yosina 1.1.3 → 3.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 506016d955f814d8eb4ee28c64ac510b251f3336681bcf93f116a929dcf2540b
-  data.tar.gz: 492d9fd8286a536a1dd07858d8c7c789cb30e0b78ba11adf3c76ba43c2035ac2
+  metadata.gz: 4d31d3fa007ce64078273418d9372e42e5aae7bacec0682019ce4fab8d2e93f1
+  data.tar.gz: c641e3563f3baf54e781660b34f76806d90983d03598133f4efbf1bbeb29531c
 SHA512:
-  metadata.gz: 6b7691ee1fd50819b2cc7fb6a48f53bb0524c0817279104f2b6bfe731d794c80a99d2f44b7da733d60c3d28f29c83ba859200987356d807550ba6adc45fbfc9a
-  data.tar.gz: 100771ea6fa26397a3ba462842f7427b739a2d5f2cb3f580982ef76896adfc532b3ebe8cbc89a760927ffd774f8417a7c5559d5996f2333c06cb49a2a9f5b8ab
+  metadata.gz: '05648b23304aa9cf4668826f2b94bb27c29d7d74eb8832d910ce5d1ff237157c6814af181804dd2cc692d0c7322201daa8c38cbc52cb19c5bb9a20115ec5c497'
+  data.tar.gz: 64de351469c65778f19759b834ee516af59cdf8dc42a104cbee1b4e16c627704b13ba421eb3513f0b407b7df2146c6f7a66e288d06eec56e3679bfd1bb6057a1

data/lib/yosina/recipes.rb

@@ -271,7 +271,11 @@ module Yosina
 
     def apply_replace_suspicious_hyphens_to_prolonged_sound_marks(ctx)
       if @replace_suspicious_hyphens_to_prolonged_sound_marks
-        ctx.insert_middle([:prolonged_sound_marks, { replace_prolonged_marks_following_alnums: true }])
+        ctx.insert_middle([:prolonged_sound_marks, {
+                            replace_prolonged_marks_following_alnums: true,
+                            replace_prolonged_marks_between_non_kanas:
+                              @replace_suspicious_hyphens_to_prolonged_sound_marks == 'aggressive'
+                          }])
       else
         ctx
       end

data/lib/yosina/transliterators/circled_or_squared_data.rb

@@ -416,7 +416,6 @@ module Yosina
         "\u{1f1fe}" => { rendering: 'Y', type: 's', emoji: false },
         "\u{1f1ff}" => { rendering: 'Z', type: 's', emoji: false },
         "\u{1f6be}" => { rendering: 'WC', type: 's', emoji: true },
-        "\u{1f200}" => { rendering: "\u{307b}\u{304b}", type: 's', emoji: false },
         "\u{1f201}" => { rendering: "\u{30b3}\u{30b3}", type: 's', emoji: true },
         "\u{1f202}" => { rendering: "\u{30b5}", type: 's', emoji: false },
         "\u{1f210}" => { rendering: "\u{624b}", type: 's', emoji: false },

data/lib/yosina/transliterators/hyphens_data.rb

@@ -13,7 +13,7 @@ module Yosina
 
       # Generated mapping data
       HYPHENS_MAPPINGS = {
-        '-' => HyphensRecord.new(ascii: '-', jisx0201: '-', jisx0208_90: "\u{2212}", jisx0208_90_windows: "\u{2212}"),
+        '-' => HyphensRecord.new(ascii: '-', jisx0201: '-', jisx0208_90: "\u{2212}", jisx0208_90_windows: "\u{ff0d}"),
         '|' => HyphensRecord.new(ascii: '|', jisx0201: '|', jisx0208_90: "\u{ff5c}", jisx0208_90_windows: "\u{ff5c}"),
         '~' => HyphensRecord.new(ascii: '~', jisx0201: '~', jisx0208_90: "\u{301c}", jisx0208_90_windows: "\u{ff5e}"),
         "\u{a2}" => HyphensRecord.new(jisx0208_90: "\u{a2}", jisx0208_90_windows: "\u{ffe0}"),

data/lib/yosina/transliterators/prolonged_sound_marks.rb

@@ -71,6 +71,12 @@ module Yosina
           [0x30fc, 0xff70].include?(char_code)
         end
 
+        def kana?(char_code)
+          hiragana?(char_code) || katakana?(char_code) ||
+            hatsuon?(char_code) || sokuon?(char_code) ||
+            prolonged_sound_mark?(char_code)
+        end
+
         def prolongable?(char_code)
           prolonged_sound_mark?(char_code) || hiragana?(char_code) || katakana?(char_code)
         end
@@ -83,7 +89,8 @@ module Yosina
       # Transliterator for prolonged sound marks
       class Transliterator < Yosina::BaseTransliterator
         attr_reader :skip_already_transliterated_chars, :allow_prolonged_hatsuon,
-                    :allow_prolonged_sokuon, :replace_prolonged_marks_following_alnums
+                    :allow_prolonged_sokuon, :replace_prolonged_marks_following_alnums,
+                    :replace_prolonged_marks_between_non_kanas
 
