@cogitator-ai/rag 0.1.10 → 0.3.0

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 (81) hide show
  1. package/README.md +66 -31
  2. package/dist/chunkers/create-chunker.js.map +1 -1
  3. package/dist/chunkers/index.d.ts.map +1 -1
  4. package/dist/chunkers/index.js.map +1 -1
  5. package/dist/chunkers/recursive-chunker.d.ts +1 -2
  6. package/dist/chunkers/recursive-chunker.d.ts.map +1 -1
  7. package/dist/chunkers/recursive-chunker.js +68 -58
  8. package/dist/chunkers/recursive-chunker.js.map +1 -1
  9. package/dist/chunkers/semantic-chunker.d.ts.map +1 -1
  10. package/dist/chunkers/semantic-chunker.js +5 -2
  11. package/dist/chunkers/semantic-chunker.js.map +1 -1
  12. package/dist/index.d.ts +2 -2
  13. package/dist/index.d.ts.map +1 -1
  14. package/dist/index.js +4 -2
  15. package/dist/index.js.map +1 -1
  16. package/dist/loaders/csv-loader.d.ts.map +1 -1
  17. package/dist/loaders/csv-loader.js +9 -4
  18. package/dist/loaders/csv-loader.js.map +1 -1
  19. package/dist/loaders/html-loader.d.ts +4 -0
  20. package/dist/loaders/html-loader.d.ts.map +1 -1
  21. package/dist/loaders/html-loader.js +26 -2
  22. package/dist/loaders/html-loader.js.map +1 -1
  23. package/dist/loaders/index.d.ts.map +1 -1
  24. package/dist/loaders/index.js.map +1 -1
  25. package/dist/loaders/json-loader.js +1 -1
  26. package/dist/loaders/json-loader.js.map +1 -1
  27. package/dist/loaders/markdown-loader.d.ts.map +1 -1
  28. package/dist/loaders/markdown-loader.js +7 -3
  29. package/dist/loaders/markdown-loader.js.map +1 -1
  30. package/dist/loaders/pdf-loader.d.ts +3 -3
  31. package/dist/loaders/pdf-loader.d.ts.map +1 -1
  32. package/dist/loaders/pdf-loader.js +45 -42
  33. package/dist/loaders/pdf-loader.js.map +1 -1
  34. package/dist/loaders/text-loader.js +1 -1
  35. package/dist/loaders/text-loader.js.map +1 -1
  36. package/dist/loaders/web-loader.d.ts +29 -1
  37. package/dist/loaders/web-loader.d.ts.map +1 -1
  38. package/dist/loaders/web-loader.js +227 -92
  39. package/dist/loaders/web-loader.js.map +1 -1
  40. package/dist/rag-builder.d.ts +10 -1
  41. package/dist/rag-builder.d.ts.map +1 -1
  42. package/dist/rag-builder.js +70 -7
  43. package/dist/rag-builder.js.map +1 -1
  44. package/dist/rag-pipeline.d.ts +1 -0
  45. package/dist/rag-pipeline.d.ts.map +1 -1
  46. package/dist/rag-pipeline.js +55 -19
  47. package/dist/rag-pipeline.js.map +1 -1
  48. package/dist/rerankers/index.d.ts.map +1 -1
  49. package/dist/rerankers/index.js.map +1 -1
  50. package/dist/retrievers/chunk-indexer.d.ts +19 -0
  51. package/dist/retrievers/chunk-indexer.d.ts.map +1 -0
  52. package/dist/retrievers/chunk-indexer.js +8 -0
  53. package/dist/retrievers/chunk-indexer.js.map +1 -0
  54. package/dist/retrievers/create-retriever.d.ts +1 -0
  55. package/dist/retrievers/create-retriever.d.ts.map +1 -1
  56. package/dist/retrievers/create-retriever.js +1 -0
  57. package/dist/retrievers/create-retriever.js.map +1 -1
  58. package/dist/retrievers/hybrid-retriever.d.ts +8 -1
  59. package/dist/retrievers/hybrid-retriever.d.ts.map +1 -1
  60. package/dist/retrievers/hybrid-retriever.js +18 -5
  61. package/dist/retrievers/hybrid-retriever.js.map +1 -1
  62. package/dist/retrievers/index.d.ts +2 -0
  63. package/dist/retrievers/index.d.ts.map +1 -1
  64. package/dist/retrievers/index.js +1 -0
  65. package/dist/retrievers/index.js.map +1 -1
  66. package/dist/retrievers/mmr-retriever.d.ts +5 -0
  67. package/dist/retrievers/mmr-retriever.d.ts.map +1 -1
  68. package/dist/retrievers/mmr-retriever.js +24 -2
  69. package/dist/retrievers/mmr-retriever.js.map +1 -1
  70. package/dist/retrievers/multi-query-retriever.d.ts +12 -1
  71. package/dist/retrievers/multi-query-retriever.d.ts.map +1 -1
  72. package/dist/retrievers/multi-query-retriever.js +46 -7
  73. package/dist/retrievers/multi-query-retriever.js.map +1 -1
  74. package/dist/retrievers/similarity-retriever.d.ts.map +1 -1
  75. package/dist/retrievers/similarity-retriever.js +2 -1
  76. package/dist/retrievers/similarity-retriever.js.map +1 -1
  77. package/dist/tools.d.ts +11 -2
  78. package/dist/tools.d.ts.map +1 -1
  79. package/dist/tools.js +39 -7
  80. package/dist/tools.js.map +1 -1
  81. package/package.json +15 -14
@@ -1,4 +1,11 @@
1
1
  import { RAGPipelineConfigSchema } from './schema.js';
2
+ import { isChunkIndexer } from './retrievers/chunk-indexer.js';
3
+ const STORE_CONCURRENCY = 16;
4
+ function definedEntries(value) {
5
+ if (!value)
6
+ return {};
7
+ return Object.fromEntries(Object.entries(value).filter(([, v]) => v !== undefined));
8
+ }
2
9
  export class RAGPipeline {
3
10
  config;
4
11
  deps;
@@ -10,24 +17,28 @@ export class RAGPipeline {
10
17
  async ingest(source) {
11
18
  const documents = await this.deps.loader.load(source);
12
19
  let totalChunks = 0;
20
+ let ingestedDocuments = 0;
13
21
  for (const doc of documents) {
14
22
  const chunks = await this.chunkDocument(doc);
15
- if (chunks.length === 0)
16
- continue;
17
- const texts = chunks.map((c) => c.content);
18
- const vectors = await this.deps.embeddingService.embedBatch(texts);
19
- if (vectors.length !== chunks.length) {
20
- throw new Error(`Embedding count mismatch: got ${vectors.length} vectors for ${chunks.length} chunks`);
23
+ if (chunks.length > 0) {
24
+ const vectors = await this.deps.embeddingService.embedBatch(chunks.map((c) => c.content));
25
+ if (vectors.length !== chunks.length) {
26
+ throw new Error(`Embedding count mismatch: got ${vectors.length} vectors for ${chunks.length} chunks`);
27
+ }
28
+ await this.storeChunks(chunks, vectors, doc);
29
+ totalChunks += chunks.length;
30
+ this.stats.chunksStored += chunks.length;
21
31
  }
22
- await Promise.all(chunks.map((chunk, i) => this.storeChunk(chunk, vectors[i], doc)));
23
