@gondola-lyrics/kit 0.5.0 → 0.6.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gondola-lyrics/kit",
-  "version": "0.5.0",
+  "version": "0.6.0",
   "description": "Gondola lyrics toolkit",
   "keywords": [
     "gondola",

package/dist/language.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"language.js","names":[],"sources":["../src/utils/array.ts","../src/utils/language.ts"],"sourcesContent":["/**\n * Preserves the input array when every element keeps its identity, so unchanged parent nodes can also be reused.\n *\n * The callback must leave inputs untouched and return the original element when unchanged.\n */\nexport const mapShared = <T>(items: T[], transform: (item: T) => T): T[] => {\n  let result = items\n  for (const [index, item] of items.entries()) {\n    const transformed = transform(item)\n    if (transformed !== item) {\n      if (result === items) {\n        result = [...items]\n      }\n      result[index] = transformed\n    }\n  }\n  return result\n}\n\n/**\n * Finds the first value meeting a lower bound so range queries can skip earlier candidates.\n *\n * @param values Values sorted in ascending order.\n * @param target Inclusive lower bound.\n *\n * @returns The first matching index, or the array length when every value is smaller.\n */\nexport const lowerBound = (values: readonly number[], target: number): number => {\n  let left = 0\n  let right = values.length\n  while (left < right) {\n    const middle = Math.floor((left + right) / 2)\n    if (values[middle]! < target) {\n      left = middle + 1\n    } else {\n      right = middle\n    }\n  }\n  return left\n}\n","const enum ScriptGroup {\n  NEUTRAL,\n  EAST_ASIAN,\n  LATIN,\n  HANGUL,\n  CYRILLIC,\n  GREEK,\n  ARABIC,\n  HEBREW,\n  DEVANAGARI,\n  THAI,\n}\n\nconst SCRIPT_PATTERNS: readonly (readonly [RegExp, ScriptGroup])[] = [\n  [/[\\p{Script=Han}\\p{Script=Hiragana}\\p{Script=Katakana}]/u, ScriptGroup.EAST_ASIAN],\n  [/\\p{Script=Latin}/u, ScriptGroup.LATIN],\n  [/\\p{Script=Hangul}/u, ScriptGroup.HANGUL],\n  [/\\p{Script=Cyrillic}/u, ScriptGroup.CYRILLIC],\n  [/\\p{Script=Greek}/u, ScriptGroup.GREEK],\n  [/\\p{Script=Arabic}/u, ScriptGroup.ARABIC],\n  [/\\p{Script=Hebrew}/u, ScriptGroup.HEBREW],\n  [/\\p{Script=Devanagari}/u, ScriptGroup.DEVANAGARI],\n  [/\\p{Script=Thai}/u, ScriptGroup.THAI],\n]\n\nconst getOtherScript = (text: string, start: number, width: number): ScriptGroup => {\n  const unit = text.charCodeAt(start)\n  if (unit === 0x30fb || unit === 0x30fc) {\n    return ScriptGroup.NEUTRAL\n  }\n  if ((unit >= 0x0400 && unit <= 0x0481) || (unit >= 0x048a && unit <= 0x052f)) {\n    return