@elanlanguages/bridge-anonymization 0.1.3 → 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.
Files changed (54) hide show
  1. package/README.md +255 -94
  2. package/dist/crypto/pii-map-crypto.d.ts +50 -36
  3. package/dist/crypto/pii-map-crypto.d.ts.map +1 -1
  4. package/dist/crypto/pii-map-crypto.js +133 -68
  5. package/dist/crypto/pii-map-crypto.js.map +1 -1
  6. package/dist/index.d.ts +3 -1
  7. package/dist/index.d.ts.map +1 -1
  8. package/dist/index.js +48 -9
  9. package/dist/index.js.map +1 -1
  10. package/dist/ner/model-manager.d.ts +20 -11
  11. package/dist/ner/model-manager.d.ts.map +1 -1
  12. package/dist/ner/model-manager.js +147 -76
  13. package/dist/ner/model-manager.js.map +1 -1
  14. package/dist/ner/ner-model.d.ts +1 -1
  15. package/dist/ner/ner-model.d.ts.map +1 -1
  16. package/dist/ner/ner-model.js +40 -27
  17. package/dist/ner/ner-model.js.map +1 -1
  18. package/dist/ner/onnx-runtime.d.ts +6 -5
  19. package/dist/ner/onnx-runtime.d.ts.map +1 -1
  20. package/dist/ner/onnx-runtime.js +55 -24
  21. package/dist/ner/onnx-runtime.js.map +1 -1
  22. package/dist/ner/tokenizer.d.ts +5 -0
  23. package/dist/ner/tokenizer.d.ts.map +1 -1
  24. package/dist/ner/tokenizer.js +15 -2
  25. package/dist/ner/tokenizer.js.map +1 -1
  26. package/dist/pipeline/semantic-data-loader.d.ts +24 -16
  27. package/dist/pipeline/semantic-data-loader.d.ts.map +1 -1
  28. package/dist/pipeline/semantic-data-loader.js +169 -176
  29. package/dist/pipeline/semantic-data-loader.js.map +1 -1
  30. package/dist/pipeline/semantic-enricher.d.ts +12 -2
  31. package/dist/pipeline/semantic-enricher.d.ts.map +1 -1
  32. package/dist/pipeline/semantic-enricher.js +85 -35
  33. package/dist/pipeline/semantic-enricher.js.map +1 -1
  34. package/dist/utils/index.d.ts +2 -0
  35. package/dist/utils/index.d.ts.map +1 -1
  36. package/dist/utils/index.js +2 -0
  37. package/dist/utils/index.js.map +1 -1
  38. package/dist/utils/path.d.ts +34 -0
  39. package/dist/utils/path.d.ts.map +1 -0
  40. package/dist/utils/path.js +96 -0
  41. package/dist/utils/path.js.map +1 -0
  42. package/dist/utils/storage-browser.d.ts +51 -0
  43. package/dist/utils/storage-browser.d.ts.map +1 -0
  44. package/dist/utils/storage-browser.js +381 -0
  45. package/dist/utils/storage-browser.js.map +1 -0
  46. package/dist/utils/storage-node.d.ts +43 -0
  47. package/dist/utils/storage-node.d.ts.map +1 -0
  48. package/dist/utils/storage-node.js +93 -0
  49. package/dist/utils/storage-node.js.map +1 -0
  50. package/dist/utils/storage.d.ts +70 -0
  51. package/dist/utils/storage.d.ts.map +1 -0
  52. package/dist/utils/storage.js +69 -0
  53. package/dist/utils/storage.js.map +1 -0
  54. package/package.json +2 -1
@@ -3,22 +3,23 @@
3
3
  * ONNX Runtime integration for Named Entity Recognition
4
4
  * Supports both onnxruntime-node and onnxruntime-web
5
5
  */
6
- import { loadRuntime } from './onnx-runtime.js';
7
- import { WordPieceTokenizer, loadVocabFromFile, } from './tokenizer.js';
8
- import { decodeBIOTags, convertToSpanMatches, cleanupSpanBoundaries, mergeAdjacentSpans, } from './bio-decoder.js';
6
+ import { loadRuntime } from "./onnx-runtime.js";
7
+ import { WordPieceTokenizer, loadVocabFromFile, } from "./tokenizer.js";
8
+ import { decodeBIOTags, convertToSpanMatches, cleanupSpanBoundaries, mergeAdjacentSpans, } from "./bio-decoder.js";
9
+ import { getStorageProvider, isBrowser } from "../utils/storage.js";
9
10
  /**
10
11
  * Default label map for common NER models (CoNLL-style)
11
12
  */
12
13
  export const DEFAULT_LABEL_MAP = [
13
- 'O',
14
- 'B-PER',
15
- 'I-PER',
16
- 'B-ORG',
17
- 'I-ORG',
18
- 'B-LOC',
19
- 'I-LOC',
20
- 'B-MISC',
21
- 'I-MISC',
14
+ "O",
15
+ "B-PER",
16
+ "I-PER",
17
+ "B-ORG",
18
+ "I-ORG",
19
+ "B-LOC",
20
+ "I-LOC",
21
+ "B-MISC",
22
+ "I-MISC",
22
23
  ];
23
24
  /**
24
25
  * NER Model wrapper for ONNX inference
@@ -41,8 +42,19 @@ export class NERModel {
41
42
  // Load ONNX runtime (auto-detects best runtime for environment)
42
43
  this.ort = await loadRuntime();
43
44
  // Load ONNX model
44
- this.session = await this.ort.InferenceSession.create(this.config.modelPath);
45
- // Load tokenizer vocabulary
45
+ // In browsers, we need to load the model as ArrayBuffer since file paths don't work
46
+ // onnxruntime-web accepts ArrayBuffer/Uint8Array, while onnxruntime-node accepts file paths
47
+ if (isBrowser()) {
48
+ const storage = await getStorageProvider();
49
+ const modelData = await storage.readFile(this.config.modelPath);
50
+ // onnxruntime-web accepts Uint8Array directly
51
+ this.session = await this.ort.InferenceSession.create(modelData);
52
+ }
53
+ else {
54
+ // In Node.js, we can use the file path directly
55
+ this.session = await this.ort.InferenceSession.create(this.config.modelPath);
56
+ }
57
+ // Load tokenizer vocabulary (already uses storage abstraction internally)
46
58
  const vocab = await loadVocabFromFile(this.config.vocabPath);
47
59
  this.tokenizer = new WordPieceTokenizer(vocab, {
48
60
  maxLength: this.config.maxLength,
@@ -56,7 +68,7 @@ export class NERModel {
56
68
  async predict(text, policy) {
57
69
  const startTime = performance.now();
58
70
  if (!this.isLoaded || this.session === null || this.tokenizer === null) {
59
- throw new Error('Model not loaded. Call load() first.');
71
+ throw new Error("Model not loaded. Call load() first.");
60
72
  }
61
73
  // Tokenize input
62
74
  const tokenization = this.tokenizer.tokenize(text);
@@ -72,7 +84,8 @@ export class NERModel {
72
84
  spans = mergeAdjacentSpans(spans, text);
73
85
  // Filter by enabled types in policy
74
86
  if (policy !== undefined) {
75
- spans = spans.filter((span) => policy.enabledTypes.has(span.type) && policy.nerEnabledTypes.has(span.type));
87
+ spans = spans.filter((span) => policy.enabledTypes.has(span.type) &&
88
+ policy.nerEnabledTypes.has(span.type));
76
89
  }
77
90
  const endTime = performance.now();
78
91
  return {
@@ -86,14 +99,14 @@ export class NERModel {
86
99
  */
87
100
  async runInference(tokenization) {
88
101
  if (this.session === null || this.ort === null) {
89
- throw new Error('Session not initialized');
102
+ throw new Error("Session not initialized");
90
103
  }
91
104
  const session = this.session;
92
105
  const seqLength = tokenization.inputIds.length;
93
106
  // Create tensors
94
- const inputIdsTensor = new this.ort.Tensor('int64', BigInt64Array.from(tokenization.inputIds.map(BigInt)), [1, seqLength]);
95
- const attentionMaskTensor = new this.ort.Tensor('int64', BigInt64Array.from(tokenization.attentionMask.map(BigInt)), [1, seqLength]);
96
- const tokenTypeIdsTensor = new this.ort.Tensor('int64', BigInt64Array.from(tokenization.tokenTypeIds.map(BigInt)), [1, seqLength]);
107
+ const inputIdsTensor = new this.ort.Tensor("int64", BigInt64Array.from(tokenization.inputIds.map(BigInt)), [1, seqLength]);
108
+ const attentionMaskTensor = new this.ort.Tensor("int64", BigInt64Array.from(tokenization.attentionMask.map(BigInt)), [1, seqLength]);
109
+ const tokenTypeIdsTensor = new this.ort.Tensor("int64", BigInt64Array.from(tokenization.tokenTypeIds.map(BigInt)), [1, seqLength]);
97
110
  // Run inference
98
111
  const feeds = {
99
112
  input_ids: inputIdsTensor,
@@ -101,18 +114,18 @@ export class NERModel {
101
114
  };
102
115
  // Some models also need token_type_ids
103
116
  const inputNames = session.inputNames;
104
- if (inputNames.includes('token_type_ids')) {
105
- feeds['token_type_ids'] = tokenTypeIdsTensor;
117
+ if (inputNames.includes("token_type_ids")) {
118
+ feeds["token_type_ids"] = tokenTypeIdsTensor;
106
119
  }
107
120
  const results = await session.run(feeds);
108
121
  // Get logits output
109
122
  const outputName = session.outputNames[0];
110
123
  if (outputName === undefined) {
111
- throw new Error('No output from model');
124
+ throw new Error("No output from model");
112
125
  }
113
126
  const logits = results[outputName];
114
127
  if (logits === undefined) {
115
- throw new Error('Logits output not found');
128
+ throw new Error("Logits output not found");
116
129
  }
117
130
  // Process logits to get labels and confidences
118
131
  return this.processLogits(logits, seqLength);
@@ -142,7 +155,7 @@ export class NERModel {
142
155
  maxIdx = j;
143
156
  }
144
157
  }
145
- labels.push(this.config.labelMap[maxIdx] ?? 'O');
158
+ labels.push(this.config.labelMap[maxIdx] ?? "O");
146
159
  confidences.push(maxProb);
147
160
  }
148
161
  return { labels, confidences };
@@ -205,7 +218,7 @@ export function createNERModel(config) {
205
218
  labelMap: config.labelMap ?? DEFAULT_LABEL_MAP,
206
219
  maxLength: config.maxLength ?? 512,
207
220
  doLowerCase: config.doLowerCase ?? false, // XLM-RoBERTa is cased
