@besales/anonymizer 0.1.0 → 0.2.0
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.
- package/README.md +11 -3
- package/dist/detectors/detection-pipeline.d.ts.map +1 -1
- package/dist/detectors/detection-pipeline.js +47 -1
- package/dist/detectors/detection-pipeline.js.map +1 -1
- package/dist/detectors/dictionary.detector.d.ts +2 -0
- package/dist/detectors/dictionary.detector.d.ts.map +1 -1
- package/dist/detectors/dictionary.detector.js +50 -4
- package/dist/detectors/dictionary.detector.js.map +1 -1
- package/dist/detectors/regex.detector.d.ts.map +1 -1
- package/dist/detectors/regex.detector.js +3 -0
- package/dist/detectors/regex.detector.js.map +1 -1
- package/dist/locales/locale.interface.d.ts +1 -0
- package/dist/locales/locale.interface.d.ts.map +1 -1
- package/dist/locales/ru/index.d.ts.map +1 -1
- package/dist/locales/ru/index.js +82 -0
- package/dist/locales/ru/index.js.map +1 -1
- package/dist/types/entity-type.d.ts +7 -0
- package/dist/types/entity-type.d.ts.map +1 -1
- package/dist/types/entity-type.js +7 -0
- package/dist/types/entity-type.js.map +1 -1
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -61,9 +61,17 @@ const { messages: restored } = deanonymizeMessages(llmResponseMessages, mappingD
|
|
|
61
61
|
|
|
62
62
|
## Детектируемые типы
|
|
63
63
|
|
|
64
|
-
`PERSON`, `
|
|
65
|
-
`ADDRESS`, `COMPANY`, `ACCOUNT
|
|
66
|
-
|
|
64
|
+
`PERSON`, `FIRST_NAME`, `LAST_NAME`, `PATRONYMIC`, `PHONE`, `EMAIL`, `INN`, `PASSPORT`,
|
|
65
|
+
`CARD`, `AMOUNT`, `DATE_OF_BIRTH`, `ADDRESS`, `COMPANY`, `ACCOUNT`, `SNILS`, `OGRN`,
|
|
66
|
+
`OGRNIP`, `SOCIAL_ID`. Локали: `ru` (имена — словарь; phone/email/card — языконезависимый
|
|
67
|
+
regex; СНИЛС/ОГРН/ОГРНИП — по контрольной сумме, ловятся без слова-метки; SOCIAL_ID —
|
|
68
|
+
VK/Telegram/Max @-хэндлы и ссылки на профиль). `excludeEntityTypes` сужает набор под профиль.
|
|
69
|
+
|
|
70
|
+
Имена, распознанные словарём, раскладываются на `FIRST_NAME` / `LAST_NAME` /
|
|
71
|
+
`PATRONYMIC` (нераспознанные — остаются `PERSON`). Это даёт **частичный pass-through**:
|
|
72
|
+
исключив `FIRST_NAME`, вы отдаёте реальное имя в LLM (модель склоняет его в нужный
|
|
73
|
+
падеж сама), а фамилию/отчество/контакты по-прежнему токенизируете. Выключение типа =
|
|
74
|
+
осознанная передача этих данных провайдеру модели.
|
|
67
75
|
|
|
68
76
|
## Потребители
|
|
69
77
|
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"detection-pipeline.d.ts","sourceRoot":"","sources":["../../src/detectors/detection-pipeline.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;AAC9D,OAAO,
|
|
1
|
+
{"version":3,"file":"detection-pipeline.d.ts","sourceRoot":"","sources":["../../src/detectors/detection-pipeline.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;AAC9D,OAAO,EAAe,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAgD/D,wBAAgB,oBAAoB,CAClC,IAAI,EAAE,MAAM,EACZ,OAAO,EAAE,gBAAgB,EACzB,MAAM,CAAC,EAAE,MAAM,EACf,kBAAkB,CAAC,EAAE,SAAS,UAAU,EAAE,GACzC,cAAc,EAAE,CAqBlB;AASD,wBAAgB,aAAa,CAAC,QAAQ,EAAE,cAAc,EAAE,GAAG,cAAc,EAAE,CA8D1E"}
|
|
@@ -3,6 +3,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
3
3
|
exports.runDetectionPipeline = runDetectionPipeline;
|
|
4
4
|
exports.mergeEntities = mergeEntities;
|
|
5
5
|
const constants_1 = require("../constants");
|
|
6
|
+
const entity_type_1 = require("../types/entity-type");
|
|
6
7
|
const locales_1 = require("../locales");
|
|
7
8
|
const contextual_detector_1 = require("./contextual.detector");
|
|
8
9
|
const dictionary_detector_1 = require("./dictionary.detector");
|
|
@@ -36,7 +37,7 @@ function runDetectionPipeline(text, context, locale, excludeEntityTypes) {
|
|
|
36
37
|
...pipeline.dict.detect(text, context),
|
|
37
38
|
...pipeline.ctx.detect(text, context),
|
|
38
39
|
];
|
|
39
|
-
const merged = mergeEntities(all);
|
|
40
|
+
const merged = refinePersonRoles(mergeEntities(all), loc);
|
|
40
41
|
const excluded = excludeEntityTypes && excludeEntityTypes.length > 0
|
|
41
42
|
? new Set(excludeEntityTypes)
|
|
42
43
|
: null;
|
|
@@ -99,4 +100,49 @@ function pickWinner(a, b) {
|
|
|
99
100
|
}
|
|
100
101
|
return a.confidence >= b.confidence ? a : b;
|
|
101
102
|
}
|
|
103
|
+
function refinePersonRoles(entities, loc) {
|
|
104
|
+
const result = [];
|
|
105
|
+
for (const entity of entities) {
|
|
106
|
+
if (entity.type !== entity_type_1.ENTITY_TYPE.PERSON) {
|
|
107
|
+
result.push(entity);