- totalChunks += chunks.length;
32
+ ingestedDocuments++;
33
+ this.stats.documentsIngested++;
24
34
  }
25
- this.stats.documentsIngested += documents.length;
26
- this.stats.chunksStored += totalChunks;
27
- return { documents: documents.length, chunks: totalChunks };
35
+ return { documents: ingestedDocuments, chunks: totalChunks };
28
36
  }
29
37
  async query(text, options) {
30
- const merged = { ...this.config.retrieval, ...options };
38
+ const merged = {
39
+ ...this.config.retrieval,
40
+ ...definedEntries(options),
41
+ };
31
42
  const results = await this.deps.retriever.retrieve(text, merged);
32
43
  this.stats.queriesProcessed++;
33
44
  if (this.deps.reranker && this.config.reranking?.enabled) {
@@ -39,24 +50,49 @@ export class RAGPipeline {
39
50
  return { ...this.stats };
40
51
  }
41
52
  async chunkDocument(doc) {
42
- return this.deps.chunker.chunk(doc.content, doc.id);
53
+ const chunks = await this.deps.chunker.chunk(doc.content, doc.id);
54
+ return chunks.filter((chunk) => chunk.content.trim().length > 0);
55
+ }
56
+ async storeChunks(chunks, vectors, doc) {
57
+ let next = 0;
58
+ const worker = async () => {
59
+ while (next < chunks.length) {
60
+ const index = next++;
61
+ await this.storeChunk(chunks[index], vectors[index], doc);
62
+ }
63
+ };
64
+ await Promise.all(Array.from({ length: Math.min(STORE_CONCURRENCY, chunks.length) }, () => worker()));
43
65
  }
44
66
  async storeChunk(chunk, vector, doc) {
67
+ const metadata = {
68
+ ...doc.metadata,
69
+ ...chunk.metadata,
70
+ documentId: doc.id,
71
+ source: doc.source,
72
+ sourceType: doc.sourceType,
73
+ order: chunk.order,
74
+ startOffset: chunk.startOffset,
75
+ endOffset: chunk.endOffset,
76
+ };
45
77
  const result = await this.deps.embeddingAdapter.addEmbedding({
46
78
  sourceId: chunk.id,
47
79
  sourceType: 'document',
48
80
  vector,
49
81
  content: chunk.content,
50
- metadata: {
51
- ...chunk.metadata,
52
- documentId: doc.id,
53
- source: doc.source,
54
- order: chunk.order,
55
- },
82
+ metadata,
56
83
  });
57
84
  if (!result.success) {
58
85
  throw new Error(result.error);
59
86
  }
87
+ if (isChunkIndexer(this.deps.retriever)) {
88
+ this.deps.retriever.indexChunk({
89
+ embeddingId: result.data.id,
90
+ chunkId: chunk.id,
91
+ documentId: doc.id,
92
+ content: chunk.content,
93
+ metadata,
94
+ });
95
+ }
60
96
  }
61
97
  }
62
98
  //# sourceMappingURL=rag-pipeline.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"rag-pipeline.js","sourceRoot":"","sources":["../src/rag-pipeline.ts"],"names":[],"mappings":"AAcA,OAAO,EAAE,uBAAuB,EAAE,MAAM,UAAU,CAAC;AAWnD,MAAM,OAAO,WAAW;IACL,MAAM,CAAoB;IAC1B,IAAI,CAAkB;IAC/B,KAAK,GAAG,EAAE,iBAAiB,EAAE,CAAC,EAAE,YAAY,EAAE,CAAC,EAAE,gBAAgB,EAAE,CAAC,EAAE,CAAC;IAE/E,YAAY,MAAyB,EAAE,IAAqB,EAAE,aAAa,GAAG,KAAK;QACjF,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,uBAAuB,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QAC7E,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC;IACnB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,MAAc;QACzB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACtD,IAAI,WAAW,GAAG,CAAC,CAAC;QAEpB,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;YAC5B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC;YAC7C,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAElC,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YAC3C,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YAEnE,IAAI,OAAO,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM,EAAE,CAAC;gBACrC,MAAM,IAAI,KAAK,CACb,iCAAiC,OAAO,CAAC,MAAM,gBAAgB,MAAM,CAAC,MAAM,SAAS,CACtF,CAAC;YACJ,CAAC;YAED,MAAM,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;YAErF,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;QAC/B,CAAC;QAED,IAAI,CAAC,KAAK,CAAC,iBAAiB,IAAI,SAAS,CAAC,MAAM,CAAC;QACjD,IAAI,CAAC,KAAK,CAAC,YAAY,IAAI,WAAW,CAAC;QAEvC,OAAO,EAAE,SAAS,EAAE,SAAS,CAAC,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAC9D,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,IAAY,EAAE,OAAkC;QAC1D,MAAM,MAAM,GAA6B,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,GAAG,OAAO,EAAE,CAAC;QAClF,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;QACjE,IAAI,CAAC,KAAK,CAAC,gBAAgB,EAAE,CAAC;QAE9B,IAAI,IAAI,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,OAAO,EAAE,CAAC;YACzD,OAAO,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAC9E,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAED,QAAQ;QACN,OAAO,EAAE,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC;IAC3B,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,GAAgB;QAC1C,OAAO,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;IACtD,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,KAAoB,EAAE,MAAgB,EAAE,GAAgB;QAC/E,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,YAAY,CAAC;YAC3D,QAAQ,EAAE,KAAK,CAAC,EAAE;YAClB,UAAU,EAAE,UAAU;YACtB,MAAM;YACN,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,QAAQ,EAAE;gBACR,GAAG,KAAK,CAAC,QAAQ;gBACjB,UAAU,EAAE,GAAG,CAAC,EAAE;gBAClB,MAAM,EAAE,GAAG,CAAC,MAAM;gBAClB,KAAK,EAAE,KAAK,CAAC,KAAK;aACnB;SACF,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;IACH,CAAC;CACF"}
1