208
- modelVersion: config.modelVersion ?? '1.0.0',
221
+ modelVersion: config.modelVersion ?? "1.0.0",
209
222
  };
210
223
  return new NERModel(fullConfig);
211
224
  }
@@ -214,7 +227,7 @@ export function createNERModel(config) {
214
227
  * Returns empty results - useful for regex-only mode
215
228
  */
216
229
  export class NERModelStub {
217
- version = 'stub-1.0.0';
230
+ version = "stub-1.0.0";
218
231
  loaded = true;
219
232
  async load() {
220
233
  // No-op
@@ -1 +1 @@
1
- {"version":3,"file":"ner-model.js","sourceRoot":"","sources":["../../src/ner/ner-model.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,WAAW,EAAmB,MAAM,mBAAmB,CAAC;AAEjE,OAAO,EACL,kBAAkB,EAClB,iBAAiB,GAElB,MAAM,gBAAgB,CAAC;AACxB,OAAO,EACL,aAAa,EACb,oBAAoB,EACpB,qBAAqB,EACrB,kBAAkB,GACnB,MAAM,kBAAkB,CAAC;AAgC1B;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG;IAC/B,GAAG;IACH,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,QAAQ;IACR,QAAQ;CACT,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,QAAQ;IACX,GAAG,GAAsB,IAAI,CAAC;IAC9B,OAAO,GAAY,IAAI,CAAC;IACxB,SAAS,GAA8B,IAAI,CAAC;IAC5C,MAAM,CAAiB;IACvB,QAAQ,GAAG,KAAK,CAAC;IAEzB,YAAY,MAAsB;QAChC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAE1B,gEAAgE;QAChE,IAAI,CAAC,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;QAE/B,kBAAkB;QAClB,IAAI,CAAC,OAAO,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QAE7E,4BAA4B;QAC5B,MAAM,KAAK,GAAG,MAAM,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QAC7D,IAAI,CAAC,SAAS,GAAG,IAAI,kBAAkB,CAAC,KAAK,EAAE;YAC7C,SAAS,EAAE,IAAI,CAAC,MAAM,CAAC,SAAS;YAChC,WAAW,EAAE,IAAI,CAAC,MAAM,CAAC,WAAW;SACrC,CAAC,CAAC;QAEH,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,OAAO,CACX,IAAY,EACZ,MAA4B;QAE5B,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QAEpC,IAAI,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,SAAS,KAAK,IAAI,EAAE,CAAC;YACvE,MAAM,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAC;QAC1D,CAAC;QAED,iBAAiB;QACjB,MAAM,YAAY,GAAG,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAEnD,gBAAgB;QAChB,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC;QAEtE,8BAA8B;QAC9B,MAAM,WAAW,GAAG,aAAa,CAC/B,YAAY,CAAC,MAAM,EACnB,MAAM,EACN,WAAW,EACX,IAAI,CACL,CAAC;QAEF,wDAAwD;QACxD,MAAM,aAAa,GAAG,IAAI,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAC;QACpD,IAAI,KAAK,GAAG,oBAAoB,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;QAE7D,qBAAqB;QACrB,KAAK,GAAG,qBAAqB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAC3C,KAAK,GAAG,kBAAkB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAExC,oCAAoC;QACpC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,KAAK,GAAG,KAAK,CAAC,MAAM,CAClB,CAAC,IAAI,EAAE,EAAE,CACP,MAAM,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,eAAe,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAC9E,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QAElC,OAAO;YACL,KAAK;YACL,gBAAgB,EAAE,OAAO,GAAG,SAAS;YACrC,YAAY,EAAE,IAAI,CAAC,MAAM,CAAC,YAAY;SACvC,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,YAAY,CACxB,YAAgC;QAEhC,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC;YAC/C,MAAM,IAAI,KAAK,CAAC,yBAAyB,CAAC,CAAC;QAC7C,CAAC;QAED,MAAM,OAAO,GAAG,IAAI,CAAC,OAIpB,CAAC;QAEF,MAAM,SAAS,GAAG,YAAY,CAAC,QAAQ,CAAC,MAAM,CAAC;QAE/C,iBAAiB;QACjB,MAAM,cAAc,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CACxC,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EACrD,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,MAAM,mBAAmB,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAC7C,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EAC1D,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,MAAM,kBAAkB,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAC5C,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EACzD,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,gBAAgB;QAChB,MAAM,KAAK,GAA4B;YACrC,SAAS,EAAE,cAAc;YACzB,cAAc,EAAE,mBAAmB;SACpC,CAAC;QAEF,uCAAuC;QACvC,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC;QACtC,IAAI,UAAU,CAAC,QAAQ,CAAC,gBAAgB,CAAC,EAAE,CAAC;YAC1C,KAAK,CAAC,gBAAgB,CAAC,GAAG,kBAAkB,CAAC;QAC/C,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAEzC,oBAAoB;QACpB,MAAM,UAAU,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC1C,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;YAC7B,MAAM,IAAI,KAAK,CAAC,sBAAsB,CAAC,CAAC;QAC1C,CAAC;QAED,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;QACnC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,IAAI,KAAK,CAAC,yBAAyB,CAAC,CAAC;QAC7C,CAAC;QAED,+CAA+C;QAC/C,OAAO,IAAI,CAAC,aAAa,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC/C,CAAC;IAED;;OAEG;IACK,aAAa,CACnB,MAA8B,EAC9B,SAAiB;QAEjB,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC;QACzB,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC;QAE9C,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,MAAM,WAAW,GAAa,EAAE,CAAC;QAEjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,4BAA4B;YAC5B,MAAM,WAAW,GAAa,EAAE,CAAC;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;gBACnC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,SAAS,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;YACjD,CAAC;YAED,gBAAgB;YAChB,MAAM,KAAK,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;YAEnC,aAAa;YACb,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,IAAI,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,OAAO,EAAE,CAAC;oBAC9B,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;oBACxB,MAAM,GAAG,CAAC,CAAC;gBACb,CAAC;YACH,CAAC;YAED,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,CAAC;YACjD,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAC5B,CAAC;QAED,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IACjC,CAAC;IAED;;OAEG;IACK,gBAAgB,CAAC,MAA4B;QACnD,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,GAAG,CAAC;QAErC,yCAAyC;QACzC,IAAI,YAAY,GAAG,GAAG,CAAC;QACvB,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,eAAe,EAAE,CAAC;YAC1C,MAAM,SAAS,GAAG,MAAM,CAAC,oBAAoB,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC;YAC/D,IAAI,SAAS,GAAG,YAAY,EAAE,CAAC;gBAC7B,YAAY,GAAG,SAAS,CAAC;YAC3B,CAAC;QACH,CAAC;QAED,OAAO,YAAY,CAAC;IACtB,CAAC;IAED;;OAEG;IACH,IAAI,OAAO;QACT,OAAO,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC;IAClC,CAAC;IAED;;OAEG;IACH,IAAI,MAAM;QACR,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,OAAO;QACL,+EAA+E;QAC/E,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC;QACpB,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC;QACtB,IAAI,CAAC,QAAQ,GAAG,KAAK,CAAC;QACtB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;CACF;AAED;;GAEG;AACH,SAAS,OAAO,CAAC,MAAgB;IAC/B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC;IACrC,MAAM,SAAS,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC;IAC5D,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;IACpD,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC;AAC1C,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,MAA0E;IACvG,MAAM,UAAU,GAAmB;QACjC,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,iBAAiB;QAC9C,SAAS,EAAE,MAAM,CAAC,SAAS,IAAI,GAAG;QAClC,WAAW,EAAE,MAAM,CAAC,WAAW,IAAI,KAAK,EAAE,uBAAuB;QACjE,YAAY,EAAE,MAAM,CAAC,YAAY,IAAI,OAAO;KAC7C,CAAC;IAEF,OAAO,IAAI,QAAQ,CAAC,UAAU,CAAC,CAAC;AAClC,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,YAAY;IACd,OAAO,GAAG,YAAY,CAAC;IACvB,MAAM,GAAG,IAAI,CAAC;IAEvB,KAAK,CAAC,IAAI;QACR,QAAQ;IACV,CAAC;IAED,OAAO,CAAC,KAAa,EAAE,OAA6B;QAClD,OAAO,OAAO,CAAC,OAAO,CAAC;YACrB,KAAK,EAAE,EAAE;YACT,gBAAgB,EAAE,CAAC;YACnB,YAAY,EAAE,IAAI,CAAC,OAAO;SAC3B,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,QAAQ;QACR,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;CACF;AAED;;GAEG;AACH,MAAM,UAAU,kBAAkB;IAChC,OAAO,IAAI,YAAY,EAAE,CAAC;AAC5B,CAAC"}
1