|
|
108
|
+
continue;
|
|
109
|
+
}
|
|
110
|
+
const words = splitWordsWithOffsets(entity.value, loc.dictionary.tokenSplitPattern);
|
|
111
|
+
const roles = words.length > 0
|
|
112
|
+
? (0, dictionary_detector_1.classifyNameTokens)(loc.code, loc.dictionary, words.map((w) => w.value))
|
|
113
|
+
: null;
|
|
114
|
+
if (!roles) {
|
|
115
|
+
result.push(entity);
|
|
116
|
+
continue;
|
|
117
|
+
}
|
|
118
|
+
for (let i = 0; i < words.length; i++) {
|
|
119
|
+
const word = words[i];
|
|
120
|
+
result.push({
|
|
121
|
+
type: roles[i],
|
|
122
|
+
value: word.value,
|
|
123
|
+
position: {
|
|
124
|
+
start: entity.position.start + word.start,
|
|
125
|
+
end: entity.position.start + word.end,
|
|
126
|
+
},
|
|
127
|
+
detectedBy: entity.detectedBy,
|
|
128
|
+
confidence: entity.confidence,
|
|
129
|
+
});
|
|
130
|
+
}
|
|
131
|
+
}
|
|
132
|
+
return result;
|
|
133
|
+
}
|
|
134
|
+
function splitWordsWithOffsets(text, splitPattern) {
|
|
135
|
+
const tokens = [];
|
|
136
|
+
let searchFrom = 0;
|
|
137
|
+
for (const part of text.split(splitPattern)) {
|
|
138
|
+
if (!part)
|
|
139
|
+
continue;
|
|
140
|
+
const idx = text.indexOf(part, searchFrom);
|
|
141
|
+
if (idx === -1)
|
|
142
|
+
continue;
|
|
143
|
+
tokens.push({ value: part, start: idx, end: idx + part.length });
|
|
144
|
+
searchFrom = idx + part.length;
|
|
145
|
+
}
|
|
146
|
+
return tokens;
|
|
147
|
+
}
|
|
102
148
|
//# sourceMappingURL=detection-pipeline.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"detection-pipeline.js","sourceRoot":"","sources":["../../src/detectors/detection-pipeline.ts"],"names":[],"mappings":";;AAmDA,oDA0BC;AASD,sCA8DC;AApJD,4CAAwD;
|
|
1
|
+
{"version":3,"file":"detection-pipeline.js","sourceRoot":"","sources":["../../src/detectors/detection-pipeline.ts"],"names":[],"mappings":";;AAmDA,oDA0BC;AASD,sCA8DC;AApJD,4CAAwD;AAGxD,sDAA+D;AAC/D,wCAAuC;AAEvC,+DAA2D;AAC3D,+DAA+E;AAC/E,qDAAiD;AAGjD,MAAM,iBAAiB,GAAiD;IACtE,KAAK,EAAE,CAAC;IACR,UAAU,EAAE,CAAC;IACb,UAAU,EAAE,CAAC;CACd,CAAC;AASF,MAAM,cAAc,GAAG,IAAI,GAAG,EAA4B,CAAC;AAE3D,SAAS,aAAa,CAAC,MAAoB;IACzC,OAAO;QACL,KAAK,EAAE,IAAI,8BAAa,CAAC,MAAM,CAAC,UAAU,CAAC;QAC3C,IAAI,EAAE,IAAI,wCAAkB,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,UAAU,CAAC;QAC5D,GAAG,EAAE,IAAI,wCAAkB,CAAC,MAAM,CAAC,eAAe,CAAC;KACpD,CAAC;AACJ,CAAC;AAED,SAAS,WAAW,CAAC,UAAkB;IACrC,IAAI,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IAC9C,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,QAAQ,GAAG,aAAa,CAAC,IAAA,mBAAS,EAAC,UAAU,CAAC,CAAC,CAAC;QAChD,cAAc,CAAC,GAAG,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAC3C,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAUD,SAAgB,oBAAoB,CAClC,IAAY,EACZ,OAAyB,EACzB,MAAe,EACf,kBAA0C;IAE1C,MAAM,GAAG,GAAG,IAAA,mBAAS,EAAC,MAAM,CAAC,CAAC;IAC9B,MAAM,QAAQ,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAEvC,MAAM,GAAG,GAAG;QACV,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC;QACvC,GAAG,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC;QACtC,GAAG,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC;KACtC,CAAC;IAEF,MAAM,MAAM,GAAG,iBAAiB,CAAC,aAAa,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC;IAC1D,MAAM,QAAQ,GACZ,kBAAkB,IAAI,kBAAkB,CAAC,MAAM,GAAG,CAAC;QACjD,CAAC,CAAC,IAAI,GAAG,CAAa,kBAAkB,CAAC;QACzC,CAAC,CAAC,IAAI,CAAC;IAEX,OAAO,MAAM,CAAC,MAAM,CAClB,CAAC,CAAC,EAAE,EAAE,CACJ,CAAC,CAAC,UAAU,IAAI,oCAAwB;QACxC,CAAC,QAAQ,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAC/C,CAAC;AACJ,CAAC;AASD,SAAgB,aAAa,CAAC,QAA0B;IACtD,IAAI,QAAQ,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,CAAC,GAAG,QAAQ,CAAC,CAAC;IAG/C,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,EAAE,CAAC;IAEpC,KAAK,MAAM,MAAM,IAAI,MAAM,EAAE,CAAC;QAC5B,IAAI,WAAW,GAAG,KAAK,CAAC;QACxB,IAAI,aAAa,GAAG,CAAC,CAAC,CAAC;QAEvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YAG3B,IACE,MAAM,CAAC,QAAQ,CAAC,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,GAAG;gBAC7C,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,QAAQ,CAAC,QAAQ,CAAC,KAAK,EAC7C,CAAC;gBACD,WAAW,GAAG,IAAI,CAAC;gBACnB,aAAa,GAAG,CAAC,CAAC;gBAClB,MAAM;YACR,CAAC;QACH,CAAC;QAED,IAAI,CAAC,WAAW,EAAE,CAAC;YACjB,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpB,SAAS;QACX,CAAC;QAED,MAAM,QAAQ,GAAG,MAAM,CAAC,aAAa,CAAC,CAAC;QAGvC,IACE,MAAM,CAAC,QAAQ,CAAC,KAAK,IAAI,QAAQ,CAAC,QAAQ,CAAC,KAAK;YAChD,MAAM,CAAC,QAAQ,CAAC,GAAG,IAAI,QAAQ,CAAC,QAAQ,CAAC,GAAG,EAC5C,CAAC;YACD,SAAS;QACX,CAAC;QAGD,IACE,QAAQ,CAAC,QAAQ,CAAC,KAAK,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK;YAChD,QAAQ,CAAC,QAAQ,CAAC,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,GAAG,EAC5C,CAAC;YACD,MAAM,CAAC,aAAa,CAAC,GAAG,MAAM,CAAC;YAC/B,SAAS;QACX,CAAC;QAGD,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;QAC5C,MAAM,CAAC,aAAa,CAAC,GAAG,MAAM,CAAC;IACjC,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,UAAU,CAAC,CAAiB,EAAE,CAAiB;IACtD,MAAM,SAAS,GAAG,iBAAiB,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAClD,MAAM,SAAS,GAAG,iBAAiB,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAGlD,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC5B,OAAO,SAAS,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC;IAED,MAAM,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC;IAC/C,MAAM,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC;IAE/C,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;QAClB,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7B,CAAC;IAED,OAAO,CAAC,CAAC,UAAU,IAAI,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC9C,CAAC;AAcD,SAAS,iBAAiB,CACxB,QAA0B,EAC1B,GAAiB;IAEjB,MAAM,MAAM,GAAqB,EAAE,CAAC;IAEpC,KAAK,MAAM,MAAM,IAAI,QAAQ,EAAE,CAAC;QAC9B,IAAI,MAAM,CAAC,IAAI,KAAK,yBAAW,CAAC,MAAM,EAAE,CAAC;YACvC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpB,SAAS;QACX,CAAC;QAED,MAAM,KAAK,GAAG,qBAAqB,CACjC,MAAM,CAAC,KAAK,EACZ,GAAG,CAAC,UAAU,CAAC,iBAAiB,CACjC,CAAC;QACF,MAAM,KAAK,GACT,KAAK,CAAC,MAAM,GAAG,CAAC;YACd,CAAC,CAAC,IAAA,wCAAkB,EAChB,GAAG,CAAC,IAAI,EACR,GAAG,CAAC,UAAU,EACd,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAC1B;YACH,CAAC,CAAC,IAAI,CAAC;QAEX,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACpB,SAAS;QACX,CAAC;QAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACtB,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;gBACd,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,QAAQ,EAAE;oBACR,KAAK,EAAE,MAAM,CAAC,QAAQ,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK;oBACzC,GAAG,EAAE,MAAM,CAAC,QAAQ,CAAC,KAAK,GAAG,IAAI,CAAC,GAAG;iBACtC;gBACD,UAAU,EAAE,MAAM,CAAC,UAAU;gBAC7B,UAAU,EAAE,MAAM,CAAC,UAAU;aAC9B,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAGD,SAAS,qBAAqB,CAC5B,IAAY,EACZ,YAAoB;IAEpB,MAAM,MAAM,GAAoD,EAAE,CAAC;IACnE,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC;QAC5C,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;QACzB,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC;QACjE,UAAU,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC;IACjC,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -1,8 +1,10 @@
|
|
|
1
1
|
import { DetectedEntity } from '../types/detected-entity.interface';
|
|
2
2
|
import { DetectionContext } from '../types/detection-context';
|
|
3
|
+
import { EntityType } from '../types/entity-type';
|
|
3
4
|
import { LocaleDictionaryConfig } from '../locales/locale.interface';
|
|
4
5
|
import { DetectorPort } from './detector.interface';
|
|
5
6
|
export declare function resetDictionaryCache(): void;
|
|
7
|
+
export declare function classifyNameTokens(localeCode: string, config: LocaleDictionaryConfig, words: string[]): EntityType[] | null;
|
|
6
8
|
export declare class DictionaryDetector implements DetectorPort {
|
|
7
9
|
private readonly localeCode;
|
|
8
10
|
private readonly config;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"dictionary.detector.d.ts","sourceRoot":"","sources":["../../src/detectors/dictionary.detector.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;
|
|
1
|
+
{"version":3,"file":"dictionary.detector.d.ts","sourceRoot":"","sources":["../../src/detectors/dictionary.detector.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAC;AAC9D,OAAO,EAAe,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAC/D,OAAO,EAAE,sBAAsB,EAAE,MAAM,6BAA6B,CAAC;AACrE,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AA+DpD,wBAAgB,oBAAoB,IAAI,IAAI,CAE3C;AAcD,wBAAgB,kBAAkB,CAChC,UAAU,EAAE,MAAM,EAClB,MAAM,EAAE,sBAAsB,EAC9B,KAAK,EAAE,MAAM,EAAE,GACd,UAAU,EAAE,GAAG,IAAI,CA2CrB;AAwCD,qBAAa,kBAAmB,YAAW,YAAY;IAEnD,OAAO,CAAC,QAAQ,CAAC,UAAU;IAC3B,OAAO,CAAC,QAAQ,CAAC,MAAM;gBADN,UAAU,EAAE,MAAM,EAClB,MAAM,EAAE,sBAAsB;IAGjD,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,gBAAgB,GAAG,cAAc,EAAE;CAqHnE"}
|
|
@@ -35,6 +35,7 @@ var __importStar = (this && this.__importStar) || (function () {
|
|
|
35
35
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
36
36
|
exports.DictionaryDetector = void 0;
