Add the east Asian language transliteration tweak we discussed.

I tested it with the Gutenberg book "苦悶の欄" by Earl Derr Biggers (https://www.gutenberg.org/ebooks/39287). With the tweak empty the title was transliterated to "Ku Men noLan". With the tweak set to 'ja' it was transliterated to "Kumon no Ran". Is this transliteration correct? I don't know.

resources/default_tweaks.py

@@ -580,3 +580,13 @@ openers_by_scheme = {}
 # It must be one of the values Default, Sunday, Monday, Tuesday, Wednesday,
 # Thursday, Friday, or Saturday, all in English, spelled exactly as shown.
 calendar_start_day_of_week = 'Default'
+
+#: East Asian language to use for transliteration
+# Setting this tweak will make calibre use the specified language as the "base"
+# language when transliterating East Asian languages to ASCII. This might be
+# useful if you run calibre in English but want text transliterated to
+# Japanese. The valid values are 'ja', 'kr', 'vn', 'zh', and '' (empty string).
+# The empty string means use the user interface language as the base language.
+# Any value not in the above list will be treated as the empty string.
+# Example: east_asian_base_language = 'ja'
+east_asian_base_language = ''

src/calibre/utils/localization.py

@@ -548,7 +548,10 @@ def get_udc():
     global _udc
     if _udc is None:
         from calibre.ebooks.unihandecode import Unihandecoder
-        _udc = Unihandecoder(lang=get_lang())
+        from calibre.utils.config_base import tweaks
+        lang = tweaks.get('east_asian_base_language')
+        lang = lang if lang in ('ja', 'kr', 'vn', 'zh') else get_lang()
+        _udc = Unihandecoder(lang=lang)
     return _udc