Parse Tibetan Syllables (attempt)

parse.ts

const PREFIXES = ['ག', 'ད', 'བ', 'མ', 'འ']
const SUPERSCRIPTS = ['ར', 'ལ', 'ས']
const EXTENDED_ROOT_LETTERS = [
  'ཫ',
  'ཬ',
  'ཁ༹',
  'ག༹',
  'ཕ༹',
  'བ༹',
  'གྷ',
  'ཛྷ',
  'ཊ',
  'ཋ',
  'ཌ',
  'ཌྷ',
  'ཎ',
  'དྷ',
  'བྷ',
  'ཥ',
  'ཀྵ',
]
const SPECIAL = ['ༀ']
const ROOT_LETTERS = [
  'ཀ',
  'ཁ',
  'ག',
  'ང',
  'ཅ',
  'ཆ',
  'ཇ',
  'ཉ',
  'ཏ',
  'ཐ',
  'ད',
  'ན',
  'པ',
  'ཕ',
  'བ',
  'མ',
  'ཙ',
  'ཚ',
  'ཛ',
  'ཝ',
  'ཞ',
  'ཟ',
  'འ',
  'ཡ',
  'ར',
  'ལ',
  'ཤ',
  'ས',
  'ཧ',
  'ཨ',
].concat(EXTENDED_ROOT_LETTERS)

const SUBSCRIPTS = [
  'ྐ',
  'ྑ',
  'ྒ',
  'ྒྷ',
  'ྔ',
  'ྕ',
  'ྖ',
  'ྗ',
  'ྙ',
  'ྚ',
  'ྛ',
  'ྜ',
  'ྜྷ',
  'ྞ',
  'ྟ',
  'ྠ',
  'ྡ',
  'ྡྷ',
  'ྣ',
  'ྤ',
  'ྥ',
  'ྦ',
  'ྦྷ',
  'ྨ',
  'ྩ',
  'ྪ',
  'ྫ',
  'ྫྷ',
  'ྭ',
  'ྮ',
  'ྯ',
  'ྰ',
  'ྱ',
  'ྲ',
  'ླ',
  'ྴ',
  'ྵ',
  'ྶ',
  'ྷ',
  'ྸ',
  'ྐྵ',
  'ྺ',
  'ྻ',
  'ྼ',
]
const VOWEL_SIGNS = [
  'ཱ',
  'ི',
  'ཱི',
  'ུ',
  'ཱུ',
  'ྲྀ',
  'ཷ',
  'ླྀ',
  'ཹ',
  'ྀ',
  'ཱྀ',
]
const SUFFIXES = ['ག', 'ན', 'བ', 'ད', 'མ', 'འ', 'ར', 'ང', 'ས', 'ལ']
const POST_SUFFIXES = ['ད', 'ས']
const IGNORED = [
  '༄',
  'ེ',
  'ཻ',
  'ོ',
  'ཽ',
  'ཾ',
  'ཿ',
  'ྂ',
  'ྃ',
  '྄',
  '྅',
  '྆',
  '྇',
  'ྈ',
  'ྉ',
  'ྊ',
  'ྋ',
  'ྌ',
  'ྍ',
  'ྎ',
  'ྏ',
  '\u0f89',
]

let errorI = 1

function splitSyllable(candidate: string) {
  if (SPECIAL.includes(candidate)) {
    return [candidate]
  }
  const parts = [...candidate]
  if (parts[0] === parts[1]) {
    splitSyllable(parts.slice(1).join(''))
  }

  const output = []
  let current = []
  let state = 'new'
  while (parts.length) {
    const next = parts.shift()
    if (IGNORED.includes(next)) {
      if (current.length) {
        output.push(current.join(''))
      }
      current = []
      state = 'new'
      continue
    }
    switch (state) {
      case 'new': {
        if (PREFIXES.includes(next)) {
          const peek = parts[0]
          if (SUPERSCRIPTS.includes(peek)) {
            const peek2 = parts[1]
            if (ROOT_LETTERS.includes(peek2)) {
              parts.shift()
              parts.shift()
              current.push(`${next}${peek}${peek2}`)
            } else {
              parts.shift()
              current.push(`${next}${peek}`)
            }
          } else if (ROOT_LETTERS.includes(peek)) {
            parts.shift()
            current.push(`${next}${peek}`)
          } else {
            current.push(next)
          }
          state = 'root'
        } else if (ROOT_LETTERS.includes(next)) {
          current.push(next)
          state = 'root'
        } else {
          console.log(
            errorI++,
            candidate +
              ' ' +
              JSON.stringify(
                [...candidate].map(x =>
                  x.codePointAt(0).toString(16).padStart(4, '0'),
                ),
              ),
          )
          return []
          throw new Error(
            'New state: ' +
              candidate +
              ' ' +
              JSON.stringify(
                [...candidate].map(x =>
                  x.codePointAt(0).toString(16).padStart(4, '0'),
                ),
              ),
          )
        }
        break
      }
      case 'root': {
        if (SUBSCRIPTS.includes(next)) {
          current.push(next)
          state = 'subscript'
        } else if (VOWEL_SIGNS.includes(next)) {
          current.push(next)
          state = 'vowel'
        } else if (SUFFIXES.includes(next)) {
          current.push(next)
          state = 'suffix'
        } else {
          parts.unshift(next)
          output.push(current.join(''))
          current = []
          state = 'new'
        }
        break
      }
      case 'subscript': {
        if (VOWEL_SIGNS.includes(next)) {
          current.push(next)
          state = 'vowel'
        } else if (SUFFIXES.includes(next)) {
          current.push(next)
          state = 'suffix'
        } else {
          parts.unshift(next)
          output.push(current.join(''))
          current = []
          state = 'new'
        }
        break
      }
      case 'vowel': {
        if (SUFFIXES.includes(next)) {
          current.push(next)
          state = 'suffix'
        } else {
          parts.unshift(next)
          output.push(current.join(''))
          current = []
          state = 'new'
        }
        break
      }
      case 'suffix': {
        if (POST_SUFFIXES.includes(next)) {
          current.push(next)
          output.push(current.join(''))
          current = []
          state = 'new'
        } else {
          parts.unshift(next)
          output.push(current.join(''))
          current = []
          state = 'new'
        }
        break
      }
    }
  }

  if (current.length) {
    output.push(current.join(''))
  }
  return output
}