+ {"version":3,"file":"rag-pipeline.js","sourceRoot":"","sources":["../src/rag-pipeline.ts"],"names":[],"mappings":"AAcA,OAAO,EAAE,uBAAuB,EAAE,MAAM,aAAa,CAAC;AACtD,OAAO,EAAE,cAAc,EAAE,MAAM,+BAA+B,CAAC;AAW/D,MAAM,iBAAiB,GAAG,EAAE,CAAC;AAE7B,SAAS,cAAc,CAAmB,KAAoB;IAC5D,IAAI,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IACtB,OAAO,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC,CAAe,CAAC;AACpG,CAAC;AAED,MAAM,OAAO,WAAW;IACL,MAAM,CAAoB;IAC1B,IAAI,CAAkB;IAC/B,KAAK,GAAG,EAAE,iBAAiB,EAAE,CAAC,EAAE,YAAY,EAAE,CAAC,EAAE,gBAAgB,EAAE,CAAC,EAAE,CAAC;IAE/E,YAAY,MAAyB,EAAE,IAAqB,EAAE,aAAa,GAAG,KAAK;QACjF,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,uBAAuB,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QAC7E,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC;IACnB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,MAAc;QACzB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACtD,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,IAAI,iBAAiB,GAAG,CAAC,CAAC;QAE1B,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;YAC5B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC;YAC7C,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACtB,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,UAAU,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;gBAE1F,IAAI,OAAO,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM,EAAE,CAAC;oBACrC,MAAM,IAAI,KAAK,CACb,iCAAiC,OAAO,CAAC,MAAM,gBAAgB,MAAM,CAAC,MAAM,SAAS,CACtF,CAAC;gBACJ,CAAC;gBAED,MAAM,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,OAAO,EAAE,GAAG,CAAC,CAAC;gBAC7C,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;gBAC7B,IAAI,CAAC,KAAK,CAAC,YAAY,IAAI,MAAM,CAAC,MAAM,CAAC;YAC3C,CAAC;YACD,iBAAiB,EAAE,CAAC;YACpB,IAAI,CAAC,KAAK,CAAC,iBAAiB,EAAE,CAAC;QACjC,CAAC;QAED,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAC/D,CAAC;IAED,KAAK,CAAC,KAAK,CAAC,IAAY,EAAE,OAAkC;QAC1D,MAAM,MAAM,GAA6B;YACvC,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS;YACxB,GAAG,cAAc,CAAC,OAAO,CAAC;SAC3B,CAAC;QACF,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;QACjE,IAAI,CAAC,KAAK,CAAC,gBAAgB,EAAE,CAAC;QAE9B,IAAI,IAAI,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,OAAO,EAAE,CAAC;YACzD,OAAO,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAC9E,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAED,QAAQ;QACN,OAAO,EAAE,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC;IAC3B,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,GAAgB;QAC1C,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QAClE,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACnE,CAAC;IAEO,KAAK,CAAC,WAAW,CACvB,MAAuB,EACvB,OAAmB,EACnB,GAAgB;QAEhB,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,MAAM,MAAM,GAAG,KAAK,IAAI,EAAE;YACxB,OAAO,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;gBAC5B,MAAM,KAAK,GAAG,IAAI,EAAE,CAAC;gBACrB,MAAM,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,KAAK,CAAE,EAAE,OAAO,CAAC,KAAK,CAAE,EAAE,GAAG,CAAC,CAAC;YAC9D,CAAC;QACH,CAAC,CAAC;QACF,MAAM,OAAO,CAAC,GAAG,CACf,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,iBAAiB,EAAE,MAAM,CAAC,MAAM,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,MAAM,EAAE,CAAC,CACnF,CAAC;IACJ,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,KAAoB,EAAE,MAAgB,EAAE,GAAgB;QAC/E,MAAM,QAAQ,GAA4B;YACxC,GAAG,GAAG,CAAC,QAAQ;YACf,GAAG,KAAK,CAAC,QAAQ;YACjB,UAAU,EAAE,GAAG,CAAC,EAAE;YAClB,MAAM,EAAE,GAAG,CAAC,MAAM;YAClB,UAAU,EAAE,GAAG,CAAC,UAAU;YAC1B,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,WAAW,EAAE,KAAK,CAAC,WAAW;YAC9B,SAAS,EAAE,KAAK,CAAC,SAAS;SAC3B,CAAC;QAEF,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,YAAY,CAAC;YAC3D,QAAQ,EAAE,KAAK,CAAC,EAAE;YAClB,UAAU,EAAE,UAAU;YACtB,MAAM;YACN,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,QAAQ;SACT,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC;gBAC7B,WAAW,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;gBAC3B,OAAO,EAAE,KAAK,CAAC,EAAE;gBACjB,UAAU,EAAE,GAAG,CAAC,EAAE;gBAClB,OAAO,EAAE,KAAK,CAAC,OAAO;gBACtB,QAAQ;aACT,CAAC,CAAC;QACL,CAAC;IACH,CAAC;CACF"}
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/rerankers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,KAAK,iBAAiB,EAAE,MAAM,gBAAgB,CAAC;AACrE,OAAO,EAAE,cAAc,EAAE,KAAK,oBAAoB,EAAE,MAAM,mBAAmB,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/rerankers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,KAAK,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AACxE,OAAO,EAAE,cAAc,EAAE,KAAK,oBAAoB,EAAE,MAAM,sBAAsB,CAAC"}
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/rerankers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAA0B,MAAM,gBAAgB,CAAC;AACrE,OAAO,EAAE,cAAc,EAA6B,MAAM,mBAAmB,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/rerankers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAA0B,MAAM,mBAAmB,CAAC;AACxE,OAAO,EAAE,cAAc,EAA6B,MAAM,sBAAsB,CAAC"}
@@ -0,0 +1,19 @@
1
+ import type { Retriever } from '@cogitator-ai/types';
2
+ export interface IndexedChunk {
3
+ embeddingId: string;
4
+ chunkId: string;
5
+ documentId: string;
6
+ content: string;
7
+ metadata: Record<string, unknown>;
8
+ }
9
+ /**
10
+ * A retriever that maintains its own index alongside the vector store
11
+ * (e.g. a local BM25 index). `RAGPipeline.ingest()` calls `indexChunk()` for
12
+ * every stored chunk.
13
+ */
14
+ export interface ChunkIndexer {
15
+ indexChunk(chunk: IndexedChunk): void;
16
+ }
17
+ export declare function isChunkIndexer(retriever: Retriever): retriever is Retriever & ChunkIndexer;
18
+ export declare function resultSource(metadata: Record<string, unknown> | undefined, fallback: string): string;
19
+ //# sourceMappingURL=chunk-indexer.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"chunk-indexer.d.ts","sourceRoot":"","sources":["../../src/retrievers/chunk-indexer.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,qBAAqB,CAAC;AAErD,MAAM,WAAW,YAAY;IAC3B,WAAW,EAAE,MAAM,CAAC;IACpB,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,EAAE,MAAM,CAAC;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACnC;AAED;;;;GAIG;AACH,MAAM,WAAW,YAAY;IAC3B,UAAU,CAAC,KAAK,EAAE,YAAY,GAAG,IAAI,CAAC;CACvC;AAED,wBAAgB,cAAc,CAAC,SAAS,EAAE,SAAS,GAAG,SAAS,IAAI,SAAS,GAAG,YAAY,CAE1F;AAED,wBAAgB,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,EAAE,QAAQ,EAAE,MAAM,UAG3F"}
@@ -0,0 +1,8 @@
1
+ export function isChunkIndexer(retriever) {
2
+ return typeof retriever.indexChunk === 'function';
3
+ }
4
+ export function resultSource(metadata, fallback) {
5
+ const source = metadata?.source;
6
