@besales/anonymizer 0.1.0

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 (117) hide show
  1. package/LICENSE +21 -0
  2. package/README.md +81 -0
  3. package/dist/anonymizer.d.ts +31 -0
  4. package/dist/anonymizer.d.ts.map +1 -0
  5. package/dist/anonymizer.js +175 -0
  6. package/dist/anonymizer.js.map +1 -0
  7. package/dist/constants.d.ts +11 -0
  8. package/dist/constants.d.ts.map +1 -0
  9. package/dist/constants.js +16 -0
  10. package/dist/constants.js.map +1 -0
  11. package/dist/detectors/contextual.detector.d.ts +10 -0
  12. package/dist/detectors/contextual.detector.d.ts.map +1 -0
  13. package/dist/detectors/contextual.detector.js +56 -0
  14. package/dist/detectors/contextual.detector.js.map +1 -0
  15. package/dist/detectors/detection-pipeline.d.ts +6 -0
  16. package/dist/detectors/detection-pipeline.d.ts.map +1 -0
  17. package/dist/detectors/detection-pipeline.js +102 -0
  18. package/dist/detectors/detection-pipeline.js.map +1 -0
  19. package/dist/detectors/detector.interface.d.ts +6 -0
  20. package/dist/detectors/detector.interface.d.ts.map +1 -0
  21. package/dist/detectors/detector.interface.js +3 -0
  22. package/dist/detectors/detector.interface.js.map +1 -0
  23. package/dist/detectors/dictionary.detector.d.ts +12 -0
  24. package/dist/detectors/dictionary.detector.d.ts.map +1 -0
  25. package/dist/detectors/dictionary.detector.js +222 -0
  26. package/dist/detectors/dictionary.detector.js.map +1 -0
  27. package/dist/detectors/regex.detector.d.ts +10 -0
  28. package/dist/detectors/regex.detector.d.ts.map +1 -0
  29. package/dist/detectors/regex.detector.js +41 -0
  30. package/dist/detectors/regex.detector.js.map +1 -0
  31. package/dist/index.d.ts +20 -0
  32. package/dist/index.d.ts.map +1 -0
  33. package/dist/index.js +27 -0
  34. package/dist/index.js.map +1 -0
  35. package/dist/llm-protocol/response-validator.d.ts +5 -0
  36. package/dist/llm-protocol/response-validator.d.ts.map +1 -0
  37. package/dist/llm-protocol/response-validator.js +96 -0
  38. package/dist/llm-protocol/response-validator.js.map +1 -0
  39. package/dist/llm-protocol/system-prompt-builder.d.ts +2 -0
  40. package/dist/llm-protocol/system-prompt-builder.d.ts.map +1 -0
  41. package/dist/llm-protocol/system-prompt-builder.js +8 -0
  42. package/dist/llm-protocol/system-prompt-builder.js.map +1 -0
  43. package/dist/locales/index.d.ts +9 -0
  44. package/dist/locales/index.d.ts.map +1 -0
  45. package/dist/locales/index.js +15 -0
  46. package/dist/locales/index.js.map +1 -0
  47. package/dist/locales/locale.interface.d.ts +35 -0
  48. package/dist/locales/locale.interface.d.ts.map +1 -0
  49. package/dist/locales/locale.interface.js +3 -0
  50. package/dist/locales/locale.interface.js.map +1 -0
  51. package/dist/locales/ru/data/names.csv +1291 -0
  52. package/dist/locales/ru/data/stopwords.csv +236 -0
  53. package/dist/locales/ru/index.d.ts +3 -0
  54. package/dist/locales/ru/index.d.ts.map +1 -0
  55. package/dist/locales/ru/index.js +227 -0
  56. package/dist/locales/ru/index.js.map +1 -0
  57. package/dist/tokenizer/tokenizer.d.ts +13 -0
  58. package/dist/tokenizer/tokenizer.d.ts.map +1 -0
  59. package/dist/tokenizer/tokenizer.js +50 -0
  60. package/dist/tokenizer/tokenizer.js.map +1 -0
  61. package/dist/traversers/completions.traverser.d.ts +7 -0
  62. package/dist/traversers/completions.traverser.d.ts.map +1 -0
  63. package/dist/traversers/completions.traverser.js +111 -0
  64. package/dist/traversers/completions.traverser.js.map +1 -0
  65. package/dist/traversers/json.traverser.d.ts +12 -0
  66. package/dist/traversers/json.traverser.d.ts.map +1 -0
  67. package/dist/traversers/json.traverser.js +94 -0
  68. package/dist/traversers/json.traverser.js.map +1 -0
  69. package/dist/traversers/responses.traverser.d.ts +8 -0
  70. package/dist/traversers/responses.traverser.d.ts.map +1 -0
  71. package/dist/traversers/responses.traverser.js +86 -0
  72. package/dist/traversers/responses.traverser.js.map +1 -0
  73. package/dist/traversers/text.traverser.d.ts +6 -0
  74. package/dist/traversers/text.traverser.d.ts.map +1 -0
  75. package/dist/traversers/text.traverser.js +15 -0
  76. package/dist/traversers/text.traverser.js.map +1 -0
  77. package/dist/traversers/traverser.interface.d.ts +9 -0
  78. package/dist/traversers/traverser.interface.d.ts.map +1 -0
  79. package/dist/traversers/traverser.interface.js +3 -0
  80. package/dist/traversers/traverser.interface.js.map +1 -0
  81. package/dist/types/anonymize-result.interface.d.ts +23 -0
  82. package/dist/types/anonymize-result.interface.d.ts.map +1 -0
  83. package/dist/types/anonymize-result.interface.js +3 -0
  84. package/dist/types/anonymize-result.interface.js.map +1 -0
  85. package/dist/types/deanonymize-result.interface.d.ts +18 -0
  86. package/dist/types/deanonymize-result.interface.d.ts.map +1 -0
  87. package/dist/types/deanonymize-result.interface.js +3 -0
  88. package/dist/types/deanonymize-result.interface.js.map +1 -0
  89. package/dist/types/detected-entity.interface.d.ts +23 -0
  90. package/dist/types/detected-entity.interface.d.ts.map +1 -0
  91. package/dist/types/detected-entity.interface.js +3 -0
  92. package/dist/types/detected-entity.interface.js.map +1 -0
  93. package/dist/types/detection-context.d.ts +10 -0
  94. package/dist/types/detection-context.d.ts.map +1 -0
  95. package/dist/types/detection-context.js +12 -0
  96. package/dist/types/detection-context.js.map +1 -0
  97. package/dist/types/entity-type.d.ts +16 -0
