X-Git-Url: http://git.veekun.com/zzz-pokedex.git/blobdiff_plain/c4b6240ff0e07f2365d0a4e7de3a879bd80dc01c..f79f5d23ea1fb067fadfe19fb7456c690edaef56:/pokedex/roomaji.py?ds=sidebyside

diff --git a/pokedex/roomaji.py b/pokedex/roomaji.py
index 95d559a..e91c960 100644
--- a/pokedex/roomaji.py
+++ b/pokedex/roomaji.py
@@ -37,6 +37,7 @@ _roomaji_kana = {
     u'ダ': 'da',    u'ヂ': 'ji',    u'ヅ': 'dzu',   u'デ': 'de',    u'ド': 'do',
     u'バ': 'ba',    u'ビ': 'bi',    u'ブ': 'bu',    u'ベ': 'be',    u'ボ': 'bo',
     u'パ': 'pa',    u'ピ': 'pi',    u'プ': 'pu',    u'ペ': 'pe',    u'ポ': 'po',
+                                    u'ヴ': 'vu',
 }
 
 _roomaji_youon = {
@@ -59,15 +60,18 @@ _roomaji_small_kana_combos = {
     # particularly weird, but it seems to be what GF intends
 
     # Simple vowel replacement
-                    u'ウィ': 'wi',  u'ウゥ': 'wu',  u'ウェ': 'we',
+                    u'ウィ': 'wi',  u'ウゥ': 'wu',  u'ウェ': 'we',  u'ウォ': 'wo',
+    u'ヴァ': 'va',  u'ヴィ': 'vi',                  u'ヴェ': 've',  u'ヴォ': 'vo',
                                                     u'チェ': 'che',
                                                     u'シェ': 'she',
+                                                    u'ジェ': 'je',
     u'テァ': 'tha', u'ティ': 'ti',  u'テゥ': 'thu', u'テェ': 'tye', u'テォ': 'tho',
     u'デァ': 'dha', u'ディ': 'di',  u'デゥ': 'dhu', u'デェ': 'dye', u'デォ': 'dho',
     u'ファ': 'fa',  u'フィ': 'fi',  u'ホゥ': 'hu',  u'フェ': 'fe',  u'フォ': 'fo',
 
     # Not so much
     u'シィ': 'sy',
+    u'ミィ': 'my',
     u'ビィ': 'by',
     u'ピィ': 'py',
 }
@@ -82,11 +86,12 @@ def romanize(string):
     last_char = None  # Used for small kana combos
     for char in string:
         # Full-width Latin
-        if ord(char) >= 0xff11 and ord(char) <= 0xff5e:
+        if 0xff01 <= ord(char) <= 0xff5e:
             if last_kana == 'sokuon':
                 raise ValueError("Sokuon cannot precede Latin characters.")
 
-            char = chr(ord(char) - 0xff11 + 0x31)
+            # XXX Real Unicode decomposition would be nicer
+            char = chr(ord(char) - 0xff01 + 0x21)
             characters.append(char)
 
             last_kana = None