+ return typeof source === 'string' ? source : fallback;
7
+ }
8
+ //# sourceMappingURL=chunk-indexer.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"chunk-indexer.js","sourceRoot":"","sources":["../../src/retrievers/chunk-indexer.ts"],"names":[],"mappings":"AAmBA,MAAM,UAAU,cAAc,CAAC,SAAoB;IACjD,OAAO,OAAQ,SAAmC,CAAC,UAAU,KAAK,UAAU,CAAC;AAC/E,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,QAA6C,EAAE,QAAgB;IAC1F,MAAM,MAAM,GAAG,QAAQ,EAAE,MAAM,CAAC;IAChC,OAAO,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC;AACxD,CAAC"}
@@ -27,6 +27,7 @@ interface MultiQueryDeps {
27
27
  baseRetriever: Retriever;
28
28
  expandQuery: (query: string) => Promise<string[]>;
29
29
  topK?: number;
30
+ maxQueries?: number;
30
31
  }
31
32
  export type CreateRetrieverConfig = SimilarityDeps | MMRDeps | HybridDeps | MultiQueryDeps;
32
33
  export declare function createRetriever(config: CreateRetrieverConfig): Retriever;
@@ -1 +1 @@
1
- {"version":3,"file":"create-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,SAAS,EACT,gBAAgB,EAChB,gBAAgB,EAChB,mBAAmB,EACpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAMzD,UAAU,iBAAiB;IACzB,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;CACpC;AAED,UAAU,cAAe,SAAQ,iBAAiB;IAChD,QAAQ,EAAE,YAAY,CAAC;IACvB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,OAAQ,SAAQ,iBAAiB;IACzC,QAAQ,EAAE,KAAK,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,UAAU;IAClB,QAAQ,EAAE,QAAQ,CAAC;IACnB,YAAY,EAAE,YAAY,CAAC;IAC3B,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,cAAc;IACtB,QAAQ,EAAE,aAAa,CAAC;IACxB,aAAa,EAAE,SAAS,CAAC;IACzB,WAAW,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAClD,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,MAAM,MAAM,qBAAqB,GAAG,cAAc,GAAG,OAAO,GAAG,UAAU,GAAG,cAAc,CAAC;AAE3F,wBAAgB,eAAe,CAAC,MAAM,EAAE,qBAAqB,GAAG,SAAS,CAuCxE"}
1
+ {"version":3,"file":"create-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,SAAS,EACT,gBAAgB,EAChB,gBAAgB,EAChB,mBAAmB,EACpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAMzD,UAAU,iBAAiB;IACzB,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;CACpC;AAED,UAAU,cAAe,SAAQ,iBAAiB;IAChD,QAAQ,EAAE,YAAY,CAAC;IACvB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,OAAQ,SAAQ,iBAAiB;IACzC,QAAQ,EAAE,KAAK,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,UAAU;IAClB,QAAQ,EAAE,QAAQ,CAAC;IACnB,YAAY,EAAE,YAAY,CAAC;IAC3B,OAAO,CAAC,EAAE,mBAAmB,CAAC;IAC9B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,cAAc;IACtB,QAAQ,EAAE,aAAa,CAAC;IACxB,aAAa,EAAE,SAAS,CAAC;IACzB,WAAW,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAClD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,MAAM,qBAAqB,GAAG,cAAc,GAAG,OAAO,GAAG,UAAU,GAAG,cAAc,CAAC;AAE3F,wBAAgB,eAAe,CAAC,MAAM,EAAE,qBAAqB,GAAG,SAAS,CAwCxE"}
@@ -31,6 +31,7 @@ export function createRetriever(config) {
31
31
  baseRetriever: config.baseRetriever,
32
32
  expandQuery: config.expandQuery,
33
33
  defaultTopK: config.topK,
34
+ defaultMaxQueries: config.maxQueries,
34
35
  });
35
36
  default: {
36
37
  const _exhaustive = config;
@@ -1 +1 @@
1
- {"version":3,"file":"create-retriever.js","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,OAAO,EAAE,mBAAmB,EAAE,MAAM,yBAAyB,CAAC;AAqC9D,MAAM,UAAU,eAAe,CAAC,MAA6B;IAC3D,QAAQ,MAAM,CAAC,QAAQ,EAAE,CAAC;QACxB,KAAK,YAAY;YACf,OAAO,IAAI,mBAAmB,CAAC;gBAC7B,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,KAAK;YACR,OAAO,IAAI,YAAY,CAAC;gBACtB,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,aAAa,EAAE,MAAM,CAAC,MAAM;gBAC5B,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,QAAQ;YACX,OAAO,IAAI,eAAe,CAAC;gBACzB,YAAY,EAAE,MAAM,CAAC,YAAY;gBACjC,cAAc,EAAE,MAAM,CAAC,OAAO;gBAC9B,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,aAAa;YAChB,OAAO,IAAI,mBAAmB,CAAC;gBAC7B,aAAa,EAAE,MAAM,CAAC,aAAa;gBACnC,WAAW,EAAE,MAAM,CAAC,WAAW;gBAC/B,WAAW,EAAE,MAAM,CAAC,IAAI;aACzB,CAAC,CAAC;QAEL,OAAO,CAAC,CAAC,CAAC;YACR,MAAM,WAAW,GAAU,MAAM,CAAC;YAClC,MAAM,IAAI,KAAK,CAAC,+BAA+B,IAAI,CAAC,SAAS,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;QAChF,CAAC;IACH,CAAC;AACH,CAAC"}
1
+ {"version":3,"file":"create-retriever.js","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AACxD,OAAO,EAAE,mBAAmB,EAAE,MAAM,4BAA4B,CAAC;AAsCjE,MAAM,UAAU,eAAe,CAAC,MAA6B;IAC3D,QAAQ,MAAM,CAAC,QAAQ,EAAE,CAAC;QACxB,KAAK,YAAY;YACf,OAAO,IAAI,mBAAmB,CAAC;gBAC7B,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,KAAK;YACR,OAAO,IAAI,YAAY,CAAC;gBACtB,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;gBACzC,aAAa,EAAE,MAAM,CAAC,MAAM;gBAC5B,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,QAAQ;YACX,OAAO,IAAI,eAAe,CAAC;gBACzB,YAAY,EAAE,MAAM,CAAC,YAAY;gBACjC,cAAc,EAAE,MAAM,CAAC,OAAO;gBAC9B,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,gBAAgB,EAAE,MAAM,CAAC,SAAS;aACnC,CAAC,CAAC;QAEL,KAAK,aAAa;YAChB,OAAO,IAAI,mBAAmB,CAAC;gBAC7B,aAAa,EAAE,MAAM,CAAC,aAAa;gBACnC,WAAW,EAAE,MAAM,CAAC,WAAW;gBAC/B,WAAW,EAAE,MAAM,CAAC,IAAI;gBACxB,iBAAiB,EAAE,MAAM,CAAC,UAAU;aACrC,CAAC,CAAC;QAEL,OAAO,CAAC,CAAC,CAAC;YACR,MAAM,WAAW,GAAU,MAAM,CAAC;YAClC,MAAM,IAAI,KAAK,CAAC,+BAA+B,IAAI,CAAC,SAAS,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC;QAChF,CAAC;IACH,CAAC;AACH,CAAC"}
@@ -1,17 +1,24 @@
1
1
  import type { Retriever, RetrievalConfig, RetrievalResult, HybridSearchWeights } from '@cogitator-ai/types';
2
2
  import type { HybridSearch } from '@cogitator-ai/memory';
3
+ import { type ChunkIndexer, type IndexedChunk } from './chunk-indexer.js';
3
4
  export interface HybridRetrieverConfig {
4
5
  hybridSearch: HybridSearch;
5
6
  defaultWeights?: HybridSearchWeights;
6
7
  defaultTopK?: number;
7
8
  defaultThreshold?: number;
8
9
  }
9
- export declare class HybridRetriever implements Retriever {
10
+ /**
11
+ * Vector + BM25 retrieval fused with reciprocal rank fusion. Chunks ingested through
12
+ * `RAGPipeline` are added to the HybridSearch keyword index automatically.