+ {"version":3,"file":"ner-model.js","sourceRoot":"","sources":["../../src/ner/ner-model.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,WAAW,EAAmB,MAAM,mBAAmB,CAAC;AAEjE,OAAO,EACL,kBAAkB,EAClB,iBAAiB,GAElB,MAAM,gBAAgB,CAAC;AACxB,OAAO,EACL,aAAa,EACb,oBAAoB,EACpB,qBAAqB,EACrB,kBAAkB,GACnB,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,kBAAkB,EAAE,SAAS,EAAE,MAAM,qBAAqB,CAAC;AAgCpE;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG;IAC/B,GAAG;IACH,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,QAAQ;IACR,QAAQ;CACT,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,QAAQ;IACX,GAAG,GAAsB,IAAI,CAAC;IAC9B,OAAO,GAAY,IAAI,CAAC;IACxB,SAAS,GAA8B,IAAI,CAAC;IAC5C,MAAM,CAAiB;IACvB,QAAQ,GAAG,KAAK,CAAC;IAEzB,YAAY,MAAsB;QAChC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,IAAI;QACR,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAE1B,gEAAgE;QAChE,IAAI,CAAC,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;QAE/B,kBAAkB;QAClB,oFAAoF;QACpF,4FAA4F;QAC5F,IAAI,SAAS,EAAE,EAAE,CAAC;YAChB,MAAM,OAAO,GAAG,MAAM,kBAAkB,EAAE,CAAC;YAC3C,MAAM,SAAS,GAAG,MAAM,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;YAChE,8CAA8C;YAC9C,IAAI,CAAC,OAAO,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QACnE,CAAC;aAAM,CAAC;YACN,gDAAgD;YAChD,IAAI,CAAC,OAAO,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CACnD,IAAI,CAAC,MAAM,CAAC,SAAS,CACtB,CAAC;QACJ,CAAC;QAED,0EAA0E;QAC1E,MAAM,KAAK,GAAG,MAAM,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;QAC7D,IAAI,CAAC,SAAS,GAAG,IAAI,kBAAkB,CAAC,KAAK,EAAE;YAC7C,SAAS,EAAE,IAAI,CAAC,MAAM,CAAC,SAAS;YAChC,WAAW,EAAE,IAAI,CAAC,MAAM,CAAC,WAAW;SACrC,CAAC,CAAC;QAEH,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,OAAO,CACX,IAAY,EACZ,MAA4B;QAE5B,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QAEpC,IAAI,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,SAAS,KAAK,IAAI,EAAE,CAAC;YACvE,MAAM,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAC;QAC1D,CAAC;QAED,iBAAiB;QACjB,MAAM,YAAY,GAAG,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAEnD,gBAAgB;QAChB,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC;QAEtE,8BAA8B;QAC9B,MAAM,WAAW,GAAG,aAAa,CAC/B,YAAY,CAAC,MAAM,EACnB,MAAM,EACN,WAAW,EACX,IAAI,CACL,CAAC;QAEF,wDAAwD;QACxD,MAAM,aAAa,GAAG,IAAI,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAC;QACpD,IAAI,KAAK,GAAG,oBAAoB,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;QAE7D,qBAAqB;QACrB,KAAK,GAAG,qBAAqB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAC3C,KAAK,GAAG,kBAAkB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAExC,oCAAoC;QACpC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,KAAK,GAAG,KAAK,CAAC,MAAM,CAClB,CAAC,IAAI,EAAE,EAAE,CACP,MAAM,CAAC,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC;gBAClC,MAAM,CAAC,eAAe,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CACxC,CAAC;QACJ,CAAC;QAED,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QAElC,OAAO;YACL,KAAK;YACL,gBAAgB,EAAE,OAAO,GAAG,SAAS;YACrC,YAAY,EAAE,IAAI,CAAC,MAAM,CAAC,YAAY;SACvC,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,YAAY,CACxB,YAAgC;QAEhC,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC;YAC/C,MAAM,IAAI,KAAK,CAAC,yBAAyB,CAAC,CAAC;QAC7C,CAAC;QAED,MAAM,OAAO,GAAG,IAAI,CAAC,OAMpB,CAAC;QAEF,MAAM,SAAS,GAAG,YAAY,CAAC,QAAQ,CAAC,MAAM,CAAC;QAE/C,iBAAiB;QACjB,MAAM,cAAc,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CACxC,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EACrD,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,MAAM,mBAAmB,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAC7C,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EAC1D,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,MAAM,kBAAkB,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAC5C,OAAO,EACP,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EACzD,CAAC,CAAC,EAAE,SAAS,CAAC,CACf,CAAC;QAEF,gBAAgB;QAChB,MAAM,KAAK,GAA4B;YACrC,SAAS,EAAE,cAAc;YACzB,cAAc,EAAE,mBAAmB;SACpC,CAAC;QAEF,uCAAuC;QACvC,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC;QACtC,IAAI,UAAU,CAAC,QAAQ,CAAC,gBAAgB,CAAC,EAAE,CAAC;YAC1C,KAAK,CAAC,gBAAgB,CAAC,GAAG,kBAAkB,CAAC;QAC/C,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAEzC,oBAAoB;QACpB,MAAM,UAAU,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC;QAC1C,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;YAC7B,MAAM,IAAI,KAAK,CAAC,sBAAsB,CAAC,CAAC;QAC1C,CAAC;QAED,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;QACnC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,IAAI,KAAK,CAAC,yBAAyB,CAAC,CAAC;QAC7C,CAAC;QAED,+CAA+C;QAC/C,OAAO,IAAI,CAAC,aAAa,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC/C,CAAC;IAED;;OAEG;IACK,aAAa,CACnB,MAA8B,EAC9B,SAAiB;QAEjB,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC;QACzB,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC;QAE9C,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,MAAM,WAAW,GAAa,EAAE,CAAC;QAEjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,4BAA4B;YAC5B,MAAM,WAAW,GAAa,EAAE,CAAC;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;gBACnC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,SAAS,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;YACjD,CAAC;YAED,gBAAgB;YAChB,MAAM,KAAK,GAAG,OAAO,CAAC,WAAW,CAAC,CAAC;YAEnC,aAAa;YACb,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,IAAI,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,OAAO,EAAE,CAAC;oBAC9B,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;oBACxB,MAAM,GAAG,CAAC,CAAC;gBACb,CAAC;YACH,CAAC;YAED,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,CAAC;YACjD,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAC5B,CAAC;QAED,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IACjC,CAAC;IAED;;OAEG;IACK,gBAAgB,CAAC,MAA4B;QACnD,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,GAAG,CAAC;QAErC,yCAAyC;QACzC,IAAI,YAAY,GAAG,GAAG,CAAC;QACvB,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,eAAe,EAAE,CAAC;YAC1C,MAAM,SAAS,GAAG,MAAM,CAAC,oBAAoB,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC;YAC/D,IAAI,SAAS,GAAG,YAAY,EAAE,CAAC;gBAC7B,YAAY,GAAG,SAAS,CAAC;YAC3B,CAAC;QACH,CAAC;QAED,OAAO,YAAY,CAAC;IACtB,CAAC;IAED;;OAEG;IACH,IAAI,OAAO;QACT,OAAO,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC;IAClC,CAAC;IAED;;OAEG;IACH,IAAI,MAAM;QACR,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,OAAO;QACL,+EAA+E;QAC/E,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC;QACpB,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC;QACtB,IAAI,CAAC,QAAQ,GAAG,KAAK,CAAC;QACtB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;CACF;AAED;;GAEG;AACH,SAAS,OAAO,CAAC,MAAgB;IAC/B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC;IACrC,MAAM,SAAS,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC;IAC5D,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;IACpD,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC;AAC1C,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,MAA0E;IAE1E,MAAM,UAAU,GAAmB;QACjC,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,iBAAiB;QAC9C,SAAS,EAAE,MAAM,CAAC,SAAS,IAAI,GAAG;QAClC,WAAW,EAAE,MAAM,CAAC,WAAW,IAAI,KAAK,EAAE,uBAAuB;QACjE,YAAY,EAAE,MAAM,CAAC,YAAY,IAAI,OAAO;KAC7C,CAAC;IAEF,OAAO,IAAI,QAAQ,CAAC,UAAU,CAAC,CAAC;AAClC,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,YAAY;IACd,OAAO,GAAG,YAAY,CAAC;IACvB,MAAM,GAAG,IAAI,CAAC;IAEvB,KAAK,CAAC,IAAI;QACR,QAAQ;IACV,CAAC;IAED,OAAO,CACL,KAAa,EACb,OAA6B;QAE7B,OAAO,OAAO,CAAC,OAAO,CAAC;YACrB,KAAK,EAAE,EAAE;YACT,gBAAgB,EAAE,CAAC;YACnB,YAAY,EAAE,IAAI,CAAC,OAAO;SAC3B,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,QAAQ;QACR,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;CACF;AAED;;GAEG;AACH,MAAM,UAAU,kBAAkB;IAChC,OAAO,IAAI,YAAY,EAAE,CAAC;AAC5B,CAAC"}
@@ -1,6 +1,8 @@
1
1
  /**
2
2
  * ONNX Runtime Abstraction
3
3
  * Allows switching between onnxruntime-node and onnxruntime-web
4
+ *
5
+ * In browsers without a bundler, automatically loads onnxruntime-web from CDN
4
6
  */
5
7
  export interface OrtTensor {
6
8
  data: Float32Array | BigInt64Array | Int32Array;
@@ -12,7 +14,7 @@ export interface OrtSession {
12
14
  run(feeds: Record<string, OrtTensor>): Promise<Record<string, OrtTensor>>;
13
15
  }
14
16
  export interface OrtInferenceSession {
15
- create(path: string, options?: unknown): Promise<OrtSession>;
17
+ create(pathOrBuffer: string | ArrayBuffer | Uint8Array, options?: unknown): Promise<OrtSession>;
16
18
  }
17
19
  export interface OrtTensorConstructor {
18
20
  new (type: string, data: Float32Array | BigInt64Array | Int32Array | number[] | bigint[], dims: number[]): OrtTensor;
@@ -24,15 +26,15 @@ export interface OrtRuntime {
24
26
  /**
25
27
  * Detects the best ONNX runtime for the current environment
26
28
  */
27
- export declare function detectRuntime(): 'node' | 'web';
29
+ export declare function detectRuntime(): "node" | "web";
28
30
  /**
29
31
  * Loads the appropriate ONNX runtime
30
32
  */
31
- export declare function loadRuntime(preferredRuntime?: 'node' | 'web'): Promise<OrtRuntime>;
33
+ export declare function loadRuntime(preferredRuntime?: "node" | "web"): Promise<OrtRuntime>;
32
34
  /**
33
35
  * Gets the currently loaded runtime type
34
36
  */
35
- export declare function getRuntimeType(): 'node' | 'web' | null;
37
+ export declare function getRuntimeType(): "node" | "web" | null;
36
38
  /**
37
39
  * Resets the runtime (useful for testing)
38
40
  */
@@ -40,6 +42,5 @@ export declare function resetRuntime(): void;
40
42
  declare global {
41
43
  var Bun: unknown;
42
44
  var Deno: unknown;
43
- var window: unknown;
44
45
  }
45
46
