@lokascript/i18n 2.11.0 → 3.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (42) hide show
  1. package/CHANGELOG.md +13 -0
  2. package/dist/browser.cjs +5 -1690
  3. package/dist/browser.cjs.map +1 -1
  4. package/dist/browser.d.cts +2 -2
  5. package/dist/browser.d.ts +2 -2
  6. package/dist/browser.js +6 -1685
  7. package/dist/browser.js.map +1 -1
  8. package/dist/dictionaries/index.cjs +5 -1
  9. package/dist/dictionaries/index.cjs.map +1 -1
  10. package/dist/dictionaries/index.js +5 -1
  11. package/dist/dictionaries/index.js.map +1 -1
  12. package/dist/{transformer-CsOeqayN.d.cts → index-BykxjYST.d.cts} +1 -232
  13. package/dist/{transformer-DWCTG1DQ.d.ts → index-DuIef8O7.d.ts} +1 -232
  14. package/dist/index.cjs +16 -1878
  15. package/dist/index.cjs.map +1 -1
  16. package/dist/index.d.cts +35 -35
  17. package/dist/index.d.ts +35 -35
  18. package/dist/index.js +17 -1873
  19. package/dist/index.js.map +1 -1
  20. package/dist/lokascript-i18n.min.js +1 -1
  21. package/dist/lokascript-i18n.min.js.map +1 -1
  22. package/dist/lokascript-i18n.mjs +49 -2680
  23. package/dist/lokascript-i18n.mjs.map +1 -1
  24. package/dist/plugins/vite.cjs +5 -1
  25. package/dist/plugins/vite.cjs.map +1 -1
  26. package/dist/plugins/vite.js +5 -1
  27. package/dist/plugins/vite.js.map +1 -1
  28. package/dist/plugins/webpack.cjs +5 -1
  29. package/dist/plugins/webpack.cjs.map +1 -1
  30. package/dist/plugins/webpack.js +5 -1
  31. package/dist/plugins/webpack.js.map +1 -1
  32. package/package.json +4 -4
  33. package/src/browser.ts +0 -7
  34. package/src/compatibility/browser-tests/grammar-demo.spec.ts +22 -8
  35. package/src/constants.ts +1 -0
  36. package/src/dictionaries/bn.ts +5 -1
  37. package/src/grammar/index.ts +15 -9
  38. package/src/grammar/profiles.test.ts +440 -0
  39. package/src/index.ts +0 -7
  40. package/src/lexicon-parity.test.ts +77 -0
  41. package/src/grammar/grammar.test.ts +0 -2751
  42. package/src/grammar/transformer.ts +0 -2737
package/dist/index.js CHANGED
@@ -4523,7 +4523,11 @@ var bengaliDictionary = {
4523
4523
  random: "\u098F\u09B2\u09CB\u09AE\u09C7\u09B2\u09CB",
4524
4524
  length: "\u09A6\u09C8\u09B0\u09CD\u0998\u09CD\u09AF",
4525
4525
  index: "\u09B8\u09C2\u099A\u0995",
4526
- empty: "\u0996\u09BE\u09B2\u09BF-\u0995\u09B0\u09C1\u09A8",
4526
+ // The EXPRESSION `empty` is the state predicate (`if my value is empty`),
4527
+ // not the command: `খালি-করুন` is the imperative "empty it!" and belongs to
4528
+ // the `empty` COMMAND, which keeps it. Kept in step with the semantic
4529
+ // lexicon by `lexicon-parity.test.ts`.
4530
+ empty: "\u0996\u09BE\u09B2\u09BF",
4527
4531
  "starts with": "\u09A6\u09BF\u09AF\u09BC\u09C7_\u09B6\u09C1\u09B0\u09C1",
4528
4532
  "ends with": "\u09A6\u09BF\u09AF\u09BC\u09C7_\u09B6\u09C7\u09B7",
4529
4533
  "ignoring case": "\u0995\u09C7\u09B8_\u0989\u09AA\u09C7\u0995\u09CD\u09B7\u09BE",
@@ -7435,25 +7439,25 @@ var HyperscriptTranslator = class {
7435
7439
  }
7436
7440
  translateWithDetails(text, options) {
7437
7441
  const fromLocale = options.from || this.detectLanguage(text);
7438
- const toLocale2 = options.to;
7439
- if (fromLocale === toLocale2) {
7442
+ const toLocale = options.to;
7443
+ if (fromLocale === toLocale) {
7440
7444
  return {
7441
7445
  translated: text,
7442
7446
  original: text,
7443
7447
  tokens: [],
7444
- locale: { from: fromLocale, to: toLocale2 }
7448
+ locale: { from: fromLocale, to: toLocale }
7445
7449
  };
7446
7450
  }
7447
7451
  const fromDict = this.getDictionary(fromLocale);
7448
- const toDict = this.getDictionary(toLocale2);
7452
+ const toDict = this.getDictionary(toLocale);
7449
7453
  if (!fromDict || !toDict) {
7450
- throw new Error(`Missing dictionary for locale: ${!fromDict ? fromLocale : toLocale2}`);
7454
+ throw new Error(`Missing dictionary for locale: ${!fromDict ? fromLocale : toLocale}`);
7451
7455
  }
7452
7456
  const tokens = tokenize(text, fromLocale);
7453
- const translatedTokens = this.translateTokens(tokens, fromLocale, toLocale2);
7457
+ const translatedTokens = this.translateTokens(tokens, fromLocale, toLocale);
7454
7458
  const translated = this.reconstructText(translatedTokens);
7455
7459
  if (options.validate && toDict) {
7456
- const validation = validate(toDict, toLocale2);
7460
+ const validation = validate(toDict, toLocale);
7457
7461
  if (!validation.valid) {
7458
7462
  console.warn("Translation validation warnings:", validation.warnings);
7459
7463
  }
@@ -7461,7 +7465,7 @@ var HyperscriptTranslator = class {
7461
7465
  const result = {
7462
7466
  translated,
7463
7467
  tokens: translatedTokens,
7464
- locale: { from: fromLocale, to: toLocale2 },
7468
+ locale: { from: fromLocale, to: toLocale },
7465
7469
  warnings: []
7466
7470
  };
7467
7471
  if (options.preserveOriginal) {
@@ -7469,9 +7473,9 @@ var HyperscriptTranslator = class {
7469
7473
  }
7470
7474
  return result;
7471
7475
  }
7472
- translateTokens(tokens, fromLocale, toLocale2) {
7476
+ translateTokens(tokens, fromLocale, toLocale) {
7473
7477
  const fromDict = this.getDictionary(fromLocale);
7474
- const toDict = this.getDictionary(toLocale2);
7478
+ const toDict = this.getDictionary(toLocale);
7475
7479
  const reverseFromDict = this.getReverseDictionary(fromLocale);
7476
7480
  const emptyDict = {
7477
7481
  commands: {},
@@ -7486,7 +7490,7 @@ var HyperscriptTranslator = class {
7486
7490
  return tokens.map((token) => {
7487
7491
  let translated = token.value;
7488
7492
  if (this.isTranslatableToken(token)) {
7489
- if (fromLocale !== "en" && toLocale2 !== "en") {
7493
+ if (fromLocale !== "en" && toLocale !== "en") {
7490
7494
  const english = this.findTranslation(token.value, fromDict || emptyDict, reverseFromDict);
7491
7495
  if (english) {
7492
7496
  translated = this.findTranslation(english, toDict || emptyDict, /* @__PURE__ */ new Map()) || token.value;
@@ -7617,39 +7621,6 @@ var HyperscriptTranslator = class {
7617
7621
  new HyperscriptTranslator({ locale: "en" });
7618
7622
 
7619
7623
  // src/constants.ts
7620
- var ENGLISH_MODIFIER_ROLES = {
7621
- to: "destination",
7622
- into: "destination",
7623
- from: "source",
7624
- with: "style",
7625
- by: "quantity",
7626
- as: "method",
7627
- on: "event",
7628
- over: "duration",
7629
- for: "duration"
7630
- };
7631
- var COMMAND_PRIMARY_ROLES = {
7632
- set: "destination",
7633
- on: "event",
7634
- trigger: "event",
7635
- send: "event",
7636
- wait: "duration",
7637
- fetch: "source",
7638
- get: "source",
7639
- if: "condition",
7640
- unless: "condition",
7641
- while: "condition",
7642
- repeat: "loopType",
7643
- go: "destination",
7644
- scroll: "destination",
7645
- tell: "destination",
7646
- default: "destination",
7647
- swap: "destination",
7648
- // morph deliberately absent: its schema primaryRole is `patient` (the
7649
- // element being morphed — aligned with the transformer's patient marking
7650
- // in the session-9 role-layout swap), and patient is the default.
