ya-kansuji 0.1.0

package/src/groups.ts

@@ -0,0 +1,31 @@
+import { UNIT_EXP4 } from './constants.js'
+
+const GROUP_BASE = 10_000n
+const GROUP_UNITS = ['', ...UNIT_EXP4] as const
+const GROUP_COUNT = GROUP_UNITS.length
+const MAX_EXCLUSIVE = GROUP_BASE ** BigInt(GROUP_COUNT)
+const MAX_DIGITS = GROUP_COUNT * 4
+
+// 有効な最上位から4桁ごとのグループを返す。途中のゼログループも含む。
+export function groups4(num: bigint): Array<[number, string]> {
+  if (num < 0n || num >= MAX_EXCLUSIVE) {
+    throw new RangeError(`Kansuji formatters require a bigint between 0 and 10^${MAX_DIGITS} - 1`)
+  }
+
+  const ret: Array<[number, string]> = []
+  let value = num
+  let unitIndex = 0
+  do {
+    const unit = GROUP_UNITS[unitIndex]
+    if (unit === undefined) {
+      throw new RangeError(
+        `Kansuji formatters require a bigint between 0 and 10^${MAX_DIGITS} - 1`,
+      )
+    }
+    ret.push([Number(value % GROUP_BASE), unit])
+    value /= GROUP_BASE
+    unitIndex++
+  } while (value > 0n)
+
+  return ret.reverse()
+}

package/src/index.ts

@@ -0,0 +1,13 @@
+export { toBigInt, toNumber } from './parse.js'
+export { UNIT_EXP3, UNIT_EXP4 } from './constants.js'
+export {
+  getFormatter,
+  registerFormatter,
+  toKan,
+  type KansujiFormatter,
+  type KansujiFormatterOptions,
+} from './formatters.js'
+export { simple } from './formatter/simple.js'
+export { gov } from './formatter/gov.js'
+export { lawyer } from './formatter/lawyer.js'
+export { judicH, judicV } from './formatter/judic.js'