  //# sourceMappingURL=onnx-runtime.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"onnx-runtime.d.ts","sourceRoot":"","sources":["../../src/ner/onnx-runtime.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,YAAY,GAAG,aAAa,GAAG,UAAU,CAAC;IAChD,IAAI,EAAE,SAAS,MAAM,EAAE,CAAC;CACzB;AAED,MAAM,WAAW,UAAU;IACzB,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IAC9B,WAAW,EAAE,SAAS,MAAM,EAAE,CAAC;IAC/B,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,CAAC;CAC3E;AAED,MAAM,WAAW,mBAAmB;IAClC,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;CAC9D;AAED,MAAM,WAAW,oBAAoB;IACnC,KACE,IAAI,EAAE,MAAM,EACZ,IAAI,EAAE,YAAY,GAAG,aAAa,GAAG,UAAU,GAAG,MAAM,EAAE,GAAG,MAAM,EAAE,EACrE,IAAI,EAAE,MAAM,EAAE,GACb,SAAS,CAAC;CACd;AAED,MAAM,WAAW,UAAU;IACzB,gBAAgB,EAAE,mBAAmB,CAAC;IACtC,MAAM,EAAE,oBAAoB,CAAC;CAC9B;AAQD;;GAEG;AACH,wBAAgB,aAAa,IAAI,MAAM,GAAG,KAAK,CA4B9C;AAED;;GAEG;AACH,wBAAsB,WAAW,CAAC,gBAAgB,CAAC,EAAE,MAAM,GAAG,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,CA8CxF;AAED;;GAEG;AACH,wBAAgB,cAAc,IAAI,MAAM,GAAG,KAAK,GAAG,IAAI,CAEtD;AAED;;GAEG;AACH,wBAAgB,YAAY,IAAI,IAAI,CAGnC;AAGD,OAAO,CAAC,MAAM,CAAC;IAEb,IAAI,GAAG,EAAE,OAAO,CAAC;IAEjB,IAAI,IAAI,EAAE,OAAO,CAAC;IAElB,IAAI,MAAM,EAAE,OAAO,CAAC;CACrB"}
1
+ {"version":3,"file":"onnx-runtime.d.ts","sourceRoot":"","sources":["../../src/ner/onnx-runtime.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAQH,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,YAAY,GAAG,aAAa,GAAG,UAAU,CAAC;IAChD,IAAI,EAAE,SAAS,MAAM,EAAE,CAAC;CACzB;AAED,MAAM,WAAW,UAAU;IACzB,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IAC9B,WAAW,EAAE,SAAS,MAAM,EAAE,CAAC;IAC/B,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,CAAC;CAC3E;AAED,MAAM,WAAW,mBAAmB;IAClC,MAAM,CACJ,YAAY,EAAE,MAAM,GAAG,WAAW,GAAG,UAAU,EAC/C,OAAO,CAAC,EAAE,OAAO,GAChB,OAAO,CAAC,UAAU,CAAC,CAAC;CACxB;AAED,MAAM,WAAW,oBAAoB;IACnC,KACE,IAAI,EAAE,MAAM,EACZ,IAAI,EAAE,YAAY,GAAG,aAAa,GAAG,UAAU,GAAG,MAAM,EAAE,GAAG,MAAM,EAAE,EACrE,IAAI,EAAE,MAAM,EAAE,GACb,SAAS,CAAC;CACd;AAED,MAAM,WAAW,UAAU;IACzB,gBAAgB,EAAE,mBAAmB,CAAC;IACtC,MAAM,EAAE,oBAAoB,CAAC;CAC9B;AAQD;;GAEG;AACH,wBAAgB,aAAa,IAAI,MAAM,GAAG,KAAK,CA4B9C;AAiCD;;GAEG;AACH,wBAAsB,WAAW,CAC/B,gBAAgB,CAAC,EAAE,MAAM,GAAG,KAAK,GAChC,OAAO,CAAC,UAAU,CAAC,CA2CrB;AAED;;GAEG;AACH,wBAAgB,cAAc,IAAI,MAAM,GAAG,KAAK,GAAG,IAAI,CAEtD;AAED;;GAEG;AACH,wBAAgB,YAAY,IAAI,IAAI,CAGnC;AAGD,OAAO,CAAC,MAAM,CAAC;IAEb,IAAI,GAAG,EAAE,OAAO,CAAC;IAEjB,IAAI,IAAI,EAAE,OAAO,CAAC;CACnB"}
@@ -1,7 +1,12 @@
1
1
  /**
2
2
  * ONNX Runtime Abstraction
3
3
  * Allows switching between onnxruntime-node and onnxruntime-web
4
+ *
5
+ * In browsers without a bundler, automatically loads onnxruntime-web from CDN
4
6
  */
7
+ // CDN URL for onnxruntime-web (used when bare import fails in browser)
8
+ // Using the bundled ESM version that includes WebAssembly backend
9
+ const ONNX_WEB_CDN_URL = "https://cdn.jsdelivr.net/npm/onnxruntime-web@1.19.2/dist/ort.bundle.min.mjs";
5
10
  /**
6
11
  * Runtime detection and loading
7
12
  */
@@ -12,28 +17,57 @@ let _runtimeType = null;
12
17
  */
13
18
  export function detectRuntime() {
14
19
  // Check if we're in Bun
15
- const isBun = typeof globalThis.Bun !== 'undefined';
20
+ const isBun = typeof globalThis.Bun !== "undefined";
16
21
  // Check if we're in a browser-like environment
17
22
  // eslint-disable-next-line @typescript-eslint/no-unnecessary-condition
18
- const isBrowser = typeof globalThis.window !== 'undefined';
23
+ const isBrowser = typeof globalThis.window !== "undefined";
19
24
  // Check if we're in Deno
20
- const isDeno = typeof globalThis.Deno !== 'undefined';
25
+ const isDeno = typeof globalThis.Deno !== "undefined";
21
26
  if (isBrowser || isDeno) {
22
- return 'web';
27
+ return "web";
23
28
  }
24
29
  // For Bun, try node first, fall back to web
25
30
  if (isBun) {
26
31
  try {
27
32
  // Quick check if onnxruntime-node is loadable
28
- require.resolve('onnxruntime-node');
29
- return 'node';
33
+ require.resolve("onnxruntime-node");
34
+ return "node";
30
35
  }
31
36
  catch {
32
- return 'web';
37
+ return "web";
33
38
  }
34
39
  }
35
40
  // Default to node for Node.js
36
- return 'node';
41
+ return "node";
42
+ }
43
+ /**
44
+ * Attempts to load onnxruntime-web, first via bare import, then via CDN
45
+ */
46
+ async function loadOnnxWeb() {
47
+ // First try bare import (works with bundlers or import maps)
48
+ try {
49
+ // eslint-disable-next-line @typescript-eslint/ban-ts-comment
50
+ // @ts-ignore - onnxruntime-web may not be installed
51
+ const ort = (await import("onnxruntime-web"));
52
+ return ort;
53
+ }
54
+ catch {
55
+ // Bare import failed, try CDN (for browsers without bundlers)
56
+ // eslint-disable-next-line @typescript-eslint/no-unnecessary-condition
57
+ const isBrowser = typeof globalThis.window !== "undefined";
58
+ if (isBrowser) {
59
+ try {
60
+ // Dynamic import from CDN URL
61
+ const ort = (await import(
62
+ /* webpackIgnore: true */ ONNX_WEB_CDN_URL));
63
+ return ort;
64
+ }
65
+ catch (cdnError) {
66
+ throw new Error(`Failed to load onnxruntime-web from CDN: ${String(cdnError)}`);
67
+ }
68
+ }
69
+ throw new Error("onnxruntime-web is not available");
70
+ }
37
71
  }
38
72
  /**
39
73
  * Loads the appropriate ONNX runtime
@@ -44,36 +78,33 @@ export async function loadRuntime(preferredRuntime) {
44
78
  }
45
79
  const runtimeType = preferredRuntime ?? detectRuntime();
46
80
  try {
47
- if (runtimeType === 'node') {
81
+ if (runtimeType === "node") {
48
82
  // Dynamic import for onnxruntime-node
49
- const ort = await import('onnxruntime-node');
83
+ const ort = (await import("onnxruntime-node"));
50
84
  _runtime = ort;
51
- _runtimeType = 'node';
85
+ _runtimeType = "node";
52
86
  }
53
87
  else {
54
- // Dynamic import for onnxruntime-web
55
- // eslint-disable-next-line @typescript-eslint/ban-ts-comment
56
- // @ts-ignore - onnxruntime-web may not be installed
57
- const ort = await import('onnxruntime-web');
88
+ // Load onnxruntime-web (with CDN fallback for browsers)
89
+ const ort = await loadOnnxWeb();
58
90
  _runtime = ort;
59
- _runtimeType = 'web';
91
+ _runtimeType = "web";
60
92
  }
61
93
  }
62
94
  catch (e) {
63
95
  // If preferred runtime fails, try the other
64
- const fallbackType = runtimeType === 'node' ? 'web' : 'node';
96
+ const fallbackType = runtimeType === "node" ? "web" : "node";
65
97
  try {
66
- if (fallbackType === 'node') {
67
- const ort = await import('onnxruntime-node');
98
+ if (fallbackType === "node") {
99
+ const ort = (await import("onnxruntime-node"));
68
100
  _runtime = ort;
69
- _runtimeType = 'node';
101
+ _runtimeType = "node";
70
102
  }
71
103
  else {
72
- // eslint-disable-next-line @typescript-eslint/ban-ts-comment
73
- // @ts-ignore - onnxruntime-web may not be installed
74
- const ort = await import('onnxruntime-web');
104
+ // Load onnxruntime-web (with CDN fallback for browsers)
105
+ const ort = await loadOnnxWeb();
75
106
  _runtime = ort;
76
- _runtimeType = 'web';
107
+ _runtimeType = "web";
77
108
  }
78
109
  }
79
110
  catch {
@@ -1 +1 @@
1
- {"version":3,"file":"onnx-runtime.js","sourceRoot":"","sources":["../../src/ner/onnx-runtime.ts"],"names":[],"mappings":"AAAA;;;GAGG;AA+BH;;GAEG;AACH,IAAI,QAAQ,GAAsB,IAAI,CAAC;AACvC,IAAI,YAAY,GAA0B,IAAI,CAAC;AAE/C;;GAEG;AACH,MAAM,UAAU,aAAa;IAC3B,wBAAwB;IACxB,MAAM,KAAK,GAAG,OAAO,UAAU,CAAC,GAAG,KAAK,WAAW,CAAC;IAEpD,+CAA+C;IAC/C,uEAAuE;IACvE,MAAM,SAAS,GAAG,OAAO,UAAU,CAAC,MAAM,KAAK,WAAW,CAAC;IAE3D,yBAAyB;IACzB,MAAM,MAAM,GAAG,OAAO,UAAU,CAAC,IAAI,KAAK,WAAW,CAAC;IAEtD,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;QACxB,OAAO,KAAK,CAAC;IACf,CAAC;IAED,4CAA4C;IAC5C,IAAI,KAAK,EAAE,CAAC;QACV,IAAI,CAAC;YACH,8CAA8C;YAC9C,OAAO,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC;YACpC,OAAO,MAAM,CAAC;QAChB,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,KAAK,CAAC;QACf,CAAC;IACH,CAAC;IAED,8BAA8B;IAC9B,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,gBAAiC;IACjE,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;QACtB,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,MAAM,WAAW,GAAG,gBAAgB,IAAI,aAAa,EAAE,CAAC;IAExD,IAAI,CAAC;QACH,IAAI,WAAW,KAAK,MAAM,EAAE,CAAC;YAC3B,sCAAsC;YACtC,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAe,CAAC;YAC3D,QAAQ,GAAG,GAAG,CAAC;YACf,YAAY,GAAG,MAAM,CAAC;QACxB,CAAC;aAAM,CAAC;YACN,qCAAqC;YACrC,6DAA6D;YAC7D,oDAAoD;YACpD,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,iBAAiB,CAAe,CAAC;YAC1D,QAAQ,GAAG,GAAG,CAAC;YACf,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;IACH,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,4CAA4C;QAC5C,MAAM,YAAY,GAAG,WAAW,KAAK,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC;QAE7D,IAAI,CAAC;YACH,IAAI,YAAY,KAAK,MAAM,EAAE,CAAC;gBAC5B,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAe,CAAC;gBAC3D,QAAQ,GAAG,GAAG,CAAC;gBACf,YAAY,GAAG,MAAM,CAAC;YACxB,CAAC;iBAAM,CAAC;gBACN,6DAA6D;gBAC7D,oDAAoD;gBACpD,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,iBAAiB,CAAe,CAAC;gBAC1D,QAAQ,GAAG,GAAG,CAAC;gBACf,YAAY,GAAG,KAAK,CAAC;YACvB,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,KAAK,CACb,wFAAwF;gBACxF,mBAAmB,MAAM,CAAC,CAAC,CAAC,EAAE,CAC/B,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc;IAC5B,OAAO,YAAY,CAAC;AACtB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY;IAC1B,QAAQ,GAAG,IAAI,CAAC;IAChB,YAAY,GAAG,IAAI,CAAC;AACtB,CAAC"}
