isodoc-i18n 1.0.7 → 1.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ce8cb0d922cdc28fe16fc3605db3d205fe66c12d0224282f6597d9510d44ee6d
-  data.tar.gz: 6fe8fc95e446fdc6b57eb9172d6f4d0665fd2f48ed0c29877819c3d8398805dc
+  metadata.gz: 8f00540677aae6f53c0c517b376529ce6eb5094c6cebd1d1a9b31ce20e346266
+  data.tar.gz: 69dd44b3aedc560540c5de52f05f0ae11b1332edccb4565817845648f23a9631
 SHA512:
-  metadata.gz: 75991361380709948d84f9707bb999e3637f634433ebd83ee3f0b24e783381131326fa824e1897f8900a6cf8c30522a56589887fc93def7458480eadaaa60c21
-  data.tar.gz: 0e1a2149f1a06190962dda186978f01677d35ef25caf711dd058753741dca79b74c3de98cb04cbe0be4d1ba770b5f649589d7fd0b9f12e6dafb04b58372fa7fe
+  metadata.gz: 8b3a53fe9058b2cba6f929938e61147adcb6eaa0782705ca2097cf448f63a76500abbe4b4d3513b9d0cfca04456547d8127e278ef7a521ec3b1cff6de8302f58
+  data.tar.gz: 24ff8195e4172769e91fc4f879fe6835b4f24170041e16198ab3a405625f07ed56bce3a702604b5cb25102b4d6f21ce87d5d92889122cd524a4694549ba8d0c8

data/lib/isodoc/i18n.rb

@@ -54,9 +54,10 @@ module IsoDoc
       @labels[key] = val
     end
 
-    def initialize(lang, script, i18nyaml: nil, i18nhash: nil)
+    def initialize(lang, script, locale: nil, i18nyaml: nil, i18nhash: nil)
       @lang = lang
       @script = script
+      @locale = locale
       y = load_yaml(lang, script, i18nyaml, i18nhash)
       @labels = y
       @labels["language"] = @lang
@@ -66,17 +67,16 @@ module IsoDoc
       end
     end
 
-    def self.l10n(text, lang = @lang, script = @script)
-      l10n(text, lang, script)
+    def self.l10n(text, lang = @lang, script = @script, locale = @locale)
+      l10n(text, lang, script, locale)
     end
 
-    # TODO: move to localization file
     # function localising spaces and punctuation.
     # Not clear if period needs to be localised for zh
-    def l10n(text, lang = @lang, script = @script)
-      if lang == "zh" && script == "Hans" then l10n_zh(text)
-      else bidiwrap(text, lang, script)
-      end
+    def l10n(text, lang = @lang, script = @script, locale = @locale)
+      lang == "zh" && script == "Hans" and text = l10n_zh(text)
+      lang == "fr" && text = l10n_fr(text, locale || "FR")
+      bidiwrap(text, lang, script)
     end
 
     def bidiwrap(text, lang, script)
@@ -107,23 +107,53 @@ module IsoDoc
       xml.to_xml.gsub(/<b>/, "").gsub("</b>", "").gsub(/<\?[^>]+>/, "")
     end
 
+    def l10n_fr(text, locale)
+      xml = Nokogiri::HTML::DocumentFragment.parse(text)
+      xml.traverse do |n|
+        next unless n.text?
+
+        n.replace(cleanup_entities(l10n_fr1(n.text, locale), is_xml: false))
+      end
+      xml.to_xml
+    end
+
     ZH_CHAR = "\\p{Han}|\\p{In CJK Symbols And Punctuation}|"\
               "\\p{In Halfwidth And Fullwidth Forms}".freeze
 