  98. package/dist/types/entity-type.d.ts.map +1 -0
  99. package/dist/types/entity-type.js +18 -0
  100. package/dist/types/entity-type.js.map +1 -0
  101. package/dist/types/llm-message.interface.d.ts +23 -0
  102. package/dist/types/llm-message.interface.d.ts.map +1 -0
  103. package/dist/types/llm-message.interface.js +3 -0
  104. package/dist/types/llm-message.interface.js.map +1 -0
  105. package/dist/types/llm-validation.interface.d.ts +8 -0
  106. package/dist/types/llm-validation.interface.d.ts.map +1 -0
  107. package/dist/types/llm-validation.interface.js +3 -0
  108. package/dist/types/llm-validation.interface.js.map +1 -0
  109. package/dist/types/malformed-token.interface.d.ts +6 -0
  110. package/dist/types/malformed-token.interface.d.ts.map +1 -0
  111. package/dist/types/malformed-token.interface.js +3 -0
  112. package/dist/types/malformed-token.interface.js.map +1 -0
  113. package/dist/types/mapping-data.interface.d.ts +9 -0
  114. package/dist/types/mapping-data.interface.d.ts.map +1 -0
  115. package/dist/types/mapping-data.interface.js +12 -0
  116. package/dist/types/mapping-data.interface.js.map +1 -0
  117. package/package.json +48 -0
@@ -0,0 +1,222 @@
1
+ "use strict";
2
+ var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
+ if (k2 === undefined) k2 = k;
4
+ var desc = Object.getOwnPropertyDescriptor(m, k);
5
+ if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
+ desc = { enumerable: true, get: function() { return m[k]; } };
7
+ }
8
+ Object.defineProperty(o, k2, desc);
9
+ }) : (function(o, m, k, k2) {
10
+ if (k2 === undefined) k2 = k;
11
+ o[k2] = m[k];
12
+ }));
13
+ var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
+ Object.defineProperty(o, "default", { enumerable: true, value: v });
15
+ }) : function(o, v) {
16
+ o["default"] = v;
17
+ });
18
+ var __importStar = (this && this.__importStar) || (function () {
19
+ var ownKeys = function(o) {
20
+ ownKeys = Object.getOwnPropertyNames || function (o) {
21
+ var ar = [];
22
+ for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
23
+ return ar;
24
+ };
25
+ return ownKeys(o);
26
+ };
27
+ return function (mod) {
28
+ if (mod && mod.__esModule) return mod;
29
+ var result = {};
30
+ if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
31
+ __setModuleDefault(result, mod);
32
+ return result;
33
+ };
34
+ })();
35
+ Object.defineProperty(exports, "__esModule", { value: true });
36
+ exports.DictionaryDetector = void 0;
37
+ exports.resetDictionaryCache = resetDictionaryCache;
38
+ const fs = __importStar(require("fs"));
39
+ const entity_type_1 = require("../types/entity-type");
40
+ const LOCALE_CACHE = new Map();
41
+ function loadCsv(filePath, filterType) {
42
+ const content = fs.readFileSync(filePath, 'utf-8');
43
+ const lines = content.split('\n').slice(1);
44
+ const result = new Set();
45
+ for (const line of lines) {
46
+ const trimmed = line.trim();
47
+ if (!trimmed)
48
+ continue;
49
+ if (filterType) {
50
+ const [value, type] = trimmed.split(',');
51
+ if (type?.trim() === filterType && value?.trim()) {
52
+ result.add(value.trim());
53
+ }
54
+ }
55
+ else {
56
+ result.add(trimmed);
57
+ }
58
+ }
59
+ return result;
60
+ }
61
+ function ensureCache(localeCode, config) {
62
+ const cached = LOCALE_CACHE.get(localeCode);
63
+ if (cached)
64
+ return cached;
65
+ const result = {
66
+ firstNames: loadCsv(config.namesPath, 'first_name'),
67
+ lastNames: loadCsv(config.namesPath, 'last_name'),
68
+ patronymics: loadCsv(config.namesPath, 'patronymic'),
69
+ stopwords: loadCsv(config.stopwordsPath),
70
+ };
71
+ LOCALE_CACHE.set(localeCode, result);
72
+ return result;
73
+ }
74
+ function resetDictionaryCache() {
75
+ LOCALE_CACHE.clear();
76
+ }
77
+ function tokenizeText(text, splitPattern) {
78
+ const tokens = [];
79
+ const parts = text.split(splitPattern);
80
+ let searchFrom = 0;
81
+ for (const part of parts) {
82
+ if (!part)
83
+ continue;
84
+ const idx = text.indexOf(part, searchFrom);
85
+ if (idx === -1)
86
+ continue;
87
+ tokens.push({
88
+ value: part,
89
+ start: idx,
90
+ end: idx + part.length,
91
+ });
92
+ searchFrom = idx + part.length;
93
+ }
94
+ return tokens;
95
+ }
96
+ function isCapitalized(value) {
97
+ if (!value)
98
+ return false;
99
+ const first = value[0];
100
+ return first !== first.toLowerCase();
101
+ }
102
+ const CONFIDENCE_FIO_FULL = 0.95;
103
+ const CONFIDENCE_FIO_TWO_DICT = 0.90;
104
+ const CONFIDENCE_FIO_NAME_PATRONYMIC = 0.85;
105
+ const CONFIDENCE_FIO_HEURISTIC = 0.60;
106
+ class DictionaryDetector {
107
+ localeCode;
108
+ config;
109
+ constructor(localeCode, config) {
110
+ this.localeCode = localeCode;
111
+ this.config = config;
112
+ }
113
+ detect(text, _context) {
114
+ const cache = ensureCache(this.localeCode, this.config);
115
+ const tokens = tokenizeText(text, this.config.tokenSplitPattern);
116
+ const results = [];
117
+ const consumed = new Set();
118
+ for (let i = 0; i < tokens.length; i++) {
119
+ if (consumed.has(i))
120
+ continue;
121
+ const t0 = tokens[i];
122
+ const t1 = tokens[i + 1];
123
+ const t2 = tokens[i + 2];
124
+ const t0Lower = t0.value.toLowerCase();
125
+ const t1Lower = t1?.value.toLowerCase();
126
+ const t2Lower = t2?.value.toLowerCase();
127
+ if (t1 && t2 &&
128
+ isCapitalized(t0.value) &&
129
+ isCapitalized(t1.value) &&
130
+ isCapitalized(t2.value) &&
131
+ cache.lastNames.has(t0Lower) &&
132
+ cache.firstNames.has(t1Lower) &&
133
+ cache.patronymics.has(t2Lower)) {
134
