script_detector_2 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 76df1da987c76f07c1116e360f2686bcb177d5be42d5faf8e8898561d9975f7a
-  data.tar.gz: 81e79124e3d9e1e9283891c12970a84ab09dd5b2f6aaa4c3e626e07f852d2696
+  metadata.gz: 9709f6c83f82a3bf7073bb8b9cfc95a1edea6885e0f596ef6e93f01a62c26dbb
+  data.tar.gz: 3ed583c8487617e9687b3c776116f0cc22140993ef9debf4f8542057e9f9f232
 SHA512:
-  metadata.gz: 34d104036ce6fd3aa8140cfb659d745e1d786996b7139d489cd352c1a4cb835991a4e45a8634c28d6ba9cdc65fe39d3c3a3da3bddd19dd3ae799f741059eb484
-  data.tar.gz: d7ddc49d2a0a835e1549632f7b0f7a07b6983b345798e491ed819f93a30c72ef6b90a0b1063105def6ee22f8defc3d92ebd7d19511d5b0cbc1ba29d55fc2fa29
+  metadata.gz: 5d17c19eee4868b540af844e08c95133415de8ccf0293dd208c11a29c14315b716c1563c45fd22081fbbf604147c837fa14bc98772a4103b2df080e4d09f89fa
+  data.tar.gz: 4a7894f23f49494f9debcafc967217ccb00549234014166f9fe4f1fe3d13b12b15fbe63eac426b164dfa763acbc9b4e15dfa8cd616de84720f1b2ea3249abfde

data/.rubocop.yml

@@ -1,5 +1,9 @@
 inherit_from: .rubocop_todo.yml
 
+inherit_mode:
+  merge:
+    - Exclude
+
 AllCops:
   TargetRubyVersion: 2.5
   SuggestExtensions: false

data/.rubocop_todo.yml

@@ -1,6 +1,6 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2021-08-22 13:16:05 UTC using RuboCop version 1.19.1.
+# on 2021-10-11 07:56:31 UTC using RuboCop version 1.21.0.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
@@ -9,9 +9,9 @@
 # Offense count: 1
 # Configuration parameters: IgnoredMethods.
 Metrics/CyclomaticComplexity:
-  Max: 8
+  Max: 10
 
-# Offense count: 1
+# Offense count: 2
 # Configuration parameters: CountComments, CountAsOne, ExcludedMethods, IgnoredMethods.
 Metrics/MethodLength:
   Max: 14
@@ -19,4 +19,4 @@ Metrics/MethodLength:
 # Offense count: 1
 # Configuration parameters: IgnoredMethods.
 Metrics/PerceivedComplexity:
-  Max: 10
+  Max: 11

data/.ruby-version

@@ -0,0 +1 @@
+2.7.4

data/CHANGELOG.md

@@ -1,5 +1,12 @@
 ## [Unreleased]
 
+## [0.3.0] - 2021-10-13
+
+- Add `kana?` and `hangul?` methods
+- Improve accuracy of `identify_script` method
+- `chinese?` method now returns actual boolean instead of merely something
+  truthy
+
 ## [0.2.0] - 2021-08-23
 
 - Slight optimization of script-matching regexps

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    script_detector_2 (0.2.0)
+    script_detector_2 (0.3.0)
 
 GEM
   remote: https://rubygems.org/
@@ -18,9 +18,9 @@ GEM
     kramdown-parser-gfm (1.1.0)
       kramdown (~> 2.0)
     minitest (5.14.4)
-    nokogiri (1.12.3-x86_64-darwin)
+    nokogiri (1.12.5-x86_64-darwin)
       racc (~> 1.4)
-    parallel (1.20.1)
+    parallel (1.21.0)
     parser (3.0.2.0)
       ast (~> 2.4.1)
     racc (1.5.2)
@@ -30,20 +30,20 @@ GEM
     reverse_markdown (2.0.0)
       nokogiri
     rexml (3.2.5)
-    rubocop (1.19.1)
+    rubocop (1.22.1)
       parallel (~> 1.10)
       parser (>= 3.0.0.0)
       rainbow (>= 2.2.2, < 4.0)
       regexp_parser (>= 1.8, < 3.0)
       rexml
-      rubocop-ast (>= 1.9.1, < 2.0)
+      rubocop-ast (>= 1.12.0, < 2.0)
       ruby-progressbar (~> 1.7)
       unicode-display_width (>= 1.4.0, < 3.0)
-    rubocop-ast (1.10.0)
+    rubocop-ast (1.12.0)
       parser (>= 3.0.1.1)
     ruby-progressbar (1.11.0)
     rubyzip (2.3.2)
-    solargraph (0.43.0)
+    solargraph (0.44.0)
       backport (~> 1.2)
       benchmark
       bundler (>= 1.17.2)
@@ -60,7 +60,7 @@ GEM
       yard (~> 0.9, >= 0.9.24)
     thor (1.1.0)
     tilt (2.0.10)
-    unicode-display_width (2.0.0)
+    unicode-display_width (2.1.0)
     yard (0.9.26)
 