         # Initialize the transliterator with options
         #
@@ -96,12 +103,15 @@ module Yosina
         #     Default: false.
         # @option options [Boolean] :replace_prolonged_marks_following_alnums Replace prolonged marks after alphanum
         #    with hyphens. Default: false.
+        # @option options [Boolean] :replace_prolonged_marks_between_non_kanas Replace prolonged marks between
+        #    non-kana characters with hyphens. Default: false.
         def initialize(options = {})
           super()
           @skip_already_transliterated_chars = options.fetch(:skip_already_transliterated_chars, false)
           @allow_prolonged_hatsuon = options.fetch(:allow_prolonged_hatsuon, false)
           @allow_prolonged_sokuon = options.fetch(:allow_prolonged_sokuon, false)
           @replace_prolonged_marks_following_alnums = options.fetch(:replace_prolonged_marks_following_alnums, false)
+          @replace_prolonged_marks_between_non_kanas = options.fetch(:replace_prolonged_marks_between_non_kanas, false)
         end
 
         # Convert hyphen-like characters to appropriate prolonged sound marks
@@ -125,13 +135,28 @@ module Yosina
                 prev_non_prolonged_char = last_non_prolonged_char
                 last_non_prolonged_char = char
 
-                if (prev_non_prolonged_char.nil? || alphanumeric?(prev_non_prolonged_char.c.ord)) && (
+                prev_type = prev_non_prolonged_char&.c&.ord
+                following_type = last_non_prolonged_char.c.empty? ? nil : last_non_prolonged_char.c.ord
+
+                replace_by_alnum = @replace_prolonged_marks_following_alnums &&
+                                   (prev_non_prolonged_char.nil? || alphanumeric?(prev_type))
+                replace_by_non_kana = @replace_prolonged_marks_between_non_kanas &&
+                                      (prev_non_prolonged_char.nil? || !kana?(prev_type)) &&
+                                      (following_type.nil? || !kana?(following_type))
+
+                if (replace_by_alnum || replace_by_non_kana) && (
                   !@skip_already_transliterated_chars || !processed_char_in_lookahead
                 )
-                  halfwidth = halfwidth?(
-                    prev_non_prolonged_char.nil? ? last_non_prolonged_char.c.ord : prev_non_prolonged_char.c.ord
-                  )
-                  replacement = halfwidth ? "\u002d" : "\uff0d"
+                  if replace_by_non_kana
+                    prev_half = prev_non_prolonged_char.nil? || halfwidth?(prev_type)
+                    next_half = !following_type.nil? && halfwidth?(following_type)
+                    replacement = !prev_half && !next_half ? "\uff0d" : "\u002d"
+                  else
+                    halfwidth = halfwidth?(
+                      prev_non_prolonged_char.nil? ? following_type : prev_type
+                    )
+                    replacement = halfwidth ? "\u002d" : "\uff0d"
+                  end
                   lookahead_buf.each do |buffered_char|
                     y << Char.new(c: replacement, offset: offset, source: buffered_char)
                     offset += replacement.length
@@ -158,7 +183,8 @@ module Yosina
                     y << Char.new(c: replacement, offset: offset, source: char)
                     offset += replacement.length
                     next
-                  elsif @replace_prolonged_marks_following_alnums && alphanumeric?(last_non_prolonged_char.c.ord)
+                  elsif (@replace_prolonged_marks_following_alnums && alphanumeric?(last_non_prolonged_char.c.ord)) ||
+                        (@replace_prolonged_marks_between_non_kanas && !kana?(last_non_prolonged_char.c.ord))
                     lookahead_buf << char
                     next
                   end

data/lib/yosina/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Yosina
-  VERSION = '1.1.3'
+  VERSION = '3.0.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: yosina
 version: !ruby/object:Gem::Version
-  version: 1.1.3
+  version: 3.0.0
 platform: ruby
 authors:
 - Moriyoshi Koizumi
-autorequire: 
+autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-04-13 00:00:00.000000000 Z
+date: 2026-04-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: minitest
@@ -143,7 +143,7 @@ metadata:
   homepage_uri: https://github.com/yosina-lib/yosina
   source_code_uri: https://github.com/yosina-lib/yosina
   changelog_uri: https://github.com/yosina-lib/yosina/releases
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -158,8 +158,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3.1
-signing_key: 
+rubygems_version: 3.5.11
+signing_key:
 specification_version: 4
 summary: Japanese text transliteration library
 test_files: []