|
|
37
37
|
exports.resetDictionaryCache = resetDictionaryCache;
|
|
38
|
+
exports.classifyNameTokens = classifyNameTokens;
|
|
38
39
|
const fs = __importStar(require("fs"));
|
|
39
40
|
const entity_type_1 = require("../types/entity-type");
|
|
40
41
|
const LOCALE_CACHE = new Map();
|
|
@@ -74,6 +75,47 @@ function ensureCache(localeCode, config) {
|
|
|
74
75
|
function resetDictionaryCache() {
|
|
75
76
|
LOCALE_CACHE.clear();
|
|
76
77
|
}
|
|
78
|
+
function classifyNameTokens(localeCode, config, words) {
|
|
79
|
+
const cache = ensureCache(localeCode, config);
|
|
80
|
+
const lower = words.map((w) => w.toLowerCase());
|
|
81
|
+
const isFirst = (w) => cache.firstNames.has(w);
|
|
82
|
+
const isLast = (w) => cache.lastNames.has(w);
|
|
83
|
+
const isPatronymic = (w) => cache.patronymics.has(w);
|
|
84
|
+
if (lower.length === 3) {
|
|
85
|
+
const [w0, w1, w2] = lower;
|
|
86
|
+
if (isLast(w0) && isFirst(w1) && isPatronymic(w2)) {
|
|
87
|
+
return [entity_type_1.ENTITY_TYPE.LAST_NAME, entity_type_1.ENTITY_TYPE.FIRST_NAME, entity_type_1.ENTITY_TYPE.PATRONYMIC];
|
|
88
|
+
}
|
|
89
|
+
if (isFirst(w0) && isPatronymic(w1) && isLast(w2)) {
|
|
90
|
+
return [entity_type_1.ENTITY_TYPE.FIRST_NAME, entity_type_1.ENTITY_TYPE.PATRONYMIC, entity_type_1.ENTITY_TYPE.LAST_NAME];
|
|
91
|
+
}
|
|
92
|
+
return null;
|
|
93
|
+
}
|
|
94
|
+
if (lower.length === 2) {
|
|
95
|
+
const [w0, w1] = lower;
|
|
96
|
+
if (isLast(w0) && isFirst(w1)) {
|
|
97
|
+
return [entity_type_1.ENTITY_TYPE.LAST_NAME, entity_type_1.ENTITY_TYPE.FIRST_NAME];
|
|
98
|
+
}
|
|
99
|
+
if (isFirst(w0) && isLast(w1)) {
|
|
100
|
+
return [entity_type_1.ENTITY_TYPE.FIRST_NAME, entity_type_1.ENTITY_TYPE.LAST_NAME];
|
|
101
|
+
}
|
|
102
|
+
if (isFirst(w0) && isPatronymic(w1)) {
|
|
103
|
+
return [entity_type_1.ENTITY_TYPE.FIRST_NAME, entity_type_1.ENTITY_TYPE.PATRONYMIC];
|
|
104
|
+
}
|
|
105
|
+
return null;
|
|
106
|
+
}
|
|
107
|
+
if (lower.length === 1) {
|
|
108
|
+
const [w0] = lower;
|
|
109
|
+
if (isFirst(w0))
|
|
110
|
+
return [entity_type_1.ENTITY_TYPE.FIRST_NAME];
|
|
111
|
+
if (isLast(w0))
|
|
112
|
+
return [entity_type_1.ENTITY_TYPE.LAST_NAME];
|
|
113
|
+
if (isPatronymic(w0))
|
|
114
|
+
return [entity_type_1.ENTITY_TYPE.PATRONYMIC];
|
|
115
|
+
return null;
|
|
116
|
+
}
|
|
117
|
+
return null;
|
|
118
|
+
}
|
|
77
119
|
function tokenizeText(text, splitPattern) {
|
|
78
120
|
const tokens = [];
|
|
79
121
|
const parts = text.split(splitPattern);
|
|
@@ -124,13 +166,17 @@ class DictionaryDetector {
|
|
|
124
166
|
const t0Lower = t0.value.toLowerCase();
|
|
125
167
|
const t1Lower = t1?.value.toLowerCase();
|
|
126
168
|
const t2Lower = t2?.value.toLowerCase();
|
|
127
|
-
|
|
169
|
+
const isFullFio = !!t1 && !!t2 &&
|
|
128
170
|
isCapitalized(t0.value) &&
|
|
129
171
|
isCapitalized(t1.value) &&
|
|
130
172
|
isCapitalized(t2.value) &&
|
|
131
|
-
cache.lastNames.has(t0Lower) &&
|
|
132
|
-
|
|
133
|
-
|
|
173
|
+
((cache.lastNames.has(t0Lower) &&
|
|
174
|
+
cache.firstNames.has(t1Lower) &&
|
|
175
|
+
cache.patronymics.has(t2Lower)) ||
|
|
176
|
+
(cache.firstNames.has(t0Lower) &&
|
|
177
|
+
cache.patronymics.has(t1Lower) &&
|
|
178
|
+
cache.lastNames.has(t2Lower)));
|
|
179
|
+
if (isFullFio && t2) {
|
|
134
180
|
const value = text.slice(t0.start, t2.end);
|
|
135
181
|
results.push({
|
|
136
182
|
type: entity_type_1.ENTITY_TYPE.PERSON,
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"dictionary.detector.js","sourceRoot":"","sources":["../../src/detectors/dictionary.detector.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAoEA,oDAEC;
|
|
1
|
+
{"version":3,"file":"dictionary.detector.js","sourceRoot":"","sources":["../../src/detectors/dictionary.detector.