tanuki_emoji 0.3.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c67213658b22404d9912d1a6704748c1106602013ea3263e819dff1c05ee4dc8
-  data.tar.gz: 205335429c995e7fc884867b50caed6beb2f1bec90c9a4fa3bf31afdc364aa64
+  metadata.gz: f515e18873314d0d4fc8d490645c7c60c13c842016f8efa9a7bc559fc5c00ae9
+  data.tar.gz: ff2e0ad3a8991e2091c56051ae9a0005cb079ae2a2052915302668119b56a303
 SHA512:
-  metadata.gz: a32137a26584a98e2e59954829e769cfdacb126986debc443118477cae7c5bd16cde127a75cb606c9674e31545102df12792ac58ff25bab479d49676553be297
-  data.tar.gz: 4fc240c5cd68e53f5721063f74f32b4bd2e4f60840882fb55dc31cb23163615f0096d4616bed8c730ac198322d87a5328c9b6cf7b0ab7e1e2b95fb0cdad43207
+  metadata.gz: f45be46c767ccb611da4908fe0f474c1ef2f2a83802dead03256a99fb4f5b5a25eefe812dd51f6602536ed9dca74e5757ee16d0498737b9247c846642e4cbc3c
+  data.tar.gz: cccefbe3827f60f3e1b764068647e464f0a4ff1ff93d12d83225396017a33fcffd2dc6e7106dfa45708cc576d8596d4f5cf198b0026438a7360f82319a322e40

data/.ruby-version

@@ -1 +1 @@
-2.7
+2.7.4

data/CHANGELOG.md

@@ -6,6 +6,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.4.0] - 2021-09-07
+### Added
+- Index can return the `alpha_code_pattern` and `codepoints_pattern` to be used as text extraction sources
+
+### Fixed
+- Fixed `registered sign`, `copyright sign` and `trade mark sign` codepoints from gemojione index
+
 ## [0.3.0] - 2021-08-26
 ### Modified
 - Characters can be compared and will be considered equal when all of its attributes matches
@@ -34,3 +41,5 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 [0.2.0]:https://gitlab.com/gitlab-org/tanuki_emoji/-/releases/v0.2.0
 [0.2.1]:https://gitlab.com/gitlab-org/tanuki_emoji/-/releases/v0.2.1
 [0.2.2]:https://gitlab.com/gitlab-org/tanuki_emoji/-/releases/v0.2.2
+[0.3.0]:https://gitlab.com/gitlab-org/tanuki_emoji/-/releases/v0.3.0
+[0.4.0]:https://gitlab.com/gitlab-org/tanuki_emoji/-/releases/v0.4.0

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    tanuki_emoji (0.3.0)
+    tanuki_emoji (0.4.0)
 
 GEM
   remote: https://rubygems.org/
@@ -54,7 +54,9 @@ GEM
     unicode-display_width (2.0.0)
 
 PLATFORMS
+  ruby
   x86_64-darwin-20
+  x86_64-linux
 
 DEPENDENCIES
   rake (~> 13.0)
@@ -66,4 +68,4 @@ DEPENDENCIES
   tanuki_emoji!
 
 BUNDLED WITH
-   2.2.15
+   2.2.22

data/lib/tanuki_emoji/character.rb

@@ -9,6 +9,7 @@ module TanukiEmoji
     IMAGE_PREFIX = 'emoji_u'
     IMAGE_EXTENSION = '.png'
     FLAG_REGEXP = /[🇦-🇿]{2}/u.freeze
+    ALPHA_CODE_REGEXP = /:(?<alpha_text>[_+\-a-z0-9]+):/.freeze
 
     # This denotes a "color" or "emoji" version
     EMOJI_VARIATION_SELECTOR = 0xFE0F
@@ -25,7 +26,7 @@ module TanukiEmoji
     # @param [String] codepoints
     # @param [String] alpha_code
     def initialize(name, codepoints:, alpha_code:, description:)
-      @name = name
+      @name = self.class.format_name(name)
       @codepoints = codepoints
       @codepoints_alternates = []
       @alpha_code = self.class.format_alpha_code(alpha_code)
@@ -91,14 +92,18 @@ module TanukiEmoji
         description == other.description
     end
 