7651
- bind: "destination"
7652
- };
7653
7624
  var ENGLISH_MODIFIERS = /* @__PURE__ */ new Set([
7654
7625
  "to",
7655
7626
  "from",
@@ -7882,69 +7853,6 @@ var ENGLISH_EXPRESSION_KEYWORDS = /* @__PURE__ */ new Set([
7882
7853
  "starts with",
7883
7854
  "ends with"
7884
7855
  ]);
7885
- var CONDITIONAL_KEYWORDS = /* @__PURE__ */ new Set([
7886
- // English
7887
- "if",
7888
- "unless",
7889
- "when",
7890
- "where",
7891
- // Japanese
7892
- "\u3082\u3057",
7893
- "\u6642\u306B",
7894
- "\u3068\u304D\u306B",
7895
- "\u3069\u3053\u3067",
7896
- // Chinese
7897
- "\u5982\u679C",
7898
- "\u5F53",
7899
- // Arabic
7900
- "\u0625\u0630\u0627",
7901
- "\u0639\u0646\u062F\u0645\u0627",
7902
- "\u062D\u064A\u062B",
7903
- // Spanish
7904
- "si",
7905
- "cuando",
7906
- "donde",
7907
- // German
7908
- "wenn",
7909
- "wann",
7910
- "wo",
7911
- // French
7912
- "quand",
7913
- "lorsque",
7914
- "o\xF9",
7915
- // Portuguese
7916
- "quando",
7917
- "onde",
7918
- // Turkish
7919
- "e\u011Fer",
7920
- "zaman",
7921
- "nerede",
7922
- // Indonesian
7923
- "ketika",
7924
- "saat",
7925
- "dimana",
7926
- // Korean
7927
- "\uB54C",
7928
- "\uC5B4\uB514\uC11C",
7929
- // Quechua
7930
- "maypi",
7931
- // Swahili
7932
- "wakati",
7933
- "wapi"
7934
- ]);
7935
- var THEN_KEYWORDS = /* @__PURE__ */ new Set([
7936
- "then",
7937
- "\u305D\u308C\u304B\u3089",
7938
- "\u90A3\u4E48",
7939
- "\u062B\u0645",
7940
- "entonces",
7941
- "alors",
7942
- "dann",
7943
- "sonra",
7944
- "lalu",
7945
- "chayqa",
7946
- "kisha"
7947
- ]);
7948
7856
 
7949
7857
  // src/parser/create-provider.ts
7950
7858
  function createKeywordProvider(dictionary, locale, options = {}) {
@@ -14177,1774 +14085,10 @@ async function createEnhancedI18n(locale, options) {
14177
14085
  }
14178
14086
  var enhancedI18nImplementation = new TypedI18nContextImplementation();
14179
14087
 
14180
- // src/grammar/direct-mappings.ts
14181
- function reverseMapping(mapping) {
14182
- const reversedWords = {};
14183
- for (const [source, target] of Object.entries(mapping.words)) {
14184
- reversedWords[target] = source;
14185
- }
14186
- const reversedCategories = {};
14187
- if (mapping.categories) {
14188
- for (const [category, words] of Object.entries(mapping.categories)) {
14189
- reversedCategories[category] = {};
14190
- for (const [source, target] of Object.entries(words)) {
14191
- reversedCategories[category][target] = source;
14192
- }
14193
- }
14194
- }
14195
- const result = {
14196
- source: mapping.target,
14197
- target: mapping.source,
14198
- words: reversedWords
14199
- };
14200
- if (Object.keys(reversedCategories).length > 0) {
14201
- result.categories = reversedCategories;
14202
- }
14203
- return result;
14204
- }
14205
- function buildMappingFromDictionaries(sourceDict, targetDict, sourceCode, targetCode) {
14206
- const words = {};
14207
- const categories = {};
14208
- for (const category of [
14209
- "commands",
14210
- "modifiers",
14211
- "events",
14212
- "logical",
14213
- "temporal",
14214
- "values",
14215
- "attributes"
14216
- ]) {
14217
- const sourceCategory = sourceDict[category];
14218
- const targetCategory = targetDict[category];
14219
- if (sourceCategory && targetCategory) {
14220
- categories[category] = {};
14221
- for (const [englishKey, sourceWord] of Object.entries(sourceCategory)) {
14222
- const targetWord = targetCategory[englishKey];
14223
- if (targetWord && sourceWord !== targetWord) {
14224
- words[sourceWord] = targetWord;
14225
- categories[category][sourceWord] = targetWord;
14226
- }
14227
- }
14228
- }
14229
- }
14230
- return {
14231
- source: sourceCode,
14232
- target: targetCode,
14233
- words,
14234
- categories
14235
- };
14236
- }
14237
- var jaZhMapping = buildMappingFromDictionaries(ja, zh, "ja", "zh");
14238
- var zhJaMapping = reverseMapping(jaZhMapping);
14239
- var koJaMapping = buildMappingFromDictionaries(ko, ja, "ko", "ja");
14240
- var jaKoMapping = reverseMapping(koJaMapping);
14241
- var esPtMapping = {
14242
- source: "es",
14243
- target: "pt",
14244
- words: {
14245
- // Commands - many are cognates
14246
- en: "em",
14247
- // on (event)
14248
- decir: "dizer",
14249
- // tell
14250
- disparar: "disparar",
14251
- // trigger
14252
- enviar: "enviar",
14253
- // send
14254
- tomar: "pegar",
14255
- // take
14256
- poner: "colocar",
14257
- // put
14258
- establecer: "definir",
14259
- // set
14260
- obtener: "obter",
14261
- // get
14262
- agregar: "adicionar",
14263
- // add
14264
- quitar: "remover",
14265
- // remove
14266
- alternar: "alternar",
14267
- // toggle
14268
- ocultar: "esconder",
14269
- // hide
14270
- mostrar: "mostrar",
14271
- // show
14272
- si: "se",
14273
- // if
14274
- menos: "a menos",
14275
- // unless
14276
- repetir: "repetir",
14277
- // repeat
14278
- para: "para",
14279
- // for
14280
- mientras: "enquanto",
14281
- // while
14282
- hasta: "at\xE9",
14283
- // until
14284
- continuar: "continuar",
14285
- // continue
14286
- romper: "parar",
14287
- // break
14288
- detener: "parar",
14289
- // halt
14290
- esperar: "esperar",
14291
- // wait
14292
- buscar: "buscar",
14293
- // fetch
14294
- llamar: "chamar",
14295
- // call
14296
- retornar: "retornar",
14297
- // return
14298
- hacer: "fazer",
14299
- // make
14300
- registrar: "registrar",
14301
- // log
14302
- lanzar: "lan\xE7ar",
14303
- // throw
14304
- atrapar: "capturar",
14305
- // catch
14306
- medir: "medir",
14307
- // measure
14308
- transici\u00F3n: "transi\xE7\xE3o",
14309
- // transition
14310
- incrementar: "incrementar",
14311
- // increment
14312
- decrementar: "decrementar",
14313
- // decrement
14314
- predeterminar: "padr\xE3o",
14315
- // default
14316
- ir: "ir",
14317
- // go
14318
- copiar: "copiar",
14319
- // copy
14320
- escoger: "escolher",
14321
- // pick
14322
- intercambiar: "trocar",
14323
- // swap
14324
- transformar: "transformar",
14325
- // morph
14326
- a\u00F1adir: "anexar",
14327
- // append
14328
- salir: "sair",
14329
- // exit
14330
- // Modifiers
14331
- a: "para",
14332
- // to
14333
- de: "de",
14334
- // from
14335
- dentro: "em",
14336
- // into
14337
- con: "com",
14338
- // with
14339
- como: "como",
14340
- // as
14341
- por: "por",
14342
- // by
14343
- // Events
14344
- clic: "clique",
14345
- // click
14346
- cambio: "mudan\xE7a",
14347
- // change
14348
- entrada: "entrada",
14349
- // input
14350
- env\u00EDo: "envio",
14351
- // submit
14352
- carga: "carregar",
14353
- // load
14354
- enfoque: "foco",
14355
- // focus
14356
- desenfoque: "desfoque",
14357
- // blur
14358
- // Logical
14359
- verdadero: "verdadeiro",
14360
- // true
14361
- falso: "falso",
14362
- // false
14363
- y: "e",
14364
- // and
14365
- o: "ou",
14366
- // or
14367
- no: "n\xE3o",
14368
- // not
14369
- es: "\xE9",
14370
- // is
14371
- // Temporal
14372
- segundos: "segundos",
14373
- // seconds
14374
- milisegundos: "milissegundos",
14375
- // milliseconds
14376
- // Values
14377
- nulo: "nulo",
14378
- // null
14379
- indefinido: "indefinido",
14380
- // undefined
14381
- vac\u00EDo: "vazio"
14382
- // empty
14383
- },
14384
- categories: {
14385
- commands: {
14386
- en: "em",
14387
- decir: "dizer",
14388
- tomar: "pegar",
14389
- poner: "colocar",
14390
- establecer: "definir",
14391
- obtener: "obter",
14392
- agregar: "adicionar",
14393
- quitar: "remover",
14394
- ocultar: "esconder",
14395
- si: "se",
14396
- mientras: "enquanto",
14397
- hasta: "at\xE9",
14398
- romper: "parar",
14399
- detener: "parar",
14400
- llamar: "chamar",
14401
- hacer: "fazer",
14402
- lanzar: "lan\xE7ar",
14403
- atrapar: "capturar",
14404
- escoger: "escolher",
14405
- intercambiar: "trocar",
14406
- a\u00F1adir: "anexar",
14407
- salir: "sair"
14408
- },
14409
- events: {
14410
- clic: "clique",
14411
- cambio: "mudan\xE7a",
14412
- carga: "carregar",
14413
- enfoque: "foco",
14414
- desenfoque: "desfoque"
14415
- },
14416
- logical: {
14417
- verdadero: "verdadeiro",
14418
- y: "e",
14419
- o: "ou",
14420
- no: "n\xE3o",
14421
- es: "\xE9"
14422
- }
14423
- }
14424
- };
14425
- var ptEsMapping = reverseMapping(esPtMapping);
14426
- var directMappings = /* @__PURE__ */ new Map([
14427
- // Japanese ↔ Chinese
14428
- ["ja->zh", jaZhMapping],
14429
- ["zh->ja", zhJaMapping],
14430
- // Korean ↔ Japanese
14431
- ["ko->ja", koJaMapping],
14432
- ["ja->ko", jaKoMapping],
14433
- // Spanish ↔ Portuguese
14434
- ["es->pt", esPtMapping],
14435
- ["pt->es", ptEsMapping]
14436
- ]);
14437
- function hasDirectMapping(source, target) {
14438
- return directMappings.has(`${source}->${target}`);
14439
- }
14440
- function getDirectMapping(source, target) {
14441
- return directMappings.get(`${source}->${target}`);
14442
- }
14443
-
14444
- // src/grammar/transformer.ts
14445
- function getCommandKeywordsForLocale(locale) {
14446
- const keywords = new Set(ENGLISH_COMMANDS);
14447
- const dict = dictionaries[locale];
14448
- if (dict?.commands) {
14449
- Object.values(dict.commands).forEach((cmd) => {
14450
- if (typeof cmd === "string") {
14451
- keywords.add(cmd.toLowerCase());
14452
- }
14453
- });
14454
- }
14455
- return keywords;
14456
- }
14457
- var EN_COPULAS = ["is", "are", "was", "were", "am", "be"];
14458
- function getCopulasForLocale(locale) {
14459
- const copulas = new Set(EN_COPULAS);
14460
- if (locale !== "en") {
14461
- for (const form of EN_COPULAS) {
14462
- copulas.add(translateWord(form, "en", locale).toLowerCase());
14463
- }
14464
- }
14465
- return copulas;
14466
- }
14467
- function isPredicateAdjectivePosition(tokens, i, copulas) {
14468
- const prev = tokens[i - 1]?.toLowerCase();
14469
- return !!prev && copulas.has(prev);
14470
- }
14471
- function getForLoopWordsForLocale(locale) {
14472
- const forWords = /* @__PURE__ */ new Set(["for"]);
14473
- const inWords = /* @__PURE__ */ new Set(["in"]);
14474
- if (locale !== "en") {
14475
- forWords.add(translateWord("for", "en", locale).toLowerCase());
14476
- inWords.add(translateWord("in", "en", locale).toLowerCase());
14477
- }
14478
- return { forWords, inWords };
14479
- }
14480
- function isLoopHeadFor(tokens, i, inWords, commandKeywords) {
14481
- for (let j = i + 1; j < tokens.length; j++) {
14482
- const lt = tokens[j].toLowerCase();
14483
- if (inWords.has(lt)) return true;
14484
- if (commandKeywords.has(lt)) return false;
14485
- }
14486
- return false;
14487
- }
14488
- function repairHebrewFrontedAccusative(text) {
14489
- const ACC = "\u05D0\u05EA";
14490
- const verbs = getCommandKeywordsForLocale("he");
14491
- const tokens = text.split(/\s+/);
14492
- let changed = false;
14493
- for (let i = 0; i + 1 < tokens.length; i++) {
14494
- if (tokens[i] === ACC && verbs.has(tokens[i + 1].toLowerCase())) {
14495
- [tokens[i], tokens[i + 1]] = [tokens[i + 1], tokens[i]];
14496
- changed = true;
14497
- i++;
14498
- }
14499
- }
14500
- return changed ? tokens.join(" ") : text;
14501
- }
14502
- function extractBlockStructure(input, sourceLocale) {
14503
- const tokens = input.split(/\s+/);
14504
- const head = tokens[0]?.toLowerCase();
14505
- if (!head || !BLOCK_HEAD_KEYWORDS.has(head)) return null;
14506
- let depth = 1;
14507
- let endIdx = -1;
14508
- for (let i = 1; i < tokens.length; i++) {
14509
- const t = tokens[i].toLowerCase();
14510
- if (BLOCK_HEAD_KEYWORDS.has(t)) depth++;
14511
- else if (t === "end") {
14512
- depth--;
14513
- if (depth === 0) {
14514
- endIdx = i;
14515
- break;
14516
- }
14517
- }
14518
- }
14519
- if (endIdx !== -1 && endIdx !== tokens.length - 1) return null;
14520
- const inner = endIdx !== -1 ? tokens.slice(1, endIdx) : tokens.slice(1);
14521
- const base = { headKeyword: tokens[0], body: "" };
14522
- if (endIdx !== -1) base.tailKeyword = tokens[endIdx];
14523
- if (head === "live") {
