TextHalo Japanese and Mandarin Kokoro pronunciation support

Misaki phone mappings and kana conversion adapted into Rust; modified to use
native dictionaries. Upstream revision: fba1236595f2d2bf21d414ba6e57d25256afada3
https://github.com/hexgrad/misaki
License: Apache-2.0 (see misaki-Apache-2.0.txt).
Cutlet is adapted from polm/cutlet, MIT (see cutlet-MIT.txt).
Mandarin transcription is adapted from stefantaubert/pinyin-to-ipa, MIT
(see pinyin-to-ipa-MIT.txt).

Japanese readings and normalization: JPreprocess 0.15.0, BSD-3-Clause
(see jpreprocess-BSD-3-Clause.txt).
Japanese dictionary tokenizer: Lindera 3.0.7, MIT (see lindera-MIT.txt).

Embedded Japanese dictionary: jpreprocess/naist-jdic v0.1.3,
including NAIST, UniDic Consortium, Open JTalk and JPreprocess notices
(see naist-jdic-BSD.txt).

Mandarin character and phrase readings: pinyin_pro 3.29.4, MIT
(see pinyin-pro-MIT.txt).
Mandarin word segmentation and embedded dictionary: jieba-rs 0.10.3 and
fxsjy/jieba, MIT (see jieba-rs-MIT.txt and jieba-MIT.txt).

These third-party licenses remain independent of TextHalo's own terms.