unicode-emoji 2.9.0 → 3.1.1

data/lib/unicode/emoji.rb

@@ -3,212 +3,30 @@
 require "unicode/version"
 
 require_relative "emoji/constants"
-require_relative "emoji/index"
 
 module Unicode
   module Emoji
-    PROPERTY_NAMES = {
-      E: "Emoji",
-      B: "Emoji_Modifier_Base",
-      M: "Emoji_Modifier",
-      C: "Emoji_Component",
-      P: "Emoji_Presentation",
-      X: "Extended_Pictographic",
-    }
-
-    EMOJI_VARIATION_SELECTOR      = 0xFE0F
-    TEXT_VARIATION_SELECTOR       = 0xFE0E
-    EMOJI_TAG_BASE_FLAG           = 0x1F3F4
-    CANCEL_TAG                    = 0xE007F
-    TAGS                          = [*0xE0020..0xE007E]
-    EMOJI_KEYCAP_SUFFIX           = 0x20E3
-    ZWJ                           = 0x200D
-    REGIONAL_INDICATORS           = [*0x1F1E6..0x1F1FF]
-
-    EMOJI_CHAR                    = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:E) }.keys.freeze
-    EMOJI_PRESENTATION            = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:P) }.keys.freeze
-    TEXT_PRESENTATION             = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:E) && !props.include?(:P) }.keys.freeze
-    EMOJI_COMPONENT               = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:C) }.keys.freeze
-    EMOJI_MODIFIER_BASES          = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:B) }.keys.freeze
-    EMOJI_MODIFIERS               = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:M) }.keys.freeze
-
-    EXTENDED_PICTOGRAPHIC         = INDEX[:PROPERTIES].select{ |ord, props| props.include?(:X) }.keys.freeze
-    EXTENDED_PICTOGRAPHIC_NO_EMOJI= INDEX[:PROPERTIES].select{ |ord, props| props.include?(:X) && !props.include?(:E) }.keys.freeze
-    EMOJI_KEYCAPS                 = INDEX[:KEYCAPS].freeze
-    VALID_REGION_FLAGS            = INDEX[:FLAGS].freeze
-    VALID_SUBDIVISIONS            = INDEX[:SD].freeze
-    RECOMMENDED_SUBDIVISION_FLAGS = INDEX[:TAGS].freeze
-    RECOMMENDED_ZWJ_SEQUENCES     = INDEX[:ZWJ].freeze
-
-    LIST                          = INDEX[:LIST].freeze.each_value(&:freeze)
-    LIST_REMOVED_KEYS             = [
-      "Smileys & People",
-    ]
-
-    pack = ->(ord){ Regexp.escape(Array(ord).pack("U*")) }
-    join = -> (*strings){ "(?:" + strings.join("|") + ")" }
-    pack_and_join = ->(ords){  join[*ords.map{ |ord| pack[ord] }] }
-
-    if EMOJI_VERSION == Unicode::Version.emoji_version
-      emoji_character     = "\\p{Emoji}"
-      emoji_modifier      = "\\p{Emoji Modifier}"
-      emoji_modifier_base = "\\p{Emoji Modifier Base}"
-      emoji_component     = "\\p{Emoji Component}"
-      emoji_presentation  = "\\p{Emoji Presentation}"
-      picto               = "\\p{Extended Pictographic}"
-      picto_no_emoji      = "\\p{Extended Pictographic}(?<!\\p{Emoji})"
-    else
-      emoji_character     = pack_and_join[EMOJI_CHAR]
-      emoji_modifier      = pack_and_join[EMOJI_MODIFIERS]
-      emoji_modifier_base = pack_and_join[EMOJI_MODIFIER_BASES]
-      emoji_component     = pack_and_join[EMOJI_COMPONENT]
-      emoji_presentation  = pack_and_join[EMOJI_PRESENTATION]
-      picto               = pack_and_join[EXTENDED_PICTOGRAPHIC]
-      picto_no_emoji      = pack_and_join[EXTENDED_PICTOGRAPHIC_NO_EMOJI]
+    autoload :INDEX, File.expand_path('emoji/index', __dir__)
+
+    %w[
+      EMOJI_CHAR EMOJI_CHAR EMOJI_PRESENTATION TEXT_PRESENTATION EMOJI_COMPONENT EMOJI_MODIFIER_BASES
+      EMOJI_MODIFIERS EXTENDED_PICTOGRAPHIC EXTENDED_PICTOGRAPHIC_NO_EMOJI EMOJI_KEYCAPS VALID_REGION_FLAGS
+      VALID_SUBDIVISIONS RECOMMENDED_SUBDIVISION_FLAGS RECOMMENDED_ZWJ_SEQUENCES LIST LIST_REMOVED_KEYS
+    ].each do |const_name|
+      autoload const_name, File.expand_path('emoji/lazy_constants', __dir__)
     end
 