14524
- return { ...base, body: inner.join(" ") };
14525
- }
14526
- if (head === "when") {
14527
- const idx = inner.findIndex((t) => t.toLowerCase() === "changes");
14528
- if (idx >= 0) {
14529
- return {
14530
- ...base,
14531
- prefixExpr: inner.slice(0, idx).join(" "),
14532
- connector: inner[idx],
14533
- body: inner.slice(idx + 1).join(" ")
14534
- };
14535
- }
14536
- return null;
14537
- }
14538
- const commands = getCommandKeywordsForLocale(sourceLocale);
14539
- const copulas = getCopulasForLocale(sourceLocale);
14540
- let bodyStart = -1;
14541
- for (let i = 0; i < inner.length; i++) {
14542
- if (commands.has(inner[i].toLowerCase()) && !isPredicateAdjectivePosition(inner, i, copulas)) {
14543
- bodyStart = i;
14544
- break;
14545
- }
14546
- }
14547
- if (bodyStart <= 0) return null;
14548
- return {
14549
- ...base,
14550
- prefixExpr: inner.slice(0, bodyStart).join(" "),
14551
- body: inner.slice(bodyStart).join(" ")
14552
- };
14553
- }
14554
- function splitCompoundStatement(input, sourceLocale) {
14555
- const lines = input.split(/\n/).map((line) => line.trim()).filter((line) => line.length > 0);
14556
- const parts = [];
14557
- for (const line of lines) {
14558
- const lineParts = splitOnThen(line, sourceLocale);
14559
- for (const part of lineParts) {
14560
- const commandParts = splitOnCommandBoundaries(part, sourceLocale);
14561
- parts.push(...commandParts);
14562
- }
14563
- }
14564
- return parts;
14565
- }
14566
- function splitCompoundStatementWithMetadata(input, sourceLocale) {
14567
- const rawLines = input.split("\n");
14568
- const lineMetadata = [];
14569
- const parts = [];
14570
- const partToLineIndex = [];
14571
- for (let lineIndex = 0; lineIndex < rawLines.length; lineIndex++) {
14572
- const rawLine = rawLines[lineIndex];
14573
- const indentMatch = rawLine.match(/^(\s*)/);
14574
- const originalIndent = indentMatch ? indentMatch[1] : "";
14575
- const trimmed = rawLine.trim();
14576
- lineMetadata.push({
14577
- content: trimmed,
14578
- originalIndent,
14579
- isBlank: trimmed.length === 0
14580
- });
14581
- if (trimmed.length > 0) {
14582
- const lineParts = splitOnThen(trimmed, sourceLocale);
14583
- for (const part of lineParts) {
14584
- const commandParts = splitOnCommandBoundaries(part, sourceLocale);
14585
- for (const cmdPart of commandParts) {
14586
- parts.push(cmdPart);
14587
- partToLineIndex.push(lineIndex);
14588
- }
14589
- }
14590
- }
14591
- }
14592
- return { parts, lineMetadata, partToLineIndex };
14593
- }
14594
- function normalizeIndentation(lineMetadata) {
14595
- const indentedLines = lineMetadata.filter((m) => !m.isBlank && m.originalIndent.length > 0);
14596
- if (indentedLines.length === 0) {
14597
- return lineMetadata.map(() => "");
14598
- }
14599
- const indentLengths = indentedLines.map((m) => {
14600
- const normalized = m.originalIndent.replace(/\t/g, " ");
14601
- return normalized.length;
14602
- });
14603
- const minIndent = Math.min(...indentLengths);
14604
- const baseUnit = minIndent > 0 ? minIndent : 4;
14605
- return lineMetadata.map((meta) => {
14606
- if (meta.isBlank) {
14607
- return "";
14608
- }
14609
- if (meta.originalIndent.length === 0) {
14610
- return "";
14611
- }
14612
- const normalized = meta.originalIndent.replace(/\t/g, " ");
14613
- const level = Math.round(normalized.length / baseUnit);
14614
- return " ".repeat(level);
14615
- });
14616
- }
14617
- function reconstructWithLineStructure(transformedParts, lineMetadata, partToLineIndex, targetThen) {
14618
- const nonBlankCount = lineMetadata.filter((m) => !m.isBlank).length;
14619
- if (nonBlankCount <= 1 && transformedParts.length <= 1) {
14620
- const normalizedIndents2 = normalizeIndentation(lineMetadata);
14621
- const result2 = [];
14622
- for (let i = 0; i < lineMetadata.length; i++) {
14623
- if (lineMetadata[i].isBlank) {
14624
- result2.push("");
14625
- } else if (transformedParts.length > 0) {
14626
- result2.push(normalizedIndents2[i] + transformedParts[0]);
14627
- }
14628
- }
14629
- return result2.join("\n");
14630
- }
14631
- const normalizedIndents = normalizeIndentation(lineMetadata);
14632
- const partsPerLine = /* @__PURE__ */ new Map();
14633
- for (let i = 0; i < transformedParts.length; i++) {
14634
- const lineIdx = partToLineIndex[i];
14635
- if (!partsPerLine.has(lineIdx)) {
14636
- partsPerLine.set(lineIdx, []);
14637
- }
14638
- partsPerLine.get(lineIdx).push(transformedParts[i]);
14639
- }
14640
- const result = [];
14641
- for (let i = 0; i < lineMetadata.length; i++) {
14642
- const meta = lineMetadata[i];
14643
- const indent = normalizedIndents[i];
14644
- if (meta.isBlank) {
14645
- result.push("");
14646
- } else {
14647
- const lineParts = partsPerLine.get(i) || [];
14648
- if (lineParts.length > 0) {
14649
- const lineContent = lineParts.join(` ${targetThen} `);
14650
- result.push(indent + lineContent);
14651
- }
14652
- }
14653
- }
14654
- return result.join("\n");
14655
- }
14656
- var BOUNDARY_MODIFIERS = /* @__PURE__ */ new Set([
14657
- "to",
14658
- "into",
14659
- "from",
14660
- "with",
14661
- "by",
14662
- "as",
14663
- "at",
14664
- "in",
14665
- "on",
14666
- "of",
14667
- "over"
14668
- ]);
14669
- var boundaryModifiersCache = /* @__PURE__ */ new Map();
14670
- function getBoundaryModifiersForLocale(locale) {
14671
- const cached = boundaryModifiersCache.get(locale);
14672
- if (cached) return cached;
14673
- const modifiers = new Set(BOUNDARY_MODIFIERS);
14674
- const profile = getProfile(locale);
14675
- profile?.markers.forEach((marker) => {
14676
- const form = marker.form.replace(/^-|-$/g, "").toLowerCase();
14677
- if (form) modifiers.add(form);
14678
- marker.alternatives?.forEach((alt) => {
14679
- const altForm = alt.replace(/^-|-$/g, "").toLowerCase();
14680
- if (altForm) modifiers.add(altForm);
14681
- });
14682
- });
14683
- boundaryModifiersCache.set(locale, modifiers);
14684
- return modifiers;
14685
- }
14686
- var ON_TARGET_COMMANDS = /* @__PURE__ */ new Set(["toggle", "add", "remove", "trigger", "send"]);
14687
- function commandVerbOf(tokens, commandKeywords) {
14688
- for (const token of tokens) {
14689
- const lt = token.toLowerCase();
14690
- if (commandKeywords.has(lt) && !BOUNDARY_MODIFIERS.has(lt) && !EVENT_KEYWORDS.has(lt)) {
14691
- return lt;
14692
- }
14693
- }
14694
- return null;
14695
- }
14696
- var BLOCK_HEAD_KEYWORDS = /* @__PURE__ */ new Set(["live", "when", "unless"]);
14697
- var BLOCK_BODY_KEYWORDS = /* @__PURE__ */ new Set(["if", "repeat", "unless", "while", "for"]);
14698
- var UNLESS_GUARD_OBJECT_MARKING_LOCALES = /* @__PURE__ */ new Set(["he", "zh"]);
14699
- function splitOnCommandBoundaries(input, sourceLocale) {
14700
- const commandKeywords = getCommandKeywordsForLocale(sourceLocale);
14701
- const boundaryModifiers = getBoundaryModifiersForLocale(sourceLocale);
14702
- const { forWords, inWords } = getForLoopWordsForLocale(sourceLocale);
14703
- const tokens = input.split(/\s+/);
14704
- if (tokens.length === 0) return [input];
14705
- const parts = [];
14706
- let currentPart = [];
14707
- const firstTokenLower = tokens[0]?.toLowerCase();
14708
- const isEventHandler = EVENT_KEYWORDS.has(firstTokenLower);
14709
- let seenFirstCommand = !isEventHandler;
14710
- let blockDepth = 0;
14711
- for (let i = 0; i < tokens.length; i++) {
14712
- const token = tokens[i];
14713
- const lowerToken = token.toLowerCase();
14714
- if (BLOCK_HEAD_KEYWORDS.has(lowerToken)) {
14715
- blockDepth++;
14716
- } else if (lowerToken === "end" && blockDepth > 0) {
14717
- blockDepth--;
14718
- }
14719
- if (commandKeywords.has(lowerToken) && currentPart.length > 0) {
14720
- const prevToken = currentPart[currentPart.length - 1];
14721
- const prevLower = prevToken.toLowerCase();
14722
- if (!seenFirstCommand) {
14723
- seenFirstCommand = true;
14724
- currentPart.push(token);
14725
- continue;
14726
- }
14727
- if (blockDepth > 0) {
14728
- currentPart.push(token);
14729
- continue;
14730
- }
14731
- if (forWords.has(lowerToken) && !isLoopHeadFor(tokens, i, inWords, commandKeywords)) {
14732
- currentPart.push(token);
14733
- continue;
14734
- }
14735
- if (BOUNDARY_MODIFIERS.has(lowerToken)) {
14736
- const verb = commandVerbOf(currentPart, commandKeywords);
14737
- if (verb && ON_TARGET_COMMANDS.has(verb)) {
14738
- currentPart.push(token);
14739
- continue;
14740
- }
14741
- if (lowerToken === "on" && verb === "set") {
14742
- const nextTok = tokens[i + 1];
14743
- const nextLower = nextTok?.toLowerCase();
14744
- const scopeLike = !!nextTok && (/^[#.<@[]/.test(nextTok) || nextLower === "me" || nextLower === "it" || nextLower === "you");
14745
- if (scopeLike) {
14746
- currentPart.push(token);
14747
- continue;
14748
- }
14749
- }
14750
- }
14751
- if (!boundaryModifiers.has(prevLower) && !commandKeywords.has(prevLower)) {
14752
- parts.push(currentPart.join(" "));
14753
- currentPart = [token];
14754
- continue;
14755
- }
14756
- }
14757
- currentPart.push(token);
14758
- }
14759
- if (currentPart.length > 0) {
14760
- parts.push(currentPart.join(" "));
14761
- }
14762
- return parts.filter((p) => p.length > 0);
14763
- }
14764
- function splitOnThen(input, sourceLocale) {
14765
- const thenKeywords = Array.from(THEN_KEYWORDS);
14766
- const sourceDict = sourceLocale === "en" ? null : dictionaries[sourceLocale];
14767
- if (sourceDict?.modifiers?.then) {
14768
- thenKeywords.push(sourceDict.modifiers.then);
14769
- }
14770
- if (sourceDict?.logical?.then) {
14771
- thenKeywords.push((sourceDict?.logical).then);
14772
- }
14773
- const escapedKeywords = thenKeywords.map((k) => k.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"));
14774
- const pattern = new RegExp(`\\s+(${escapedKeywords.join("|")})\\s+`, "gi");
14775
- const parts = input.split(pattern).filter((part) => {
14776
- const lowerPart = part.toLowerCase().trim();
14777
- return lowerPart && !thenKeywords.some((k) => k.toLowerCase() === lowerPart);
14778
- });
14779
- return parts.map((p) => p.trim()).filter((p) => p.length > 0);
14780
- }
14781
- function getTargetThenKeyword(targetLocale) {
14782
- if (targetLocale === "en") return "then";
14783
- const targetDict = dictionaries[targetLocale];
14784
- if (!targetDict) return "then";
14785
- return targetDict.modifiers?.then || targetDict.logical?.then || "then";
14786
- }
14787
- function deriveEventKeywordsFromProfiles() {
14788
- const keywords = /* @__PURE__ */ new Set();
14789
- keywords.add("on");
14790
- for (const profile of Object.values(profiles)) {
14791
- for (const marker of profile.markers) {
14792
- if (marker.role === "event") {
14793
- const form = marker.form.replace(/^-|-$/g, "").toLowerCase();
14794
- if (form) keywords.add(form);
14795
- marker.alternatives?.forEach((alt) => {
14796
- const altForm = alt.replace(/^-|-$/g, "").toLowerCase();
14797
- if (altForm) keywords.add(altForm);
14798
- });
14799
- }
14800
- }
14801
- }
14802
- return keywords;
14803
- }
14804
- var EVENT_KEYWORDS = deriveEventKeywordsFromProfiles();
14805
- var EVENT_CONJUNCTIONS = /* @__PURE__ */ new Set(["or"]);
14806
- var BODY_MODIFIER_KEYWORDS = /* @__PURE__ */ new Set([
14807
- "async",
14808
- "once",
14809
- "debounced",
14810
- "debounce",
14811
- "throttled",
14812
- "throttle"
14813
- ]);
14814
- function generateModifierMap(profile) {
14815
- const map = {};
14816
- profile.markers.forEach((marker) => {
14817
- const form = marker.form.replace(/^-|-$/g, "").toLowerCase();
14818
- if (form) {
14819
- map[form] = marker.role;
14820
- }
14821
- marker.alternatives?.forEach((alt) => {
14822
- const altForm = alt.replace(/^-|-$/g, "").toLowerCase();
14823
- if (altForm) {
14824
- map[altForm] = marker.role;
14825
- }
14826
- });
14827
- });
14828