+ const value = text.slice(t0.start, t2.end);
135
+ results.push({
136
+ type: entity_type_1.ENTITY_TYPE.PERSON,
137
+ value,
138
+ position: { start: t0.start, end: t2.end },
139
+ detectedBy: 'dictionary',
140
+ confidence: CONFIDENCE_FIO_FULL,
141
+ });
142
+ consumed.add(i);
143
+ consumed.add(i + 1);
144
+ consumed.add(i + 2);
145
+ continue;
146
+ }
147
+ if (t1 && isCapitalized(t0.value) && isCapitalized(t1.value)) {
148
+ let confidence = null;
149
+ if (cache.lastNames.has(t0Lower) && cache.firstNames.has(t1Lower)) {
150
+ confidence = CONFIDENCE_FIO_TWO_DICT;
151
+ }
152
+ else if (cache.firstNames.has(t0Lower) && cache.lastNames.has(t1Lower)) {
153
+ confidence = CONFIDENCE_FIO_TWO_DICT;
154
+ }
155
+ else if (cache.firstNames.has(t0Lower) && cache.patronymics.has(t1Lower)) {
156
+ confidence = CONFIDENCE_FIO_NAME_PATRONYMIC;
157
+ }
158
+ if (confidence !== null) {
159
+ const value = text.slice(t0.start, t1.end);
160
+ results.push({
161
+ type: entity_type_1.ENTITY_TYPE.PERSON,
162
+ value,
163
+ position: { start: t0.start, end: t1.end },
164
+ detectedBy: 'dictionary',
165
+ confidence,
166
+ });
167
+ consumed.add(i);
168
+ consumed.add(i + 1);
169
+ continue;
170
+ }
171
+ }
172
+ if (t1 &&
173
+ isCapitalized(t0.value) &&
174
+ isCapitalized(t1.value) &&
175
+ !consumed.has(i + 1)) {
176
+ const isT0InDict = cache.firstNames.has(t0Lower) || cache.lastNames.has(t0Lower);
177
+ const isT1InDict = cache.firstNames.has(t1Lower) || cache.lastNames.has(t1Lower);
178
+ if (isT0InDict || isT1InDict) {
179
+ const isT0Stop = cache.stopwords.has(t0Lower);
180
+ const isT1Stop = cache.stopwords.has(t1Lower);
181
+ if (!isT0Stop && !isT1Stop) {
182
+ const value = text.slice(t0.start, t1.end);
183
+ results.push({
184
+ type: entity_type_1.ENTITY_TYPE.PERSON,
185
+ value,
186
+ position: { start: t0.start, end: t1.end },
187
+ detectedBy: 'dictionary',
188
+ confidence: CONFIDENCE_FIO_HEURISTIC,
189
+ });
190
+ consumed.add(i);
191
+ consumed.add(i + 1);
192
+ continue;
193
+ }
194
+ }
195
+ }
196
+ }
197
+ return deduplicateResults(results);
198
+ }
199
+ }
200
+ exports.DictionaryDetector = DictionaryDetector;
201
+ function deduplicateResults(entities) {
202
+ if (entities.length <= 1)
203
+ return entities;
204
+ const sorted = [...entities].sort((a, b) => {
205
+ const startDiff = a.position.start - b.position.start;
206
+ if (startDiff !== 0)
207
+ return startDiff;
208
+ const aLen = a.position.end - a.position.start;
209
+ const bLen = b.position.end - b.position.start;
210
+ return bLen - aLen;
211
+ });
212
+ const result = [sorted[0]];
213
+ for (let i = 1; i < sorted.length; i++) {
214
+ const current = sorted[i];
215
+ const last = result[result.length - 1];
216
+ if (current.position.start >= last.position.end) {
217
+ result.push(current);
218
+ }
219
+ }
220
+ return result;
221
+ }
222
+ //# sourceMappingURL=dictionary.detector.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"dictionary.detector.js","sourceRoot":"","sources":["../../src/detectors/dictionary.detector.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAoEA,oDAEC;AAtED,uCAAyB;AAGzB,sDAAmD;AAqBnD,MAAM,YAAY,GAAG,IAAI,GAAG,EAA2B,CAAC;AAExD,SAAS,OAAO,CACd,QAAgB,EAChB,UAAmB;IAEnB,MAAM,OAAO,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IACnD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAC3C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;IAEjC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,OAAO;YAAE,SAAS;QAEvB,IAAI,UAAU,EAAE,CAAC;YACf,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;YACzC,IAAI,IAAI,EAAE,IAAI,EAAE,KAAK,UAAU,IAAI,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC;gBACjD,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;YAC3B,CAAC;QACH,CAAC;aAAM,CAAC;YAEN,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,WAAW,CAAC,UAAkB,EAAE,MAA8B;IACrE,MAAM,MAAM,GAAG,YAAY,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IAC5C,IAAI,MAAM;QAAE,OAAO,MAAM,CAAC;IAE1B,MAAM,MAAM,GAAoB;QAC9B,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,SAAS,EAAE,YAAY,CAAC;QACnD,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,SAAS,EAAE,WAAW,CAAC;QACjD,WAAW,EAAE,OAAO,CAAC,MAAM,CAAC,SAAS,EAAE,YAAY,CAAC;QACpD,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,aAAa,CAAC;KACzC,CAAC;IAEF,YAAY,CAAC,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;IACrC,OAAO,MAAM,CAAC;AAChB,CAAC;AAGD,SAAgB,oBAAoB;IAClC,YAAY,CAAC,KAAK,EAAE,CAAC;AACvB,CAAC;AAID,SAAS,YAAY,CAAC,IAAY,EAAE,YAAoB;IACtD,MAAM,MAAM,GAAgB,EAAE,CAAC;IAE/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;IACvC,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;QAC3C,IAAI,GAAG,KAAK,CAAC,CAAC;YAAE,SAAS;QAEzB,MAAM,CAAC,IAAI,CAAC;YACV,KAAK,EAAE,IAAI;YACX,KAAK,EAAE,GAAG;YACV,GAAG,EAAE,GAAG,GAAG,IAAI,CAAC,MAAM;SACvB,CAAC,CAAC;QACH,UAAU,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;IACjC,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAGD,SAAS,aAAa,CAAC,KAAa;IAClC,IAAI,CAAC,KAAK;QAAE,OAAO,KAAK,CAAC;IACzB,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACvB,OAAO,KAAK,KAAK,KAAK,CAAC,WAAW,EAAE,CAAC;AACvC,CAAC;AAID,MAAM,mBAAmB,GAAG,IAAI,CAAC;AACjC,MAAM,uBAAuB,GAAG,IAAI,CAAC;AACrC,MAAM,8BAA8B,GAAG,IAAI,CAAC;AAC5C,MAAM,wBAAwB,GAAG,IAAI,CAAC;AAEtC,MAAa,kBAAkB;IAEV;IACA;IAFnB,YACmB,UAAkB,EAClB,MAA8B;QAD9B,eAAU,GAAV,UAAU,CAAQ;QAClB,WAAM,GAAN,MAAM,CAAwB;IAC9C,CAAC;IAEJ,MAAM,CAAC,IAAY,EAAE,QAA0B;QAC7C,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;QACxD,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,iBAAiB,CAAC,CAAC;QACjE,MAAM,OAAO