居戎氏
對了。Rime 讓用戶自定義詞典，沒有固定的詞彙表。 那麼語言模型的數據庫就不能存儲詞的 ID 之類而是用文字檢索二元組。 具體輸入方案的詞典分詞的粒度不同，有一些實現細節上的挑戰。 總之不是非常標準的實現。要參考還是參考更典型的 [昇陽拼音](https://github.com/sunpinyin/sunpinyin) 和 BYVoid 的 [習作](https://byvoid.com/zht/blog/slm_based_pinyin_ime/)。
詞組編碼那裏應該是有問題的。先不管他。 正則表達式這個 `[AB]` 語法是枚舉字符，因爲 `xform` 規則是 UTF-8 編碼的字符串，因此目前這樣寫無法枚舉 Unicode 字符。 可以改寫成 `(A|B)` 的形式，或者改寫成支持 Unicode 字符的 `xlit/AB/aa` 規則。
既是異體字，爲何不同音？
> derive/([aei])n$/$1ng/ 這是一條 YAML 字符串，字符串的內容是 RIME 自定義的格式，詳見 [拼寫運算](https://github.com/rime/home/wiki/SpellingAlgebra#%E6%8B%BC%E5%AF%AB%E9%81%8B%E7%AE%97%E7%9A%84%E7%AE%97%E5%BC%8F)
I think there is no substantial feature update since last release?
Rebuilding ibus-rime is not needed for updated features in librime to take effect, as long as the new features don't introduce new APIs. This is because librime is independently released...
I'm not familiar with Array IME. @archerindigo can you comment on this? The described behaviour looks like special logic from other Array IME software. Is that a crucial feature? Rime...
Now I understand the issue better. The space key is equivalent to page down function in the context of entering array symbols. I don't know if space key is also...
> I'm not sure whether it is supported to let the main input (array30) method to respect the key_binder options set in the secondary method (array30_wsymbols) when the main method...