ScriptGroup.CYRILLIC\n  }\n  if (unit >= 0x03a3 && unit <= 0x03e1) {\n    return ScriptGroup.GREEK\n  }\n  if ((unit >= 0x0620 && unit <= 0x063f) || (unit >= 0x0641 && unit <= 0x064a)) {\n    return ScriptGroup.ARABIC\n  }\n  if (unit >= 0x05d0 && unit <= 0x05ea) {\n    return ScriptGroup.HEBREW\n  }\n  if ((unit >= 0x0900 && unit <= 0x0950) || (unit >= 0x0955 && unit <= 0x0963) || (unit >= 0x0966 && unit <= 0x097f)) {\n    return ScriptGroup.DEVANAGARI\n  }\n  if ((unit >= 0x0e01 && unit <= 0x0e3a) || (unit >= 0x0e40 && unit <= 0x0e5b)) {\n    return ScriptGroup.THAI\n  }\n\n  const character = text.slice(start, start + width)\n  for (const [pattern, script] of SCRIPT_PATTERNS) {\n    if (pattern.test(character)) {\n      return script\n    }\n  }\n  return ScriptGroup.NEUTRAL\n}\n\n/**\n * Finds a boundary between supported scripts without allocating a segment list.\n * Han and kana share a group so Japanese labels stay intact; Hangul, Latin, Cyrillic, Greek, Arabic, Hebrew, Devanagari and Thai remain separate.\n * Punctuation, digits, inherited marks and unsupported scripts stay in the surrounding run.\n *\n * @param start A UTF-16 offset at a code-point boundary, usually the end returned by the previous call.\n *\n * @returns The exclusive UTF-16 end of the run, or text.length when no later boundary exists.\n */\nexport const findScriptRunEnd = (text: string, start = 0): number => {\n  let script = ScriptGroup.NEUTRAL\n  let index = start\n\n  while (index < text.length) {\n    const unit = text.charCodeAt(index)\n    let nextScript = ScriptGroup.NEUTRAL\n    let width = 1\n\n    if (unit <= 0x7f) {\n      const letter = unit | 0x20\n      if (letter >= 0x61 && letter <= 0x7a) {\n        if (script !== ScriptGroup.NEUTRAL && script !== ScriptGroup.LATIN) {\n          return index\n        }\n        script = ScriptGroup.LATIN\n        do {\n          index += 1\n          const nextLetter = text.charCodeAt(index) | 0x20\n          if (nextLetter < 0x61 || nextLetter > 0x7a) {\n            break\n          }\n        } while (index < text.length)\n        continue\n      }\n    } else if (unit >= 0x4e00 && unit <= 0x9fff) {\n      nextScript = ScriptGroup.EAST_ASIAN\n    } else if ((unit >= 0x3041 && unit <= 0x3096) || (unit >= 0x30a1 && unit <= 0x30fa)) {\n      nextScript = ScriptGroup.EAST_ASIAN\n    } else if (unit >= 0xac00 && unit <= 0xd7a3) {\n      nextScript = ScriptGroup.HANGUL\n    } else {\n      width = unit >= 0xd800 && unit <= 0xdbff && text.codePointAt(index)! > 0xffff ? 2 : 1\n      nextScript = getOtherScript(text, index, width)\n    }\n\n    if (nextScript !