- for (const [key, role] of Object.entries(ENGLISH_MODIFIER_ROLES)) {
14829
- if (!(key in map)) {
14830
- map[key] = role;
14831
- }
14832
- }
14833
- return map;
14834
- }
14835
- function buildArgumentModifierMap(profile, actionVerb) {
14836
- const map = generateModifierMap(profile);
14837
- const verb = actionVerb?.toLowerCase();
14838
- if (profile.wordOrder !== "SVO" || !verb || !ON_TARGET_COMMANDS.has(verb)) {
14839
- return map;
14840
- }
14841
- const remapped = {};
14842
- for (const [form, role] of Object.entries(map)) {
14843
- remapped[form] = role === "event" ? "destination" : role;
14844
- }
14845
- return remapped;
14846
- }
14847
- function parseStatement(input, sourceLocale = "en") {
14848
- const profile = getProfile(sourceLocale);
14849
- if (!profile) return null;
14850
- const tokens = tokenize2(input, profile);
14851
- const statementType = identifyStatementType(tokens, profile);
14852
- switch (statementType) {
14853
- case "event-handler":
14854
- return parseEventHandler(tokens, profile);
14855
- case "command":
14856
- return parseCommand(tokens, profile);
14857
- case "conditional":
14858
- return parseConditional(tokens);
14859
- default:
14860
- return null;
14861
- }
14862
- }
14863
- var ATTACHED_SUFFIXES = {
14864
- // Chinese: 时 (time/when) often attaches to events like 点击时 (when clicking)
14865
- zh: ["\u65F6", "\u7684", "\u5730", "\u5F97"],
14866
- // Japanese: Some particles may attach in casual writing
14867
- ja: [],
14868
- // Korean: Particles sometimes written without spaces
14869
- ko: []
14870
- };
14871
- var ATTACHED_PREFIXES = {
14872
- // Chinese: 当 (when) sometimes written attached
14873
- zh: ["\u5F53"],
14874
- // Arabic: Prepositions that attach
14875
- ar: ["\u0628\u0640", "\u0643\u0640", "\u0648"]
14876
- };
14877
- function splitAttachedAffixes(tokens, locale) {
14878
- const suffixes = ATTACHED_SUFFIXES[locale] || [];
14879
- const prefixes = ATTACHED_PREFIXES[locale] || [];
14880
- if (suffixes.length === 0 && prefixes.length === 0) {
14881
- return tokens;
14882
- }
14883
- const result = [];
14884
- for (const token of tokens) {
14885
- if (/^[#.<@]/.test(token) || /^\d+/.test(token)) {
14886
- result.push(token);
14887
- continue;
14888
- }
14889
- let processed = token;
14890
- let prefix = "";
14891
- let suffix = "";
14892
- for (const p of prefixes) {
14893
- if (processed.startsWith(p) && processed.length > p.length) {
14894
- prefix = p;
14895
- processed = processed.slice(p.length);
14896
- break;
14897
- }
14898
- }
14899
- for (const s of suffixes) {
14900
- if (processed.endsWith(s) && processed.length > s.length) {
14901
- suffix = s;
14902
- processed = processed.slice(0, -s.length);
14903
- break;
14904
- }
14905
- }
14906
- if (prefix) result.push(prefix);
14907
- if (processed) result.push(processed);
14908
- if (suffix) result.push(suffix);
14909
- }
14910
- return result;
14911
- }
14912
- function tokenize2(input, profile) {
14913
- const tokens = [];
14914
- let current = "";
14915
- let inSelector = false;
14916
- let selectorDepth = 0;
14917
- let bracketDepth = 0;
14918
- let parenDepth = 0;
14919
- for (let i = 0; i < input.length; i++) {
14920
- const char = input[i];
14921
- if (char === "<") {
14922
- inSelector = true;
14923
- selectorDepth++;
14924
- } else if (char === ">" && inSelector) {
14925
- selectorDepth--;
14926
- if (selectorDepth === 0) inSelector = false;
14927
- }
14928
- if (char === "[") {
14929
- bracketDepth++;
14930
- } else if (char === "]" && bracketDepth > 0) {
14931
- bracketDepth--;
14932
- }
14933
- if (char === "(") {
14934
- parenDepth++;
14935
- } else if (char === ")" && parenDepth > 0) {
14936
- parenDepth--;
14937
- }
14938
- if (/\s/.test(char) && !inSelector && bracketDepth === 0 && parenDepth === 0) {
14939
- if (current) {
14940
- tokens.push(current);
14941
- current = "";
14942
- }
14943
- } else {
14944
- current += char;
14945
- }
14946
- }
14947
- if (current) {
14948
- tokens.push(current);
14949
- }
14950
- return splitAttachedAffixes(tokens, profile.code);
14951
- }
14952
- function identifyStatementType(tokens, profile) {
14953
- if (tokens.length === 0) return "unknown";
14954
- const firstToken = tokens[0].toLowerCase();
14955
- const eventMarker = profile.markers.find((m) => m.role === "event" && m.position === "preposition");
14956
- if (eventMarker && firstToken === eventMarker.form.toLowerCase()) {
14957
- return "event-handler";
14958
- }
14959
- if (EVENT_KEYWORDS.has(firstToken)) {
14960
- return "event-handler";
14961
- }
14962
- if (CONDITIONAL_KEYWORDS.has(firstToken)) {
14963
- return "conditional";
14964
- }
14965
- return "command";
14966
- }
14967
- function parseEventHandler(tokens, profile) {
14968
- const roles = /* @__PURE__ */ new Map();
14969
- let startIndex = EVENT_KEYWORDS.has(tokens[0]?.toLowerCase()) ? 1 : 0;
14970
- if (tokens[startIndex]) {
14971
- const eventTokens = [tokens[startIndex]];
14972
- startIndex++;
14973
- while (tokens[startIndex] && EVENT_CONJUNCTIONS.has(tokens[startIndex].toLowerCase()) && tokens[startIndex + 1]) {
14974
- eventTokens.push(tokens[startIndex], tokens[startIndex + 1]);
14975
- startIndex += 2;
14976
- }
14977
- roles.set("event", {
14978
- role: "event",
14979
- value: eventTokens.join(" ")
14980
- });
14981
- }
14982
- if (tokens[startIndex] && tokens[startIndex].toLowerCase() === "from" && tokens[startIndex + 1]) {
14983
- startIndex++;
14984
- const sourceValue = [];
14985
- while (tokens[startIndex]) {
14986
- if (ENGLISH_COMMANDS.has(tokens[startIndex].toLowerCase())) break;
14987
- sourceValue.push(tokens[startIndex]);
14988
- startIndex++;
14989
- }
14990
- if (sourceValue.length > 0) {
14991
- const value = sourceValue.join(" ");
14992
- roles.set("source", {
14993
- role: "source",
14994
- value,
14995
- isSelector: /^[#.<@]/.test(value)
14996
- });
14997
- }
14998
- }
14999
- if (tokens[startIndex]) {
15000
- roles.set("action", {
15001
- role: "action",
15002
- value: tokens[startIndex]
15003
- });
15004
- startIndex++;
15005
- }
15006
- if (tokens[startIndex]) {
15007
- const modifierMap = buildArgumentModifierMap(profile, roles.get("action")?.value);
15008
- let currentRole = "patient";
15009
- let currentValue = [];
15010
- for (let i = startIndex; i < tokens.length; i++) {
15011
- const token = tokens[i];
15012
- const mappedRole = modifierMap[token.toLowerCase()];
15013
- if (mappedRole) {
15014
- if (currentValue.length > 0) {
15015
- const value = currentValue.join(" ");
15016
- roles.set(currentRole, {
15017
- role: currentRole,
15018
- value,
15019
- isSelector: /^[#.<@]/.test(value)
15020
- });
15021
- }
15022
- currentRole = mappedRole;
15023
- currentValue = [];
15024
- } else {
15025
- currentValue.push(token);
15026
- }
15027
- }
15028
- if (currentValue.length > 0) {
15029
- const value = currentValue.join(" ");
15030
- roles.set(currentRole, {
15031
- role: currentRole,
15032
- value,
15033
- isSelector: /^[#.<@]/.test(value)
15034
- });
15035
- }
15036
- }
15037
- return {
15038
- type: "event-handler",
15039
- roles,
15040
- original: tokens.join(" ")
15041
- };
15042
- }
15043
- function parseCommand(tokens, profile) {
15044
- const roles = /* @__PURE__ */ new Map();
15045
- if (tokens.length === 0) {
15046
- return { type: "command", roles, original: "" };
15047
- }
15048
- roles.set("action", {
15049
- role: "action",
15050
- value: tokens[0]
15051
- });
15052
- const modifierMap = buildArgumentModifierMap(profile, tokens[0]);
15053
- let currentRole = "patient";
15054
- let currentValue = [];
15055
- for (let i = 1; i < tokens.length; i++) {
15056
- const token = tokens[i];
15057
- const mappedRole = modifierMap[token.toLowerCase()];
15058
- if (mappedRole) {
15059
- if (currentValue.length > 0) {
15060
- const value = currentValue.join(" ");
15061
- roles.set(currentRole, {
15062
- role: currentRole,
15063
- value,
15064
- isSelector: /^[#.<@]/.test(value)
15065
- });
15066
- }
15067
- currentRole = mappedRole;
15068
- currentValue = [];
15069
- } else {
15070
- currentValue.push(token);
15071
- }
15072
- }
15073
- if (currentValue.length > 0) {
15074
- const value = currentValue.join(" ");
15075
- roles.set(currentRole, {
15076
- role: currentRole,
15077
- value,
15078
- isSelector: /^[#.<@]/.test(value)
15079
- });
15080
- }
15081
- return {
15082
- type: "command",
15083
- roles,
15084
- original: tokens.join(" ")
15085
- };
15086
- }
15087
- var LITERAL_PRIMARY_ROLES = /* @__PURE__ */ new Set([
15088
- "duration",
15089
- "quantity"
15090
- ]);
15091
- function applyPrimaryRole(parsed, targetProfile) {
15092
- if (parsed.type !== "command") return;
15093
- const action = parsed.roles.get("action")?.value;
15094
- if (!action) return;
15095
- const primaryRole = COMMAND_PRIMARY_ROLES[action.toLowerCase()];
15096
- if (!primaryRole || !LITERAL_PRIMARY_ROLES.has(primaryRole)) return;
15097
- const patientEl = parsed.roles.get("patient");
15098
- if (!patientEl || parsed.roles.has(primaryRole)) return;
15099
- if (targetProfile.markers.some((m) => m.role === primaryRole)) return;
15100
- parsed.roles.delete("patient");
15101
- parsed.roles.set(primaryRole, { ...patientEl, role: primaryRole });
15102
- }
15103
- function parseConditional(tokens, _profile) {
15104
- const roles = /* @__PURE__ */ new Map();
15105
- roles.set("action", {
15106
- role: "action",
15107
- value: tokens[0]
15108
- });
15109
- const thenIndex = tokens.findIndex((t) => THEN_KEYWORDS.has(t.toLowerCase()));
15110
- if (thenIndex > 1) {
15111
- const conditionValue = tokens.slice(1, thenIndex).join(" ");
15112
- roles.set("condition", {
15113
- role: "condition",
15114
- value: conditionValue
15115
- });
15116
- } else if (thenIndex === -1 && tokens.length > 1) {
15117
- roles.set("condition", {
15118
- role: "condition",
15119
- value: tokens.slice(1).join(" ")
15120
- });
15121
- }
15122
- return {
15123
- type: "conditional",
15124
- roles,
15125
- original: tokens.join(" ")
15126
- };
15127
- }
15128
- function translateWord(word, sourceLocale, targetLocale) {
15129
- if (/^[#.<@]/.test(word)) {
15130
- return word;
15131
- }
15132
- if (/^\d+/.test(word)) {
15133
- return word;
15134
- }
15135
- if (/\s/.test(word) && word.startsWith("(")) {
15136
- return word.split(/\s+/).map((w) => translateWord(w, sourceLocale, targetLocale)).join(" ");
15137
- }
15138
- if (word.length > 1 && (word.startsWith("(") || word.endsWith(")"))) {
15139
- const m = word.match(/^(\(*)([^()]+)(\)*)$/);
15140
- if (m && (m[1] || m[3])) {
15141
- return m[1] + translateWord(m[2], sourceLocale, targetLocale) + m[3];
15142
- }
15143
- }
15144
- const sourceDict = sourceLocale === "en" ? null : dictionaries[sourceLocale];
15145
- const targetDict = dictionaries[targetLocale];
15146
- if (!targetDict) return word;
15147
- let englishWord = word;
15148
- if (sourceDict) {
15149
- const found = findInDictionary(sourceDict, word);
15150
- if (found) {
15151
- englishWord = found.englishKey;
15152
- }
15153
- }
15154
- const translated = translateFromEnglish(targetDict, englishWord);
15155
- return translated ?? word;
15156
- }
15157
- var POSSESSIVE_MARKERS = {
15158
- en: { type: "suffix", marker: "'s" },
15159
- es: { type: "preposition", marker: "de" },
15160
- pt: { type: "preposition", marker: "de" },
15161
- fr: { type: "preposition", marker: "de" },
15162
- de: { type: "preposition", marker: "von" },
15163
- ja: { type: "suffix", marker: "\u306E" },
15164
- ko: { type: "suffix", marker: "\uC758" },
15165
- zh: { type: "suffix", marker: "\u7684" },
15166
- ar: { type: "preposition", marker: "\u0644\u0640" },
15167
- // Spaced genitive particle (not the glued `'ın`), so the tokenizer can split
15168
- // it off the selector — consistent with Turkish's other spaced case markers.