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAoEA,oDAEC;AAcD,gDA+CC;AAnID,uCAAyB;AAGzB,sDAA+D;AAqB/D,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;AAcD,SAAgB,kBAAkB,CAChC,UAAkB,EAClB,MAA8B,EAC9B,KAAe;IAEf,MAAM,KAAK,GAAG,WAAW,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;IAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IAChD,MAAM,OAAO,GAAG,CAAC,CAAS,EAAW,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,MAAM,GAAG,CAAC,CAAS,EAAW,EAAE,CAAC,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IAC9D,MAAM,YAAY,GAAG,CAAC,CAAS,EAAW,EAAE,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IAEtE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,GAAG,KAAK,CAAC;QAE3B,IAAI,MAAM,CAAC,EAAE,CAAC,IAAI,OAAO,CAAC,EAAE,CAAC,IAAI,YAAY,CAAC,EAAE,CAAC,EAAE,CAAC;YAClD,OAAO,CAAC,yBAAW,CAAC,SAAS,EAAE,yBAAW,CAAC,UAAU,EAAE,yBAAW,CAAC,UAAU,CAAC,CAAC;QACjF,CAAC;QAED,IAAI,OAAO,CAAC,EAAE,CAAC,IAAI,YAAY,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,CAAC;YAClD,OAAO,CAAC,yBAAW,CAAC,UAAU,EAAE,yBAAW,CAAC,UAAU,EAAE,yBAAW,CAAC,SAAS,CAAC,CAAC;QACjF,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,CAAC,EAAE,EAAE,EAAE,CAAC,GAAG,KAAK,CAAC;QACvB,IAAI,MAAM,CAAC,EAAE,CAAC,IAAI,OAAO,CAAC,EAAE,CAAC,EAAE,CAAC;YAC9B,OAAO,CAAC,yBAAW,CAAC,SAAS,EAAE,yBAAW,CAAC,UAAU,CAAC,CAAC;QACzD,CAAC;QACD,IAAI,OAAO,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,CAAC;YAC9B,OAAO,CAAC,yBAAW,CAAC,UAAU,EAAE,yBAAW,CAAC,SAAS,CAAC,CAAC;QACzD,CAAC;QACD,IAAI,OAAO,CAAC,EAAE,CAAC,IAAI,YAAY,CAAC,EAAE,CAAC,EAAE,CAAC;YACpC,OAAO,CAAC,yBAAW,CAAC,UAAU,EAAE,yBAAW,CAAC,UAAU,CAAC,CAAC;QAC1D,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC;QACnB,IAAI,OAAO,CAAC,EAAE,CAAC;YAAE,OAAO,CAAC,yBAAW,CAAC,UAAU,CAAC,CAAC;QACjD,IAAI,MAAM,CAAC,EAAE,CAAC;YAAE,OAAO,CAAC,yBAAW,CAAC,SAAS,CAAC,CAAC;QAC/C,IAAI,YAAY,CAAC,EAAE,CAAC;YAAE,OAAO,CAAC,yBAAW,CAAC,UAAU,CAAC,CAAC;QACtD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACd,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;YAKxC,MAAM,SAAS,GACb,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,EAAE;gBACZ,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBACvB,aAAa,CAAC,EAAE,CAAC,KAAK,CAAC;gBAEvB,CAAC,CAAC,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC;oBAC5B,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC;oBAC7B,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;oBAE/B,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC;wBAC5B,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC;wBAC9B,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;YAErC,IAAI,SAAS,IAAI,EAAE,EAAE,CAAC;gBACpB,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;AA3HD,gDA2HC;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"}
|
|
@@ -1 +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;
|
|
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;CAkCnE"}
|
|
@@ -1 +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;
|
|
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;gBAGD,IAAI,IAAI,CAAC,QAAQ,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;oBAC3C,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;AAzCD,sCAyCC"}
|
|
@@ -1 +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;
|
|
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;IAMnB,QAAQ,CAAC,EAAE,CAAC,GAAG,EAAE,MAAM,KAAK,OAAO,CAAC;CACrC;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"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/locales/ru/index.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/locales/ru/index.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAgInD,eAAO,MAAM,SAAS,EAAE,YAsMvB,CAAC"}
|
package/dist/locales/ru/index.js
CHANGED
|
@@ -50,6 +50,42 @@ const AMOUNT_PATTERN = /\b\d[\d\s.,]{0,20}?\s*(?:руб|₽|\$|€|[KкКk]|мл
|
|
|
50
50
|
const AMOUNT_CONTEXT = /сделк|бюджет|чек|оплат|стоимост|цен|сумм|платёж|платеж/i;
|
|
51
51
|
const ACCOUNT_PATTERN = /\b\d{20}\b/g;
|
|
52
52
|
const ACCOUNT_CONTEXT = /сч[её]т|р[/.]с|расч[её]тн/i;
|
|
53
|
+
const SNILS_PATTERN = /\b\d{3}[\s-]?\d{3}[\s-]?\d{3}[\s-]?\d{2}\b/g;
|
|
54
|
+
const OGRN_PATTERN = /\b\d{13}\b/g;
|
|
55
|
+
const OGRNIP_PATTERN = /\b\d{15}\b/g;
|
|
56
|
+
const SOCIAL_HANDLE_PATTERN = /(?<![\w@/.])@[a-zA-Z][a-zA-Z0-9_]{3,31}\b/g;
|
|
57
|
+
const SOCIAL_URL_PATTERN = /(?:https?:\/\/)?(?:t\.me|telegram\.me|vk\.com|vk\.ru|max\.ru)\/[A-Za-z0-9_.+-]+/gi;
|
|
58
|
+
function modStr(digits, m) {
|
|
59
|
+
let r = 0;
|
|
60
|
+
for (let i = 0; i < digits.length; i++) {
|
|
61
|
+
r = (r * 10 + (digits.charCodeAt(i) - 48)) % m;
|
|
62
|
+
}
|
|
63
|
+
return r;
|
|
64
|
+
}
|
|
65
|
+
function validateSnils(raw) {
|
|
66
|
+
const d = raw.replace(/\D/g, '');
|
|
67
|
+
if (d.length !== 11)
|
|
68
|
+
return false;
|
|
69
|
+
let sum = 0;
|
|
70
|
+
for (let i = 0; i < 9; i++)
|
|
71
|
+
sum += (d.charCodeAt(i) - 48) * (9 - i);
|
|
72
|
+
let control = sum % 101;
|
|
73
|
+
if (control === 100)
|
|
74
|
+
control = 0;
|
|
75
|
+
return control === Number(d.slice(9));
|
|
76
|
+
}
|
|
77
|
+
function validateOgrn(raw) {
|
|
78
|
+
const d = raw.replace(/\D/g, '');
|
|
79
|
+