13
+ */
14
+ export declare class HybridRetriever implements Retriever, ChunkIndexer {
10
15
  private readonly hybridSearch;
11
16
  private readonly defaultWeights?;
12
17
  private readonly defaultTopK;
13
18
  private readonly defaultThreshold;
19
+ private readonly indexed;
14
20
  constructor(config: HybridRetrieverConfig);
21
+ indexChunk(chunk: IndexedChunk): void;
15
22
  retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
16
23
  private toRetrievalResult;
17
24
  }
@@ -1 +1 @@
1
- {"version":3,"file":"hybrid-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/hybrid-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,SAAS,EACT,eAAe,EACf,eAAe,EACf,mBAAmB,EAEpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAEzD,MAAM,WAAW,qBAAqB;IACpC,YAAY,EAAE,YAAY,CAAC;IAC3B,cAAc,CAAC,EAAE,mBAAmB,CAAC;IACrC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAKD,qBAAa,eAAgB,YAAW,SAAS;IAC/C,OAAO,CAAC,QAAQ,CAAC,YAAY,CAAe;IAC5C,OAAO,CAAC,QAAQ,CAAC,cAAc,CAAC,CAAsB;IACtD,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;gBAE9B,MAAM,EAAE,qBAAqB;IAOnC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAgB7F,OAAO,CAAC,iBAAiB;CAiB1B"}
1
+ {"version":3,"file":"hybrid-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/hybrid-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,SAAS,EACT,eAAe,EACf,eAAe,EACf,mBAAmB,EAEpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AACzD,OAAO,EAAgB,KAAK,YAAY,EAAE,KAAK,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAExF,MAAM,WAAW,qBAAqB;IACpC,YAAY,EAAE,YAAY,CAAC;IAC3B,cAAc,CAAC,EAAE,mBAAmB,CAAC;IACrC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAKD;;;GAGG;AACH,qBAAa,eAAgB,YAAW,SAAS,EAAE,YAAY;IAC7D,OAAO,CAAC,QAAQ,CAAC,YAAY,CAAe;IAC5C,OAAO,CAAC,QAAQ,CAAC,cAAc,CAAC,CAAsB;IACtD,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;IAC1C,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAmC;gBAE/C,MAAM,EAAE,qBAAqB;IAOzC,UAAU,CAAC,KAAK,EAAE,YAAY,GAAG,IAAI;IAK/B,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAgB7F,OAAO,CAAC,iBAAiB;CAoB1B"}
@@ -1,16 +1,26 @@
1
+ import { resultSource } from './chunk-indexer.js';
1
2
  const DEFAULT_TOP_K = 10;
2
3
  const DEFAULT_THRESHOLD = 0.0;
4
+ /**
5
+ * Vector + BM25 retrieval fused with reciprocal rank fusion. Chunks ingested through
6
+ * `RAGPipeline` are added to the HybridSearch keyword index automatically.
7
+ */
3
8
  export class HybridRetriever {
4
9
  hybridSearch;
5
10
  defaultWeights;
6
11
  defaultTopK;
7
12
  defaultThreshold;
13
+ indexed = new Map();
8
14
  constructor(config) {
9
15
  this.hybridSearch = config.hybridSearch;
10
16
  this.defaultWeights = config.defaultWeights;
11
17
  this.defaultTopK = config.defaultTopK ?? DEFAULT_TOP_K;
12
18
  this.defaultThreshold = config.defaultThreshold ?? DEFAULT_THRESHOLD;
13
19
  }
20
+ indexChunk(chunk) {
21
+ this.indexed.set(chunk.embeddingId, chunk);
22
+ this.hybridSearch.indexDocument(chunk.embeddingId, chunk.content);
23
+ }
14
24
  async retrieve(query, options) {
15
25
  const result = await this.hybridSearch.search({
16
26
  query,
@@ -25,16 +35,19 @@ export class HybridRetriever {
25
35
  return result.data.map((entry) => this.toRetrievalResult(entry));
26
36
  }
27
37
  toRetrievalResult(entry) {
28
- const rawDocId = entry.metadata?.documentId;
29
- const documentId = typeof rawDocId === 'string' ? rawDocId : entry.sourceId;
38
+ const known = this.indexed.get(entry.id);
39
+ const metadata = { ...known?.metadata, ...entry.metadata };
40
+ const rawDocId = metadata.documentId;
41
+ const chunkId = known?.chunkId ?? entry.sourceId;
42
+ const documentId = typeof rawDocId === 'string' ? rawDocId : chunkId;
30
43
  return {
31
- chunkId: entry.sourceId,
44
+ chunkId,
32
45
  documentId,
33
46
  content: entry.content,
34
47
  score: entry.score,
35
- source: entry.sourceType,
48
+ source: resultSource(metadata, entry.sourceType),
36
49
  metadata: {
37
- ...entry.metadata,
50
+ ...metadata,
38
51
  vectorScore: entry.vectorScore,
39
52
  keywordScore: entry.keywordScore,
40
53
  },
@@ -1 +1 @@
1
- {"version":3,"file":"hybrid-retriever.js","sourceRoot":"","sources":["../../src/retrievers/hybrid-retriever.ts"],"names":[],"mappings":"AAgBA,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAE9B,MAAM,OAAO,eAAe;IACT,YAAY,CAAe;IAC3B,cAAc,CAAuB;IACrC,WAAW,CAAS;IACpB,gBAAgB,CAAS;IAE1C,YAAY,MAA6B;QACvC,IAAI,CAAC,YAAY,GAAG,MAAM,CAAC,YAAY,CAAC;QACxC,IAAI,CAAC,cAAc,GAAG,MAAM,CAAC,cAAc,CAAC;QAC5C,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC;YAC5C,KAAK;YACL,QAAQ,EAAE,QAAQ;YAClB,OAAO,EAAE,IAAI,CAAC,cAAc;YAC5B,KAAK,EAAE,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW;YACxC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB;SACvD,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IACnE,CAAC;IAEO,iBAAiB,CAAC,KAAmB;QAC3C,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;QAC5C,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC;QAE5E,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,QAAQ;YACvB,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,KAAK,CAAC,UAAU;YACxB,QAAQ,EAAE;gBACR,GAAG,KAAK,CAAC,QAAQ;gBACjB,WAAW,EAAE,KAAK,CAAC,WAAW;gBAC9B,YAAY,EAAE,KAAK,CAAC,YAAY;aACjC;SACF,CAAC;IACJ,CAAC;CACF"}
1
+ {"version":3,"file":"hybrid-retriever.js","sourceRoot":"","sources":["../../src/retrievers/hybrid-retriever.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,YAAY,EAAwC,MAAM,oBAAoB,CAAC;AASxF,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAE9B;;;GAGG;AACH,MAAM,OAAO,eAAe;IACT,YAAY,CAAe;IAC3B,cAAc,CAAuB;IACrC,WAAW,CAAS;IACpB,gBAAgB,CAAS;IACzB,OAAO,GAAG,IAAI,GAAG,EAAwB,CAAC;IAE3D,YAAY,MAA6B;QACvC,IAAI,CAAC,YAAY,GAAG,MAAM,CAAC,YAAY,CAAC;QACxC,IAAI,CAAC,cAAc,GAAG,MAAM,CAAC,cAAc,CAAC;QAC5C,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,UAAU,CAAC,KAAmB;QAC5B,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,CAAC,CAAC;QAC3C,IAAI,CAAC,YAAY,CAAC,aAAa,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IACpE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC;YAC5C,KAAK;YACL,QAAQ,EAAE,QAAQ;YAClB,OAAO,EAAE,IAAI,CAAC,cAAc;YAC5B,KAAK,EAAE,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW;YACxC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB;SACvD,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IACnE,CAAC;IAEO,iBAAiB,CAAC,KAAmB;QAC3C,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QACzC,MAAM,QAAQ,GAAG,EAAE,GAAG,KAAK,EAAE,QAAQ,EAAE,GAAG,KAAK,CAAC,QAAQ,EAAE,CAAC;QAC3D,MAAM,QAAQ,GAAG,QAAQ,CAAC,UAAU,CAAC;QACrC,MAAM,OAAO,GAAG,KAAK,EAAE,OAAO,IAAI,KAAK,CAAC,QAAQ,CAAC;QACjD,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;QAErE,OAAO;YACL,OAAO;YACP,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,YAAY,CAAC,QAAQ,EAAE,KAAK,CAAC,UAAU,CAAC;YAChD,QAAQ,EAAE;gBACR,GAAG,QAAQ;gBACX,WAAW,EAAE,KAAK,CAAC,WAAW;gBAC9B,YAAY,EAAE,KAAK,CAAC,YAAY;aACjC;SACF,CAAC;IACJ,CAAC;CACF"}
@@ -8,4 +8,6 @@ export { MultiQueryRetriever } from './multi-query-retriever.js';
8
8
  export type { MultiQueryRetrieverConfig } from './multi-query-retriever.js';
9
9
  export { createRetriever } from './create-retriever.js';
10
10
  export type { CreateRetrieverConfig } from './create-retriever.js';
11
+ export { isChunkIndexer } from './chunk-indexer.js';
12
+ export type { ChunkIndexer, IndexedChunk } from './chunk-indexer.js';
11
13