15169
- tr: { type: "particle", marker: "\u0131n" },
15170
- id: { type: "preposition", marker: "dari" },
15171
- // Latin-script genitive: must be a *spaced* particle (`#picker pa`), since a
15172
- // glued `#pickerpa` can't be split from the selector by the tokenizer the
15173
- // way a non-Latin suffix (の/의/র) can.
15174
- qu: { type: "particle", marker: "pa" },
15175
- // Bengali SOV postposition genitive, like ja/ko — a spaced suffix the
15176
- // tokenizer splits off as a particle. Previously absent, so it fell back to
15177
- // the English `'s` marker and its possessive property paths never parsed.
15178
- // (Hindi `का` is intentionally omitted: its `bind` lacks a verb-final
15179
- // grammar rule, so fixing its possessive alone yields a wrong `on` parse —
15180
- // tracked as separate follow-up.)
15181
- bn: { type: "suffix", marker: "\u09B0" },
15182
- sw: { type: "preposition", marker: "ya" }
15183
- };
15184
- function translatePossessive(token, sourceLocale, targetLocale) {
15185
- const possessiveMatch = token.match(/^(.+)'s$/i);
15186
- if (!possessiveMatch) {
15187
- return token;
15188
- }
15189
- const owner = possessiveMatch[1];
15190
- const targetMarker = POSSESSIVE_MARKERS[targetLocale] || POSSESSIVE_MARKERS.en;
15191
- const pronounPossessives = {
15192
- me: "my",
15193
- it: "its",
15194
- you: "your"
15195
- };
15196
- const lowerOwner = owner.toLowerCase();
15197
- if (pronounPossessives[lowerOwner]) {
15198
- const possessiveForm = pronounPossessives[lowerOwner];
15199
- return translateWord(possessiveForm, "en", targetLocale);
15200
- }
15201
- const translatedOwner = translateWord(owner, sourceLocale, targetLocale);
15202
- switch (targetMarker.type) {
15203
- case "suffix":
15204
- return `${translatedOwner}${targetMarker.marker}`;
15205
- case "particle":
15206
- return `${translatedOwner} ${targetMarker.marker}`;
15207
- case "preposition":
15208
- return `__POSS__${targetMarker.marker}__${translatedOwner}__POSS__`;
15209
- default:
15210
- return `${translatedOwner}'s`;
15211
- }
15212
- }
15213
- var POSSESSIVE_DOT_REGEX = /^(my|its|your|me|it|you)(\??\..+)$/i;
15214
- var POSSESSIVE_DOT_PRONOUNS = {
15215
- me: "my",
15216
- it: "its",
15217
- you: "your",
15218
- my: "my",
15219
- its: "its",
15220
- your: "your"
15221
- };
15222
- function translatePossessiveDotNotation(value, sourceLocale, targetLocale) {
15223
- const match = value.match(POSSESSIVE_DOT_REGEX);
15224
- if (!match) return null;
15225
- const possessiveWord = match[1].toLowerCase();
15226
- const propertySuffix = match[2];
15227
- const possessiveKey = POSSESSIVE_DOT_PRONOUNS[possessiveWord] || possessiveWord;
15228
- const translated = translateWord(possessiveKey, sourceLocale, targetLocale);
15229
- if (translated.includes(" ")) return null;
15230
- if (translated !== possessiveKey) {
15231
- return translated + propertySuffix;
15232
- }
15233
- if (possessiveWord !== possessiveKey) {
15234
- const alt = translateWord(possessiveWord, sourceLocale, targetLocale);
15235
- if (alt !== possessiveWord && !alt.includes(" ")) {
15236
- return alt + propertySuffix;
15237
- }
15238
- }
15239
- return null;
15240
- }
15241
- function translateMultiWordValue(value, sourceLocale, targetLocale) {
15242
- if (value.includes("[")) {
15243
- const guards = [];
15244
- const masked = value.replace(/\[[^\]]*\]/g, (match) => {
15245
- guards.push(match);
15246
- return `\uE000${guards.length - 1}\uE001`;
15247
- });
15248
- if (guards.length > 0) {
15249
- const translated2 = translateMultiWordValue(masked, sourceLocale, targetLocale);
15250
- return translated2.replace(/(\d+)/g, (_, n) => guards[Number(n)]);
15251
- }
15252
- }
15253
- if (!value.includes(" ")) {
15254
- if (value.includes("'s")) {
15255
- return translatePossessive(value, sourceLocale, targetLocale);
15256
- }
15257
- const dotResult = translatePossessiveDotNotation(value, sourceLocale, targetLocale);
15258
- if (dotResult !== null) return dotResult;
15259
- return translateWord(value, sourceLocale, targetLocale);
15260
- }
15261
- const words = value.split(/\s+/);
15262
- const translated = [];
15263
- let i = 0;
15264
- while (i < words.length) {
15265
- const word = words[i];
15266
- if (word.includes("'s")) {
15267
- const possessiveResult = translatePossessive(word, sourceLocale, targetLocale);
15268
- const prepMatch = possessiveResult.match(/^__POSS__(.+)__(.+)__POSS__$/);
15269
- if (prepMatch && i + 1 < words.length) {
15270
- const marker = prepMatch[1];
15271
- const owner = prepMatch[2];
15272
- const property = words[i + 1];
15273
- const translatedProperty = translateWord(property, sourceLocale, targetLocale);
15274
- translated.push(`${translatedProperty} ${marker} ${owner}`);
15275
- i += 2;
15276
- continue;
15277
- } else if (prepMatch) {
15278
- const marker = prepMatch[1];
15279
- const owner = prepMatch[2];
15280
- translated.push(`${marker} ${owner}`);
15281
- i++;
15282
- continue;
15283
- }
15284
- translated.push(possessiveResult);
15285
- i++;
15286
- continue;
15287
- }
15288
- if (/^[#.<@]/.test(word) || /^\d+/.test(word)) {
15289
- translated.push(word);
15290
- i++;
15291
- continue;
15292
- }
15293
- if (/^["'].*["']$/.test(word)) {
15294
- translated.push(word);
15295
- i++;
15296
- continue;
15297
- }
15298
- const dotResult = translatePossessiveDotNotation(word, sourceLocale, targetLocale);
15299
- if (dotResult !== null) {
15300
- translated.push(dotResult);
15301
- i++;
15302
- continue;
15303
- }
15304
- translated.push(translateWord(word, sourceLocale, targetLocale));
15305
- i++;
15306
- }
15307
- return translated.join(" ");
15308
- }
15309
- function translateElements(parsed, sourceLocale, targetLocale) {
15310
- for (const [_role, element] of parsed.roles) {
15311
- if (element.value.includes("'s")) {
15312
- element.translated = translateMultiWordValue(element.value, sourceLocale, targetLocale);
15313
- } else if (!element.isSelector && !element.isLiteral) {
15314
- element.translated = translateMultiWordValue(element.value, sourceLocale, targetLocale);
15315
- } else {
15316
- element.translated = element.value;
15317
- }
15318
- }
15319
- }
15320
- var CARET_SCOPE_OPEN = "\uE000";
15321
- var CARET_SCOPE_CLOSE = "\uE001";
15322
- var CARET_SCOPE_RE = /(\^[A-Za-z_][\w-]*)(\s+on\s+(?:[#.][\w-]+|<[^>]*\/>|\[[^\]]+\]))/g;
15323
- function maskCaretScopes(input) {
15324
- const scopes = [];
15325
- const masked = input.replace(CARET_SCOPE_RE, (_m, varTok, scope) => {
15326
- const idx = scopes.length;
15327
- scopes.push(scope);
15328
- return `${varTok}${CARET_SCOPE_OPEN}${idx}${CARET_SCOPE_CLOSE}`;
15329
- });
15330
- return scopes.length > 0 ? { masked, scopes } : null;
15331
- }
15332
- function restoreCaretScopes(input, scopes) {
15333
- return input.replace(
15334
- new RegExp(`${CARET_SCOPE_OPEN}(\\d+)${CARET_SCOPE_CLOSE}`, "g"),
15335
- (_m, n) => scopes[Number(n)] ?? ""
15336
- );
15337
- }
15338
- var VIEW_TAIL_OPEN = "\uE002";
15339
- var VIEW_TAIL_CLOSE = "\uE003";
15340
- var VIEW_TAIL_RE = /\busing\s+view\b(?:\s+(?!then\b)[A-Za-z][\w-]*)?/gi;
15341
- var VIEW_TAIL_TOKEN_RE = new RegExp(`^${VIEW_TAIL_OPEN}(\\d+)${VIEW_TAIL_CLOSE}$`);
15342
- function maskViewTransitionTails(input) {
15343
- const tails = [];
15344
- const masked = input.replace(VIEW_TAIL_RE, (match) => {
15345
- const idx = tails.length;
15346
- tails.push(match);
15347
- return `${VIEW_TAIL_OPEN}${idx}${VIEW_TAIL_CLOSE}`;
15348
- });
15349
- return tails.length > 0 ? { masked, tails } : null;
15350
- }
15351
- function restoreViewTransitionTails(input, tails) {
15352
- return input.replace(
15353
- new RegExp(`${VIEW_TAIL_OPEN}(\\d+)${VIEW_TAIL_CLOSE}`, "g"),
15354
- (_m, n) => tails[Number(n)] ?? ""
15355
- );
15356
- }
15357
- var GrammarTransformer = class {
15358
- constructor(sourceLocale = "en", targetLocale) {
15359
- const source = getProfile(sourceLocale);
15360
- const target = getProfile(targetLocale);
15361
- if (!source) throw new Error(`Unknown source locale: ${sourceLocale}`);
15362
- if (!target) throw new Error(`Unknown target locale: ${targetLocale}`);
15363
- this.sourceProfile = source;
15364
- this.targetProfile = target;
15365
- }
15366
- /**
15367
- * Transform a hyperscript statement from source to target language.
15368
- * Handles compound statements with "then" by splitting, transforming each part,
15369
- * and rejoining with the target language's "then" keyword.
15370
- *
15371
- * For multi-line input, preserves line structure (indentation, blank lines).
