[v0.4.2] Redefine cloud data models

[anggoran]

Context

There were some inconsistencies in how we defined hanzis and pinyins.

Criteria

• [x] Database design:
  • [x] pinyins (id, name, latin, tone)
  • [x] hanzis (id, form, meaning, type, etymology)
  • [x] hanzis_pinyins (id, hanzi_id, pinyin_id)
• [x] Database manipulation:
  • [x] Generate new supabase-pinyin.csv (pinyins.json, refer to cedict.txt)
  • [x] Generate new supabase-hanzi.csv (unihan.txt, refer to cedict.txt)
  • [x] Generate supabase-hanzi_pinyin.csv (exclude unknown pinyins)
  • [x] Import to Supabase