1
+ {"version":3,"file":"onnx-runtime.js","sourceRoot":"","sources":["../../src/ner/onnx-runtime.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,uEAAuE;AACvE,kEAAkE;AAClE,MAAM,gBAAgB,GACpB,6EAA6E,CAAC;AAkChF;;GAEG;AACH,IAAI,QAAQ,GAAsB,IAAI,CAAC;AACvC,IAAI,YAAY,GAA0B,IAAI,CAAC;AAE/C;;GAEG;AACH,MAAM,UAAU,aAAa;IAC3B,wBAAwB;IACxB,MAAM,KAAK,GAAG,OAAO,UAAU,CAAC,GAAG,KAAK,WAAW,CAAC;IAEpD,+CAA+C;IAC/C,uEAAuE;IACvE,MAAM,SAAS,GAAG,OAAO,UAAU,CAAC,MAAM,KAAK,WAAW,CAAC;IAE3D,yBAAyB;IACzB,MAAM,MAAM,GAAG,OAAO,UAAU,CAAC,IAAI,KAAK,WAAW,CAAC;IAEtD,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;QACxB,OAAO,KAAK,CAAC;IACf,CAAC;IAED,4CAA4C;IAC5C,IAAI,KAAK,EAAE,CAAC;QACV,IAAI,CAAC;YACH,8CAA8C;YAC9C,OAAO,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC;YACpC,OAAO,MAAM,CAAC;QAChB,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,KAAK,CAAC;QACf,CAAC;IACH,CAAC;IAED,8BAA8B;IAC9B,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,KAAK,UAAU,WAAW;IACxB,6DAA6D;IAC7D,IAAI,CAAC;QACH,6DAA6D;QAC7D,oDAAoD;QACpD,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,iBAAiB,CAAC,CAAe,CAAC;QAC5D,OAAO,GAAG,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACP,8DAA8D;QAC9D,uEAAuE;QACvE,MAAM,SAAS,GAAG,OAAO,UAAU,CAAC,MAAM,KAAK,WAAW,CAAC;QAC3D,IAAI,SAAS,EAAE,CAAC;YACd,IAAI,CAAC;gBACH,8BAA8B;gBAC9B,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM;gBACvB,yBAAyB,CAAC,gBAAgB,CAC3C,CAAe,CAAC;gBACjB,OAAO,GAAG,CAAC;YACb,CAAC;YAAC,OAAO,QAAQ,EAAE,CAAC;gBAClB,MAAM,IAAI,KAAK,CACb,4CAA4C,MAAM,CAAC,QAAQ,CAAC,EAAE,CAC/D,CAAC;YACJ,CAAC;QACH,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,kCAAkC,CAAC,CAAC;IACtD,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,gBAAiC;IAEjC,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;QACtB,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,MAAM,WAAW,GAAG,gBAAgB,IAAI,aAAa,EAAE,CAAC;IAExD,IAAI,CAAC;QACH,IAAI,WAAW,KAAK,MAAM,EAAE,CAAC;YAC3B,sCAAsC;YACtC,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAe,CAAC;YAC7D,QAAQ,GAAG,GAAG,CAAC;YACf,YAAY,GAAG,MAAM,CAAC;QACxB,CAAC;aAAM,CAAC;YACN,wDAAwD;YACxD,MAAM,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;YAChC,QAAQ,GAAG,GAAG,CAAC;YACf,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;IACH,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,4CAA4C;QAC5C,MAAM,YAAY,GAAG,WAAW,KAAK,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC;QAE7D,IAAI,CAAC;YACH,IAAI,YAAY,KAAK,MAAM,EAAE,CAAC;gBAC5B,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAe,CAAC;gBAC7D,QAAQ,GAAG,GAAG,CAAC;gBACf,YAAY,GAAG,MAAM,CAAC;YACxB,CAAC;iBAAM,CAAC;gBACN,wDAAwD;gBACxD,MAAM,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;gBAChC,QAAQ,GAAG,GAAG,CAAC;gBACf,YAAY,GAAG,KAAK,CAAC;YACvB,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,KAAK,CACb,wFAAwF;gBACtF,mBAAmB,MAAM,CAAC,CAAC,CAAC,EAAE,CACjC,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc;IAC5B,OAAO,YAAY,CAAC;AACtB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY;IAC1B,QAAQ,GAAG,IAAI,CAAC;IAChB,YAAY,GAAG,IAAI,CAAC;AACtB,CAAC"}
@@ -96,8 +96,13 @@ export declare class WordPieceTokenizer {
96
96
  }
97
97
  /**
98
98
  * Loads vocabulary from a file (supports tokenizer.json and vocab.txt)
99
+ * Uses storage abstraction for browser compatibility
99
100
  */
100
101
  export declare function loadVocabFromFile(filePath: string): Promise<Map<string, number>>;
102
+ /**
103
+ * Loads vocabulary from content string (for when content is already available)
104
+ */
105
+ export declare function loadVocabFromContent(content: string, format?: 'json' | 'txt'): Map<string, number>;
101
106
  /**
102
107
  * Parses HuggingFace tokenizer.json format
103
108
  */
@@ -1 +1 @@
1
- {"version":3,"file":"tokenizer.d.ts","sourceRoot":"","sources":["../../src/ner/tokenizer.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH;;GAEG;AACH,MAAM,WAAW,KAAK;IACpB,6BAA6B;IAC7B,EAAE,EAAE,MAAM,CAAC;IACX,mBAAmB;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,8CAA8C;IAC9C,KAAK,EAAE,MAAM,CAAC;IACd,4CAA4C;IAC5C,GAAG,EAAE,MAAM,CAAC;IACZ,2CAA2C;IAC3C,cAAc,EAAE,OAAO,CAAC;IACxB,sCAAsC;IACtC,SAAS,EAAE,OAAO,CAAC;CACpB;AAED;;GAEG;AACH,MAAM,WAAW,kBAAkB;IACjC,sBAAsB;IACtB,MAAM,EAAE,KAAK,EAAE,CAAC;IAChB,0BAA0B;IAC1B,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,qBAAqB;IACrB,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,6CAA6C;IAC7C,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB,8DAA8D;IAC9D,eAAe,EAAE,KAAK,CAAC,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;CACjD;AAsBD;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,8BAA8B;IAC9B,SAAS,EAAE,MAAM,CAAC;IAClB,iCAAiC;IACjC,WAAW,EAAE,OAAO,CAAC;CACtB;AAED;;GAEG;AACH,eAAO,MAAM,wBAAwB,EAAE,eAGtC,CAAC;AAEF;;GAEG;AACH,qBAAa,kBAAkB;IAC7B,OAAO,CAAC,KAAK,CAAsB;IACnC,OAAO,CAAC,YAAY,CAAsB;IAC1C,OAAO,CAAC,MAAM,CAAkB;IAChC,OAAO,CAAC,WAAW,CAA0B;IAG7C,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAG1B,OAAO,CAAC,QAAQ,CAAiB;IACjC,OAAO,CAAC,QAAQ,CAAkB;IAClC,OAAO,CAAC,QAAQ,CAAmB;IACnC,OAAO,CAAC,QAAQ,CAAmB;gBAEvB,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,GAAE,OAAO,CAAC,eAAe,CAAM;IAiB7E;;OAEG;IACH,OAAO,CAAC,mBAAmB;IAyB3B;;OAEG;IACH,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,kBAAkB;IAsF1C;;OAEG;IACH,OAAO,CAAC,aAAa;IA0CrB;;OAEG;IACH,MAAM,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,MAAM;IAmBlC;;OAEG;IACH,IAAI,SAAS,IAAI,MAAM,CAEtB;IAED;;OAEG;IACH,UAAU,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS;IAI7C;;OAEG;IACH,QAAQ,CAAC,EAAE,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS;CAGzC;AAED;;GAEG;AACH,wBAAsB,iBAAiB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAUtF;AAED;;GAEG;AACH,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAmCzE;AAED;;GAEG;AACH,wBAAgB,UAAU,CAAC,OAAO,EAAE,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAY/D;AAED;;GAEG;AACH,wBAAgB,eAAe,IAAI,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAuBrD"}
1
+ {"version":3,"file":"tokenizer.d.ts","sourceRoot":"","sources":["../../src/ner/tokenizer.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH;;GAEG;AACH,MAAM,WAAW,KAAK;IACpB,6BAA6B;IAC7B,EAAE,EAAE,MAAM,CAAC;IACX,mBAAmB;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,8CAA8C;IAC9C,KAAK,EAAE,MAAM,CAAC;IACd,4CAA4C;IAC5C,GAAG,EAAE,MAAM,CAAC;IACZ,2CAA2C;IAC3C,cAAc,EAAE,OAAO,CAAC;IACxB,sCAAsC;IACtC,SAAS,EAAE,OAAO,CAAC;CACpB;AAED;;GAEG;AACH,MAAM,WAAW,kBAAkB;IACjC,sBAAsB;IACtB,MAAM,EAAE,KAAK,EAAE,CAAC;IAChB,0BAA0B;IAC1B,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,qBAAqB;IACrB,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,6CAA6C;IAC7C,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB,8DAA8D;IAC9D,eAAe,EAAE,KAAK,CAAC,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;CACjD;AAsBD;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,8BAA8B;IAC9B,SAAS,EAAE,MAAM,CAAC;IAClB,iCAAiC;IACjC,WAAW,EAAE,OAAO,CAAC;CACtB;AAED;;GAEG;AACH,eAAO,MAAM,wBAAwB,EAAE,eAGtC,CAAC;AAEF;;GAEG;AACH,qBAAa,kBAAkB;IAC7B,OAAO,CAAC,KAAK,CAAsB;IACnC,OAAO,CAAC,YAAY,CAAsB;IAC1C,OAAO,CAAC,MAAM,CAAkB;IAChC,OAAO,CAAC,WAAW,CAA0B;IAG7C,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAC1B,OAAO,CAAC,KAAK,CAAa;IAG1B,OAAO,CAAC,QAAQ,CAAiB;IACjC,OAAO,CAAC,QAAQ,CAAkB;IAClC,OAAO,CAAC,QAAQ,CAAmB;IACnC,OAAO,CAAC,QAAQ,CAAmB;gBAEvB,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,GAAE,OAAO,CAAC,eAAe,CAAM;IAiB7E;;OAEG;IACH,OAAO,CAAC,mBAAmB;IAyB3B;;OAEG;IACH,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,kBAAkB;IAsF1C;;OAEG;IACH,OAAO,CAAC,aAAa;IA0CrB;;OAEG;IACH,MAAM,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,MAAM;IAmBlC;;OAEG;IACH,IAAI,SAAS,IAAI,MAAM,CAEtB;IAED;;OAEG;IACH,UAAU,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS;IAI7C;;OAEG;IACH,QAAQ,CAAC,EAAE,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS;CAGzC;AAED;;;GAGG;AACH,wBAAsB,iBAAiB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAWtF;AAED;;GAEG;AACH,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,MAAM,EAAE,MAAM,GAAE,MAAM,GAAG,KAAc,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAM1G;AAED;;GAEG;AACH,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAmCzE;AAED;;GAEG;AACH,wBAAgB,UAAU,CAAC,OAAO,EAAE,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAY/D;AAED;;GAEG;AACH,wBAAgB,eAAe,IAAI,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAuBrD"}
@@ -228,10 +228,12 @@ export class WordPieceTokenizer {
228
228
  }
229
229
  /**
230
230
  * Loads vocabulary from a file (supports tokenizer.json and vocab.txt)
231
+ * Uses storage abstraction for browser compatibility
231
232
  */
232
233
  export async function loadVocabFromFile(filePath) {
233
- const fs = await import('fs/promises');
234