if (d.length !== 13)
|
|
80
|
+
return false;
|
|
81
|
+
return modStr(d.slice(0, 12), 11) % 10 === d.charCodeAt(12) - 48;
|
|
82
|
+
}
|
|
83
|
+
function validateOgrnip(raw) {
|
|
84
|
+
const d = raw.replace(/\D/g, '');
|
|
85
|
+
if (d.length !== 15)
|
|
86
|
+
return false;
|
|
87
|
+
return modStr(d.slice(0, 14), 13) % 10 === d.charCodeAt(14) - 48;
|
|
88
|
+
}
|
|
53
89
|
const CONTEXT_RADIUS_INN = 30;
|
|
54
90
|
const CONTEXT_RADIUS_PASSPORT = 50;
|
|
55
91
|
const CONTEXT_RADIUS_DOB = 50;
|
|
@@ -61,6 +97,9 @@ const CONFIDENCE_PASSPORT = 0.85;
|
|
|
61
97
|
const CONFIDENCE_DOB = 0.85;
|
|
62
98
|
const CONFIDENCE_AMOUNT = 0.80;
|
|
63
99
|
const CONFIDENCE_ACCOUNT = 0.90;
|
|
100
|
+
const CONFIDENCE_SNILS = 0.92;
|
|
101
|
+
const CONFIDENCE_OGRN = 0.92;
|
|
102
|
+
const CONFIDENCE_SOCIAL = 0.85;
|
|
64
103
|
const MAX_EXAMPLE_TOKENS = 5;
|
|
65
104
|
function buildRuInstruction(activeTokens) {
|
|
66
105
|
if (activeTokens.length === 0)
|
|
@@ -144,6 +183,49 @@ exports.RU_LOCALE = {
|
|
|
144
183
|
contextDirection: 'both',
|
|
145
184
|
confidence: CONFIDENCE_ACCOUNT,
|
|
146
185
|
},
|
|
186
|
+
{
|
|
187
|
+
type: entity_type_1.ENTITY_TYPE.SNILS,
|
|
188
|
+
pattern: SNILS_PATTERN,
|
|
189
|
+
contextPattern: null,
|
|
190
|
+
contextRadius: 0,
|
|
191
|
+
contextDirection: 'both',
|
|
192
|
+
confidence: CONFIDENCE_SNILS,
|
|
193
|
+
validate: validateSnils,
|
|
194
|
+
},
|
|
195
|
+
{
|
|
196
|
+
type: entity_type_1.ENTITY_TYPE.OGRN,
|
|
197
|
+
pattern: OGRN_PATTERN,
|
|
198
|
+
contextPattern: null,
|
|
199
|
+
contextRadius: 0,
|
|
200
|
+
contextDirection: 'both',
|
|
201
|
+
confidence: CONFIDENCE_OGRN,
|
|
202
|
+
validate: validateOgrn,
|
|
203
|
+
},
|
|
204
|
+
{
|
|
205
|
+
type: entity_type_1.ENTITY_TYPE.OGRNIP,
|
|
206
|
+
pattern: OGRNIP_PATTERN,
|
|
207
|
+
contextPattern: null,
|
|
208
|
+
contextRadius: 0,
|
|
209
|
+
contextDirection: 'both',
|
|
210
|
+
confidence: CONFIDENCE_OGRN,
|
|
211
|
+
validate: validateOgrnip,
|
|
212
|
+
},
|
|
213
|
+
{
|
|
214
|
+
type: entity_type_1.ENTITY_TYPE.SOCIAL_ID,
|
|
215
|
+
pattern: SOCIAL_HANDLE_PATTERN,
|
|
216
|
+
contextPattern: null,
|
|
217
|
+
contextRadius: 0,
|
|
218
|
+
contextDirection: 'both',
|
|
219
|
+
confidence: CONFIDENCE_SOCIAL,
|
|
220
|
+
},
|
|
221
|
+
{
|
|
222
|
+
type: entity_type_1.ENTITY_TYPE.SOCIAL_ID,
|
|
223
|
+
pattern: SOCIAL_URL_PATTERN,
|
|
224
|
+
contextPattern: null,
|
|
225
|
+
contextRadius: 0,
|
|
226
|
+
contextDirection: 'both',
|
|
227
|
+
confidence: CONFIDENCE_SOCIAL,
|
|
228
|
+
},
|
|
147
229
|
],
|
|
148
230
|
contextualRules: [
|
|
149
231
|
{
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/locales/ru/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,2CAA6B;AAC7B,qEAAkE;AAClE,yDAAsD;AAMtD,MAAM,aAAa,GACjB,sEAAsE,CAAC;AAGzE,MAAM,aAAa,GACjB,8DAA8D,CAAC;AAGjE,MAAM,WAAW,GAAG,uBAAuB,CAAC;AAC5C,MAAM,WAAW,GAAG,cAAc,CAAC;AAGnC,MAAM,YAAY,GAAG,6CAA6C,CAAC;AAGnE,MAAM,gBAAgB,GAAG,4BAA4B,CAAC;AACtD,MAAM,gBAAgB,GAAG,yBAAyB,CAAC;AAGnD,MAAM,WAAW,GAAG,8BAA8B,CAAC;AACnD,MAAM,WAAW,GAAG,wBAAwB,CAAC;AAG7C,MAAM,cAAc,GAClB,qFAAqF,CAAC;AACxF,MAAM,cAAc,GAClB,yDAAyD,CAAC;AAG5D,MAAM,eAAe,GAAG,aAAa,CAAC;AACtC,MAAM,eAAe,GAAG,4BAA4B,CAAC;
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/locales/ru/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,2CAA6B;AAC7B,qEAAkE;AAClE,yDAAsD;AAMtD,MAAM,aAAa,GACjB,sEAAsE,CAAC;AAGzE,MAAM,aAAa,GACjB,8DAA8D,CAAC;AAGjE,MAAM,WAAW,GAAG,uBAAuB,CAAC;AAC5C,MAAM,WAAW,GAAG,cAAc,CAAC;AAGnC,MAAM,YAAY,GAAG,6CAA6C,CAAC;AAGnE,MAAM,gBAAgB,GAAG,4BAA4B,CAAC;AACtD,MAAM,gBAAgB,GAAG,yBAAyB,CAAC;AAGnD,MAAM,WAAW,GAAG,8BAA8B,CAAC;AACnD,MAAM,WAAW,GAAG,wBAAwB,CAAC;AAG7C,MAAM,cAAc,GAClB,qFAAqF,CAAC;AACxF,MAAM,cAAc,GAClB,yDAAyD,CAAC;AAG5D,MAAM,eAAe,GAAG,aAAa,CAAC;AACtC,MAAM,eAAe,GAAG,4BAA4B,CAAC;AAGrD,MAAM,aAAa,GAAG,6CAA6C,CAAC;AAEpE,MAAM,YAAY,GAAG,aAAa,CAAC;AACnC,MAAM,cAAc,GAAG,aAAa,CAAC;AAErC,MAAM,qBAAqB,GAAG,4CAA4C,CAAC;AAC3E,MAAM,kBAAkB,GACtB,mFAAmF,CAAC;AAGtF,SAAS,MAAM,CAAC,MAAc,EAAE,CAAS;IACvC,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC;IACjD,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAGD,SAAS,aAAa,CAAC,GAAW;IAChC,MAAM,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACjC,IAAI,CAAC,CAAC,MAAM,KAAK,EAAE;QAAE,OAAO,KAAK,CAAC;IAClC,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;QAAE,GAAG,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACpE,IAAI,OAAO,GAAG,GAAG,GAAG,GAAG,CAAC;IACxB,IAAI,OAAO,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,CAAC;IACjC,OAAO,OAAO,KAAK,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;AACxC,CAAC;AAGD,SAAS,YAAY,CAAC,GAAW;IAC/B,MAAM,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACjC,IAAI,CAAC,CAAC,MAAM,KAAK,EAAE;QAAE,OAAO,KAAK,CAAC;IAClC,OAAO,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,UAAU,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;AACnE,CAAC;AAGD,SAAS,cAAc,CAAC,GAAW;IACjC,MAAM,CAAC,GAAG,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACjC,IAAI,CAAC,CAAC,MAAM,KAAK,EAAE;QAAE,OAAO,KAAK,CAAC;IAClC,OAAO,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,UAAU,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;AACnE,CAAC;AAED,MAAM,kBAAkB,GAAG,EAAE,CAAC;AAC9B,MAAM,uBAAuB,GAAG,EAAE,CAAC;AACnC,MAAM,kBAAkB,GAAG,EAAE,CAAC;AAC9B,MAAM,qBAAqB,GAAG,EAAE,CAAC;AACjC,MAAM,sBAAsB,GAAG,EAAE,CAAC;AAElC,MAAM,eAAe,GAAG,IAAI,CAAC;AAC7B,MAAM,cAAc,GAAG,IAAI,CAAC;AAC5B,MAAM,mBAAmB,GAAG,IAAI,CAAC;AACjC,MAAM,cAAc,GAAG,IAAI,CAAC;AAC5B,MAAM,iBAAiB,GAAG,IAAI,CAAC;AAC/B,MAAM,kBAAkB,GAAG,IAAI,CAAC;AAChC,MAAM,gBAAgB,GAAG,IAAI,CAAC;AAC9B,MAAM,eAAe,GAAG,IAAI,CAAC;AAC7B,MAAM,iBAAiB,GAAG,IAAI,CAAC;AAY/B,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAE7B,SAAS,kBAAkB,CAAC,YAAsB;IAChD,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEzC,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAEtE,OAAO;;;kCAGyB,QAAQ;;;;;;IAMtC,CAAC;AACL,CAAC;AAIY,QAAA,SAAS,GAAiB;IACrC,IAAI,EAAE,IAAI;IAEV,UAAU,EAAE;QACV;YACE,IAAI,EAAE,yBAAW,CAAC,KAAK;YACvB,OAAO,EAAE,aAAa;YACtB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,eAAe;SAC5B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,KAAK;YACvB,OAAO,EAAE,aAAa;YACtB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,eAAe;SAC5B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,GAAG;YACrB,OAAO,EAAE,WAAW;YACpB,cAAc,EAAE,WAAW;YAC3B,aAAa,EAAE,kBAAkB;YACjC,gBAAgB,EAAE,QAAQ;YAC1B,UAAU,EAAE,cAAc;SAC3B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,IAAI;YACtB,OAAO,EAAE,YAAY;YACrB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,eAAe;SAC5B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,QAAQ;YAC1B,OAAO,EAAE,gBAAgB;YACzB,cAAc,EAAE,gBAAgB;YAChC,aAAa,EAAE,uBAAuB;YACtC,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,mBAAmB;SAChC;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,aAAa;YAC/B,OAAO,EAAE,WAAW;YACpB,cAAc,EAAE,WAAW;YAC3B,aAAa,EAAE,kBAAkB;YACjC,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,cAAc;SAC3B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,MAAM;YACxB,OAAO,EAAE,cAAc;YACvB,cAAc,EAAE,cAAc;YAC9B,aAAa,EAAE,qBAAqB;YACpC,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,iBAAiB;SAC9B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,OAAO;YACzB,OAAO,EAAE,eAAe;YACxB,cAAc,EAAE,eAAe;YAC/B,aAAa,EAAE,sBAAsB;YACrC,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,kBAAkB;SAC/B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,KAAK;YACvB,OAAO,EAAE,aAAa;YACtB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,gBAAgB;YAC5B,QAAQ,EAAE,aAAa;SACxB;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,IAAI;YACtB,OAAO,EAAE,YAAY;YACrB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,eAAe;YAC3B,QAAQ,EAAE,YAAY;SACvB;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,MAAM;YACxB,OAAO,EAAE,cAAc;YACvB,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,eAAe;YAC3B,QAAQ,EAAE,cAAc;SACzB;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,SAAS;YAC3B,OAAO,EAAE,qBAAqB;YAC9B,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,iBAAiB;SAC9B;QACD;YACE,IAAI,EAAE,yBAAW,CAAC,SAAS;YAC3B,OAAO,EAAE,kBAAkB;YAC3B,cAAc,EAAE,IAAI;YACpB,aAAa,EAAE,CAAC;YAChB,gBAAgB,EAAE,MAAM;YACxB,UAAU,EAAE,iBAAiB;SAC9B;KACF;IAED,eAAe,EAAE;QAEf;YACE,EAAE,EAAE,IAAI;YACR,QAAQ,EAAE;gBACR,qCAAiB,CAAC,qBAAqB;gBACvC,qCAAiB,CAAC,aAAa;gBAC/B,qCAAiB,CAAC,iBAAiB;aACpC;YACD,UAAU,EAAE,yBAAW,CAAC,MAAM;YAC9B,cAAc,EAAE,uCAAuC;YACvD,cAAc,EAAE,yCAAyC;YACzD,SAAS,EAAE,OAAO;YAClB,WAAW,EAAE,EAAE;YACf,UAAU,EAAE,IAAI;SACjB;QAED;YACE,EAAE,EAAE,IAAI;YACR,QAAQ,EAAE;gBACR,qCAAiB,CAAC,qBAAqB;gBACvC,qCAAiB,CAAC,gBAAgB;gBAClC,qCAAiB,CAAC,aAAa;gBAC/B,qCAAiB,CAAC,iBAAiB;aACpC;YACD,UAAU,EAAE,yBAAW,CAAC,OAAO;YAC/B,cAAc,EAAE,8CAA8C;YAC9D,cAAc,EAAE,yBAAyB;YACzC,SAAS,EAAE,OAAO;YAClB,WAAW,EAAE,EAAE;YACf,UAAU,EAAE,IAAI;SACjB;QAED;YACE,EAAE,EAAE,IAAI;YACR,QAAQ,EAAE;gBACR,qCAAiB,CAAC,gBAAgB;gBAClC,qCAAiB,CAAC,iBAAiB;aACpC;YACD,UAAU,EAAE,yBAAW,CAAC,MAAM;YAC9B,cAAc,EAAE,uCAAuC;YACvD,cAAc,EAAE,oDAAoD;YACpE,SAAS,EAAE,OAAO;YAClB,WAAW,EAAE,EAAE;YACf,UAAU,EAAE,IAAI;SACjB;QAED;YACE,EAAE,EAAE,IAAI;YACR,QAAQ,EAAE;gBACR,qCAAiB,CAAC,qBAAqB;gBACvC,qCAAiB,CAAC,YAAY;gBAC9B,qCAAiB,CAAC,iBAAiB;aACpC;YACD,UAAU,EAAE,yBAAW,CAAC,MAAM;YAC9B,cAAc,EAAE,8BAA8B;YAC9C,cAAc,EAAE,oCAAoC;YACpD,SAAS,EAAE,OAAO;YAClB,WAAW,EAAE,EAAE;YACf,UAAU,EAAE,IAAI;SACjB;QAED;YACE,EAAE,EAAE,IAAI;YACR,QAAQ,EAAE;gBACR,qCAAiB,CAAC,aAAa;aAChC;YACD,UAAU,EAAE,yBAAW,CAAC,MAAM;YAC9B,cAAc,EAAE,2BAA2B;YAC3C,cAAc,EAAE,yCAAyC;YACzD,SAAS,EAAE,OAAO;YAClB,WAAW,EAAE,EAAE;YACf,UAAU,EAAE,IAAI;SACjB;KACF;IAED,UAAU,EAAE;QACV,SAAS,EAAE,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,MAAM,EAAE,WAAW,CAAC;QACpD,aAAa,EAAE,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,MAAM,EAAE,eAAe,CAAC;QAC5D,iBAAiB,EAAE,qBAAqB;KACzC;IAED,YAAY,EAAE;QACZ,gBAAgB,EAAE,kBAAkB;KACrC;CACF,CAAC"}
|
|
@@ -1,5 +1,8 @@
|
|
|
1
1
|
export declare const ENTITY_TYPE: {
|
|
2
2
|
readonly PERSON: "PERSON";
|
|
3
|
+
readonly FIRST_NAME: "FIRST_NAME";
|
|
4
|
+
readonly LAST_NAME: "LAST_NAME";
|
|
5
|
+
readonly PATRONYMIC: "PATRONYMIC";
|
|
3
6
|
readonly PHONE: "PHONE";
|
|
4
7
|
readonly EMAIL: "EMAIL";
|
|
5
8
|
readonly INN: "INN";
|
|
@@ -10,6 +13,10 @@ export declare const ENTITY_TYPE: {
|
|
|
10
13
|
readonly ADDRESS: "ADDRESS";
|
|
11
14
|
readonly COMPANY: "COMPANY";
|
|
12
15
|
readonly ACCOUNT: "ACCOUNT";
|
|
16
|
+
readonly SNILS: "SNILS";
|
|
17
|
+
readonly OGRN: "OGRN";
|
|
18
|
+
readonly OGRNIP: "OGRNIP";
|
|
19
|
+
readonly SOCIAL_ID: "SOCIAL_ID";
|
|
13
20
|
};
|
|
14
21
|
export type EntityType = (typeof ENTITY_TYPE)[keyof typeof ENTITY_TYPE];
|
|
15
22
|
export declare const ENTITY_TYPES: readonly EntityType[];
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"entity-type.d.ts","sourceRoot":"","sources":["../../src/types/entity-type.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,WAAW
|
|
1
|
+
{"version":3,"file":"entity-type.d.ts","sourceRoot":"","sources":["../../src/types/entity-type.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,WAAW;;;;;;;;;;;;;;;;;;;CAmBd,CAAC;AAEX,MAAM,MAAM,UAAU,GAAG,CAAC,OAAO,WAAW,CAAC,CAAC,MAAM,OAAO,WAAW,CAAC,CAAC;AACxE,eAAO,MAAM,YAAY,EAAE,SAAS,UAAU,EAA+B,CAAC"}
|
|
@@ -3,6 +3,9 @@ Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
3
3
|
exports.ENTITY_TYPES = exports.ENTITY_TYPE = void 0;
|
|
4
4
|
exports.ENTITY_TYPE = {
|
|
5
5
|
PERSON: 'PERSON',
|
|
6
|
+
FIRST_NAME: 'FIRST_NAME',
|
|
7
|
+
LAST_NAME: 'LAST_NAME',
|
|
8
|
+
PATRONYMIC: 'PATRONYMIC',
|
|
6
9
|
PHONE: 'PHONE',
|
|
7
10
|
EMAIL: 'EMAIL',
|
|
8
11
|
INN: 'INN',
|
|
@@ -13,6 +16,10 @@ exports.ENTITY_TYPE = {
|
|
|
13
16
|
ADDRESS: 'ADDRESS',
|
|
14
17
|
COMPANY: 'COMPANY',
|
|
15
18
|
ACCOUNT: 'ACCOUNT',
|
|
19
|
+
SNILS: 'SNILS',
|
|
20
|
+
OGRN: 'OGRN',
|
|
21
|
+
OGRNIP: 'OGRNIP',
|
|
22
|
+
SOCIAL_ID: 'SOCIAL_ID',
|
|
16
23
|
};
|
|
17
24
|
exports.ENTITY_TYPES = Object.values(exports.ENTITY_TYPE);
|
|
18
25
|
//# sourceMappingURL=entity-type.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"entity-type.js","sourceRoot":"","sources":["../../src/types/entity-type.ts"],"names":[],"mappings":";;;AAAa,QAAA,WAAW,GAAG;IACzB,MAAM,EAAE,QAAQ;IAChB,KAAK,EAAE,OAAO;IACd,KAAK,EAAE,OAAO;IACd,GAAG,EAAE,KAAK;IACV,QAAQ,EAAE,UAAU;IACpB,IAAI,EAAE,MAAM;IACZ,MAAM,EAAE,QAAQ;IAChB,aAAa,EAAE,eAAe;IAC9B,OAAO,EAAE,SAAS;IAClB,OAAO,EAAE,SAAS;IAClB,OAAO,EAAE,SAAS;
|
|
1
|
+
{"version":3,"file":"entity-type.js","sourceRoot":"","sources":["../../src/types/entity-type.ts"],"names":[],"mappings":";;;AAAa,QAAA,WAAW,GAAG;IACzB,MAAM,EAAE,QAAQ;IAChB,UAAU,EAAE,YAAY;IACxB,SAAS,EAAE,WAAW;IACtB,UAAU,EAAE,YAAY;IACxB,KAAK,EAAE,OAAO;IACd,KAAK,EAAE,OAAO;IACd,GAAG,EAAE,KAAK;IACV,QAAQ,EAAE,UAAU;IACpB,IAAI,EAAE,MAAM;IACZ,MAAM,EAAE,QAAQ;IAChB,aAAa,EAAE,eAAe;IAC9B,OAAO,EAAE,SAAS;IAClB,OAAO,EAAE,SAAS;IAClB,OAAO,EAAE,SAAS;IAClB,KAAK,EAAE,OAAO;IACd,IAAI,EAAE,MAAM;IACZ,MAAM,EAAE,QAAQ;IAChB,SAAS,EAAE,WAAW;CACd,CAAC;AAGE,QAAA,YAAY,GAA0B,MAAM,CAAC,MAAM,CAAC,mBAAW,CAAC,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@besales/anonymizer",
|
|
3
|
-
"version": "0.
|
|
3
|
+
"version": "0.2.0",
|
|
4
4
|
"description": "Stateless обратимая псевдонимизация PII (detect → tokenize → mapping → detokenize) для besales/animaly. Чистые функции, без БД/сети/фреймворков.",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"main": "dist/index.js",
|