 PLATFORMS
@@ -76,4 +76,4 @@ DEPENDENCIES
   solargraph
 
 BUNDLED WITH
-   2.2.26
+   2.2.29

data/README.md

@@ -12,7 +12,7 @@ Unlike the original script_detector, this gem:
 - Uses the
   [kUnihanCore2020](https://www.unicode.org/reports/tr38/#kUnihanCore2020)
   property of the Unicode Unihan database to determine which characters belong
-  to which script (Unicode 13)
+  to which script (Unicode 14)
   ([details](http://www.unicode.org/L2/L2019/19388-unihan-core-2020.pdf))
 - Uses [ISO 15924 script names](https://en.wikipedia.org/wiki/ISO_15924) in
   symbol form as return values (instead of English strings)

data/lib/script_detector_2/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module ScriptDetector2
-  VERSION = '0.2.0'
+  VERSION = '0.3.0'
 end

data/lib/script_detector_2.rb

@@ -10,8 +10,8 @@ module ScriptDetector2
     # @param string [String]
     # @return [Boolean]
     def japanese?(string)
-      return true if string =~ /[\p{Hiragana}\p{Katakana}]/
-      return false if string =~ /\p{Hangul}/
+      return true if kana?(string)
+      return false if hangul?(string)
 
       kanji = string.scan(/\p{Han}/)
       return false unless kanji.any?
@@ -19,12 +19,18 @@ module ScriptDetector2
       kanji.all?(JAPANESE_PATTERN)
     end
 
+    # @param string [String]
+    # @return [Boolean]
+    def kana?(string)
+      /[\p{Hiragana}\p{Katakana}]/.match?(string)
+    end
+
     # @param string [String]
     # @return [Boolean]
     def chinese?(string)
       return false if string =~ /[\p{Hiragana}\p{Katakana}\p{Hangul}]/
 
-      string =~ /\p{Han}/
+      /\p{Han}/.match?(string)
     end
 
     # @param string [String]
@@ -52,8 +58,8 @@ module ScriptDetector2
     # @param string [String]
     # @return [Boolean]
     def korean?(string)
-      return true if string =~ /\p{Hangul}/
-      return false if string =~ /[\p{Hiragana}\p{Katakana}]/
+      return true if hangul?(string)
+      return false if kana?(string)
 
       hanja = string.scan(/\p{Han}/)
       return false unless hanja.any?
@@ -61,19 +67,28 @@ module ScriptDetector2
       hanja.all?(KOREAN_PATTERN)
     end
 
+    # @param string [String]
+    # @return [Boolean]
+    def hangul?(string)
+      /\p{Hangul}/.match?(string)
+    end
+
     # @param string [String]
     # @return [Symbol]
     def identify_script(string)
-      return :Jpan if japanese?(string)
-      return :Kore if korean?(string)
+      return :Jpan if kana?(string)
+      return :Kore if hangul?(string)
 
       is_hant = traditional_chinese?(string)
       is_hans = simplified_chinese?(string)
       if is_hant && is_hans then :Hani
       elsif is_hans then :Hans
       elsif is_hant then :Hant
+      elsif japanese?(string) then :Jpan
+      elsif korean?(string) then :Kore
       elsif chinese?(string) then :Hani # rubocop:disable Lint/DuplicateBranch
-      else :Zyyy
+      else
+        :Zyyy
       end
     end
   end

data/tasks/gen_src.rake

@@ -2,7 +2,7 @@
 
 TMP_DIR = 'tmp'
 UNIHAN_ZIP = File.join(TMP_DIR, 'Unihan.zip')
-UNIHAN_URL = 'https://www.unicode.org/Public/UCD/latest/ucd/Unihan.zip'
+UNIHAN_URL = 'https://www.unicode.org/Public/14.0.0/ucd/Unihan.zip'
 
 directory TMP_DIR
 

data/tasks/unihan.rb

@@ -5,15 +5,20 @@ module Unihan
   CODEPOINT_PATTERN = /U\+(?<hex>[A-F0-9]+)/.freeze
 
   class << self
-    # @param readings_data [Hash<Integer,Hash{String => String}>]
+    # @param dict_data [Hash<Integer,Hash{String => String}>]
     # @param tags [Array<String>]
     # @return [Regexp]
     def gen_unihan_core_pattern(dict_data, *tags)
-      codepoints = dict_data.select do |_, data|
+      gen_pattern(codepoints_for_tags(dict_data, tags))
+    end
+
+    # @param dict_data [Hash<Integer,Hash{String => String}>]
+    # @param tags [Array<String>]
+    # @return [Array<Integer>]
+    def codepoints_for_tags(dict_data, tags)
+      dict_data.select do |_, data|
         tags.all? { |t| data['kUnihanCore2020']&.include?(t) }
       end.keys
-
-      gen_pattern(codepoints)
     end
 
     # @param codepoints [Array<Integer>]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: script_detector_2
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Aaron Madlon-Kay
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2021-08-23 00:00:00.000000000 Z
+date: 2021-10-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: byebug
@@ -76,6 +76,7 @@ files:
 - ".dir-locals.el"
 - ".rubocop.yml"
 - ".rubocop_todo.yml"
+- ".ruby-version"
 - ".solargraph.yml"
 - CHANGELOG.md
 - Gemfile