- const content = await fs.readFile(filePath, 'utf-8');
234
+ const { getStorageProvider } = await import('../utils/storage.js');
235
+ const storage = await getStorageProvider();
236
+ const content = await storage.readTextFile(filePath);
235
237
  // Detect format
236
238
  if (filePath.endsWith('.json') || content.trim().startsWith('{')) {
237
239
  return parseHFTokenizerJson(content);
@@ -240,6 +242,17 @@ export async function loadVocabFromFile(filePath) {
240
242
  return parseVocab(content);
241
243
  }
242
244
  }
245
+ /**
246
+ * Loads vocabulary from content string (for when content is already available)
247
+ */
248
+ export function loadVocabFromContent(content, format = 'json') {
249
+ if (format === 'json' || content.trim().startsWith('{')) {
250
+ return parseHFTokenizerJson(content);
251
+ }
252
+ else {
253
+ return parseVocab(content);
254
+ }
255
+ }
243
256
  /**
244
257
  * Parses HuggingFace tokenizer.json format
245
258
  */
@@ -1 +1 @@
1
- {"version":3,"file":"tokenizer.js","sourceRoot":"","sources":["../../src/ner/tokenizer.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAkEH;;GAEG;AACH,MAAM,CAAC,MAAM,wBAAwB,GAAoB;IACvD,SAAS,EAAE,GAAG;IACd,WAAW,EAAE,KAAK,EAAE,gCAAgC;CACrD,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,kBAAkB;IACrB,KAAK,CAAsB;IAC3B,YAAY,CAAsB;IAClC,MAAM,CAAkB;IACxB,WAAW,CAA0B;IAE7C,wCAAwC;IAChC,KAAK,GAAW,CAAC,CAAC,CAAE,MAAM;IAC1B,KAAK,GAAW,CAAC,CAAC,CAAE,OAAO;IAC3B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAC5B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAEpC,wBAAwB;IAChB,QAAQ,GAAW,KAAK,CAAC;IACzB,QAAQ,GAAW,MAAM,CAAC;IAC1B,QAAQ,GAAW,OAAO,CAAC;IAC3B,QAAQ,GAAW,OAAO,CAAC;IAEnC,YAAY,KAA0B,EAAE,SAAmC,EAAE;QAC3E,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,MAAM,GAAG,EAAE,GAAG,wBAAwB,EAAE,GAAG,MAAM,EAAE,CAAC;QAEzD,sBAAsB;QACtB,IAAI,CAAC,YAAY,GAAG,IAAI,GAAG,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC;YAChC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;QACnC,CAAC;QAED,iEAAiE;QACjE,IAAI,CAAC,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QAEzF,0CAA0C;QAC1C,IAAI,CAAC,mBAAmB,EAAE,CAAC;IAC7B,CAAC;IAED;;OAEG;IACK,mBAAmB;QACzB,oBAAoB;QACpB,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAC1B,IAAI,CAAC,QAAQ,GAAG,KAAK,CAAC;YACtB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,CAAC,QAAQ,GAAG,MAAM,CAAC;YACvB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YACzC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QAC5C,CAAC;QACD,aAAa;aACR,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YACjC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;QAC9C,CAAC;IACH,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,eAAe,GAAmC,EAAE,CAAC;QAE3D,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,CAAC;YACR,GAAG,EAAE,CAAC;YACN,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,kBAAkB;QAClB,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QAE1E,sCAAsC;QACtC,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,OAAO,GAAG,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC;YAClC,kBAAkB;YAClB,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAE,CAAC,EAAE,CAAC;gBACnC,GAAG,EAAE,CAAC;gBACN,SAAS;YACX,CAAC;YAED,4DAA4D;YAC5D,MAAM,EAAE,KAAK,EAAE,EAAE,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC,aAAa,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC;YAErE,MAAM,aAAa,GAAG,GAAG,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,GAAG,CAAC,CAAE,CAAC,CAAC;YAEtE,MAAM,CAAC,IAAI,CAAC;gBACV,EAAE;gBACF,KAAK;gBACL,KAAK,EAAE,GAAG;gBACV,GAAG,EAAE,GAAG,GAAG,MAAM;gBACjB,cAAc,EAAE,CAAC,aAAa,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;gBACxD,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,eAAe,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC;YAE1C,GAAG,IAAI,MAAM,CAAC;QAChB,CAAC;QAED,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,IAAI,CAAC,MAAM;YAClB,GAAG,EAAE,IAAI,CAAC,MAAM;YAChB,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,wBAAwB;QACxB,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC;QACxC,IAAI,MAAM,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;YAC9B,MAAM,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC;YAC9B,eAAe,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC;YACvC,MAAM,CAAC,IAAI,CAAC;gBACV,EAAE,EAAE,IAAI,CAAC,KAAK;gBACd,KAAK,EAAE,IAAI,CAAC,QAAQ;gBACpB,KAAK,EAAE,IAAI,CAAC,MAAM;gBAClB,GAAG,EAAE,IAAI,CAAC,MAAM;gBAChB,cAAc,EAAE,KAAK;gBACrB,SAAS,EAAE,IAAI;aAChB,CAAC,CAAC;YACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;QAED,eAAe;QACf,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACzC,MAAM,aAAa,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAEzC,OAAO;YACL,MAAM;YACN,QAAQ;YACR,aAAa;YACb,YAAY;YACZ,eAAe;SAChB,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,aAAa,CAAC,IAAY,EAAE,QAAgB;QAClD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QAEvC,+DAA+D;QAC/D,MAAM,WAAW,GAAG,QAAQ,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAE,CAAC,CAAC;QAErE,6DAA6D;QAC7D,IAAI,WAAW,EAAE,CAAC;YAChB,0BAA0B;YAC1B,MAAM,UAAU,GAAG,GAAG,GAAG,SAAS,CAAC;YACnC,KAAK,MAAM,CAAC,UAAU,EAAE,EAAE,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;gBAChD,IAAI,UAAU,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;oBACtC,0EAA0E;oBAC1E,OAAO;wBACL,KAAK,EAAE,UAAU;wBACjB,EAAE;wBACF,MAAM,EAAE,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,uBAAuB;qBACtD,CAAC;gBACJ,CAAC;YACH,CAAC;QACH,CAAC;QAED,iCAAiC;QACjC,KAAK,MAAM,CAAC,UAAU,EAAE,EAAE,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;YAChD,8EAA8E;YAC9E,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC;gBAAE,SAAS;YACvE,IAAI,CAAC,WAAW,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC;gBAAE,SAAS;YAEzD,IAAI,SAAS,CAAC,UAAU,CAAC,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;gBACvD,MAAM,WAAW,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC;gBACxD,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;oBACpB,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;gBACxD,CAAC;YACH,CAAC;QACH,CAAC;QAED,4BAA4B;QAC5B,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,GAAG,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC;QAChF,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAChD,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,QAAkB;QACvB,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACxC,IAAI,KAAK,KAAK,SAAS;gBAAE,SAAS;YAClC,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ;gBAAE,SAAS;YAE5F,+CAA+C;YAC/C,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC1B,KAAK,CAAC,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACnC,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACpB,CAAC;QACH,CAAC;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,IAAI,SAAS;QACX,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,UAAU,CAAC,KAAa;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,EAAU;QACjB,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IACnC,CAAC;CACF;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,QAAgB;IACtD,MAAM,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAErD,gBAAgB;IAChB,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QACjE,OAAO,oBAAoB,CAAC,OAAO,CAAC,CAAC;IACvC,CAAC;SAAM,CAAC;QACN,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC;IAC7B,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,oBAAoB,CAAC,OAAe;IAClD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IAExC,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAsB,CAAC;QAExD,2BAA2B;QAC3B,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC;YACvC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAED,uBAAuB;QACvB,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;YACnE,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtC,gDAAgD;gBAChD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACnD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;oBACpC,IAAI,KAAK,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;wBAC1C,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;oBACzB,CAAC;gBACH,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,mDAAmD;gBACnD,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;oBAC7D,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;gBACvB,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,mCAAmC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAClE,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,OAAe;IACxC,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC/B,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5C,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,eAAe;IAC7B,MAAM,MAAM,GAAG;QACb,KAAK;QACL,OAAO;QACP,MAAM;QACN,OAAO;QACP,QAAQ;QACR,OAAO;QACP,QAAQ;QACR,OAAO;QACP,OAAO;QACP,OAAO;QACP,KAAK;QACL,SAAS;QACT,GAAG;KACJ,CAAC;IAEF,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC9B,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;IAC1B,CAAC,CAAC,CAAC;IAEH,OAAO,KAAK,CAAC;AACf,CAAC"}
