relaton-iso 1.18.3 → 1.18.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9a8570714304ad5ccdf68c5861a53b346f14235e0839bcaeba9b5221d93d0b4f
-  data.tar.gz: 39b230964324c2753d24440bb545233771251b2fcf2e99ef9774c9fc90f10593
+  metadata.gz: 9d84323d65b954dcb15e73fb00917f6420494f11f1dfb5dbb5bf59e8fcae15ef
+  data.tar.gz: f9b9364c9804f46af7ad56d2b1c708a3e98760c9948142686f5ace87807a628c
 SHA512:
-  metadata.gz: a5b9cec3bbd724d0c0ea0fcabe4f499e57c1effa4c9aa4abeb3323b143608997de21a6013dd89f88048c811442139e44a2b61221abdcee08af688522314f2478
-  data.tar.gz: a4c3923795a224b728a7353988fc1f51f443cfbcec330de7ef2a45ee8d568193aa928ccd9eb790b551300ee75de084f7d5a2f01286a6babb95dd9d8e0d1264c4
+  metadata.gz: b242baf862c63b243d6a4c41611cc9e6f900cc269e78ed80c36d8e44fabad7a2f2ae6abcf72b8661c65467e52193c4b922d7d15b6cf82c187f5620f5264135cc
+  data.tar.gz: fb95d361df7ed56a0c6b33cc2239a2445f44e7a41f8f217f41a7b337629a3021efe59a11a6686f951fbf79f8fec747dc741978308b1a800c94601f8d0061be36

data/lib/relaton_iso/scrapper.rb

@@ -56,7 +56,7 @@ module RelatonIso
     # @return [RelatonIsoBib::IsoBibliographicItem]
     def parse_page(path, lang = nil) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
       doc, url = get_page path
-      id = doc.at("//nav[contains(@class,'heading-condensed')]/h1").text.split(" | ").first
+      id = doc.at("//h1/span[1]").text.split(" | ").first.strip
       pubid = Pubid::Iso::Identifier.parse(id)
       # Fetch edition.
       edition = doc.at("//div[div[.='Edition']]/text()[last()]")&.text&.match(/\d+$/)&.to_s
@@ -279,7 +279,7 @@ module RelatonIso
     # @return [String, nil] ID
     #
     def item_ref(doc)
-      doc.at("//main//section/div/div/div//h1")&.text
+      doc.at("//main//section/div/div/div//h1/span[1]")&.text
     end
 
     # Fetch status.
@@ -327,7 +327,9 @@ module RelatonIso
     # @return [Array<Hash>]
     def fetch_relations(doc)
       types = ["Now", "Now under review"]
-      doc.xpath("//ul[@class='steps']/li", "//div[@class='sub-step']").reduce([]) do |a, r|
+      doc.xpath(
+        "//ul[@class='steps']/li", "//div[contains(@class, 'sub-step')]"
+      ).reduce([]) do |a, r|
         type, date = relation_type(r.at("h4", "h5").text.strip, doc)
         next a if types.include?(type)
 
@@ -347,7 +349,7 @@ module RelatonIso
       date = []
       t = case type.strip
           when "Previously", "Will be replaced by" then "obsoletes"
-          when "Corrigenda / Amendments", "Revised by", "Now confirmed"
+          when /Corrigenda|Amendments|Revised by|Now confirmed|replaced by/
             on = doc.xpath('//span[@class="stage-date"][contains(., "-")]').last
             date << { type: "circulated", on: on.text } if on
             "updates"
@@ -411,9 +413,9 @@ module RelatonIso
     end
 
     def titles(doc)
-      head = doc.at "//nav[contains(@class,'heading-condensed')]"
-      ttls = head.xpath("h2 | h3 | h4").map &:text
-      ttls = ttls[0].split " - " if ttls.size == 1
+      # head = doc.at "//nav[contains(@class,'heading-condensed')]"
+      ttls = doc.xpath("//h1[@class='stdTitle']/span[position()>1]").map(&:text)
+      ttls[0, 1] = ttls[0].split(/\s(?:-|\u2014)\s/) # if ttls.size == 1
       case ttls.size
       when 0, 1 then [nil, ttls.first, nil]
       else RelatonBib::TypedTitleString.intro_or_part ttls
@@ -489,7 +491,7 @@ module RelatonIso
     #
     def fetch_link(doc, url)
       links = [{ type: "src", content: url }]
-      obp = doc.at("//h4[contains(@class, 'h5')]/a")
+      obp = doc.at("//a[.='Read sample']")
       links << { type: "obp", content: obp[:href] } if obp
       rss = doc.at("//a[contains(@href, 'rss')]")
       links << { type: "rss", content: DOMAIN + rss[:href] } if rss

data/lib/relaton_iso/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RelatonIso
-  VERSION = "1.18.3"
+  VERSION = "1.18.4"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: relaton-iso
 version: !ruby/object:Gem::Version
-  version: 1.18.3
+  version: 1.18.4
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2024-05-20 00:00:00.000000000 Z
+date: 2024-05-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pubid