  //# sourceMappingURL=index.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/retrievers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,YAAY,EAAE,yBAAyB,EAAE,MAAM,wBAAwB,CAAC;AACxE,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,YAAY,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAC;AAC1D,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,YAAY,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAChE,OAAO,EAAE,mBAAmB,EAAE,MAAM,yBAAyB,CAAC;AAC9D,YAAY,EAAE,yBAAyB,EAAE,MAAM,yBAAyB,CAAC;AACzE,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,YAAY,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/retrievers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,YAAY,EAAE,yBAAyB,EAAE,MAAM,2BAA2B,CAAC;AAC3E,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,YAAY,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AAC7D,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AACxD,YAAY,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AACnE,OAAO,EAAE,mBAAmB,EAAE,MAAM,4BAA4B,CAAC;AACjE,YAAY,EAAE,yBAAyB,EAAE,MAAM,4BAA4B,CAAC;AAC5E,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AACxD,YAAY,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AACnE,OAAO,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AACpD,YAAY,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC"}
@@ -3,4 +3,5 @@ export { MMRRetriever } from './mmr-retriever.js';
3
3
  export { HybridRetriever } from './hybrid-retriever.js';
4
4
  export { MultiQueryRetriever } from './multi-query-retriever.js';
5
5
  export { createRetriever } from './create-retriever.js';
6
+ export { isChunkIndexer } from './chunk-indexer.js';
6
7
  //# sourceMappingURL=index.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/retrievers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAE7D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAE/C,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAErD,OAAO,EAAE,mBAAmB,EAAE,MAAM,yBAAyB,CAAC;AAE9D,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/retrievers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAEhE,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAElD,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAExD,OAAO,EAAE,mBAAmB,EAAE,MAAM,4BAA4B,CAAC;AAEjE,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAExD,OAAO,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC"}
@@ -14,6 +14,11 @@ export declare class MMRRetriever implements Retriever {
14
14
  private readonly defaultThreshold;
15
15
  constructor(config: MMRRetrieverConfig);
16
16
  retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
17
+ /**
18
+ * Some vector stores (e.g. Qdrant) do not return stored vectors from search.
19
+ * Missing candidate vectors are re-embedded so diversity can still be computed.
20
+ */
21
+ private ensureVectors;
17
22
  private selectMMR;
18
23
  private toRetrievalResult;
19
24
  }
@@ -1 +1 @@
1
- {"version":3,"file":"mmr-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/mmr-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,gBAAgB,EAChB,gBAAgB,EAEhB,SAAS,EACT,eAAe,EACf,eAAe,EAChB,MAAM,qBAAqB,CAAC;AAE7B,MAAM,WAAW,kBAAkB;IACjC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAyBD,qBAAa,YAAa,YAAW,SAAS;IAC5C,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAmB;IAC3C,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;IAC7C,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAS;IACvC,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;gBAE9B,MAAM,EAAE,kBAAkB;IAQhC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAwB7F,OAAO,CAAC,SAAS;IAuCjB,OAAO,CAAC,iBAAiB;CAa1B"}
1
+ {"version":3,"file":"mmr-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/mmr-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,gBAAgB,EAChB,gBAAgB,EAEhB,SAAS,EACT,eAAe,EACf,eAAe,EAChB,MAAM,qBAAqB,CAAC;AAG7B,MAAM,WAAW,kBAAkB;IACjC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAyBD,qBAAa,YAAa,YAAW,SAAS;IAC5C,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAmB;IAC3C,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;IAC7C,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAS;IACvC,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;gBAE9B,MAAM,EAAE,kBAAkB;IAQhC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAyB7F;;;OAGG;YACW,aAAa;IAsB3B,OAAO,CAAC,SAAS;IAuCjB,OAAO,CAAC,iBAAiB;CAa1B"}
@@ -1,3 +1,4 @@
1
+ import { resultSource } from './chunk-indexer.js';
1
2
  const DEFAULT_TOP_K = 10;
2
3
  const DEFAULT_THRESHOLD = 0.0;
3
4
  const DEFAULT_LAMBDA = 0.7;
@@ -47,9 +48,30 @@ export class MMRRetriever {
47
48
  if (candidates.length === 0)
48
49
  return [];
49
50
  const lambda = options?.mmrLambda ?? this.defaultLambda;
50
- const selected = this.selectMMR(candidates, topK, lambda);
51
+ const withVectors = await this.ensureVectors(candidates);
52
+ const selected = this.selectMMR(withVectors, topK, lambda);
51
53
  return selected.map((entry) => this.toRetrievalResult(entry));
52
54
  }
55
+ /**
56
+ * Some vector stores (e.g. Qdrant) do not return stored vectors from search.
57
+ * Missing candidate vectors are re-embedded so diversity can still be computed.
58
+ */
59
+ async ensureVectors(candidates) {
60
+ const missing = candidates
61
+ .map((candidate, index) => ({ candidate, index }))
62
+ .filter(({ candidate }) => !candidate.vector || candidate.vector.length === 0);
63
+ if (missing.length === 0)
64
+ return candidates;
65
+ const vectors = await this.embedding.embedBatch(missing.map(({ candidate }) => candidate.content));
66
+ if (vectors.length !== missing.length) {
67
+ throw new Error(`Embedding count mismatch: got ${vectors.length} vectors for ${missing.length} candidates`);
68
+ }
69
+ const result = [...candidates];
70
+ missing.forEach(({ candidate, index }, i) => {
71
+ result[index] = { ...candidate, vector: vectors[i] };
72
+ });
73
+ return result;
74
+ }
53
75
  selectMMR(candidates, topK, lambda) {
54
76
  const remaining = new Set(candidates.map((_, i) => i));
55
77
  const selected = [];
@@ -87,7 +109,7 @@ export class MMRRetriever {
87
109
  documentId,
88
110
  content: entry.content,
89
111
  score: entry.mmrScore,
90
- source: entry.sourceType,
112
+ source: resultSource(entry.metadata, entry.sourceType),
91
113
  metadata: { ...entry.metadata, originalScore: entry.score },
92
114
  };
93
115
  }
@@ -1 +1 @@
1
- {"version":3,"file":"mmr-retriever.js","sourceRoot":"","sources":["../../src/retrievers/mmr-retriever.ts"],"names":[],"mappings":"AAiBA,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAC9B,MAAM,cAAc,GAAG,GAAG,CAAC;AAC3B,MAAM,oBAAoB,GAAG,CAAC,CAAC;AAI/B,SAAS,gBAAgB,CAAC,CAAW,EAAE,CAAW;IAChD,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM,EAAE,CAAC;QAC1B,MAAM,IAAI,KAAK,CAAC,8BAA8B,CAAC,CAAC,MAAM,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACnB,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACrB,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAClD,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,KAAK,CAAC;AACvC,CAAC;AAED,MAAM,OAAO,YAAY;IACN,OAAO,CAAmB;IAC1B,SAAS,CAAmB;IAC5B,aAAa,CAAS;IACtB,WAAW,CAAS;IACpB,gBAAgB,CAAS;IAE1C,YAAY,MAA0B;QACpC,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACvC,IAAI,CAAC,SAAS,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACzC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC,aAAa,IAAI,cAAc,CAAC;QAC5D,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW,CAAC;QAC/C,MAAM,SAAS,GAAG,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB,CAAC;QAC9D,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAEtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;YACvC,MAAM,EAAE,WAAW;YACnB,KAAK;YACL,KAAK,EAAE,IAAI,GAAG,oBAAoB;YAClC,SAAS;SACV,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC;QAC/B,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEvC,MAAM,MAAM,GAAG,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,aAAa,CAAC;QACxD,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,UAAU,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAC1D,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IAChE,CAAC;IAEO,SAAS,CACf,UAA6B,EAC7B,IAAY,EACZ,MAAc;QAEd,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QACvD,MAAM,QAAQ,GAA+C,EAAE,CAAC;QAEhE,OAAO,QAAQ,CAAC,MAAM,GAAG,IAAI,IAAI,SAAS,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;YACpD,IAAI,OAAO,GAAG,CAAC,CAAC,CAAC;YACjB,IAAI,SAAS,GAAG,CAAC,QAAQ,CAAC;YAE1B,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC5B,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC;gBAExC,IAAI,cAAc,GAAG,CAAC,QAAQ,CAAC;gBAC/B,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;oBAC3B,MAAM,GAAG,GAAG,gBAAgB,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;oBACjE,IAAI,GAAG,GAAG,cAAc;wBAAE,cAAc,GAAG,GAAG,CAAC;gBACjD,CAAC;gBACD,IAAI,cAAc,KAAK,CAAC,QAAQ;oBAAE,cAAc,GAAG,CAAC,CAAC;gBAErD,MAAM,QAAQ,GAAG,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,cAAc,CAAC;gBAEpE,IAAI,QAAQ,GAAG,SAAS,EAAE,CAAC;oBACzB,SAAS,GAAG,QAAQ,CAAC;oBACrB,OAAO,GAAG,GAAG,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,IAAI,OAAO,KAAK,CAAC,CAAC;gBAAE,MAAM;YAE1B,SAAS,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAC1B,QAAQ,CAAC,IAAI,CAAC,EAAE,GAAG,UAAU,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC,CAAC;QACjE,CAAC;QAED,OAAO,QAAQ,CAAC;IAClB,CAAC;IAEO,iBAAiB,CAAC,KAA6C;QACrE,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;QAC5C,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC;QAE5E,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,QAAQ;YACvB,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,QAAQ;YACrB,MAAM,EAAE,KAAK,CAAC,UAAU;YACxB,QAAQ,EAAE,EAAE,GAAG,KAAK,CAAC,QAAQ,EAAE,aAAa,EAAE,KAAK,CAAC,KAAK,EAAE;SAC5D,CAAC;IACJ,CAAC;CACF"}
1
+ {"version":3,"file":"mmr-retriever.js","sourceRoot":"","sources":["../../src/retrievers/mmr-retriever.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAUlD,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAC9B,MAAM,cAAc,GAAG,GAAG,CAAC;AAC3B,MAAM,oBAAoB,GAAG,CAAC,CAAC;AAI/B,SAAS,gBAAgB,CAAC,CAAW,EAAE,CAAW;IAChD,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM,EAAE,CAAC;QAC1B,MAAM,IAAI,KAAK,CAAC,8BAA8B,CAAC,CAAC,MAAM,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACnB,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACrB,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAClD,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,KAAK,CAAC;AACvC,CAAC;AAED,MAAM,OAAO,YAAY;IACN,OAAO,CAAmB;IAC1B,SAAS,CAAmB;IAC5B,aAAa,CAAS;IACtB,WAAW,CAAS;IACpB,gBAAgB,CAAS;IAE1C,YAAY,MAA0B;QACpC,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACvC,IAAI,CAAC,SAAS,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACzC,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC,aAAa,IAAI,cAAc,CAAC;QAC5D,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW,CAAC;QAC/C,MAAM,SAAS,GAAG,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB,CAAC;QAC9D,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAEtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;YACvC,MAAM,EAAE,WAAW;YACnB,KAAK;YACL,KAAK,EAAE,IAAI,GAAG,oBAAoB;YAClC,SAAS;SACV,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC;QAC/B,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEvC,MAAM,MAAM,GAAG,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,aAAa,CAAC;QACxD,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,UAAU,CAAC,CAAC;QACzD,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,WAAW,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;QAC3D,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IAChE,CAAC;IAED;;;OAGG;IACK,KAAK,CAAC,aAAa,CAAC,UAA6B;QACvD,MAAM,OAAO,GAAG,UAAU;aACvB,GAAG,CAAC,CAAC,SAAS,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAC;aACjD,MAAM,CAAC,CAAC,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,MAAM,IAAI,SAAS,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC;QACjF,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,UAAU,CAAC;QAE5C,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,UAAU,CAC7C,OAAO,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,SAAS,CAAC,OAAO,CAAC,CAClD,CAAC;QACF,IAAI,OAAO,CAAC,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC;YACtC,MAAM,IAAI,KAAK,CACb,iCAAiC,OAAO,CAAC,MAAM,gBAAgB,OAAO,CAAC,MAAM,aAAa,CAC3F,CAAC;QACJ,CAAC;QAED,MAAM,MAAM,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC;QAC/B,OAAO,CAAC,OAAO,CAAC,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE;YAC1C,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAE,EAAE,CAAC;QACxD,CAAC,CAAC,CAAC;QACH,OAAO,MAAM,CAAC;IAChB,CAAC;IAEO,SAAS,CACf,UAA6B,EAC7B,IAAY,EACZ,MAAc;QAEd,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QACvD,MAAM,QAAQ,GAA+C,EAAE,CAAC;QAEhE,OAAO,QAAQ,CAAC,MAAM,GAAG,IAAI,IAAI,SAAS,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;YACpD,IAAI,OAAO,GAAG,CAAC,CAAC,CAAC;YACjB,IAAI,SAAS,GAAG,CAAC,QAAQ,CAAC;YAE1B,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC5B,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC;gBAExC,IAAI,cAAc,GAAG,CAAC,QAAQ,CAAC;gBAC/B,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;oBAC3B,MAAM,GAAG,GAAG,gBAAgB,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;oBACjE,IAAI,GAAG,GAAG,cAAc;wBAAE,cAAc,GAAG,GAAG,CAAC;gBACjD,CAAC;gBACD,IAAI,cAAc,KAAK,CAAC,QAAQ;oBAAE,cAAc,GAAG,CAAC,CAAC;gBAErD,MAAM,QAAQ,GAAG,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,cAAc,CAAC;gBAEpE,IAAI,QAAQ,GAAG,SAAS,EAAE,CAAC;oBACzB,SAAS,GAAG,QAAQ,CAAC;oBACrB,OAAO,GAAG,GAAG,CAAC;gBAChB,CAAC;YACH,CAAC;YAED,IAAI,OAAO,KAAK,CAAC,CAAC;gBAAE,MAAM;YAE1B,SAAS,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAC1B,QAAQ,CAAC,IAAI,CAAC,EAAE,GAAG,UAAU,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC,CAAC;QACjE,CAAC;QAED,OAAO,QAAQ,CAAC;IAClB,CAAC;IAEO,iBAAiB,CAAC,KAA6C;QACrE,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;QAC5C,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC;QAE5E,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,QAAQ;YACvB,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,QAAQ;YACrB,MAAM,EAAE,YAAY,CAAC,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,UAAU,CAAC;YACtD,QAAQ,EAAE,EAAE,GAAG,KAAK,CAAC,QAAQ,EAAE,aAAa,EAAE,KAAK,CAAC,KAAK,EAAE;SAC5D,CAAC;IACJ,CAAC;CACF"}
@@ -1,15 +1,26 @@
1
1
  import type { Retriever, RetrievalConfig, RetrievalResult } from '@cogitator-ai/types';
2
+ import { type ChunkIndexer, type IndexedChunk } from './chunk-indexer.js';
2
3
  export interface MultiQueryRetrieverConfig {
3
4
  baseRetriever: Retriever;
4
5
  expandQuery: (query: string) => Promise<string[]>;
5
6
  defaultTopK?: number;
7
+ /** Maximum number of generated query variants used in addition to the original query. */
8
+ defaultMaxQueries?: number;
6
9
  }
7
- export declare class MultiQueryRetriever implements Retriever {
10
+ /**
11
+ * Expands the query into variants, retrieves for each in parallel and merges results
12
+ * (best score per chunk). The original query is always searched. If expansion fails the
13
+ * original query alone is used; if every retrieval fails the first error is thrown.
