[feature request] Add pitch data

[ghost]

I would appreciate if the hiragana readings contained pitch data. Wadoku's xml contains an accent tag for many entries. Jiten compiled the data from wadoku into an easy-to-read form where ꜜ is inserted after the mora which carries the accent (or not inserted, if there is no accent) and ꜛ is inserted after the first mora, except if the accent lies on the first mora. I tried to tamper with res/raw/dict myself, but sqlite3 considered the file invalid.