15372
- */
15373
- transform(input) {
15374
- const out = this.transformInternal(input);
15375
- return this.targetProfile.code === "he" ? repairHebrewFrontedAccusative(out) : out;
15376
- }
15377
- transformInternal(input) {
15378
- const viewTails = maskViewTransitionTails(input);
15379
- if (viewTails) {
15380
- return restoreViewTransitionTails(this.transformInternal(viewTails.masked), viewTails.tails);
15381
- }
15382
- const caret = maskCaretScopes(input);
15383
- if (caret) {
15384
- return restoreCaretScopes(this.transform(caret.masked), caret.scopes);
15385
- }
15386
- const targetThen = getTargetThenKeyword(this.targetProfile.code);
15387
- if (!input.includes("\n")) {
15388
- const jsBlock = this.tryTransformJsBlock(input);
15389
- if (jsBlock !== null) return jsBlock;
15390
- const eventModifier = this.tryTransformEventWithModifierBody(input);
15391
- if (eventModifier !== null) return eventModifier;
15392
- const eventBlock = this.tryTransformEventWithBlockBody(input);
15393
- if (eventBlock !== null) return eventBlock;
15394
- const eventGuard = this.tryTransformEventWithUnlessGuard(input);
15395
- if (eventGuard !== null) return eventGuard;
15396
- }
15397
- const hasMultiLineStructure = input.includes("\n");
15398
- if (hasMultiLineStructure) {
15399
- const { parts: parts2, lineMetadata, partToLineIndex } = splitCompoundStatementWithMetadata(
15400
- input,
15401
- this.sourceProfile.code
15402
- );
15403
- const transformedParts = parts2.map((part) => this.transformSingle(part));
15404
- return reconstructWithLineStructure(
15405
- transformedParts,
15406
- lineMetadata,
15407
- partToLineIndex,
15408
- targetThen
15409
- );
15410
- }
15411
- const parts = splitCompoundStatement(input, this.sourceProfile.code);
15412
- if (parts.length > 1) {
15413
- const transformedParts = parts.map((part) => this.transformSingle(part));
15414
- return transformedParts.join(` ${targetThen} `);
15415
- }
15416
- return this.transformSingle(input);
15417
- }
15418
- /**
15419
- * Transform a single hyperscript statement (no compound "then" chains).
15420
- */
15421
- transformSingle(input) {
15422
- const block = extractBlockStructure(input, this.sourceProfile.code);
15423
- if (block) {
15424
- return this.transformBlock(block);
15425
- }
15426
- const strippedEnd = this.transformWithTrailingEnd(input);
15427
- if (strippedEnd !== null) {
15428
- return strippedEnd;
15429
- }
15430
- const setScope = this.transformSetWithScope(input);
15431
- if (setScope !== null) {
15432
- return setScope;
15433
- }
15434
- const viewTail = this.transformWithViewTransitionTail(input);
15435
- if (viewTail !== null) {
15436
- return viewTail;
15437
- }
15438
- const parsed = parseStatement(input, this.sourceProfile.code);
15439
- if (!parsed) {
15440
- return input;
15441
- }
15442
- applyPrimaryRole(parsed, this.targetProfile);
15443
- translateElements(parsed, this.sourceProfile.code, this.targetProfile.code);
15444
- const rule = this.findRule(parsed);
15445
- if (rule?.transform.custom) {
15446
- return rule.transform.custom(parsed, this.targetProfile);
15447
- }
15448
- const roleOrder = rule?.transform.roleOrder || this.targetProfile.canonicalOrder;
15449
- const reordered = reorderRoles(parsed.roles, roleOrder);
15450
- const shouldInsertMarkers = rule?.transform.insertMarkers ?? true;
15451
- if (shouldInsertMarkers) {
15452
- const result = insertMarkers(
15453
- reordered,
15454
- this.targetProfile.markers,
15455
- this.targetProfile.adpositionType
15456
- );
15457
- return joinTokens(result);
15458
- }
15459
- return joinTokens(reordered.map((e) => e.translated || e.value));
15460
- }
15461
- /**
15462
- * Clause carrying a masked `using view transition` tail: strip the opaque
15463
- * token, transform the clause alone, and re-append the token at the very end.
15464
- *
15465
- * The tail is a clause-final modifier in every word order the corpus emits:
15466
- * the semantic side matches it as the literal `using view` marker plus a value
15467
- * word, and the SOV/VSO event-handler patterns admit it as an optional
15468
- * TRAILING group (after the with-marked operand). So the target position is
15469
- * "end of the transformed clause" for all 24 languages — no per-profile
15470
- * placement decision, which is what makes this a passthrough rather than a
15471
- * role.
15472
- *
15473
- * Returns null when the clause carries no masked tail, or when the token is
15474
- * not clause-final (nothing to reposition — leaving it in place still restores
15475
- * verbatim English).
15476
- */
15477
- transformWithViewTransitionTail(input) {
15478
- const trimmed = input.trim();
15479
- const tokens = trimmed.split(/\s+/);
15480
- if (tokens.length < 2) {
15481
- return null;
15482
- }
15483
- if (!VIEW_TAIL_TOKEN_RE.test(tokens[tokens.length - 1])) {
15484
- return null;
15485
- }
15486
- const tail = tokens[tokens.length - 1];
15487
- const head = tokens.slice(0, -1).join(" ");
15488
- return `${this.transformSingle(head)} ${tail}`;
15489
- }
15490
- /**
15491
- * `<command …> end` fragments: transform the command without its stranded
15492
- * terminator, then re-append the translated terminator as a standalone
15493
- * trailing token. Fragments that open a block of their own (`if … end`,
15494
- * `repeat … end`, `js … end`) bail — their terminator belongs to them and
15495
- * their dedicated paths handle it.
15496
- */
15497
- transformWithTrailingEnd(input) {
15498
- const src = this.sourceProfile.code;
15499
- const tokens = input.trim().split(/\s+/);
15500
- if (tokens.length < 2) {
15501
- return null;
15502
- }
15503
- const sourceEnd = translateWord("end", "en", src).toLowerCase();
15504
- if (tokens[tokens.length - 1].toLowerCase() !== sourceEnd) {
15505
- return null;
15506
- }
15507
- const openers = new Set(
15508
- ["if", "repeat", "unless", "while", "when", "live", "js"].map(
15509
- (k) => translateWord(k, "en", src).toLowerCase()
15510
- )
15511
- );
15512
- if (tokens.slice(0, -1).some((t) => openers.has(t.toLowerCase()))) {
15513
- return null;
15514
- }
15515
- const inner = this.transformSingle(tokens.slice(0, -1).join(" "));
15516
- const endT = translateWord(tokens[tokens.length - 1], src, this.targetProfile.code);
15517
- return `${inner} ${endT}`;
15518
- }
15519
- /**
15520
- * Detect and transform an inline JS block (`[on <event>] js <raw js> end`).
15521
- *
15522
- * The `js ... end` body is raw JavaScript: it must not be tokenized,
15523
- * translated, or word-order reordered. We mask the whole block with a single
15524
- * opaque placeholder, run the surrounding statement (the event-handler head,
15525
- * if any) through the normal reorder pipeline so the placeholder lands in the
15526
- * correct action position, then substitute the translated `js`/`end` keywords
15527
- * around the verbatim body.
15528
- *
15529
- * Returns `null` (fall through to the normal path) when there is no js block,
15530
- * no matching `end`, or trailing content after `end` (kept tight on purpose).
15531
- */
15532
- tryTransformJsBlock(input) {
15533
- const src = this.sourceProfile.code;
15534
- const dst = this.targetProfile.code;
15535
- const sourceJs = translateWord("js", "en", src);
15536
- const sourceEnd = translateWord("end", "en", src).toLowerCase();
15537
- const tokens = input.split(/\s+/).filter((t) => t.length > 0);
15538
- const escapedJs = sourceJs.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
15539
- const jsRe = new RegExp(`^${escapedJs}(\\(.*\\))?$`, "i");
15540
- const jsIdx = tokens.findIndex((t) => jsRe.test(t));
15541
- if (jsIdx === -1) return null;
15542
- let endIdx = -1;
15543
- for (let i = jsIdx + 1; i < tokens.length; i++) {
15544
- if (tokens[i].toLowerCase() === sourceEnd) {
15545
- endIdx = i;
15546
- break;
15547
- }
15548
- }
15549
- if (endIdx === -1) return null;
15550
- if (endIdx !== tokens.length - 1) return null;
15551
- const jsToken = tokens[jsIdx];
15552
- const jsParen = jsToken.match(jsRe)?.[1] ?? "";
15553
- const jsKeywordRaw = jsParen ? jsToken.slice(0, jsToken.length - jsParen.length) : jsToken;
15554
- const body = tokens.slice(jsIdx + 1, endIdx).join(" ");
15555
- const targetJs = translateWord(jsKeywordRaw, src, dst) + jsParen;
15556
- const targetEnd = translateWord(tokens[endIdx], src, dst);
15557
- const replacement = [targetJs, body, targetEnd].filter((s) => s.length > 0).join(" ");
15558
- const before = tokens.slice(0, jsIdx);
15559
- if (before.length === 0) return replacement;
15560
- const placeholder = "JSBLOCKPLACEHOLDER";
15561
- const reordered = this.transformSingle([...before, placeholder].join(" "));
15562
- if (!reordered.includes(placeholder)) return null;
15563
- return reordered.replace(placeholder, replacement);
15564
- }
15565
- /**
15566
- * Transform an event handler whose body is a block command
15567
- * (`on <event> [from <src>] {if|repeat|unless|while|for} … end`).
15568
- *
15569
- * `parseEventHandler` would treat the block keyword as the action and sweep the
15570
- * condition/body into role values, then reorder them — shredding the block
15571
- * (`if event.shiftKey call submitAndContinue() end` → scattered tokens). Instead
15572
- * we mask the whole block as an opaque action placeholder, reorder the event
15573
- * head normally, transform the block as a self-contained unit, and restitch.
15574
- *
15575
- * Returns `null` (fall through) when the input isn't an event handler, has no
15576
- * block-keyword body, or has no closing `end`.
15577
- */
15578
- tryTransformEventWithBlockBody(input) {
15579
- const tokens = tokenize2(input, this.sourceProfile);
15580
- if (tokens.length === 0) return null;
15581
- if (!EVENT_KEYWORDS.has(tokens[0]?.toLowerCase())) return null;
15582
- let blockIdx = -1;
15583
- for (let i = 1; i < tokens.length; i++) {
15584
- if (BLOCK_BODY_KEYWORDS.has(tokens[i].toLowerCase())) {
15585
- blockIdx = i;
15586
- break;
15587
- }
15588
- }
15589
- if (blockIdx <= 0) return null;
15590
- if (tokens[tokens.length - 1].toLowerCase() !== "end") return null;
15591
- const eventHead = tokens.slice(0, blockIdx);
15592
- const blockTokens = tokens.slice(blockIdx);
15593
- if (this.targetProfile.wordOrder === "VSO" && eventHead.some((t) => t.toLowerCase() === "from")) {
15594
- return null;
15595
- }
15596
- const placeholder = "EVENTBLOCKPLACEHOLDER";
15597
- const headOut = this.transformSingle([...eventHead, placeholder].join(" "));
15598
- if (!headOut.includes(placeholder)) return null;
15599
- const blockOut = this.transformBlockBody(blockTokens);
15600
- const eventClause = headOut.replace(placeholder, "").replace(/\s+/g, " ").trim();
15601
- return [eventClause, blockOut].filter((s) => s.length > 0).join(" ");
15602
- }
15603
- /**
15604
- * Transform an event handler whose body is an inline `unless` guard with NO
15605
- * `end` (`on <event> unless <cond> <body>` — the `unless-condition` shape).
15606
- *
15607
- * Object-marking SVO targets (he, zh). `parseEventHandler` reads `unless` as the
15608
- * action and sweeps the whole `<cond> <body>` tail into a single `patient` blob;
15609
- * the target then prefixes that blob with its object marker — Hebrew's accusative
15610
- * את (`… אלא את I match .disabled מתג .selected`) or Chinese's BA particle 把
15611
- * (`… 除非 把 I match .disabled 切换 .selected`) — and the inner toggle loses its
15612
- * own marker. The semantic parser can't recover the guard from that: the marker
15613
- * ahead of the condition blocks the `unless` pattern AND the now-markerless body
15614
- * command fails its object-marked toggle pattern, so the body collapses (`unless`
15615
- * dropped). Marker-less languages (de/it/ar/pl) tolerate the same role-blob and
15616
- * stay faithful, so this is an object-marker artifact, not a general parse gap.
