fix: live2d motion & expression control based on emotion analysis

[cdfmlr]

I got a plan: mv live2ddriver/emotext muvtuberdriver/

• [ ] live2ddriver:
  • [ ] -= emotext client
  • [ ] -= API /shizuku: make everything reusable
  • [ ] += Emotion -> Motion&Expression mapper (configurable instead of hardcoded)
  • [ ] += API /emotion: input (Emotion) -> map(Emotion -> Motion&Expression) -> fwd to live2dview
• [ ] muvtuberdriver:
  • [ ] += emotext client
  • [ ] += a side flow (i.e. a sub task): (TextOut) -> emotext -> (Emotion) -> live2ddriver