,GAAqB,EAAE,CAAC;QACrC,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;QAEnC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,IAAI,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC;gBAAE,SAAS;YAE9B,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YACrB,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACzB,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAEzB,MAAM,OAAO,GAAG,EAAE,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC;YACvC,MAAM,OAAO,GAAG,EAAE,EAAE,KAAK,CAAC,WAAW,EAAE,CAAC;YACxC,MAAM,OAAO,GAAG,EAAE,EAAE,KAAK,CAAC,WAAW,EAAE,CAAC;YAGxC,IACE,EAAE,IAAI,EAAE;gBACR,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC;gBAC5B,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC;gBAC7B,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,EAC9B,CAAC;gBACD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;gBAC3C,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,yBAAW,CAAC,MAAM;oBACxB,KAAK;oBACL,QAAQ,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE;oBAC1C,UAAU,EAAE,YAAY;oBACxB,UAAU,EAAE,mBAAmB;iBAChC,CAAC,CAAC;gBACH,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;gBAChB,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBACpB,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBACpB,SAAS;YACX,CAAC;YAGD,IAAI,EAAE,IAAI,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC7D,IAAI,UAAU,GAAkB,IAAI,CAAC;gBAGrC,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;oBAClE,UAAU,GAAG,uBAAuB,CAAC;gBACvC,CAAC;qBAEI,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;oBACvE,UAAU,GAAG,uBAAuB,CAAC;gBACvC,CAAC;qBAEI,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;oBACzE,UAAU,GAAG,8BAA8B,CAAC;gBAC9C,CAAC;gBAED,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;oBACxB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;oBAC3C,OAAO,CAAC,IAAI,CAAC;wBACX,IAAI,EAAE,yBAAW,CAAC,MAAM;wBACxB,KAAK;wBACL,QAAQ,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE;wBAC1C,UAAU,EAAE,YAAY;wBACxB,UAAU;qBACX,CAAC,CAAC;oBACH,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;oBAChB,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;oBACpB,SAAS;gBACX,CAAC;YACH,CAAC;YAID,IACE,EAAE;gBACF,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,EACpB,CAAC;gBACD,MAAM,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBACjF,MAAM,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBAGjF,IAAI,UAAU,IAAI,UAAU,EAAE,CAAC;oBAE7B,MAAM,QAAQ,GAAG,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;oBAC9C,MAAM,QAAQ,GAAG,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;oBAE9C,IAAI,CAAC,QAAQ,IAAI,CAAC,QAAQ,EAAE,CAAC;wBAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;wBAC3C,OAAO,CAAC,IAAI,CAAC;4BACX,IAAI,EAAE,yBAAW,CAAC,MAAM;4BACxB,KAAK;4BACL,QAAQ,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE;4BAC1C,UAAU,EAAE,YAAY;4BACxB,UAAU,EAAE,wBAAwB;yBACrC,CAAC,CAAC;wBACH,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;wBAChB,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;wBACpB,SAAS;oBACX,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,kBAAkB,CAAC,OAAO,CAAC,CAAC;IACrC,CAAC;CACF;AAnHD,gDAmHC;AAKD,SAAS,kBAAkB,CAAC,QAA0B;IACpD,IAAI,QAAQ,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC;IAG1C,MAAM,MAAM,GAAG,CAAC,GAAG,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACzC,MAAM,SAAS,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC;QACtD,IAAI,SAAS,KAAK,CAAC;YAAE,OAAO,SAAS,CAAC;QACtC,MAAM,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC;QAC/C,MAAM,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC;QAC/C,OAAO,IAAI,GAAG,IAAI,CAAC;IACrB,CAAC,CAAC,CAAC;IAEH,MAAM,MAAM,GAAqB,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAE7C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QAC1B,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAGvC,IAAI,OAAO,CAAC,QAAQ,CAAC,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE,CAAC;YAChD,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACvB,CAAC;IAGH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
@@ -0,0 +1,10 @@
1
+ import { DetectedEntity } from '../types/detected-entity.interface';
2
+ import { DetectionContext } from '../types/detection-context';
3
+ import { LocaleRegexRule } from '../locales/locale.interface';
4
+ import { DetectorPort } from './detector.interface';
5
+ export declare class RegexDetector implements DetectorPort {
6
+ private readonly rules;
7
+ constructor(rules: LocaleRegexRule[]);
8
+ detect(text: string, _context: DetectionContext): DetectedEntity[];
9
+ }
10
+ //# sourceMappingURL=regex.detector.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"regex.detector.d.ts","sourceRoot":"","sources":["../../src/detectors/regex.detector.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;AAC9D,OAAO,EAAE,eAAe,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAuBpD,qBAAa,aAAc,YAAW,YAAY;IAChD,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAoB;gBAE9B,KAAK,EAAE,eAAe,EAAE;IAIpC,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,gBAAgB,GAAG,cAAc,EAAE;CA6BnE"}
@@ -0,0 +1,41 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.RegexDetector = void 0;
4
+ function hasContext(text, matchStart, matchEnd, contextPattern, radius, direction) {
5
+ const windowStart = Math.max(0, matchStart - radius);
6
+ const windowEnd = direction === 'before'
7
+ ? matchStart
8
+ : Math.min(text.length, matchEnd + radius);
9
+ const searchWindow = text.slice(windowStart, windowEnd);
10
+ return contextPattern.test(searchWindow);
11
+ }
12
+ class RegexDetector {
13
+ rules;
14
+ constructor(rules) {