-    emoji_presentation_sequence = \
-      join[
-        pack_and_join[TEXT_PRESENTATION] + pack[EMOJI_VARIATION_SELECTOR],
-        emoji_presentation + "(?!" + pack[TEXT_VARIATION_SELECTOR] + ")" + pack[EMOJI_VARIATION_SELECTOR] + "?",
-      ]
-
-    non_component_emoji_presentation_sequence = \
-      "(?!" + emoji_component + ")" + emoji_presentation_sequence
-
-    text_presentation_sequence = \
-      join[
-        pack_and_join[TEXT_PRESENTATION]+ "(?!" + join[emoji_modifier, pack[EMOJI_VARIATION_SELECTOR]] + ")" + pack[TEXT_VARIATION_SELECTOR] + "?",
-        emoji_presentation + pack[TEXT_VARIATION_SELECTOR]
-      ]
-
-    emoji_modifier_sequence = \
-      emoji_modifier_base + emoji_modifier
-
-    emoji_keycap_sequence = \
-      pack_and_join[EMOJI_KEYCAPS] + pack[[EMOJI_VARIATION_SELECTOR, EMOJI_KEYCAP_SUFFIX]]
-
-    emoji_valid_flag_sequence = \
-      pack_and_join[VALID_REGION_FLAGS]
-
-    emoji_well_formed_flag_sequence = \
-      "(?:" +
-        pack_and_join[REGIONAL_INDICATORS] +
-        pack_and_join[REGIONAL_INDICATORS] +
-      ")"
-
-    emoji_valid_core_sequence = \
-      join[
-        # emoji_character,
-        emoji_keycap_sequence,
-        emoji_modifier_sequence,
-        non_component_emoji_presentation_sequence,
-        emoji_valid_flag_sequence,
-      ]
-
-    emoji_well_formed_core_sequence = \
-      join[
-        # emoji_character,
-        emoji_keycap_sequence,
-        emoji_modifier_sequence,
-        non_component_emoji_presentation_sequence,
-        emoji_well_formed_flag_sequence,
-      ]
-
-    emoji_rgi_tag_sequence = \
-      pack_and_join[RECOMMENDED_SUBDIVISION_FLAGS]
-
-    emoji_valid_tag_sequence = \
-      "(?:" +
-        pack[EMOJI_TAG_BASE_FLAG] +
-        "(?:" + VALID_SUBDIVISIONS.map{ |sd| Regexp.escape(sd.tr("\u{20}-\u{7E}", "\u{E0020}-\u{E007E}"))}.join("|") + ")" +
-        pack[CANCEL_TAG] +
-      ")"
-
-    emoji_well_formed_tag_sequence = \
-      "(?:" +
-        join[
-          non_component_emoji_presentation_sequence,
-          emoji_modifier_sequence,
-        ] +
-        pack_and_join[TAGS] + "+" +
-        pack[CANCEL_TAG] +
-      ")"
-
-    emoji_rgi_zwj_sequence = \
-      pack_and_join[RECOMMENDED_ZWJ_SEQUENCES]
-
-    emoji_valid_zwj_element = \
-      join[
-        emoji_modifier_sequence,
-        emoji_presentation_sequence,
-        emoji_character,
-      ]
-
-    emoji_valid_zwj_sequence = \
-      "(?:" +
-        "(?:" + emoji_valid_zwj_element + pack[ZWJ] + ")+" + emoji_valid_zwj_element +
-      ")"
-
-    emoji_rgi_sequence = \
-      join[
-        emoji_rgi_zwj_sequence,
-        emoji_rgi_tag_sequence,
-        emoji_valid_core_sequence,
-      ]
-
-    emoji_valid_sequence = \
-      join[
-        emoji_valid_zwj_sequence,
-        emoji_valid_tag_sequence,
-        emoji_valid_core_sequence,
-      ]
-
-    emoji_well_formed_sequence = \
-      join[
-        emoji_valid_zwj_sequence,
-        emoji_well_formed_tag_sequence,
-        emoji_well_formed_core_sequence,
-      ]
-
-    # Matches basic singleton emoji and all kind of sequences, but restrict zwj and tag sequences to known sequences (rgi)
-    REGEX = Regexp.compile(emoji_rgi_sequence)
-
-    # Matches basic singleton emoji and all kind of valid sequences
-    REGEX_VALID = Regexp.compile(emoji_valid_sequence)
-
-    # Matches basic singleton emoji and all kind of sequences
-    REGEX_WELL_FORMED = Regexp.compile(emoji_well_formed_sequence)
-
-    # Matches only basic single, non-textual emoji
-    # Ignores "components" like modifiers or simple digits
-    REGEX_BASIC = Regexp.compile(
-      "(?!" + emoji_component + ")" + emoji_presentation_sequence
+    generated_constants_dirpath = File.expand_path(
+      EMOJI_VERSION == Unicode::Version.emoji_version ? "emoji/generated_native/" : "emoji/generated/",
+      __dir__
     )
 