15617
- *
15618
- * The standalone `unless <cond> <body>` path already produces the correct shape
15619
- * (`extractBlockStructure` → `transformBlock`: condition kept marker-free, body
15620
- * command keeps its marker — he `אלא I match .disabled מתג את .selected`, zh
15621
- * `除非 I match .disabled 切换 把 .selected`). So we split the event head off,
15622
- * transform the guard through that path, and emit the event clause first (he and
15623
- * zh are both SVO — event leads). Returns `null` (fall through) when the input
15624
- * isn't an object-marking event handler with an un-terminated inline `unless`
15625
- * guard.
15626
- */
15627
- tryTransformEventWithUnlessGuard(input) {
15628
- if (!UNLESS_GUARD_OBJECT_MARKING_LOCALES.has(this.targetProfile.code)) return null;
15629
- const tokens = tokenize2(input, this.sourceProfile);
15630
- if (tokens.length === 0) return null;
15631
- if (!EVENT_KEYWORDS.has(tokens[0]?.toLowerCase())) return null;
15632
- let guardIdx = -1;
15633
- for (let i = 1; i < tokens.length; i++) {
15634
- if (tokens[i].toLowerCase() === "unless") {
15635
- guardIdx = i;
15636
- break;
15637
- }
15638
- }
15639
- if (guardIdx <= 0) return null;
15640
- if (tokens[tokens.length - 1].toLowerCase() === "end") return null;
15641
- const eventHead = tokens.slice(0, guardIdx);
15642
- const guard = tokens.slice(guardIdx).join(" ");
15643
- const guardOut = this.transform(guard);
15644
- if (!guardOut) return null;
15645
- const placeholder = "EVENTGUARDPLACEHOLDER";
15646
- const headOut = this.transformSingle([...eventHead, placeholder].join(" "));
15647
- if (!headOut.includes(placeholder)) return null;
15648
- const eventClause = headOut.replace(placeholder, "").replace(/\s+/g, " ").trim();
15649
- return [eventClause, guardOut].filter((s) => s.length > 0).join(" ");
15650
- }
15651
- /**
15652
- * Transform an event handler whose body leads with a command-modifier
15653
- * (`on <event> [from <src>] {async|once|debounced [at N]|throttled [at N]} <body>`).
15654
- *
15655
- * `parseEventHandler` reads the first token after the event as the **action**, so
15656
- * a leading modifier is mistaken for the verb and the real verb (`fetch`/`add`) is
15657
- * swept into the patient. For SOV targets the reorder then surfaces that verb
15658
- * **first** (`取得 /api/data を クリック …`), and the semantic parser matches the
15659
- * leading `<verb> <patient>` with the low-priority `*-generated-verb-first`
15660
- * command pattern — returning a bare command and discarding the event + the rest
15661
- * of the body (degenerate parse).
15662
- *
15663
- * Instead, lift the modifier out, transform the modifier-free handler through the
15664
- * normal path (which keeps the body in canonical patient-first SOV order so the
15665
- * event sits mid-stream and the existing SOV event-extraction recovers it), then
15666
- * re-emit the modifier as a **leading English literal**. The semantic parser
15667
- * strips a leading `once`/`debounced`/`throttled` (`extractStandaloneModifiers`)
15668
- * and an `async` anywhere (`stripAsyncModifier`) before parsing, so the modifier
15669
- * is consumed as handler metadata rather than shadowing the body.
15670
- *
15671
- * Returns `null` (fall through) when the input isn't an event handler or the body
15672
- * doesn't lead with a modifier — leaving simple/Mode-B handlers byte-identical.
15673
- */
15674
- tryTransformEventWithModifierBody(input) {
15675
- if (this.targetProfile.wordOrder !== "SOV") return null;
15676
- const tokens = tokenize2(input, this.sourceProfile);
15677
- if (tokens.length === 0) return null;
15678
- if (!EVENT_KEYWORDS.has(tokens[0]?.toLowerCase())) return null;
15679
- let i = 1;
15680
- if (!tokens[i]) return null;
15681
- i++;
15682
- while (tokens[i] && EVENT_CONJUNCTIONS.has(tokens[i].toLowerCase()) && tokens[i + 1]) {
15683
- i += 2;
15684
- }
15685
- if (tokens[i]?.toLowerCase() === "from" && tokens[i + 1]) {
15686
- i++;
15687
- while (tokens[i] && !ENGLISH_COMMANDS.has(tokens[i].toLowerCase()) && !BODY_MODIFIER_KEYWORDS.has(tokens[i].toLowerCase())) {
15688
- i++;
15689
- }
15690
- }
15691
- const modWord = tokens[i]?.toLowerCase();
15692
- if (!modWord || !BODY_MODIFIER_KEYWORDS.has(modWord)) return null;
15693
- const modStart = i;
15694
- let modEnd = i + 1;
15695
- if (modWord !== "async" && modWord !== "once") {
15696
- if (tokens[modEnd]?.toLowerCase() === "at") modEnd++;
15697
- if (tokens[modEnd] && /^\d+(ms|s|m)?$/.test(tokens[modEnd])) modEnd++;
15698
- }
15699
- const modifierPhrase = tokens.slice(modStart, modEnd).join(" ");
15700
- const rebuilt = [...tokens.slice(0, modStart), ...tokens.slice(modEnd)].join(" ");
15701
- if (tokens.length - (modEnd - modStart) <= 2) return null;
15702
- const bodyOut = this.transform(rebuilt);
15703
- return [modifierPhrase, bodyOut].filter((s) => s.length > 0).join(" ");
15704
- }
15705
- /**
15706
- * Transform a `set <stuff> on <scope>` clause (S1 tabs-aria). The trailing
15707
- * `on <scope>` is the element(s) the attribute is set on — kept attached by
15708
- * splitOnCommandBoundaries. The semantic parser captures it as a `scope` role
15709
- * via the passthrough literal `on` (setSchema's scope markerOverride is `on`
15710
- * in every language), so `on` is emitted verbatim and only the scope *value*
15711
- * is translated (selectors pass through; `me`/`it`/`you` translate to the
15712
- * native reference, which the parser also accepts).
15713
- *
15714
- * Positioning matches where the set patterns expect the scope: at the clause
15715
- * end for verb-first orders (SVO/VSO), and immediately before the clause-final
15716
- * verb for SOV (the generated SOV pattern is `{dest} {patient} on {scope}
15717
- * {verb}`). Returns null (fall through) when there is no trailing `on <scope>`.
15718
- *
15719
- * Source is English in the sync-translations pipeline, so the `set` verb and
15720
- * `on` marker are matched as English literals.
15721
- */
15722
- transformSetWithScope(input) {
15723
- const src = this.sourceProfile.code;
15724
- const dst = this.targetProfile.code;
15725
- const m = input.match(/^(.*\bset\b.*\S)\s+on\s+([#.<@[]\S*|me|it|you)\s*$/i);
15726
- if (!m) return null;
15727
- const head = m[1];
15728
- const scopeRaw = m[2];
15729
- const headOut = this.transformSingle(head);
15730
- const scopeT = /^[#.<@[]/.test(scopeRaw) ? scopeRaw : translateWord(scopeRaw, src, dst);
15731
- if (this.targetProfile.wordOrder === "SOV") {
15732
- const verb = translateWord("set", "en", dst);
15733
- const firstTok = input.trim().split(/\s+/)[0]?.toLowerCase();
15734
- const isEventHandler = !!firstTok && EVENT_KEYWORDS.has(firstTok);
15735
- const toks = headOut.split(/\s+/).filter(Boolean);
15736
- if (!isEventHandler) {
15737
- const vIdx = toks.indexOf(verb);
15738
- if (vIdx >= 0) {
15739
- toks.splice(vIdx, 1);
15740
- toks.push("on", scopeT, verb);
15741
- return toks.join(" ");
15742
- }
15743
- } else if (toks.length > 0 && toks[toks.length - 1] === verb) {
15744
- toks.splice(toks.length - 1, 0, "on", scopeT);
15745
- return toks.join(" ");
15746
- }
15747
- return `${headOut} on ${scopeT}`;
15748
- }
15749
- return `${headOut} on ${scopeT}`;
15750
- }
15751
- /**
15752
- * Transform a self-contained block command (`{head} {clause?} {body} {end}`),
15753
- * where head ∈ {if, repeat, unless, …}. The clause (condition / `until event …`)
15754
- * runs up to the first command verb and is translated word-by-word; the body is
15755
- * recursively transformed (so its inner commands reorder for the target); the
15756
- * head/tail keywords are translated. The block is never word-order reordered as
15757
- * a whole — delimiters stay at the edges regardless of target word order.
15758
- */
15759
- transformBlockBody(blockTokens) {
15760
- const src = this.sourceProfile.code;
15761
- const dst = this.targetProfile.code;
15762
- const head = blockTokens[0];
15763
- const hasEnd = blockTokens[blockTokens.length - 1]?.toLowerCase() === "end";
15764
- const tail = hasEnd ? blockTokens[blockTokens.length - 1] : "";
15765
- const inner = blockTokens.slice(1, hasEnd ? -1 : void 0);
15766
- const commands = getCommandKeywordsForLocale(src);
15767
- const copulas = getCopulasForLocale(src);
15768
- let bodyStart = inner.findIndex(
15769
- (t, i) => commands.has(t.toLowerCase()) && !isPredicateAdjectivePosition(inner, i, copulas)
15770
- );
15771
- if (bodyStart < 0) bodyStart = inner.length;
15772
- const clause = inner.slice(0, bodyStart).join(" ");
15773
- const bodyTokens = inner.slice(bodyStart);
15774
- const headT = translateWord(head, src, dst);
15775
- const tailT = tail ? translateWord(tail, src, dst) : "";
15776
- const clauseT = clause ? translateMultiWordValue(clause, src, dst) : "";
15777
- const bodyT = this.transformConditionalBody(bodyTokens);
15778
- const POSITIONAL_BRANCH_HEADS = /* @__PURE__ */ new Set(["first", "last", "next", "previous", "closest"]);
15779
- const thenT = this.targetProfile.wordOrder === "SOV" && clauseT && POSITIONAL_BRANCH_HEADS.has(bodyTokens[1]?.toLowerCase()) && inner[bodyStart - 1]?.toLowerCase() !== "then" ? translateWord("then", src, dst) : "";
15780
- return [headT, clauseT, thenT, bodyT, tailT].filter((s) => s.length > 0).join(" ");
15781
- }
15782
- /**
15783
- * Transform an `if`/`unless` block body, splitting it at a top-level `else` into
15784
- * a then-branch and an else-branch so each is reordered as a self-contained unit
15785
- * and the `else` keyword itself is translated. Without this, the body is reordered
15786
- * as one stream: `else` rides along glued to the preceding clause (and, when that
15787
- * clause begins with a selector, is marked a selector and left *untranslated*),
15788
- * and a spurious `then` is inserted around it — both of which break the target
15789
- * text and the downstream parse. The split is depth-aware so an `else` belonging
15790
- * to a nested block is not mistaken for this block's separator. Bodies without an
15791
- * `else` transform exactly as before.
15792
- */
15793
- transformConditionalBody(bodyTokens) {
15794
- const src = this.sourceProfile.code;
15795
- const dst = this.targetProfile.code;
15796
- const sourceElse = translateWord("else", "en", src).toLowerCase();
15797
- let depth = 0;
15798
- let elseIdx = -1;
15799
- for (let i = 0; i < bodyTokens.length; i++) {
15800
- const t = bodyTokens[i].toLowerCase();
15801
- if (BLOCK_BODY_KEYWORDS.has(t)) depth++;
15802
- else if (t === "end" && depth > 0) depth--;
15803
- else if (t === sourceElse && depth === 0) {
15804
- elseIdx = i;
15805
- break;
15806
- }
15807
- }
15808
- if (elseIdx === -1) {
15809
- const body = bodyTokens.join(" ");
15810
- return body ? this.transform(body) : "";
15811
- }
15812
- const thenBranch = bodyTokens.slice(0, elseIdx).join(" ");
15813
- const elseBranch = bodyTokens.slice(elseIdx + 1).join(" ");
15814
- const elseT = translateWord(bodyTokens[elseIdx], src, dst);
15815
- return [
15816
- thenBranch ? this.transform(thenBranch) : "",
15817
- elseT,
15818
- elseBranch ? this.transform(elseBranch) : ""
15819
- ].filter((s) => s.length > 0).join(" ");
15820
- }
15821
- /**
15822
- * Translate a reactive block by translating the head/tail/connector
15823
- * via the dictionary, recursively transforming the body through the
15824
- * regular pipeline, and rejoining in source-language position order.