1
+ {"version":3,"file":"tokenizer.js","sourceRoot":"","sources":["../../src/ner/tokenizer.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAkEH;;GAEG;AACH,MAAM,CAAC,MAAM,wBAAwB,GAAoB;IACvD,SAAS,EAAE,GAAG;IACd,WAAW,EAAE,KAAK,EAAE,gCAAgC;CACrD,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,kBAAkB;IACrB,KAAK,CAAsB;IAC3B,YAAY,CAAsB;IAClC,MAAM,CAAkB;IACxB,WAAW,CAA0B;IAE7C,wCAAwC;IAChC,KAAK,GAAW,CAAC,CAAC,CAAE,MAAM;IAC1B,KAAK,GAAW,CAAC,CAAC,CAAE,OAAO;IAC3B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAC5B,KAAK,GAAW,CAAC,CAAC,CAAE,QAAQ;IAEpC,wBAAwB;IAChB,QAAQ,GAAW,KAAK,CAAC;IACzB,QAAQ,GAAW,MAAM,CAAC;IAC1B,QAAQ,GAAW,OAAO,CAAC;IAC3B,QAAQ,GAAW,OAAO,CAAC;IAEnC,YAAY,KAA0B,EAAE,SAAmC,EAAE;QAC3E,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,MAAM,GAAG,EAAE,GAAG,wBAAwB,EAAE,GAAG,MAAM,EAAE,CAAC;QAEzD,sBAAsB;QACtB,IAAI,CAAC,YAAY,GAAG,IAAI,GAAG,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC;YAChC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;QACnC,CAAC;QAED,iEAAiE;QACjE,IAAI,CAAC,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QAEzF,0CAA0C;QAC1C,IAAI,CAAC,mBAAmB,EAAE,CAAC;IAC7B,CAAC;IAED;;OAEG;IACK,mBAAmB;QACzB,oBAAoB;QACpB,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAC1B,IAAI,CAAC,QAAQ,GAAG,KAAK,CAAC;YACtB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,CAAC,QAAQ,GAAG,MAAM,CAAC;YACvB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YACzC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QAC5C,CAAC;QACD,aAAa;aACR,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YACjC,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;YAC5C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,GAAG,OAAO,CAAC;YACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC;QAC9C,CAAC;IACH,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,IAAY;QACnB,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,eAAe,GAAmC,EAAE,CAAC;QAE3D,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,CAAC;YACR,GAAG,EAAE,CAAC;YACN,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,kBAAkB;QAClB,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QAE1E,sCAAsC;QACtC,IAAI,GAAG,GAAG,CAAC,CAAC;QACZ,OAAO,GAAG,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC;YAClC,kBAAkB;YAClB,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAE,CAAC,EAAE,CAAC;gBACnC,GAAG,EAAE,CAAC;gBACN,SAAS;YACX,CAAC;YAED,4DAA4D;YAC5D,MAAM,EAAE,KAAK,EAAE,EAAE,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC,aAAa,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC;YAErE,MAAM,aAAa,GAAG,GAAG,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,GAAG,CAAC,CAAE,CAAC,CAAC;YAEtE,MAAM,CAAC,IAAI,CAAC;gBACV,EAAE;gBACF,KAAK;gBACL,KAAK,EAAE,GAAG;gBACV,GAAG,EAAE,GAAG,GAAG,MAAM;gBACjB,cAAc,EAAE,CAAC,aAAa,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;gBACxD,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,eAAe,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC;YAE1C,GAAG,IAAI,MAAM,CAAC;QAChB,CAAC;QAED,gBAAgB;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,EAAE,EAAE,IAAI,CAAC,KAAK;YACd,KAAK,EAAE,IAAI,CAAC,QAAQ;YACpB,KAAK,EAAE,IAAI,CAAC,MAAM;YAClB,GAAG,EAAE,IAAI,CAAC,MAAM;YAChB,cAAc,EAAE,KAAK;YACrB,SAAS,EAAE,IAAI;SAChB,CAAC,CAAC;QACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE3B,wBAAwB;QACxB,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC;QACxC,IAAI,MAAM,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;YAC9B,MAAM,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC;YAC9B,eAAe,CAAC,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC;YACvC,MAAM,CAAC,IAAI,CAAC;gBACV,EAAE,EAAE,IAAI,CAAC,KAAK;gBACd,KAAK,EAAE,IAAI,CAAC,QAAQ;gBACpB,KAAK,EAAE,IAAI,CAAC,MAAM;gBAClB,GAAG,EAAE,IAAI,CAAC,MAAM;gBAChB,cAAc,EAAE,KAAK;gBACrB,SAAS,EAAE,IAAI;aAChB,CAAC,CAAC;YACH,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;QAED,eAAe;QACf,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACzC,MAAM,aAAa,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QAEzC,OAAO;YACL,MAAM;YACN,QAAQ;YACR,aAAa;YACb,YAAY;YACZ,eAAe;SAChB,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,aAAa,CAAC,IAAY,EAAE,QAAgB;QAClD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QAEvC,+DAA+D;QAC/D,MAAM,WAAW,GAAG,QAAQ,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAE,CAAC,CAAC;QAErE,6DAA6D;QAC7D,IAAI,WAAW,EAAE,CAAC;YAChB,0BAA0B;YAC1B,MAAM,UAAU,GAAG,GAAG,GAAG,SAAS,CAAC;YACnC,KAAK,MAAM,CAAC,UAAU,EAAE,EAAE,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;gBAChD,IAAI,UAAU,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;oBACtC,0EAA0E;oBAC1E,OAAO;wBACL,KAAK,EAAE,UAAU;wBACjB,EAAE;wBACF,MAAM,EAAE,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,uBAAuB;qBACtD,CAAC;gBACJ,CAAC;YACH,CAAC;QACH,CAAC;QAED,iCAAiC;QACjC,KAAK,MAAM,CAAC,UAAU,EAAE,EAAE,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;YAChD,8EAA8E;YAC9E,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC;gBAAE,SAAS;YACvE,IAAI,CAAC,WAAW,IAAI,UAAU,CAAC,UAAU,CAAC,GAAG,CAAC;gBAAE,SAAS;YAEzD,IAAI,SAAS,CAAC,UAAU,CAAC,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;gBACvD,MAAM,WAAW,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC;gBACxD,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;oBACpB,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;gBACxD,CAAC;YACH,CAAC;QACH,CAAC;QAED,4BAA4B;QAC5B,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,GAAG,IAAI,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC;QAChF,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAChD,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,QAAkB;QACvB,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACxC,IAAI,KAAK,KAAK,SAAS;gBAAE,SAAS;YAClC,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,QAAQ;gBAAE,SAAS;YAE5F,+CAA+C;YAC/C,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC1B,KAAK,CAAC,IAAI,CAAC,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACnC,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACpB,CAAC;QACH,CAAC;QAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,IAAI,SAAS;QACX,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,UAAU,CAAC,KAAa;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,QAAQ,CAAC,EAAU;QACjB,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IACnC,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,QAAgB;IACtD,MAAM,EAAE,kBAAkB,EAAE,GAAG,MAAM,MAAM,CAAC,qBAAqB,CAAC,CAAC;IACnE,MAAM,OAAO,GAAG,MAAM,kBAAkB,EAAE,CAAC;IAC3C,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IAErD,gBAAgB;IAChB,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QACjE,OAAO,oBAAoB,CAAC,OAAO,CAAC,CAAC;IACvC,CAAC;SAAM,CAAC;QACN,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC;IAC7B,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,oBAAoB,CAAC,OAAe,EAAE,SAAyB,MAAM;IACnF,IAAI,MAAM,KAAK,MAAM,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QACxD,OAAO,oBAAoB,CAAC,OAAO,CAAC,CAAC;IACvC,CAAC;SAAM,CAAC;QACN,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC;IAC7B,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,oBAAoB,CAAC,OAAe;IAClD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IAExC,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAsB,CAAC;QAExD,2BAA2B;QAC3B,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC;YACvC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC;YACrC,CAAC;QACH,CAAC;QAED,uBAAuB;QACvB,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;YACnE,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtC,gDAAgD;gBAChD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACnD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;oBACpC,IAAI,KAAK,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;wBAC1C,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;oBACzB,CAAC;gBACH,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,mDAAmD;gBACnD,KAAK,MAAM,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;oBAC7D,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;gBACvB,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,mCAAmC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAClE,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,OAAe;IACxC,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC/B,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5C,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,eAAe;IAC7B,MAAM,MAAM,GAAG;QACb,KAAK;QACL,OAAO;QACP,MAAM;QACN,OAAO;QACP,QAAQ;QACR,OAAO;QACP,QAAQ;QACR,OAAO;QACP,OAAO;QACP,OAAO;QACP,KAAK;QACL,SAAS;QACT,GAAG;KACJ,CAAC;IAEF,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;IACxC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC9B,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;IAC1B,CAAC,CAAC,CAAC;IAEH,OAAO,KAAK,CAAC;AACf,CAAC"}
@@ -1,6 +1,7 @@
1
1
  /**
2
2
  * Semantic Data Loader
3
3
  * Handles automatic downloading, caching, and parsing of semantic enrichment data.