15
+ this.rules = rules;
16
+ }
17
+ detect(text, _context) {
18
+ const results = [];
19
+ for (const rule of this.rules) {
20
+ for (const match of text.matchAll(rule.pattern)) {
21
+ const value = match[0];
22
+ const start = match.index;
23
+ const end = start + value.length;
24
+ if (rule.contextPattern !== null &&
25
+ !hasContext(text, start, end, rule.contextPattern, rule.contextRadius, rule.contextDirection)) {
26
+ continue;
27
+ }
28
+ results.push({
29
+ type: rule.type,
30
+ value,
31
+ position: { start, end },
32
+ detectedBy: 'regex',
33
+ confidence: rule.confidence,
34
+ });
35
+ }
36
+ }
37
+ return results;
38
+ }
39
+ }
40
+ exports.RegexDetector = RegexDetector;
41
+ //# sourceMappingURL=regex.detector.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"regex.detector.js","sourceRoot":"","sources":["../../src/detectors/regex.detector.ts"],"names":[],"mappings":";;;AASA,SAAS,UAAU,CACjB,IAAY,EACZ,UAAkB,EAClB,QAAgB,EAChB,cAAsB,EACtB,MAAc,EACd,SAA4B;IAE5B,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,GAAG,MAAM,CAAC,CAAC;IACrD,MAAM,SAAS,GACb,SAAS,KAAK,QAAQ;QACpB,CAAC,CAAC,UAAU;QACZ,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE,QAAQ,GAAG,MAAM,CAAC,CAAC;IAC/C,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC;IACxD,OAAO,cAAc,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;AAC3C,CAAC;AAED,MAAa,aAAa;IACP,KAAK,CAAoB;IAE1C,YAAY,KAAwB;QAClC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;IACrB,CAAC;IAED,MAAM,CAAC,IAAY,EAAE,QAA0B;QAC7C,MAAM,OAAO,GAAqB,EAAE,CAAC;QAErC,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;gBAChD,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;gBACvB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAM,CAAC;gBAC3B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC;gBAGjC,IACE,IAAI,CAAC,cAAc,KAAK,IAAI;oBAC5B,CAAC,UAAU,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,cAAc,EAAE,IAAI,CAAC,aAAa,EAAE,IAAI,CAAC,gBAAgB,CAAC,EAC7F,CAAC;oBACD,SAAS;gBACX,CAAC;gBAED,OAAO,CAAC,IAAI,CAAC;oBACX,IAAI,EAAE,IAAI,CAAC,IAAI;oBACf,KAAK;oBACL,QAAQ,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;oBACxB,UAAU,EAAE,OAAO;oBACnB,UAAU,EAAE,IAAI,CAAC,UAAU;iBAC5B,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;CACF;AApCD,sCAoCC"}
@@ -0,0 +1,20 @@
1
+ export { anonymizeText, deanonymizeText, anonymizeMessages, deanonymizeMessages, anonymizePayload, deanonymizePayload, } from './anonymizer';
2
+ export type { AnonymizeTextOptions, AnonymizeMessagesOptions, AnonymizePayloadOptions, } from './anonymizer';
3
+ export { buildSystemPromptInstruction } from './llm-protocol/system-prompt-builder';
4
+ export { validateLlmResponse, autoFixMalformedTokens } from './llm-protocol/response-validator';
5
+ export { createEmptyMapping } from './types/mapping-data.interface';
6
+ export { LOCALES, DEFAULT_LOCALE, getLocale, RU_LOCALE } from './locales';
7
+ export type { Locale } from './locales';
8
+ export type { LocaleConfig, LocaleRegexRule, LocaleContextualRule, LocaleDictionaryConfig, } from './locales/locale.interface';
9
+ export { ENTITY_TYPE } from './types/entity-type';
10
+ export { DETECTION_CONTEXT } from './types/detection-context';
11
+ export type { EntityType } from './types/entity-type';
12
+ export type { DetectionContext } from './types/detection-context';
13
+ export type { MappingData } from './types/mapping-data.interface';
14
+ export type { AnonymizeTextResult, AnonymizeMessagesResult, AnonymizePayloadResult, } from './types/anonymize-result.interface';
15
+ export type { DeanonymizeTextResult, DeanonymizeMessagesResult, DeanonymizePayloadResult, } from './types/deanonymize-result.interface';
16
+ export type { DetectedEntityMeta } from './types/detected-entity.interface';
17
+ export type { MalformedToken } from './types/malformed-token.interface';
18
+ export type { CompletionsMessage, CompletionsContentPart, CompletionsToolCall, } from './types/llm-message.interface';
19
+ export type { LlmResponseValidation } from './types/llm-validation.interface';
20
+ //# sourceMappingURL=index.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAMA,OAAO,EACL,aAAa,EACb,eAAe,EACf,iBAAiB,EACjB,mBAAmB,EACnB,gBAAgB,EAChB,kBAAkB,GACnB,MAAM,cAAc,CAAC;AAGtB,YAAY,EACV,oBAAoB,EACpB,wBAAwB,EACxB,uBAAuB,GACxB,MAAM,cAAc,CAAC;AAGtB,OAAO,EAAE,4BAA4B,EAAE,MAAM,sCAAsC,CAAC;AACpF,OAAO,EAAE,mBAAmB,EAAE,sBAAsB,EAAE,MAAM,mCAAmC,CAAC;AAGhG,OAAO,EAAE,kBAAkB,EAAE,MAAM,gCAAgC,CAAC;AAGpE,OAAO,EAAE,OAAO,EAAE,cAAc,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AAC1E,YAAY,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AACxC,YAAY,EACV,YAAY,EACZ,eAAe,EACf,oBAAoB,EACpB,sBAAsB,GACvB,MAAM,4BAA4B,CAAC;AAGpC,OAAO,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAClD,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAG9D,YAAY,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AACtD,YAAY,EAAE,gBAAgB,EAAE,MAAM,2BAA2B,CAAC;AAClE,YAAY,EAAE,WAAW,EAAE,MAAM,gCAAgC,CAAC;AAClE,YAAY,EACV,mBAAmB,EACnB,uBAAuB,EACvB,sBAAsB,GACvB,MAAM,oCAAoC,CAAC;AAC5C,YAAY,EACV,qBAAqB,EACrB,yBAAyB,EACzB,wBAAwB,GACzB,MAAM,sCAAsC,CAAC;AAC9C,YAAY,EAAE,kBAAkB,EAAE,MAAM,mCAAmC,CAAC;AAC5E,YAAY,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACxE,YAAY,EACV,kBAAkB,EAClB,sBAAsB,EACtB,mBAAmB,GACpB,MAAM,+BAA+B,CAAC;AACvC,YAAY,EAAE,qBAAqB,EAAE,MAAM,kCAAkC,CAAC"}
package/dist/index.js ADDED