-    # Matches only basic single, textual emoji
-    # Ignores "components" like modifiers or simple digits
-    REGEX_TEXT = Regexp.compile(
-      "(?!" + emoji_component + ")" + text_presentation_sequence
-    )
-
-    # Matches any emoji-related codepoint - Use with caution (returns partil matches)
-    REGEX_ANY = Regexp.compile(
-      emoji_character
-    )
-
-    # Combined REGEXes which also match for TEXTUAL emoji
-    REGEX_INCLUDE_TEXT = Regexp.union(REGEX, REGEX_TEXT)
-    REGEX_VALID_INCLUDE_TEXT = Regexp.union(REGEX_VALID, REGEX_TEXT)
-    REGEX_WELL_FORMED_INCLUDE_TEXT = Regexp.union(REGEX_WELL_FORMED, REGEX_TEXT)
-
-    REGEX_PICTO = Regexp.compile(
-      picto
-    )
-
-    REGEX_PICTO_NO_EMOJI = Regexp.compile(
-      picto_no_emoji
-    )
+    %w[
+      REGEX REGEX_VALID REGEX_WELL_FORMED REGEX_BASIC REGEX_TEXT REGEX_ANY REGEX_INCLUDE_TEXT
+      REGEX_VALID_INCLUDE_TEXT REGEX_WELL_FORMED_INCLUDE_TEXT REGEX_PICTO REGEX_PICTO_NO_EMOJI
+    ].each do |const_name|
+      autoload const_name, File.join(generated_constants_dirpath, const_name.downcase)
+    end
 
     def self.properties(char)
       ord = get_codepoint_value(char)

data/spec/unicode_emoji_spec.rb

@@ -108,6 +108,13 @@ describe Unicode::Emoji do
       "🪺 nest with eggs" =~ Unicode::Emoji::REGEX
       assert_equal "🪺", $&
     end
+
+    # See gh#12 and https://github.com/matt17r/nw5k/commit/05a34d3c9211a23e5ae6853bb19fd2f224779ef4#diff-afb6f8bc3bae71b75743e00882a060863e2430cbe858ec9014e5956504dfc61cR2
+    it "matches family emoji correctly" do
+      ["👨‍👩‍👧‍👦", "👨‍👩‍👦‍👦", "👨‍👩‍👧‍👧", "👨‍👨‍👧‍👦", "👨‍👨‍👦‍👦", "👨‍👨‍👧‍👧", "👩‍👩‍👧‍👦", "👩‍👩‍👦‍👦", "👩‍👩‍👧‍👧", "👨‍👦‍👦", "👨‍👧‍👦", "👨‍👧‍👧", "👩‍👦‍👦", "👩‍👧‍👦", "👩‍👧‍👧"].each { |family|
+        assert_equal family, family[Unicode::Emoji::REGEX]
+      }
+    end
   end
 
   describe "REGEX_VALID" do
@@ -407,7 +414,7 @@ describe Unicode::Emoji do
     end
   end
 