15825
- * Block-syntactic tokens are never reordered: they're delimiters, not
15826
- * arguments, and authors expect them at start/end positions
15827
- * regardless of target word order.
15828
- */
15829
- transformBlock(block) {
15830
- const src = this.sourceProfile.code;
15831
- const dst = this.targetProfile.code;
15832
- const head = translateWord(block.headKeyword, src, dst);
15833
- const tail = block.tailKeyword ? translateWord(block.tailKeyword, src, dst) : "";
15834
- const connector = block.connector ? translateWord(block.connector, src, dst) : "";
15835
- const prefix = block.prefixExpr ? translateMultiWordValue(block.prefixExpr, src, dst) : "";
15836
- const body = this.transform(block.body);
15837
- return [head, prefix, connector, body, tail].filter((s) => s.length > 0).join(" ");
15838
- }
15839
- /**
15840
- * Find the best matching rule for this statement
15841
- */
15842
- findRule(parsed) {
15843
- if (!this.targetProfile.rules) return void 0;
15844
- const matchingRules = this.targetProfile.rules.filter((rule) => this.matchesRule(parsed, rule)).sort((a, b) => b.priority - a.priority);
15845
- return matchingRules[0];
15846
- }
15847
- /**
15848
- * Check if a parsed statement matches a rule
15849
- */
15850
- matchesRule(parsed, rule) {
15851
- const { match } = rule;
15852
- for (const role of match.requiredRoles) {
15853
- if (!parsed.roles.has(role)) {
15854
- return false;
15855
- }
15856
- }
15857
- if (match.commands && match.commands.length > 0) {
15858
- const action = parsed.roles.get("action");
15859
- if (!action) return false;
15860
- const actionValue = action.value.toLowerCase();
15861
- if (!match.commands.some((cmd) => cmd.toLowerCase() === actionValue)) {
15862
- return false;
15863
- }
15864
- }
15865
- if (match.predicate && !match.predicate(parsed)) {
15866
- return false;
15867
- }
15868
- return true;
15869
- }
15870
- };
15871
- function toLocale(input, targetLocale) {
15872
- const transformer = new GrammarTransformer("en", targetLocale);
15873
- return transformer.transform(input);
15874
- }
15875
- function toEnglish(input, sourceLocale) {
15876
- const transformer = new GrammarTransformer(sourceLocale, "en");
15877
- return transformer.transform(input);
15878
- }
15879
- function translate(input, sourceLocale, targetLocale) {
15880
- if (sourceLocale === targetLocale) return input;
15881
- if (sourceLocale === "en") return toLocale(input, targetLocale);
15882
- if (targetLocale === "en") return toEnglish(input, sourceLocale);
15883
- if (hasDirectMapping(sourceLocale, targetLocale)) {
15884
- return translateDirect(input, sourceLocale, targetLocale);
15885
- }
15886
- const english = toEnglish(input, sourceLocale);
15887
- return toLocale(english, targetLocale);
15888
- }
15889
- function translateDirect(input, sourceLocale, targetLocale) {
15890
- const mapping = getDirectMapping(sourceLocale, targetLocale);
15891
- if (!mapping) {
15892
- return toLocale(toEnglish(input, sourceLocale), targetLocale);
15893
- }
15894
- const tokens = input.split(/\s+/);
15895
- const translated = tokens.map((token) => {
15896
- if (token.startsWith("#") || token.startsWith(".") || token.startsWith("@")) {
15897
- return token;
15898
- }
15899
- if (token.startsWith('"') || token.startsWith("'")) {
15900
- return token;
15901
- }
15902
- const directTranslation = mapping.words[token];
15903
- if (directTranslation) {
15904
- return directTranslation;
15905
- }
15906
- const suffixMatch = token.match(/^(.+?)(-.+)$/);
15907
- if (suffixMatch) {
15908
- const [, base, suffix] = suffixMatch;
15909
- const translatedBase = mapping.words[base] || base;
15910
- return translatedBase + suffix;
15911
- }
15912
- return token;
15913
- });
15914
- return translated.join(" ");
15915
- }
15916
- var examples = {
15917
- english: {
15918
- eventHandler: "on click increment #count",
15919
- putInto: "put my value into #output",
15920
- toggle: "toggle .active",
15921
- wait: "wait 2 seconds"
15922
- },
15923
- // Expected outputs (approximate, for reference)
15924
- japanese: {
15925
- eventHandler: "#count \u3092 \u30AF\u30EA\u30C3\u30AF \u3067 \u5897\u52A0",
15926
- putInto: "\u79C1\u306E \u5024 \u3092 #output \u306B \u7F6E\u304F",
15927
- toggle: ".active \u3092 \u5207\u308A\u66FF\u3048",
15928
- wait: "2\u79D2 \u5F85\u3064"
15929
- },
15930
- chinese: {
15931
- eventHandler: "\u5F53 \u70B9\u51FB \u65F6 \u589E\u52A0 #count",
15932
- putInto: "\u628A \u6211\u7684\u503C \u653E \u5230 #output",
15933
- toggle: "\u5207\u6362 .active",
15934
- wait: "\u7B49\u5F85 2\u79D2"
15935
- },
15936
- arabic: {
15937
- eventHandler: "\u0632\u0650\u062F #count \u0639\u0646\u062F \u0627\u0644\u0646\u0642\u0631",
15938
- putInto: "\u0636\u0639 \u0642\u064A\u0645\u062A\u064A \u0641\u064A #output",
15939
- toggle: "\u0628\u062F\u0651\u0644 .active",
15940
- wait: "\u0627\u0646\u062A\u0638\u0631 \u062B\u0627\u0646\u064A\u062A\u064A\u0646"
15941
- }
15942
- };
15943
-
15944
14088
  // src/index.ts
15945
14089
  var defaultTranslator = new HyperscriptTranslator({ locale: "en" });
15946
14090
  var defaultRuntime = new RuntimeI18nManager({ locale: "en" });
15947
14091
 
15948
- export { DICTIONARY_CATEGORIES, DateFormatter, ENGLISH_COMMANDS, ENGLISH_KEYWORDS, EnhancedI18nInputSchema, EnhancedI18nOutputSchema, GrammarTransformer, HyperscriptI18nWebpackPlugin, HyperscriptTranslator, LANGUAGE_FAMILY_DEFAULTS, LocaleFormatter, LocaleManager, NumberFormatter, PluralAwareTranslator, RuntimeI18nManager, SSRLocaleManager, TypedI18nContextImplementation, UNIVERSAL_ENGLISH_KEYWORDS, UNIVERSAL_PATTERNS, ar2 as ar, ar as arDictionary, arKeywords, arabicProfile, bn2 as bn, bn as bnDictionary, bnKeywords, chineseProfile, createEnglishProvider, createEnhancedI18n, createExpressI18nMiddleware, createI18nContext, createKeywordProvider, de2 as de, de as deDictionary, deKeywords, defaultRuntime, defaultTranslator, detectBrowserLocale, detectLocale, dictionaries, en2 as en, englishProfile, enhancedI18nImplementation, es2 as es, es as esDictionary, esKeywords, findInDictionary, forEachCategory, formatForLocale, fr2 as fr, fr as frDictionary, frKeywords, getBrowserLocales, getDictionary, getDictionaryCategory, getFormatter, getI18n, getPlural, getProfile, getSupportedLocales, examples as grammarExamples, hi, hindiDictionary as hiDictionary, hiKeywords, hyperscriptI18nVitePlugin, id2 as id, id as idDictionary, idKeywords, indonesianProfile, initializeI18n, insertMarkers, isDictionaryCategory, isLocaleSupported, isRTL, it2 as it, it as itDictionary, itKeywords, ja2 as ja, ja as jaDictionary, jaKeywords, japaneseProfile, ko2 as ko, ko as koDictionary, koKeywords, koreanProfile, malayProfile, ms2 as ms, ms as msDictionary, msKeywords, parseStatement, pl2 as pl, pl as plDictionary, plKeywords, pluralRules, pluralTimeExpressions, profiles, pt2 as pt, pt as ptDictionary, ptKeywords, qu2 as qu, qu as quDictionary, quKeywords, quechuaProfile, reorderRoles, ru, russianDictionary as ruDictionary, ruKeywords, runtimeI18n, spanishProfile, supportedLocales, sw2 as sw, sw as swDictionary, swKeywords, swahiliProfile, th2 as th, th as thDictionary, thKeywords, tl2 as tl, tl as tlDictionary, tlKeywords, toEnglish, toLocale, tokenize, tr2 as tr, tr as trDictionary, trKeywords, transformStatement, translate, translateFromEnglish, turkishProfile, uk, ukrainianDictionary as ukDictionary, ukKeywords, validate, vi2 as vi, vi as viDictionary, viKeywords, withI18n, zh2 as zh, zh as zhDictionary, zhKeywords };
14092
+ export { DICTIONARY_CATEGORIES, DateFormatter, ENGLISH_COMMANDS, ENGLISH_KEYWORDS, EnhancedI18nInputSchema, EnhancedI18nOutputSchema, HyperscriptI18nWebpackPlugin, HyperscriptTranslator, LANGUAGE_FAMILY_DEFAULTS, LocaleFormatter, LocaleManager, NumberFormatter, PluralAwareTranslator, RuntimeI18nManager, SSRLocaleManager, TypedI18nContextImplementation, UNIVERSAL_ENGLISH_KEYWORDS, UNIVERSAL_PATTERNS, ar2 as ar, ar as arDictionary, arKeywords, arabicProfile, bn2 as bn, bn as bnDictionary, bnKeywords, chineseProfile, createEnglishProvider, createEnhancedI18n, createExpressI18nMiddleware, createI18nContext, createKeywordProvider, de2 as de, de as deDictionary, deKeywords, defaultRuntime, defaultTranslator, detectBrowserLocale, detectLocale, dictionaries, en2 as en, englishProfile, enhancedI18nImplementation, es2 as es, es as esDictionary, esKeywords, findInDictionary, forEachCategory, formatForLocale, fr2 as fr, fr as frDictionary, frKeywords, getBrowserLocales, getDictionary, getDictionaryCategory, getFormatter, getI18n, getPlural, getProfile, getSupportedLocales, hi, hindiDictionary as hiDictionary, hiKeywords, hyperscriptI18nVitePlugin, id2 as id, id as idDictionary, idKeywords, indonesianProfile, initializeI18n, insertMarkers, isDictionaryCategory, isLocaleSupported, isRTL, it2 as it, it as itDictionary, itKeywords, ja2 as ja, ja as jaDictionary, jaKeywords, japaneseProfile, ko2 as ko, ko as koDictionary, koKeywords, koreanProfile, malayProfile, ms2 as ms, ms as msDictionary, msKeywords, pl2 as pl, pl as plDictionary, plKeywords, pluralRules, pluralTimeExpressions, profiles, pt2 as pt, pt as ptDictionary, ptKeywords, qu2 as qu, qu as quDictionary, quKeywords, quechuaProfile, reorderRoles, ru, russianDictionary as ruDictionary, ruKeywords, runtimeI18n, spanishProfile, supportedLocales, sw2 as sw, sw as swDictionary, swKeywords, swahiliProfile, th2 as th, th as thDictionary, thKeywords, tl2 as tl, tl as tlDictionary, tlKeywords, tokenize, tr2 as tr, tr as trDictionary, trKeywords, transformStatement, translateFromEnglish, turkishProfile, uk, ukrainianDictionary as ukDictionary, ukKeywords, validate, vi2 as vi, vi as viDictionary, viKeywords, withI18n, zh2 as zh, zh as zhDictionary, zhKeywords };
15949
14093
  //# sourceMappingURL=index.js.map
15950
14094
  //# sourceMappingURL=index.js.map