@@ -0,0 +1,27 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.DETECTION_CONTEXT = exports.ENTITY_TYPE = exports.RU_LOCALE = exports.getLocale = exports.DEFAULT_LOCALE = exports.LOCALES = exports.createEmptyMapping = exports.autoFixMalformedTokens = exports.validateLlmResponse = exports.buildSystemPromptInstruction = exports.deanonymizePayload = exports.anonymizePayload = exports.deanonymizeMessages = exports.anonymizeMessages = exports.deanonymizeText = exports.anonymizeText = void 0;
4
+ var anonymizer_1 = require("./anonymizer");
5
+ Object.defineProperty(exports, "anonymizeText", { enumerable: true, get: function () { return anonymizer_1.anonymizeText; } });
6
+ Object.defineProperty(exports, "deanonymizeText", { enumerable: true, get: function () { return anonymizer_1.deanonymizeText; } });
7
+ Object.defineProperty(exports, "anonymizeMessages", { enumerable: true, get: function () { return anonymizer_1.anonymizeMessages; } });
8
+ Object.defineProperty(exports, "deanonymizeMessages", { enumerable: true, get: function () { return anonymizer_1.deanonymizeMessages; } });
9
+ Object.defineProperty(exports, "anonymizePayload", { enumerable: true, get: function () { return anonymizer_1.anonymizePayload; } });
10
+ Object.defineProperty(exports, "deanonymizePayload", { enumerable: true, get: function () { return anonymizer_1.deanonymizePayload; } });
11
+ var system_prompt_builder_1 = require("./llm-protocol/system-prompt-builder");
12
+ Object.defineProperty(exports, "buildSystemPromptInstruction", { enumerable: true, get: function () { return system_prompt_builder_1.buildSystemPromptInstruction; } });
13
+ var response_validator_1 = require("./llm-protocol/response-validator");
14
+ Object.defineProperty(exports, "validateLlmResponse", { enumerable: true, get: function () { return response_validator_1.validateLlmResponse; } });
15
+ Object.defineProperty(exports, "autoFixMalformedTokens", { enumerable: true, get: function () { return response_validator_1.autoFixMalformedTokens; } });
16
+ var mapping_data_interface_1 = require("./types/mapping-data.interface");
17
+ Object.defineProperty(exports, "createEmptyMapping", { enumerable: true, get: function () { return mapping_data_interface_1.createEmptyMapping; } });
18
+ var locales_1 = require("./locales");
19
+ Object.defineProperty(exports, "LOCALES", { enumerable: true, get: function () { return locales_1.LOCALES; } });
20
+ Object.defineProperty(exports, "DEFAULT_LOCALE", { enumerable: true, get: function () { return locales_1.DEFAULT_LOCALE; } });
21
+ Object.defineProperty(exports, "getLocale", { enumerable: true, get: function () { return locales_1.getLocale; } });
22
+ Object.defineProperty(exports, "RU_LOCALE", { enumerable: true, get: function () { return locales_1.RU_LOCALE; } });
23
+ var entity_type_1 = require("./types/entity-type");
24
+ Object.defineProperty(exports, "ENTITY_TYPE", { enumerable: true, get: function () { return entity_type_1.ENTITY_TYPE; } });
25
+ var detection_context_1 = require("./types/detection-context");
26
+ Object.defineProperty(exports, "DETECTION_CONTEXT", { enumerable: true, get: function () { return detection_context_1.DETECTION_CONTEXT; } });
27
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;;AAMA,2CAOsB;AANpB,2GAAA,aAAa,OAAA;AACb,6GAAA,eAAe,OAAA;AACf,+GAAA,iBAAiB,OAAA;AACjB,iHAAA,mBAAmB,OAAA;AACnB,8GAAA,gBAAgB,OAAA;AAChB,gHAAA,kBAAkB,OAAA;AAWpB,8EAAoF;AAA3E,qIAAA,4BAA4B,OAAA;AACrC,wEAAgG;AAAvF,yHAAA,mBAAmB,OAAA;AAAE,4HAAA,sBAAsB,OAAA;AAGpD,yEAAoE;AAA3D,4HAAA,kBAAkB,OAAA;AAG3B,qCAA0E;AAAjE,kGAAA,OAAO,OAAA;AAAE,yGAAA,cAAc,OAAA;AAAE,oGAAA,SAAS,OAAA;AAAE,oGAAA,SAAS,OAAA;AAUtD,mDAAkD;AAAzC,0GAAA,WAAW,OAAA;AACpB,+DAA8D;AAArD,sHAAA,iBAAiB,OAAA"}
@@ -0,0 +1,5 @@
1
+ import { LlmResponseValidation } from '../types/llm-validation.interface';
2
+ import { MalformedToken } from '../types/malformed-token.interface';
3
+ export declare function validateLlmResponse(llmResponse: string, knownTokens: string[]): LlmResponseValidation;
4
+ export declare function autoFixMalformedTokens(text: string, malformedTokens: MalformedToken[], confidenceThreshold: number): string;
5
+ //# sourceMappingURL=response-validator.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"response-validator.d.ts","sourceRoot":"","sources":["../../src/llm-protocol/response-validator.ts"],"names":[],"mappings":"AAKA,OAAO,EAAE,qBAAqB,EAAE,MAAM,mCAAmC,CAAC;AAC1E,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AASpE,wBAAgB,mBAAmB,CACjC,WAAW,EAAE,MAAM,EACnB,WAAW,EAAE,MAAM,EAAE,GACpB,qBAAqB,CAoFvB;AAUD,wBAAgB,sBAAsB,CACpC,IAAI,EAAE,MAAM,EACZ,eAAe,EAAE,cAAc,EAAE,EACjC,mBAAmB,EAAE,MAAM,GAC1B,MAAM,CAkBR"}
@@ -0,0 +1,96 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.validateLlmResponse = validateLlmResponse;
4
+ exports.autoFixMalformedTokens = autoFixMalformedTokens;
5
+ const constants_1 = require("../constants");
6
+ function validateLlmResponse(llmResponse, knownTokens) {
7
+ if (!llmResponse || knownTokens.length === 0) {
8
+ return {
9
+ isValid: true,
10
+ validTokens: [],
11
+ missingTokens: [],
12
+ malformedTokens: [],
13
+ };
14
+ }
15
+ const knownSet = new Set(knownTokens);
16
+ const foundInText = new Set();
17
+ const validTokenPattern = new RegExp(constants_1.TOKEN_PATTERN.source, 'g');
18
+ let match;
19
+ while ((match = validTokenPattern.exec(llmResponse)) !== null) {
20
+ foundInText.add(match[0]);
21
+ }
22
+ const validTokens = [...foundInText].filter((t) => knownSet.has(t));
23
+ const missingTokens = knownTokens.filter((t) => !foundInText.has(t));
24
+ const malformedTokens = [];
25
+ const seenMalformed = new Set();
26
+ const malformedPattern = new RegExp(constants_1.MALFORMED_TOKEN_PATTERN.source, 'g');
27
