[Engine] Improve the word segmentation for CJK by using ICU
The examples taken from the original report can now produce much more meaningful result:
"邓丽君 - 漫步人生路.mp3":{"mp3","邓丽君","漫步","路","人生","-"}
The examples taken from the original report can now produce much more meaningful result:
"邓丽君 - 漫步人生路.mp3":{"mp3","邓丽君","漫步","路","人生","-"}