Merge remote-tracking branch 'origin' into skin-tones

.github/workflows/test.yml

@@ -0,0 +1,24 @@
+name: Test
+
+on: [push, pull_request]
+
+jobs:
+  test:
+    strategy:
+      fail-fast: false
+      matrix:
+        ruby: ["2.7", "3.0", "3.1"]
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: Set up Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: ${{matrix.ruby}}
+        bundler-cache: true
+
+    - name: Run tests
+      run: bundle exec rake

Gemfile

@@ -2,5 +2,6 @@ source "https://rubygems.org"
 
 gem "rake", "~> 10.3.2"
 gem "minitest", "~> 5.3.5"
+gem "i18n", "~> 1.8.5"
 
 gemspec

Gemfile.lock

@@ -0,0 +1,25 @@
+PATH
+  remote: .
+  specs:
+    gemoji (4.0.0.rc3)
+
+GEM
+  remote: https://rubygems.org/
+  specs:
+    concurrent-ruby (1.1.7)
+    i18n (1.8.5)
+      concurrent-ruby (~> 1.0)
+    minitest (5.3.5)
+    rake (10.3.2)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  gemoji!
+  i18n (~> 1.8.5)
+  minitest (~> 5.3.5)
+  rake (~> 10.3.2)
+
+BUNDLED WITH
+   2.0.2

Rakefile

@@ -20,5 +20,5 @@ namespace :db do
 end
 
 file 'vendor/unicode-emoji-test.txt' do |t|
-  system 'curl', '-fsSL', 'http://unicode.org/Public/emoji/12.0/emoji-test.txt', '-o', t.name
+  system 'curl', '-fsSL', 'http://unicode.org/Public/emoji/14.0/emoji-test.txt', '-o', t.name
 end

db/dump.rb

@@ -1,9 +1,11 @@
 # frozen_string_literal: true
 
+require "i18n"
 require 'emoji'
 require 'json'
 require_relative './emoji-test-parser'
 
+I18n.config.available_locales = :en
 items = []
 
 _, categories = EmojiTestParser.parse(File.expand_path("../../vendor/unicode-emoji-test.txt", __FILE__))
@@ -14,12 +16,15 @@
     for emoji_item in sub_category[:emoji]
       raw = emoji_item[:sequences][0]
       existing_emoji = Emoji.find_by_unicode(raw) || Emoji.find_by_unicode("#{raw}\u{fe0f}")
-      # next unless existing_emoji
-      existing_emoji = nil if seen_existing.key?(existing_emoji)
-      seen_existing[existing_emoji] = true
+      if seen_existing.key?(existing_emoji)
+        existing_emoji = nil
+      else
+        seen_existing[existing_emoji] = true
+      end
+      description = emoji_item[:description].sub(/^E\d+(\.\d+)? /, '')
       output_item = {
         emoji: raw,
-        description: emoji_item[:description],
+        description: description,
         category: category[:name],
       }
       if existing_emoji
@@ -31,10 +36,10 @@
         )
       else
         output_item.update(
-          aliases: [emoji_item[:description].gsub(/\W+/, '_').downcase],
+          aliases: [I18n.transliterate(description).gsub(/\W+/, '_').downcase],
           tags: [],
-          unicode_version: "12.0",
-          ios_version: "13.0",
+          unicode_version: "14.0",
+          ios_version: "15.4",
         )
       end
       output_item[:skin_tones] = true if emoji_item[:skin_tones]

db/emoji-test-parser.rb

@@ -9,12 +9,8 @@ module EmojiTestParser
     "\u{1F3FE}", # medium-dark skin tone
     "\u{1F3FF}", # dark skin tone
   ]
-  HAIR_MODIFIERS = [
-    "\u{1F9B0}", # red-haired
-    "\u{1F9B1}", # curly-haired
-    "\u{1F9B2}", # bald
-    "\u{1F9B3}", # white-haired
-  ]
+  SKIN_TONES_RE = /(#{SKIN_TONES.join("|")})/o
+  SKIP_TYPES = ["unqualified", "component"]
 
   module_function
 
@@ -52,12 +48,12 @@ def parse_file(io)
         else
           row, desc = line.split("#", 2)
           desc = desc.strip.split(" ", 2)[1]
-          codepoints, _ = row.split(";", 2)
+          codepoints, qualification = row.split(";", 2)
+          next if SKIP_TYPES.include?(qualification.strip)
           emoji_raw = codepoints.strip.split.map { |c| c.hex }.pack("U*")
-          next if HAIR_MODIFIERS.include?(emoji_raw)
           emoji_normalized = emoji_raw
             .gsub(VARIATION_SELECTOR_16, "")
-            .gsub(/(#{SKIN_TONES.join("|")})/o, "")
+            .gsub(SKIN_TONES_RE, "")
           emoji_item = emoji_map[emoji_normalized]
           if SKIN_TONES.any? { |s| emoji_raw.include?(s) }
             emoji_item[:skin_tones] = true if emoji_item
@@ -91,7 +87,7 @@ def parse_file(io)
     html_output = true
   end
 
-  _, categories = EmojiTestParser.parse
+  _, categories = EmojiTestParser.parse(File.expand_path("../../vendor/unicode-emoji-test.txt", __FILE__))
 
   trap(:PIPE) { abort }