+ while ((match = malformedPattern.exec(llmResponse)) !== null) {
28
+ const rawCandidate = match[1] || match[2];
29
+ if (!rawCandidate)
30
+ continue;
31
+ const fullMatch = match[0];
32
+ if (foundInText.has(fullMatch))
33
+ continue;
34
+ if (seenMalformed.has(fullMatch))
35
+ continue;
36
+ seenMalformed.add(fullMatch);
37
+ const normalized = `[${rawCandidate.toUpperCase().replace(/\s+/g, '_')}]`;
38
+ if (foundInText.has(normalized))
39
+ continue;
40
+ let bestMatch = null;
41
+ let bestDist = Infinity;
42
+ for (const known of knownTokens) {
43
+ if (normalized === known) {
44
+ bestMatch = {
45
+ found: fullMatch,
46
+ suggestedMatch: known,
47
+ confidence: 0.95,
48
+ };
49
+ break;
50
+ }
51
+ const dist = levenshtein(normalized, known);
52
+ if (dist <= constants_1.MAX_LEVENSHTEIN_DISTANCE && dist < bestDist) {
53
+ bestDist = dist;
54
+ bestMatch = {
55
+ found: fullMatch,
56
+ suggestedMatch: known,
57
+ confidence: 0.7,
58
+ };
59
+ }
60
+ }
61
+ if (bestMatch) {
62
+ malformedTokens.push(bestMatch);
63
+ }
64
+ }
65
+ return {
66
+ isValid: malformedTokens.length === 0,
67
+ validTokens,
68
+ missingTokens,
69
+ malformedTokens,
70
+ };
71
+ }
72
+ function autoFixMalformedTokens(text, malformedTokens, confidenceThreshold) {
73
+ const sorted = [...malformedTokens].sort((a, b) => b.found.length - a.found.length);
74
+ let result = text;
75
+ for (const malformed of sorted) {
76
+ if (malformed.confidence >= confidenceThreshold &&
77
+ malformed.suggestedMatch !== null) {
78
+ result = result.replaceAll(malformed.found, malformed.suggestedMatch);
79
+ }
80
+ }
81
+ return result;
82
+ }
83
+ function levenshtein(a, b) {
84
+ const matrix = [];
85
+ for (let i = 0; i <= a.length; i++) {
86
+ matrix[i] = [i];
87
+ for (let j = 1; j <= b.length; j++) {
88
+ matrix[i][j] =
89
+ i === 0
90
+ ? j
91
+ : Math.min(matrix[i - 1][j] + 1, matrix[i][j - 1] + 1, matrix[i - 1][j - 1] + (a[i - 1] === b[j - 1] ? 0 : 1));
92
+ }
93
+ }
94
+ return matrix[a.length][b.length];
95
+ }
96
+ //# sourceMappingURL=response-validator.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"response-validator.js","sourceRoot":"","sources":["../../src/llm-protocol/response-validator.ts"],"names":[],"mappings":";;AAeA,kDAuFC;AAUD,wDAsBC;AAtID,4CAIsB;AAWtB,SAAgB,mBAAmB,CACjC,WAAmB,EACnB,WAAqB;IAErB,IAAI,CAAC,WAAW,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC7C,OAAO;YACL,OAAO,EAAE,IAAI;YACb,WAAW,EAAE,EAAE;YACf,aAAa,EAAE,EAAE;YACjB,eAAe,EAAE,EAAE;SACpB,CAAC;IACJ,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,WAAW,CAAC,CAAC;IAGtC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,iBAAiB,GAAG,IAAI,MAAM,CAAC,yBAAa,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAChE,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,iBAAiB,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC9D,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5B,CAAC;IAGD,MAAM,WAAW,GAAG,CAAC,GAAG,WAAW,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAGpE,MAAM,aAAa,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAGrE,MAAM,eAAe,GAAqB,EAAE,CAAC;IAC7C,MAAM,aAAa,GAAG,IAAI,GAAG,EAAU,CAAC;IACxC,MAAM,gBAAgB,GAAG,IAAI,MAAM,CAAC,mCAAuB,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAEzE,OAAO,CAAC,KAAK,GAAG,gBAAgB,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC7D,MAAM,YAAY,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;QAC1C,IAAI,CAAC,YAAY;YAAE,SAAS;QAE5B,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAG3B,IAAI,WAAW,CAAC,GAAG,CAAC,SAAS,CAAC;YAAE,SAAS;QACzC,IAAI,aAAa,CAAC,GAAG,CAAC,SAAS,CAAC;YAAE,SAAS;QAC3C,aAAa,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAG7B,MAAM,UAAU,GAAG,IAAI,YAAY,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,GAAG,CAAC;QAG1E,IAAI,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC;YAAE,SAAS;QAE1C,IAAI,SAAS,GAA0B,IAAI,CAAC;QAC5C,IAAI,QAAQ,GAAG,QAAQ,CAAC;QAExB,KAAK,MAAM,KAAK,IAAI,WAAW,EAAE,CAAC;YAChC,IAAI,UAAU,KAAK,KAAK,EAAE,CAAC;gBACzB,SAAS,GAAG;oBACV,KAAK,EAAE,SAAS;oBAChB,cAAc,EAAE,KAAK;oBACrB,UAAU,EAAE,IAAI;iBACjB,CAAC;gBACF,MAAM;YACR,CAAC;YAED,MAAM,IAAI,GAAG,WAAW,CAAC,UAAU,EAAE,KAAK,CAAC,CAAC;YAC5C,IAAI,IAAI,IAAI,oCAAwB,IAAI,IAAI,GAAG,QAAQ,EAAE,CAAC;gBACxD,QAAQ,GAAG,IAAI,CAAC;gBAChB,SAAS,GAAG;oBACV,KAAK,EAAE,SAAS;oBAChB,cAAc,EAAE,KAAK;oBACrB,UAAU,EAAE,GAAG;iBAChB,CAAC;YACJ,CAAC;QACH,CAAC;QAED,IAAI,SAAS,EAAE,CAAC;YACd,eAAe,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,OAAO;QACL,OAAO,EAAE,eAAe,CAAC,MAAM,KAAK,CAAC;QACrC,WAAW;QACX,aAAa;QACb,eAAe;KAChB,CAAC;AACJ,CAAC;AAUD,SAAgB,sBAAsB,CACpC,IAAY,EACZ,eAAiC,EACjC,mBAA2B;IAG3B,MAAM,MAAM,GAAG,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CACtC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,CAAC,MAAM,CAC1C,CAAC;IAEF,IAAI,MAAM,GAAG,IAAI,CAAC;IAElB,KAAK,MAAM,SAAS,IAAI,MAAM,EAAE,CAAC;QAC/B,IACE,SAAS,CAAC,UAAU,IAAI,mBAAmB;YAC3C,SAAS,CAAC,cAAc,KAAK,IAAI,EACjC,CAAC;YACD,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC,KAAK,EAAE,SAAS,CAAC,cAAc,CAAC,CAAC;QACxE,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAKD,SAAS,WAAW,CAAC,CAAS,EAAE,CAAS;IACvC,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACnC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QAChB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACV,CAAC,KAAK,CAAC;oBACL,CAAC,CAAC,CAAC;oBACH,CAAC,CAAC,IAAI,CAAC,GAAG,CACN,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,EACpB,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,EACpB,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CACvD,CAAC;QACV,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;AACpC,CAAC"}
