :keep_year if undated reference should
+      #   return actual reference with year
       # @return [String] Relaton XML serialisation of reference
-      def get(ref, year, opts)
+      def get(ref, year = nil, opts = {})
         opts[:ref] = ref
 
         %r{
           ^(?<code1>[^\s]+\s[^/]+) # match code
           /?
-          (?<corr>(Amd|DAmd|(CD|WD|AWI|NP)\sAmd|Cor|CD\sCor|FDAmd)\s\d+ # correction name
+          (?<corr>(Amd|DAmd|(CD|WD|AWI|NP)\sAmd|Cor|CD\sCor|FDAmd|PRF\sAmd)\s\d+ # correction name
           :?(\d{4})?(/Cor\s\d+:\d{4})?) # match correction year
         }x =~ ref
         code = code1 || ref
@@ -122,7 +117,7 @@ module RelatonIso
       # @param opts [Hash]
       def try_stages(result, corr, opts)
         res = nil
-        %w[NP WD CD DIS FDIS PRF IS AWI].each do |st| # try stages
+        %w[NP WD CD DIS FDIS PRF IS AWI TR].each do |st| # try stages
           c = yield st
           res = search_code result, c, corr, opts
           return res unless res.empty?
@@ -134,8 +129,8 @@ module RelatonIso
         result.select do |i|
           (opts[:all_parts] || i.hit["docRef"] =~ %r{^#{code}(?!-)}) && (
               corr && %r{^#{code}[\w-]*(:\d{4})?/#{corr}} =~ i.hit["docRef"] ||
-              %r{^#{code}[\w-]*(:\d{4})?/} !~ i.hit["docRef"] && !corr
-            )
+              !corr && %r{^#{code}[\w-]*(:\d{4})?/} !~ i.hit["docRef"]
+        ) # && %r{^#{code}} =~ i.hit["docRef"]
         end
       end
 
@@ -159,11 +154,15 @@ module RelatonIso
         end
         return { years: missed_years } unless hits.any?
 
-        return { ret: hits.first.fetch } if !opts[:all_parts] || hits.size == 1
+        return { ret: hits.first.fetch(opts[:lang]) } if !opts[:all_parts] || hits.size == 1
 
-        { ret: hits.to_all_parts }
+        { ret: hits.to_all_parts(opts[:lang]) }
       end
 
+      # @param code [String]
+      # @param year [String, NilClass]
+      # @param corr [String, NilClass]
+      # @param opts [Hash]
       def isobib_get1(code, year, corr, opts)
         # return iev(code) if /^IEC 60050-/.match code
         result = isobib_search_filter(code, corr, opts) || return

data/lib/relaton_iso/scrapper.rb

@@ -27,10 +27,11 @@ module RelatonIso
 
     class << self
       # Parse page.
-      # @param hit [Hash]
+      # @param hit_data [Hash]
+      # @param lang [String, NilClass]
       # @return [Hash]
       # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
-      def parse_page(hit_data)
+      def parse_page(hit_data, lang = nil)
         path = "/contents/data/standard#{hit_data["splitPath"]}/#{hit_data["csnumber"]}.html"
         doc, url = get_page path
 
@@ -38,15 +39,15 @@ module RelatonIso
         edition = doc&.xpath("//strong[contains(text(), 'Edition')]/..")&.
           children&.last&.text&.match(/\d+/)&.to_s
 
-        titles, abstract = fetch_titles_abstract(doc)
+        titles, abstract, langs = fetch_titles_abstract(doc, lang)
 
         RelatonIsoBib::IsoBibliographicItem.new(
           fetched: Date.today.to_s,
           docid: fetch_docid(hit_data["docRef"]),
           docnumber: fetch_docnumber(doc),
           edition: edition,
-          language: langs(doc).map { |l| l[:lang] },
-          script: langs(doc).map { |l| script(l[:lang]) }.uniq,
+          language: langs.map { |l| l[:lang] },
+          script: langs.map { |l| script(l[:lang]) }.uniq,
           title: titles,
           doctype: fetch_type(hit_data["docRef"]),
           docstatus: fetch_status(doc),
@@ -68,40 +69,46 @@ module RelatonIso
 
