To use your sort order in your glossary, you have to use yogosyu directive instead of glossary.
When your language have CJK Ideographs, you want to add reading to sort. In Japanese, this reading (or ‘How to read the word’) is called Yomi or Yomigana.
yogosyu directive also have the feature to add Yomigana to sort.
Put yogosyu.py in exts folder and add yogosyu into extension valuable in conf.py.
Make your sort_order_xx.py and make sure language option etc.
This work completely same as described in user_ordered_index_patch.py.
The yogosyu extends the glossary directive. As told above, you should replace the name of the directive.
When your language doesn’t have CJK Ideographs, you can skip the rest of this page ;)
To add Yomigana, use yomimark separator after each term, and place Yomigana just after yomimark.
You can give any one character and can vary it for each yogosyu directive. Of course, yomimark should be the character neither in both terms nor Yomigana-s:
.. yogosyu::
:sorted:
:yomimark: ×
You can omit sorted option. In this case, the terms don’t sorted in yogosyu glossary and sorted in index page only when user_ordered_index_patch.py used.
Note
yomimark can accept space, tab or U+hhhh form. see single_char_or_whitespace_or_unicode method in docutils.parsers.rst.directives.__init__.py.
Write follwing order; each term, the character declared with yomimark option and the Yomigana. You can omit yomimark and Yomigana for each term we don’t have to give them:
.. yogosyu::
:sorted:
:yomimark: ×
姫踊子草×Himeodorikoso
Lamium purpureum. The Purple Deadnettle in English.
キーボード
The keyboard in Katakana form.
You should not any (overadded) whitespaces before and after yomimark character.
Currently, second appearance of it in a line, all characters after that are ignored.
Note
This ignorance behavior is designed for future use. see ToDos and Issues.