-    # Convert Unicode code points to Hex format for inspection
-    #
-    # ensure alpha code is formatted with colons
+    # Ensure alpha code is formatted with colons
     #
     # @param [String] alpha_code
     # @return [String] formatted alpha code
     def self.format_alpha_code(alpha_code)
-      alpha_code.to_s.match?(/:([_+\-a-z0-9]+):/) ? alpha_code.to_s : ":#{alpha_code}:"
+      alpha_code.to_s.match?(ALPHA_CODE_REGEXP) ? alpha_code.to_s : ":#{alpha_code}:"
+    end
+
+    def self.format_name(raw_name)
+      matched = raw_name.match(ALPHA_CODE_REGEXP)
+
+      matched ? matched['alpha_text'] : raw_name
     end
 
     private

data/lib/tanuki_emoji/index.rb

@@ -36,7 +36,8 @@ module TanukiEmoji
         end
 
         emoji.aliases.each do |alpha_code|
-          @alpha_code_index[alpha_code] = emoji
+          @name_index[::TanukiEmoji::Character.format_name(alpha_code)] = emoji
+          @alpha_code_index[::TanukiEmoji::Character.format_alpha_code(alpha_code)] = emoji
         end
       end
 
@@ -75,6 +76,22 @@ module TanukiEmoji
       load_data_files if reload
     end
 
+    # Return a regular expression that can be used to search for indexed `:alpha_codes:`
+    #
+    # @return [Regexp] regular expression that matches indexed `:alpha_code:`
+    def alpha_code_pattern
+      /(?<=[^[:alnum:]:]|\n|^)
+       :(#{@name_index.keys.map { |name| Regexp.escape(name) }.join('|')}):
+       (?=[^[:alnum:]:]|$)/x
+    end
+
+    # Return a regular expression that can be used to search for emoji codepoints
+    #
+    # @return [Regexp] regular expression that matches indexed emoji codepoints
+    def codepoints_pattern
+      /(#{@codepoints_index.keys.map { |moji| Regexp.escape(moji) }.join('|')})/
+    end
+
     private
 
     def initialize

data/lib/tanuki_emoji/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module TanukiEmoji
-  VERSION = "0.3.0"
+  VERSION = "0.4.0"
 end

data/vendor/gemojione/index-3.3.0.json

@@ -5861,7 +5861,7 @@
     "moji": "👮🏿"
   },
   "copyright": {
-    "moji": "©",
+    "moji": "©️",
     "unicode": "00A9",
     "unicode_alternates": [],
     "name": "copyright sign",
@@ -25694,7 +25694,7 @@
     "moji": "🔴"
   },
   "registered": {
-    "moji": "®",
+    "moji": "®️",
     "unicode": "00AE",
     "unicode_alternates": [],
     "name": "registered sign",
@@ -30433,7 +30433,7 @@
       "tm",
       "word"
     ],
-    "moji": "™"
+    "moji": "™️"
   },
   "toilet": {
     "unicode": "1F6BD",

data/vendor/gemojione/index-4.1.0.json

@@ -7630,7 +7630,7 @@
     "moji": "👮🏿"
   },
   "copyright": {
-    "moji": "©",
+    "moji": "©️",
     "unicode": "00A9",
     "unicode_alternates": [
 
@@ -33231,7 +33231,7 @@
     "moji": "🔴"
   },
   "registered": {
-    "moji": "®",
+    "moji": "®️",
     "unicode": "00AE",
     "unicode_alternates": [
 
@@ -39373,7 +39373,7 @@
       "tm",
       "word"
     ],
-    "moji": "™"
+    "moji": "™️"
   },
   "toilet": {
     "unicode": "1F6BD",

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tanuki_emoji
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Gabriel Mazetto
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-08-26 00:00:00.000000000 Z
+date: 2021-09-08 00:00:00.000000000 Z
 dependencies: []
 description: Tanuki Emoji provides Emoji character information and metadata with support
   for Noto Emoji resources as fallback
@@ -1872,7 +1872,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.15
+rubygems_version: 3.1.6
 signing_key:
 specification_version: 4
 summary: Tanuki Emoji