-  describe "REGEX_PICTO" do
+  describe "REGEX_PICTO_NO_EMOJI" do
     it "matches codepoints with Extended_Pictograph property, but no Emoji property" do
       matches = "U+1F32D 🌭 HOT DOG, U+203C ‼ DOUBLE EXCLAMATION MARK, U+26E8 ⛨ BLACK CROSS ON SHIELD".scan(Unicode::Emoji::REGEX_PICTO_NO_EMOJI)
       assert_equal ["⛨"], matches

data/unicode-emoji.gemspec

@@ -16,6 +16,7 @@ Gem::Specification.new do |gem|
   gem.executables   = gem.files.grep(%r{^bin/}).map{ |f| File.basename(f) }
   gem.test_files    = gem.files.grep(%r{^(test|spec|features)/})
   gem.require_paths = ["lib"]
+  gem.metadata      = { "rubygems_mfa_required" => "true" }
 
   gem.required_ruby_version = ">= 2.3", "< 4.0"
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: unicode-emoji
 version: !ruby/object:Gem::Version
-  version: 2.9.0
+  version: 3.1.1
 platform: ruby
 authors:
 - Jan Lelis
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-09-15 00:00:00.000000000 Z
+date: 2022-06-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: unicode-version
@@ -42,15 +42,40 @@ files:
 - README.md
 - Rakefile
 - data/emoji.marshal.gz
+- data/generate_constants.rb
 - lib/unicode/emoji.rb
 - lib/unicode/emoji/constants.rb
+- lib/unicode/emoji/generated/regex.rb
+- lib/unicode/emoji/generated/regex_any.rb
+- lib/unicode/emoji/generated/regex_basic.rb
+- lib/unicode/emoji/generated/regex_include_text.rb
+- lib/unicode/emoji/generated/regex_picto.rb
+- lib/unicode/emoji/generated/regex_picto_no_emoji.rb
+- lib/unicode/emoji/generated/regex_text.rb
+- lib/unicode/emoji/generated/regex_valid.rb
+- lib/unicode/emoji/generated/regex_valid_include_text.rb
+- lib/unicode/emoji/generated/regex_well_formed.rb
+- lib/unicode/emoji/generated/regex_well_formed_include_text.rb
+- lib/unicode/emoji/generated_native/regex.rb
+- lib/unicode/emoji/generated_native/regex_any.rb
+- lib/unicode/emoji/generated_native/regex_basic.rb
+- lib/unicode/emoji/generated_native/regex_include_text.rb
+- lib/unicode/emoji/generated_native/regex_picto.rb
+- lib/unicode/emoji/generated_native/regex_picto_no_emoji.rb
+- lib/unicode/emoji/generated_native/regex_text.rb
+- lib/unicode/emoji/generated_native/regex_valid.rb
+- lib/unicode/emoji/generated_native/regex_valid_include_text.rb
+- lib/unicode/emoji/generated_native/regex_well_formed.rb
+- lib/unicode/emoji/generated_native/regex_well_formed_include_text.rb
 - lib/unicode/emoji/index.rb
+- lib/unicode/emoji/lazy_constants.rb
 - spec/unicode_emoji_spec.rb
 - unicode-emoji.gemspec
 homepage: https://github.com/janlelis/unicode-emoji
 licenses:
 - MIT
-metadata: {}
+metadata:
+  rubygems_mfa_required: 'true'
 post_install_message: 
 rdoc_options: []
 require_paths:
@@ -69,7 +94,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.3
+rubygems_version: 3.3.7
 signing_key: 
 specification_version: 4
 summary: Retrieve Emoji data about Unicode codepoints.