cbeta 0.4.1 → 0.4.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d6a64090261cde9c8c6439a1db03c400b6d4c686
-  data.tar.gz: cb476a970501ea55eb5cc533f2fc8415a79fda18
+  metadata.gz: 1c29d0e70255be20d828cecb8f3feff8f0d4b448
+  data.tar.gz: f90911d3955acb1f491e1e7c61fd1797617aa8f2
 SHA512:
-  metadata.gz: f07988a13bd0e60023634a26c7fab92d9da0a0b3ed7ed1872b1057a8f075b2077f2064f4b87dae542f89cc860a5b4664274202c7bbe3dba5671346ddc1d169b4
-  data.tar.gz: 998be24e731ace0358da228612ec72b5fa435cc1beb0f5f91da174329627d2ac868c06fd1f6f4cd9e1391b3f996c19207f4ce1f764a95f1665d946a9803416d5
+  metadata.gz: 2a545c0cb3fdc26d8b79a79172112b967f225aadaf03afe41b4117acfc44a67f7626cb1f07cb3262d0b6b4041978af9f87f15f9b6d4b574472f5da6c81153d21
+  data.tar.gz: fcad67ccefa4b4250a6d236671373cc4586f48d4da49b60bfddb659ae1fc0246e613f31e580ad6958b0cd0a1a358b68f9b37f34764a95efbe43f94e8711e8b7e

data/lib/cbeta.rb

@@ -20,6 +20,18 @@ class CBETA
     }
     nil
   end
+  
+  # 傳入 蘭札體 缺字碼，傳回 Unicode PUA 字元
+  def self.ranjana_pua(gid)
+    i = 0x10000 + gid[-4..-1].hex
+    [i].pack("U")
+  end
+  
+  # 傳入 悉曇字 缺字碼，傳回 Unicode PUA 字元
+  def self.siddham_pua(gid)
+    i = 0xFA000 + gid[-4..-1].hex
+    [i].pack("U")
+  end
 
 	# 載入藏經資料
   def initialize()

data/lib/cbeta/p5a_to_html.rb

@@ -27,9 +27,9 @@ class CBETA::P5aToHTML
     @gaijis = CBETA::Gaiji.new
 
     # 載入 unicode 1.1 字集列表
-    fn = File.join(File.dirname(__FILE__), 'unicode-1.1.json')
-    json = File.read(fn)
-    @unicode1 = JSON.parse(json)
+    #fn = File.join(File.dirname(__FILE__), 'unicode-1.1.json')
+    #json = File.read(fn)
+    #@unicode1 = JSON.parse(json)
   end
 
   # 將 CBETA XML P5a 轉為 HTML
@@ -175,7 +175,7 @@ class CBETA::P5aToHTML
 
   def handle_g(e, mode)
     # if 有 <mapping type="unicode">
-    #   if 在 unicode 1.1 範圍裡
+    #   if 不在 Unicode Extension C, D, E 範圍裡
     #     直接採用
     #   else
     #     預設呈現 unicode, 但仍包缺字資訊，供點選開 popup
@@ -218,10 +218,12 @@ class CBETA::P5aToHTML
    
     default = ''
     if g.has_key?('unicode')
-      if @unicode1.include?(g['unicode'])
-        return g['unicode-char'] # unicode 1.1 直接用
-      else
+      #if @unicode1.include?(g['unicode'])
+      # 如果在 unicode ext-C, ext-D, ext-E 範圍內
+      if (0x2A700..0x2CEAF).include? g['unicode'].hex
         default = g['unicode-char']
+      else
+        return g['unicode-char'] # 直接採用 unicode
       end
     end
 

data/lib/cbeta/p5a_to_simple_html.rb

@@ -98,7 +98,9 @@ class CBETA::P5aToSimpleHTML
   end
 
   def handle_g(e)
-    # if 有 <mapping type="unicode">
+    # if 悉曇字、蘭札體
+    #   使用 Unicode PUA
+    # else if 有 <mapping type="unicode">
     #   直接採用
     # else if 有 <mapping type="normal_unicode">
     #   採用 normal_unicode
@@ -111,18 +113,21 @@ class CBETA::P5aToSimpleHTML
     abort "Line:#{__LINE__} 無缺字資料:#{gid}" if g.nil?
     zzs = g['zzs']
     
-    if gid.start_with?('SD')
+    if gid.start_with?('SD') # 悉曇字
       case gid
       when 'SD-E35A'
         return '（'
       when 'SD-E35B'
         return '）'
       else
-        return g['roman']
+        return CBETA.siddham_pua(gid)
       end
     end
     
-    return g['roman'] if gid.start_with?('RJ')
+    if gid.start_with? 'RJ' # 蘭札體
+      return CBETA.ranjana_pua(gid)
+    end
+    
     return g['unicode-char'] if g.has_key?('unicode')
     return g['normal_unicode'] if g.has_key?('normal_unicode')
     return g['normal'] if g.has_key?('normal')

data/lib/cbeta/p5a_to_text.rb

@@ -191,14 +191,12 @@ class CBETA::P5aToText
       when 'SD-E35B'
         return '）'
       else
-        i = 0xFA000 + gid[-4..-1].to_i(16)
-        return [i].pack("U")
+        return CBETA.siddham_pua(gid)
       end
     end
     
-    if gid.start_with?('RJ') # 蘭札體
-      i = 0x10000 + gid[-4..-1].to_i(16)
-      return [i].pack("U")
+    if gid.start_with?('RJ') # 蘭札體      
+      return CBETA.ranjana_pua(gid)
     end
     
     return g['unicode-char'] if g.has_key?('unicode')

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: cbeta
 version: !ruby/object:Gem::Version
-  version: 0.4.1
+  version: 0.4.3
 platform: ruby
 authors:
 - Ray Chou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-08-18 00:00:00.000000000 Z
+date: 2015-08-20 00:00:00.000000000 Z
 dependencies: []
 description: Ruby gem for use Chinese Buddhist Text resources made by CBETA (http://www.cbeta.org).
 email: zhoubx@gmail.com