package/src/parse.ts

@@ -0,0 +1,136 @@
+import { NUM_ALT_CHARS, NUM_NORMALIZED_CHARS, UNIT_EXP3, UNIT_EXP4 } from './constants.js'
+
+const altChars = Array.from(NUM_ALT_CHARS)
+const normChars = Array.from(NUM_NORMALIZED_CHARS)
+const NORMALIZE_MAP = new Map(altChars.map((c, i) => [c, normChars[i] as string]))
+const SEPARATORS = new Set([
+  ',', '，', '、',
+  '\u0009', '\u000A', '\u000B', '\u000C', '\u000D', '\u0020', '\u00A0', '\u1680',
+  '\u2000', '\u2001', '\u2002', '\u2003', '\u2004', '\u2005', '\u2006', '\u2007',
+  '\u2008', '\u2009', '\u200A', '\u2028', '\u2029', '\u202F', '\u205F', '\u3000',
+  '\uFEFF',
+])
+const DIGIT_VALUES = [0n, 1n, 2n, 3n, 4n, 5n, 6n, 7n, 8n, 9n] as const
+const SPECIAL_NUMBER_VALUES = new Map<string, bigint>([
+  ['卄', 20n],
+  ['廿', 20n],
+  ['卅', 30n],
+  ['丗', 30n],
+  ['卌', 40n],
+  ['皕', 200n],
+])
+
+const SINGLE_UNITS = [...UNIT_EXP3, ...UNIT_EXP4.filter((u) => Array.from(u).length === 1)]
+const MULTI_UNITS = UNIT_EXP4.filter((u) => Array.from(u).length > 1)
+const CHAR_CLASS = [
+  ...new Set([...altChars, ...normChars, ...SINGLE_UNITS, ...SPECIAL_NUMBER_VALUES.keys()]),
+].join('')
+
+const PART_SOURCE = `${MULTI_UNITS.join('|')}|[${CHAR_CLASS}]`
+const KANSUJI_REGEXP = new RegExp(`(?:マイナス)?(?:${PART_SOURCE})+`, 'u')
+const PART_REGEXP = new RegExp(PART_SOURCE, 'gu')
+
+function createUnitValues(units: readonly string[], base: bigint): Map<string, bigint> {
+  let value = base
+  return new Map(units.map((unit) => {
+    const entry: [string, bigint] = [unit, value]
+    value *= base
+    return entry
+  }))
+}
+
+const UNIT_EXP3_VALUES = createUnitValues(UNIT_EXP3, 10n)
+const UNIT_EXP4_VALUES = createUnitValues(UNIT_EXP4, 10_000n)
+const MAX_SAFE_BIGINT = BigInt(Number.MAX_SAFE_INTEGER)
+
+function clean(str: string): string {
+  let result = ''
+  for (const c of str) {
+    const normalized = NORMALIZE_MAP.get(c) ?? c
+    if (!SEPARATORS.has(normalized)) result += normalized
+  }
+  return result
+}
+
+/**
+ * 漢数字・大字・全角数字が混じったテキストを bigint に変換する。
+ *
+ * 照合前にカンマ・読点・各種空白などの区切り文字を除去し、先頭から連続してマッチする
+ * 数値表現だけを読み取る（例: `toBigInt('五x六')` は `5n`）。マッチしなければ `0n` を返す。
+ * 先頭の「マイナス」は負号として扱うが、ASCII の `-` や U+2212（−）は負号として扱わない。
+ * 値の上限はなく、無量大数（10^68）を超える大きさもそのまま扱える。
+ *
+ * @param str 変換対象の文字列
+ * @returns テキストが表す整数値
+ * @example
+ * toBigInt('一〇二四')       // => 1024n
+ * toBigInt('マイナス千二百') // => -1200n
+ * toBigInt('一無量大数')     // => 10n ** 68n
+ */
+export function toBigInt(str: string): bigint {
+  const cleaned = clean(String(str))
+  const matched = KANSUJI_REGEXP.exec(cleaned)
+  if (!matched) return 0n
+  const matchedText = matched[0]
+  if (matchedText === undefined) return 0n
+
+  let ret3 = 0n
+  let ret4 = 0n
+  let curnum: bigint | null = null
+  for (const c of matchedText.match(PART_REGEXP) ?? []) {
+    if (c >= '1' && c <= '9') {
+      const digit = DIGIT_VALUES[c.charCodeAt(0) - 48]!
+      curnum = (curnum ?? 0n) * 10n + digit
+    } else if (c === '0') {
+      if (curnum !== null) curnum *= 10n
+    } else {
+      const specialNumber = SPECIAL_NUMBER_VALUES.get(c)
+      if (specialNumber !== undefined) {
+        ret3 += specialNumber
+        curnum = null
+      } else {
+        const unit4 = UNIT_EXP4_VALUES.get(c)
+        if (unit4 !== undefined) {
+          if (curnum !== null) {
+            ret3 += curnum
+            curnum = null
+          }
+          if (ret3 === 0n) ret3 = 1n
+          ret4 += ret3 * unit4
+          ret3 = 0n
+        } else {
+          const unit3 = UNIT_EXP3_VALUES.get(c)
+          if (unit3 !== undefined) {
+            curnum ??= 1n
+            ret3 += curnum * unit3
+            curnum = null
+          }
+        }
+      }
+    }
+  }
+  if (curnum !== null) ret3 += curnum
+  const ret = ret4 + ret3
+  return matchedText.startsWith('マイナス') ? -ret : ret
+}
+
+/**
+ * {@link toBigInt} と同じ規則でテキストを解釈し、`number` として返す。
+ *
+ * 結果の絶対値が `Number.MAX_SAFE_INTEGER`（2^53 − 1）を超える場合は `RangeError` を
+ * 投げる。安全整数を超えうる値を扱うときは {@link toBigInt} を使うこと。
+ *
+ * @param str 変換対象の文字列
+ * @returns テキストが表す数値
+ * @throws {RangeError} 値が安全整数の範囲を超える場合
+ * @example
+ * toNumber('一〇二四')   // => 1024
+ * toNumber('一無量大数') // throws RangeError
+ */
+export function toNumber(str: string): number {
+  const value = toBigInt(str)
+  if (value > MAX_SAFE_BIGINT || value < -MAX_SAFE_BIGINT) {
+    throw new RangeError(`kansuji value exceeds Number.MAX_SAFE_INTEGER: ${str}`)
+  }
+  return Number(value)
+}