       # Fetch titles and abstracts.
       # @param doc [Nokigiri::HTML::Document]
+      # @param lang [String, NilClass]
       # @return [Array<Array>]
       # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
-      def fetch_titles_abstract(doc)
+      def fetch_titles_abstract(doc, lang)
         titles   = []
         abstract = []
-        langs(doc).each do |lang|
+        langs = languages(doc, lang).reduce([]) do |s, l|
           # Don't need to get page for en. We already have it.
-          d = lang[:path] ? get_page(lang[:path])[0] : doc
-          titles << fetch_title(d, lang[:lang])
+          d = l[:path] ? get_page(l[:path])[0] : doc
+          unless d.at("//h5[@class='help-block'][.='недоступно на русском языке']")
+            s << l
+            titles << fetch_title(d, l[:lang])
 
-          # Fetch abstracts.
-          abstract_content = d.css("div[itemprop='description'] p").text
-          next if abstract_content.empty?
-
-          abstract << {
-            content: abstract_content,
-            language: lang[:lang],
-            script: script(lang[:lang]),
-            format: "text/plain",
-          }
+            # Fetch abstracts.
+            abstract_content = d.css("div[itemprop='description'] p").text
+            unless abstract_content.empty?
+              abstract << {
+                content: abstract_content,
+                language: l[:lang],
+                script: script(l[:lang]),
+                format: "text/plain",
+              }
+            end
+          end
+          s
         end
-        [titles, abstract]
+        [titles, abstract, langs]
       end
       # rubocop:enable Metrics/AbcSize, Metrics/MethodLength
 
-      # Get langs.
+      # Returns available languages.
       # @param doc [Nokogiri::HTML::Document]
+      # @pqrqm lang [String, NilClass]
       # @return [Array<Hash>]
-      def langs(doc)
+      def languages(doc, lang)
         lgs = [{ lang: "en" }]
         doc.css("li#lang-switcher ul li a").each do |lang_link|
           lang_path = lang_link.attr("href")
-          lang = lang_path.match(%r{^\/(fr)\/})
-          lgs << { lang: lang[1], path: lang_path } if lang
+          l = lang_path.match(%r{^\/(fr)\/})
+          lgs << { lang: l[1], path: lang_path } if l && (!lang || l[1] == lang)
         end
         lgs
       end
@@ -200,9 +207,12 @@ module RelatonIso
       def fetch_relations(doc)
         doc.css("ul.steps li").reduce([]) do |a, r|
           r_type = r.css("strong").text
+          date = []
           type = case r_type
                  when "Previously", "Will be replaced by" then "obsoletes"
                  when "Corrigenda/Amendments", "Revised by", "Now confirmed"
+                   date << { type: "circulated",
+                     on: doc.xpath('//span[@class="stage-date"]').last.text }
                    "updates"
                  else r_type
                  end
@@ -213,7 +223,7 @@ module RelatonIso
                 content: id.text, format: "text/plain",
               )
               bibitem = RelatonIsoBib::IsoBibliographicItem.new(
-                formattedref: fref,
+                formattedref: fref, date: date
               )
               { type: type, bibitem: bibitem }
             end
@@ -257,6 +267,7 @@ module RelatonIso
       def script(lang)
         case lang
         when "en", "fr" then "Latn"
+        # when "ru" then "Cyrl"
         end
       end
 

data/lib/relaton_iso/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonIso
-  VERSION = "0.8.0"
+  VERSION = "0.8.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iso
 version: !ruby/object:Gem::Version
-  version: 0.8.0
+  version: 0.8.1
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2020-01-11 00:00:00.000000000 Z
+date: 2020-01-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: byebug