4
+ * Browser-compatible using storage abstraction layer
4
5
  *
5
6
  * Data sources:
6
7
  * - nam_dict.txt: Name-gender mappings from gender-guesser (~40K names)
@@ -8,10 +9,9 @@
8
9
  * - countryInfo.txt: Country names and codes (~250 countries)
9
10
  * - admin1CodesASCII.txt: First-level admin divisions (~4K regions)
10
11
  *
11
- * Data is cached in the same location as NER models:
12
- * - macOS: ~/Library/Caches/bridge-anonymization/semantic-data/
13
- * - Linux: ~/.cache/bridge-anonymization/semantic-data/
14
- * - Windows: %LOCALAPPDATA%/bridge-anonymization/semantic-data/
12
+ * Data is cached using the storage abstraction layer:
13
+ * - Node.js: Platform-specific cache directories
14
+ * - Browser: IndexedDB
15
15
  */
16
16
  /**
17
17
  * Loaded name-gender data
@@ -43,13 +43,13 @@ interface SemanticData {
43
43
  }
44
44
  /**
45
45
  * Gets the cache directory for semantic data
46
- * Uses platform-specific cache location (same as NER models)
46
+ * Uses platform-specific cache location (or virtual path in browser)
47
47
  */
48
- export declare function getSemanticDataCacheDir(): string;
48
+ export declare function getSemanticDataCacheDir(): Promise<string>;
49
49
  /**
50
50
  * Gets the path to the data directory (alias for backwards compatibility)
51
51
  */
52
- export declare function getDataDirectory(): string;
52
+ export declare function getDataDirectory(): Promise<string>;
53
53
  /**
54
54
  * Semantic data file info
55
55
  */
@@ -83,9 +83,10 @@ export type SemanticDownloadProgressCallback = (progress: {
83
83
  */
84
84
  export declare function isSemanticDataDownloaded(): Promise<boolean>;
85
85
  /**
86
- * Checks if the semantic data files are available (synchronous version)
86
+ * Checks if the semantic data files are available
87
+ * Note: This is now async to work with browser storage
87
88
  */
88
- export declare function isSemanticDataAvailable(): boolean;
89
+ export declare function isSemanticDataAvailable(): Promise<boolean>;
89
90
  /**
90
91
  * Downloads all semantic data files
91
92
  */
@@ -105,11 +106,11 @@ export declare function clearSemanticDataCache(): Promise<void>;
105
106
  /**
106
107
  * Gets info about semantic data files
107
108
  */
108
- export declare function getSemanticDataInfo(): {
109
+ export declare function getSemanticDataInfo(): Promise<{
109
110
  files: SemanticDataFileInfo[];
110
111
  cacheDir: string;
111
112
  totalSize: string;
112
- };
113
+ }>;
113
114
  /**
114
115
  * Initializes semantic data (downloads if needed, then loads)
115
116
  */
@@ -122,11 +123,18 @@ export declare function initializeSemanticData(options?: {
122
123
  * Loads semantic data from cached files
123
124
  * @throws Error if required data files are not available
124
125
  */
125
- export declare function loadSemanticData(): SemanticData;
126
+ export declare function loadSemanticData(): Promise<SemanticData>;
126
127
  /**
127
128
  * Gets the loaded semantic data (loads if not already loaded)
129
+ * Note: This is now async to work with browser storage
130
+ */
131
+ export declare function getSemanticData(): Promise<SemanticData>;
132
+ /**
133
+ * Gets the loaded semantic data synchronously
134
+ * Returns null if data is not yet loaded
135
+ * Use this for performance-critical code paths after initial loading
128
136
  */
129
- export declare function getSemanticData(): SemanticData;
137
+ export declare function getSemanticDataSync(): SemanticData | null;
130
138
  /**
131
139
  * Clears the loaded semantic data (useful for testing)
132
140
  */
@@ -134,15 +142,15 @@ export declare function clearSemanticData(): void;
134
142
  /**
135
143
  * Looks up gender for a name
136
144
  */
137
- export declare function lookupGender(name: string, locale?: string): string | undefined;
145
+ export declare function lookupGender(name: string, locale?: string): Promise<string | undefined>;
138
146
  /**
139
147
  * Looks up location type (city, country, or region)
140
148
  * Priority: country > major city (pop > 500K) > region > other cities
141
149
  */
142
- export declare function lookupLocationType(location: string): {
150
+ export declare function lookupLocationType(location: string): Promise<{
143
151
  type: "city" | "country" | "region";
144
152
  countryCode?: string;
145
- } | undefined;
153
+ } | undefined>;
146
154
  /**
147
155
  * Gets statistics about loaded data
148
156
  */
@@ -1 +1 @@
1
- {"version":3,"file":"semantic-data-loader.d.ts","sourceRoot":"","sources":["../../src/pipeline/semantic-data-loader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA+EH;;GAEG;AACH,UAAU,SAAS;IACjB,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAC1C;AAED;;GAEG;AACH,UAAU,SAAS;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,WAAW;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,EAAE,MAAM,CAAC;CACd;AAED;;GAEG;AACH,UAAU,YAAY;IACpB,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC9B,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC/B,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC/B,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC;IAClC,MAAM,EAAE,OAAO,CAAC;CACjB;AASD;;;GAGG;AACH,wBAAgB,uBAAuB,IAAI,MAAM,CA0BhD;AAED;;GAEG;AACH,wBAAgB,gBAAgB,IAAI,MAAM,CAEzC;AAMD;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,eAAe;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,mBAAmB;IACnB,GAAG,EAAE,MAAM,CAAC;IACZ,+BAA+B;IAC/B,QAAQ,EAAE,OAAO,CAAC;IAClB,kBAAkB;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,uBAAuB;IACvB,IAAI,EAAE,MAAM,CAAC;CACd;AAED;;GAEG;AACH,eAAO,MAAM,mBAAmB,EAAE,oBAAoB,EA6BrD,CAAC;AAMF;;GAEG;AACH,MAAM,MAAM,gCAAgC,GAAG,CAAC,QAAQ,EAAE;IACxD,IAAI,EAAE,MAAM,CAAC;IACb,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB,KAAK,IAAI,CAAC;AAEX;;GAEG;AACH,wBAAsB,wBAAwB,IAAI,OAAO,CAAC,OAAO,CAAC,CAcjE;AAED;;GAEG;AACH,wBAAgB,uBAAuB,IAAI,OAAO,CAYjD;AAoGD;;GAEG;AACH,wBAAsB,oBAAoB,CACxC,UAAU,CAAC,EAAE,gCAAgC,EAC7C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,GAClC,OAAO,CAAC,MAAM,CAAC,CAuCjB;AAED;;GAEG;AACH,wBAAsB,kBAAkB,CACtC,OAAO,GAAE;IACP,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,UAAU,CAAC,EAAE,gCAAgC,CAAC;IAC9C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,CAAC;CAChC,GACL,OAAO,CAAC,MAAM,CAAC,CAyBjB;AAED;;GAEG;AACH,wBAAsB,sBAAsB,IAAI,OAAO,CAAC,IAAI,CAAC,CAK5D;AAED;;GAEG;AACH,wBAAgB,mBAAmB,IAAI;IACrC,KAAK,EAAE,oBAAoB,EAAE,CAAC;IAC9B,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;CACnB,CAMA;AAsQD;;GAEG;AACH,wBAAsB,sBAAsB,CAC1C,OAAO,GAAE;IACP,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,UAAU,CAAC,EAAE,gCAAgC,CAAC;IAC9C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,CAAC;CAChC,GACL,OAAO,CAAC,IAAI,CAAC,CAMf;AAED;;;GAGG;AACH,wBAAgB,gBAAgB,IAAI,YAAY,CA4B/C;AAED;;GAEG;AACH,wBAAgB,eAAe,IAAI,YAAY,CAK9C;AAED;;GAEG;AACH,wBAAgB,iBAAiB,IAAI,IAAI,CAExC;AAED;;GAEG;AACH,wBAAgB,YAAY,CAC1B,IAAI,EAAE,MAAM,EACZ,MAAM,CAAC,EAAE,MAAM,GACd,MAAM,GAAG,SAAS,CAiBpB;AAOD;;;GAGG;AACH,wBAAgB,kBAAkB,CAChC,QAAQ,EAAE,MAAM,GACf;IAAE,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,QAAQ,CAAC;IAAC,WAAW,CAAC,EAAE,MAAM,CAAA;CAAE,GAAG,SAAS,CA4B3E;AAED;;GAEG;AACH,wBAAgB,YAAY,IAAI;IAC9B,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,OAAO,CAAC;CACjB,CAYA"}
1
+ {"version":3,"file":"semantic-data-loader.d.ts","sourceRoot":"","sources":["../../src/pipeline/semantic-data-loader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA6EH;;GAEG;AACH,UAAU,SAAS;IACjB,MAAM,EAAE,MAAM,CAAC;IACf,eAAe,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAC1C;AAED;;GAEG;AACH,UAAU,SAAS;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,WAAW;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,EAAE,MAAM,CAAC;CACd;AAED;;GAEG;AACH,UAAU,YAAY;IACpB,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC9B,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAC/B,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC/B,OAAO,EAAE,GAAG,CAAC,MAAM,EAAE,WAAW,CAAC,CAAC;IAClC,MAAM,EAAE,OAAO,CAAC;CACjB;AAsBD;;;GAGG;AACH,wBAAsB,uBAAuB,IAAI,OAAO,CAAC,MAAM,CAAC,CAG/D;AAED;;GAEG;AACH,wBAAsB,gBAAgB,IAAI,OAAO,CAAC,MAAM,CAAC,CAExD;AAMD;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,eAAe;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,mBAAmB;IACnB,GAAG,EAAE,MAAM,CAAC;IACZ,+BAA+B;IAC/B,QAAQ,EAAE,OAAO,CAAC;IAClB,kBAAkB;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,uBAAuB;IACvB,IAAI,EAAE,MAAM,CAAC;CACd;AAQD;;GAEG;AACH,eAAO,MAAM,mBAAmB,EAAE,oBAAoB,EA6BrD,CAAC;AAMF;;GAEG;AACH,MAAM,MAAM,gCAAgC,GAAG,CAAC,QAAQ,EAAE;IACxD,IAAI,EAAE,MAAM,CAAC;IACb,eAAe,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB,KAAK,IAAI,CAAC;AAEX;;GAEG;AACH,wBAAsB,wBAAwB,IAAI,OAAO,CAAC,OAAO,CAAC,CAgBjE;AAED;;;GAGG;AACH,wBAAsB,uBAAuB,IAAI,OAAO,CAAC,OAAO,CAAC,CAEhE;AAuFD;;GAEG;AACH,wBAAsB,oBAAoB,CACxC,UAAU,CAAC,EAAE,gCAAgC,EAC7C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,GAClC,OAAO,CAAC,MAAM,CAAC,CA+BjB;AAED;;GAEG;AACH,wBAAsB,kBAAkB,CACtC,OAAO,GAAE;IACP,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,UAAU,CAAC,EAAE,gCAAgC,CAAC;IAC9C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,CAAC;CAChC,GACL,OAAO,CAAC,MAAM,CAAC,CAyBjB;AAED;;GAEG;AACH,wBAAsB,sBAAsB,IAAI,OAAO,CAAC,IAAI,CAAC,CAM5D;AAED;;GAEG;AACH,wBAAsB,mBAAmB,IAAI,OAAO,CAAC;IACnD,KAAK,EAAE,oBAAoB,EAAE,CAAC;IAC9B,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;CACnB,CAAC,CAOD;AAqPD;;GAEG;AACH,wBAAsB,sBAAsB,CAC1C,OAAO,GAAE;IACP,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,UAAU,CAAC,EAAE,gCAAgC,CAAC;IAC9C,QAAQ,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,IAAI,CAAC;CAChC,GACL,OAAO,CAAC,IAAI,CAAC,CAMf;AAED;;;GAGG;AACH,wBAAsB,gBAAgB,IAAI,OAAO,CAAC,YAAY,CAAC,CAqD9D;AAED;;;GAGG;AACH,wBAAsB,eAAe,IAAI,OAAO,CAAC,YAAY,CAAC,CAK7D;AAED;;;;GAIG;AACH,wBAAgB,mBAAmB,IAAI,YAAY,GAAG,IAAI,CAKzD;AAED;;GAEG;AACH,wBAAgB,iBAAiB,IAAI,IAAI,CAExC;AAED;;GAEG;AACH,wBAAsB,YAAY,CAChC,IAAI,EAAE,MAAM,EACZ,MAAM,CAAC,EAAE,MAAM,GACd,OAAO,CAAC,MAAM,GAAG,SAAS,CAAC,CAiB7B;AAOD;;;GAGG;AACH,wBAAsB,kBAAkB,CACtC,QAAQ,EAAE,MAAM,GACf,OAAO,CACR;IAAE,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,QAAQ,CAAC;IAAC,WAAW,CAAC,EAAE,MAAM,CAAA;CAAE,GAAG,SAAS,CAC1E,CA4BA;AAED;;GAEG;AACH,wBAAgB,YAAY,IAAI;IAC9B,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,OAAO,CAAC;CACjB,CAYA"}