     # note: we can't differentiate comma from enumeration comma 、
     def l10_zh1(text)
+      l10n_zh_remove_space(l10n_zh_punct(text))
+    end
+
+    def l10n_zh_punct(text)
       [":：", ",，", ".。", ")）", "]】", ":：", ";；", "?？", "!！"].each do |m|
         text = text.gsub(/(?<=#{ZH_CHAR})#{Regexp.quote m[0]}/, m[1])
+        text = text.gsub(/^#{Regexp.quote m[0]}/, m[1])
       end
       ["(（", "[【"].each do |m|
         text = text.gsub(/#{Regexp.quote m[0]}(?=#{ZH_CHAR})/, m[1])
       end
+      text
+    end
+
+    def l10n_zh_remove_space(text)
       text.gsub(/(?<=#{ZH_CHAR}) (?=#{ZH_CHAR})/o, "")
         .gsub(/(?<=\d) (?=#{ZH_CHAR})/o, "")
         .gsub(/(?<=#{ZH_CHAR}) (?=\d)/o, "")
         .gsub(/(?<=#{ZH_CHAR}) (?=[A-Za-z](#{ZH_CHAR}|$))/o, "")
     end
 
+    def l10n_fr1(text, locale)
+      text = text.gsub(/(?<=\p{Alnum})([»›;?!])(?=\s)/, "\u202f\\1")
+      text = text.gsub(/(?<=\p{Alnum})([»›;?!])$/, "\u202f\\1")
+      text = text.gsub(/^([»›;?!])/, "\u202f\\1")
+      text = text.gsub(/([«‹])/, "\\1\u202f")
+      colonsp = locale == "CH" ? "\u202f" : "\u00a0"
+      text = text.gsub(/(?<=\p{Alnum})(:)(?=\s)/, "#{colonsp}\\1")
+      text = text.gsub(/(?<=\p{Alnum})(:)$/, "#{colonsp}\\1")
+      text.gsub(/^(:\s)/, "#{colonsp}\\1")
+    end
+
     def boolean_conj(list, conn)
       case list.size
       when 0 then ""

data/lib/isodoc/version.rb

@@ -1,5 +1,5 @@
 module IsoDoc
   class I18n
-    VERSION = "1.0.7".freeze
+    VERSION = "1.1.1".freeze
   end
 end

data/spec/isodoc/base_spec.rb

@@ -38,7 +38,8 @@ RSpec.describe IsoDoc::I18n do
   end
 
   it "loads language hash overrides" do
-    c = IsoDoc::I18n.new("en", "Latn", i18nhash: YAML.load_file("spec/assets/new.yaml"))
+    c = IsoDoc::I18n.new("en", "Latn",
+                         i18nhash: YAML.load_file("spec/assets/new.yaml"))
     expect(c.text).to eq "text2"
     expect(c.at).to eq "at"
     expect(c.hash.to_s).to be_equivalent_to '{"key1"=>"val1", "key2"=>"val2"}'
@@ -85,6 +86,27 @@ RSpec.describe IsoDoc::I18n do
       .to be_equivalent_to "؜Code (hello, world.)؜"
   end
 
+  it "does French localisation" do
+    e = HTMLEntities.new
+    c = IsoDoc::I18n.new("fr", "Latn")
+    expect(e.encode(c.l10n("Code; «code» and: code!"), :hexadecimal))
+      .to be_equivalent_to "Code ; « code » "\
+                           "and : code !"
+    expect(e.encode(c.l10n("Code; «code» and: code!"), :hexadecimal))
+      .to be_equivalent_to "Code ; « code » "\
+                           "and : code !"
+    c = IsoDoc::I18n.new("fr", "Latn", locale: "FR")
+    expect(e.encode(c.l10n("Code; «code» and: code!"), :hexadecimal))
+      .to be_equivalent_to "Code ; « code » "\
+                           "and : code !"
+    c = IsoDoc::I18n.new("fr", "Latn", locale: "CH")
+    expect(e.encode(c.l10n("Code; «code» and: code!"), :hexadecimal))
+      .to be_equivalent_to "Code ; « code » "\
+                           "and : code !"
+    expect(e.encode(c.l10n("http://xyz a;b"), :hexadecimal))
+      .to be_equivalent_to "http://xyz a;b"
+  end
+
   it "does boolean conjunctions" do
     c = IsoDoc::I18n.new("en", "Latn")
     expect(c.boolean_conj([], "and")).to eq ""

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: isodoc-i18n
 version: !ruby/object:Gem::Version
-  version: 1.0.7
+  version: 1.1.1
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-09-05 00:00:00.000000000 Z
+date: 2022-09-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: htmlentities