== ScriptGroup.NEUTRAL) {\n      if (script !== ScriptGroup.NEUTRAL && script !== nextScript) {\n        return index\n      }\n      script = nextScript\n    }\n    index += width\n  }\n  return text.length\n}\n\n/**\n * Returns a language tag in the canonical form the model stores, which is lower case.\n *\n * This mirrors the model's own helper, which it does not export.\n * Comparing tags stays with the consumer, since BCP 47 matches by subtag prefix rather than by string equality.\n */\nexport const lowerTag = (tag: string | undefined): string | undefined => {\n  return tag ? tag.toLowerCase() : undefined\n}\n"],"mappings":";;;;;;AAKA,IAAa,aAAgB,OAAY,cAAmC;CAC1E,IAAI,SAAS;CACb,KAAK,MAAM,CAAC,OAAO,SAAS,MAAM,QAAQ,GAAG;EAC3C,MAAM,cAAc,UAAU,IAAI;EAClC,IAAI,gBAAgB,MAAM;GACxB,IAAI,WAAW,OACb,SAAS,CAAC,GAAG,KAAK;GAEpB,OAAO,SAAS;EAClB;CACF;CACA,OAAO;AACT;;;;;;;;;AAUA,IAAa,cAAc,QAA2B,WAA2B;CAC/E,IAAI,OAAO;CACX,IAAI,QAAQ,OAAO;CACnB,OAAO,OAAO,OAAO;EACnB,MAAM,SAAS,KAAK,OAAO,OAAO,SAAS,CAAC;EAC5C,IAAI,OAAO,UAAW,QACpB,OAAO,SAAS;OAEhB,QAAQ;CAEZ;CACA,OAAO;AACT;;;AC1BA,IAAM,kBAA+D;CACnE,CAAC,2DAAA,CAAiF;CAClF,CAAC,qBAAA,CAAsC;CACvC,CAAC,sBAAA,CAAwC;CACzC,CAAC,wBAAA,CAA4C;CAC7C,CAAC,qBAAA,CAAsC;CACvC,CAAC,sBAAA,CAAwC;CACzC,CAAC,sBAAA,CAAwC;CACzC,CAAC,0BAAA,CAAgD;CACjD,CAAC,oBAAA,CAAoC;AACvC;AAEA,IAAM,kBAAkB,MAAc,OAAe,UAA+B;CAClF,MAAM,OAAO,KAAK,WAAW,KAAK;CAClC,IAAI,SAAS,SAAU,SAAS,OAC9B,OAAA;CAEF,IAAK,QAAQ,QAAU,QAAQ,QAAY,QAAQ,QAAU,QAAQ,MACnE,OAAA;CAEF,IAAI,QAAQ,OAAU,QAAQ,KAC5B,OAAA;CAEF,IAAK,QAAQ,QAAU,QAAQ,QAAY,QAAQ,QAAU,QAAQ,MACnE,OAAA;CAEF,IAAI,QAAQ,QAAU,QAAQ,MAC5B,OAAA;CAEF,IAAK,QAAQ,QAAU,QAAQ,QAAY,QAAQ,QAAU,QAAQ,QAAY,QAAQ,QAAU,QAAQ,MACzG,OAAA;CAEF,IAAK,QAAQ,QAAU,QAAQ,QAAY,QAAQ,QAAU,QAAQ,MACnE,OAAA;CAGF,MAAM,YAAY,KAAK,MAAM,OAAO,QAAQ,KAAK;CACjD,KAAK,MAAM,CAAC,SAAS,WAAW,iBAC9B,IAAI,QAAQ,KAAK,SAAS,GACxB,OAAO;CAGX,OAAA;AACF;;;;;;;;;;AAWA,IAAa,oBAAoB,MAAc,QAAQ,MAAc;CACnE,IAAI,SAAA;CACJ,IAAI,QAAQ;CAEZ,OAAO,QAAQ,KAAK,QAAQ;EAC1B,MAAM,OAAO,KAAK,WAAW,KAAK;EAClC,IAAI,aAAA;EACJ,IAAI,QAAQ;EAEZ,IAAI,QAAQ,KAAM;GAChB,MAAM,SAAS,OAAO;GACtB,IAAI,UAAU,MAAQ,UAAU,KAAM;IACpC,IAAI,WAAA,KAAkC,WAAA,GACpC,OAAO;IAET,SAAA;IACA,GAAG;KACD,SAAS;KACT,MAAM,aAAa,KAAK,WAAW,KAAK,IAAI;KAC5C,IAAI,aAAa,MAAQ,aAAa,KACpC;IAEJ,SAAS,QAAQ,KAAK;IACtB;GACF;EACF,OAAO,IAAI,QAAQ,SAAU,QAAQ,OACnC,aAAA;OACK,IAAK,QAAQ,SAAU,QAAQ,SAAY,QAAQ,SAAU,QAAQ,OAC1E,aAAA;OACK,IAAI,QAAQ,SAAU,QAAQ,OACnC,aAAA;OACK;GACL,QAAQ,QAAQ,SAAU,QAAQ,SAAU,KAAK,YAAY,KAAK,IAAK,QAAS,IAAI;GACpF,aAAa,eAAe,MAAM,OAAO,KAAK;EAChD;EAEA,IAAI,eAAA,GAAoC;GACtC,IAAI,WAAA,KAAkC,WAAW,YAC/C,OAAO;GAET,SAAS;EACX;EACA,SAAS;CACX;CACA,OAAO,KAAK;AACd;;;;;;;AAQA,IAAa,YAAY,QAAgD;CACvE,OAAO,MAAM,IAAI,YAAY,IAAI,KAAA;AACnC"}