@@ -0,0 +1,2 @@
1
+ export declare function buildSystemPromptInstruction(activeTokens: string[]): string;
2
+ //# sourceMappingURL=system-prompt-builder.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"system-prompt-builder.d.ts","sourceRoot":"","sources":["../../src/llm-protocol/system-prompt-builder.ts"],"names":[],"mappings":"AAYA,wBAAgB,4BAA4B,CAC1C,YAAY,EAAE,MAAM,EAAE,GACrB,MAAM,CAER"}
@@ -0,0 +1,8 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.buildSystemPromptInstruction = buildSystemPromptInstruction;
4
+ const locales_1 = require("../locales");
5
+ function buildSystemPromptInstruction(activeTokens) {
6
+ return (0, locales_1.getLocale)('ru').systemPrompt.buildInstruction(activeTokens);
7
+ }
8
+ //# sourceMappingURL=system-prompt-builder.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"system-prompt-builder.js","sourceRoot":"","sources":["../../src/llm-protocol/system-prompt-builder.ts"],"names":[],"mappings":";;AAYA,oEAIC;AAhBD,wCAAuC;AAYvC,SAAgB,4BAA4B,CAC1C,YAAsB;IAEtB,OAAO,IAAA,mBAAS,EAAC,IAAI,CAAC,CAAC,YAAY,CAAC,gBAAgB,CAAC,YAAY,CAAC,CAAC;AACrE,CAAC"}
@@ -0,0 +1,9 @@
1
+ import { LocaleConfig } from './locale.interface';
2
+ export type { LocaleConfig } from './locale.interface';
3
+ export type { LocaleRegexRule, LocaleContextualRule, LocaleDictionaryConfig, } from './locale.interface';
4
+ export type Locale = 'ru';
5
+ export declare const LOCALES: Record<Locale, LocaleConfig>;
6
+ export declare const DEFAULT_LOCALE: Locale;
7
+ export declare function getLocale(code?: string): LocaleConfig;
8
+ export { RU_LOCALE } from './ru';
9
+ //# sourceMappingURL=index.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/locales/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAGlD,YAAY,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AACvD,YAAY,EACV,eAAe,EACf,oBAAoB,EACpB,sBAAsB,GACvB,MAAM,oBAAoB,CAAC;AAG5B,MAAM,MAAM,MAAM,GAAG,IAAI,CAAC;AAO1B,eAAO,MAAM,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,YAAY,CAEhD,CAAC;AAEF,eAAO,MAAM,cAAc,EAAE,MAAa,CAAC;AAO3C,wBAAgB,SAAS,CAAC,IAAI,CAAC,EAAE,MAAM,GAAG,YAAY,CAErD;AAGD,OAAO,EAAE,SAAS,EAAE,MAAM,MAAM,CAAC"}
@@ -0,0 +1,15 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.RU_LOCALE = exports.DEFAULT_LOCALE = exports.LOCALES = void 0;
4
+ exports.getLocale = getLocale;
5
+ const ru_1 = require("./ru");
6
+ exports.LOCALES = {
7
+ ru: ru_1.RU_LOCALE,
8
+ };
9
+ exports.DEFAULT_LOCALE = 'ru';
10
+ function getLocale(code) {
11
+ return exports.LOCALES[code] ?? exports.LOCALES[exports.DEFAULT_LOCALE];
12
+ }
13
+ var ru_2 = require("./ru");
14
+ Object.defineProperty(exports, "RU_LOCALE", { enumerable: true, get: function () { return ru_2.RU_LOCALE; } });
15
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/locales/index.ts"],"names":[],"mappings":";;;AA6BA,8BAEC;AA9BD,6BAAiC;AAiBpB,QAAA,OAAO,GAAiC;IACnD,EAAE,EAAE,cAAS;CACd,CAAC;AAEW,QAAA,cAAc,GAAW,IAAI,CAAC;AAO3C,SAAgB,SAAS,CAAC,IAAa;IACrC,OAAO,eAAO,CAAC,IAAc,CAAC,IAAI,eAAO,CAAC,sBAAc,CAAC,CAAC;AAC5D,CAAC;AAGD,2BAAiC;AAAxB,+FAAA,SAAS,OAAA"}
@@ -0,0 +1,35 @@
1
+ import { DetectionContext } from '../types/detection-context';
2
+ import { EntityType } from '../types/entity-type';
3
+ export interface LocaleRegexRule {
4
+ type: EntityType;
5
+ pattern: RegExp;
6
+ contextPattern: RegExp | null;
7
+ contextRadius: number;
8
+ contextDirection: 'before' | 'both';
9
+ confidence: number;
10
+ }
11
+ export interface LocaleContextualRule {
12
+ id: string;
13
+ activeIn: DetectionContext[];
14
+ entityType: EntityType;
15
+ triggerPattern: RegExp;
16
+ capturePattern: RegExp;
17
+ direction: 'after' | 'before' | 'both';
18
+ maxDistance: number;
19
+ confidence: number;
20
+ }
21
+ export interface LocaleDictionaryConfig {
22
+ namesPath: string;
23
+ stopwordsPath: string;
24
+ tokenSplitPattern: RegExp;
25
+ }
26
+ export interface LocaleConfig {
27
+ code: string;
28
+ regexRules: LocaleRegexRule[];
29
+ contextualRules: LocaleContextualRule[];
30
+ dictionary: LocaleDictionaryConfig;
31
+ systemPrompt: {
32
+ buildInstruction: (activeTokens: string[]) => string;
33
+ };
34
+ }
35
+ //# sourceMappingURL=locale.interface.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"locale.interface.d.ts","sourceRoot":"","sources":["../../src/locales/locale.interface.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;AAC9D,OAAO,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAGlD,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,UAAU,CAAC;IAEjB,OAAO,EAAE,MAAM,CAAC;IAEhB,cAAc,EAAE,MAAM,GAAG,IAAI,CAAC;IAE9B,aAAa,EAAE,MAAM,CAAC;IAEtB,gBAAgB,EAAE,QAAQ,GAAG,MAAM,CAAC;IAEpC,UAAU,EAAE,MAAM,CAAC;CACpB;AAGD,MAAM,WAAW,oBAAoB;IAEnC,EAAE,EAAE,MAAM,CAAC;IAEX,QAAQ,EAAE,gBAAgB,EAAE,CAAC;IAE7B,UAAU,EAAE,UAAU,CAAC;IAEvB,cAAc,EAAE,MAAM,CAAC;IAEvB,cAAc,EAAE,MAAM,CAAC;IAEvB,SAAS,EAAE,OAAO,GAAG,QAAQ,GAAG,MAAM,CAAC;IAEvC,WAAW,EAAE,MAAM,CAAC;IAEpB,UAAU,EAAE,MAAM,CAAC;CACpB;AAGD,MAAM,WAAW,sBAAsB;IAKrC,SAAS,EAAE,MAAM,CAAC;IAElB,aAAa,EAAE,MAAM,CAAC;IAEtB,iBAAiB,EAAE,MAAM,CAAC;CAC3B;AAGD,MAAM,WAAW,YAAY;IAE3B,IAAI,EAAE,MAAM,CAAC;IAEb,UAAU,EAAE,eAAe,EAAE,CAAC;IAE9B,eAAe,EAAE,oBAAoB,EAAE,CAAC;IAExC,UAAU,EAAE,sBAAsB,CAAC;IAEnC,YAAY,EAAE;QAEZ,gBAAgB,EAAE,CAAC,YAAY,EAAE,MAAM,EAAE,KAAK,MAAM,CAAC;KACtD,CAAC;CACH"}
@@ -0,0 +1,3 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ //# sourceMappingURL=locale.interface.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"locale.interface.js","sourceRoot":"","sources":["../../src/locales/locale.interface.ts"],"names":[],"mappings":""}