14
+ */
15
+ export declare class MultiQueryRetriever implements Retriever, ChunkIndexer {
8
16
  private readonly baseRetriever;
9
17
  private readonly expandQuery;
10
18
  private readonly defaultTopK;
19
+ private readonly defaultMaxQueries?;
11
20
  constructor(config: MultiQueryRetrieverConfig);
21
+ indexChunk(chunk: IndexedChunk): void;
12
22
  retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
23
+ private buildVariants;
13
24
  private deduplicateAndMerge;
14
25
  }
15
26
  //# sourceMappingURL=multi-query-retriever.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"multi-query-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/multi-query-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,eAAe,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAEvF,MAAM,WAAW,yBAAyB;IACxC,aAAa,EAAE,SAAS,CAAC;IACzB,WAAW,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAClD,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAID,qBAAa,mBAAoB,YAAW,SAAS;IACnD,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAY;IAC1C,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAuC;IACnE,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;gBAEzB,MAAM,EAAE,yBAAyB;IAMvC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAsB7F,OAAO,CAAC,mBAAmB;CAY5B"}
1
+ {"version":3,"file":"multi-query-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/multi-query-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,eAAe,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACvF,OAAO,EAAkB,KAAK,YAAY,EAAE,KAAK,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAE1F,MAAM,WAAW,yBAAyB;IACxC,aAAa,EAAE,SAAS,CAAC;IACzB,WAAW,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAClD,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,yFAAyF;IACzF,iBAAiB,CAAC,EAAE,MAAM,CAAC;CAC5B;AAID;;;;GAIG;AACH,qBAAa,mBAAoB,YAAW,SAAS,EAAE,YAAY;IACjE,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAY;IAC1C,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAuC;IACnE,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CAAS;gBAEhC,MAAM,EAAE,yBAAyB;IAO7C,UAAU,CAAC,KAAK,EAAE,YAAY,GAAG,IAAI;IAM/B,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;YA0B/E,aAAa;IAwB3B,OAAO,CAAC,mBAAmB;CAY5B"}
@@ -1,25 +1,64 @@
1
+ import { isChunkIndexer } from './chunk-indexer.js';
1
2
  const DEFAULT_TOP_K = 10;
3
+ /**
4
+ * Expands the query into variants, retrieves for each in parallel and merges results
5
+ * (best score per chunk). The original query is always searched. If expansion fails the
6
+ * original query alone is used; if every retrieval fails the first error is thrown.
7
+ */
2
8
  export class MultiQueryRetriever {
3
9
  baseRetriever;
4
10
  expandQuery;
5
11
  defaultTopK;
12
+ defaultMaxQueries;
6
13
  constructor(config) {
7
14
  this.baseRetriever = config.baseRetriever;
8
15
  this.expandQuery = config.expandQuery;
9
16
  this.defaultTopK = config.defaultTopK ?? DEFAULT_TOP_K;
17
+ this.defaultMaxQueries = config.defaultMaxQueries;
10
18
  }
11
- async retrieve(query, options) {
12
- const expanded = await this.expandQuery(query);
13
- const variants = expanded.length > 0 ? expanded : [query];
14
- if (!variants.includes(query)) {
15
- variants.unshift(query);
19
+ indexChunk(chunk) {
20
+ if (isChunkIndexer(this.baseRetriever)) {
21
+ this.baseRetriever.indexChunk(chunk);
16
22
  }
23
+ }
24
+ async retrieve(query, options) {
25
+ const variants = await this.buildVariants(query, options?.multiQueryCount);
17
26
  const settled = await Promise.allSettled(variants.map((variant) => this.baseRetriever.retrieve(variant, options)));
18
- const allResults = settled.flatMap((outcome) => outcome.status === 'fulfilled' ? outcome.value : []);
19
- const merged = this.deduplicateAndMerge(allResults.flat());
27
+ const fulfilled = settled.filter((outcome) => outcome.status === 'fulfilled');
28
+ if (fulfilled.length === 0) {
29
+ const firstFailure = settled.find((outcome) => outcome.status === 'rejected');
30
+ throw firstFailure?.reason instanceof Error
31
+ ? firstFailure.reason
32
+ : new Error(`MultiQueryRetriever: all retrievals failed: ${String(firstFailure?.reason)}`);
33
+ }
34
+ const merged = this.deduplicateAndMerge(fulfilled.flatMap((outcome) => outcome.value));
20
35
  const topK = options?.topK ?? this.defaultTopK;
21
36
  return merged.slice(0, topK);
22
37
  }
38
+ async buildVariants(query, requestedMax) {
39
+ let expanded;
40
+ try {
41
+ expanded = await this.expandQuery(query);
42
+ }
43
+ catch {
44
+ expanded = [];
45
+ }
46
+ const normalized = query.trim().toLowerCase();
47
+ const seen = new Set([normalized]);
48
+ const extra = [];
49
+ for (const variant of expanded) {
50
+ if (typeof variant !== 'string')
51
+ continue;
52
+ const trimmed = variant.trim();
53
+ const key = trimmed.toLowerCase();
54
+ if (!trimmed || seen.has(key))
55
+ continue;
56
+ seen.add(key);
57
+ extra.push(trimmed);
58
+ }
59
+ const max = requestedMax ?? this.defaultMaxQueries;
60
+ return [query, ...(max === undefined ? extra : extra.slice(0, max))];
61
+ }
23
62
  deduplicateAndMerge(results) {
24
63
  const best = new Map();
25
64
  for (const result of results) {
@@ -1 +1 @@
1
- {"version":3,"file":"multi-query-retriever.js","sourceRoot":"","sources":["../../src/retrievers/multi-query-retriever.ts"],"names":[],"mappings":"AAQA,MAAM,aAAa,GAAG,EAAE,CAAC;AAEzB,MAAM,OAAO,mBAAmB;IACb,aAAa,CAAY;IACzB,WAAW,CAAuC;IAClD,WAAW,CAAS;IAErC,YAAY,MAAiC;QAC3C,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC,aAAa,CAAC;QAC1C,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;QACtC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;IACzD,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC;QAC/C,MAAM,QAAQ,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;QAE1D,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;QAC1B,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,UAAU,CACtC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CACzE,CAAC;QAEF,MAAM,UAAU,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAC7C,OAAO,CAAC,MAAM,KAAK,WAAW,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CACpD,CAAC;QAEF,MAAM,MAAM,GAAG,IAAI,CAAC,mBAAmB,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC;QAC3D,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW,CAAC;QAE/C,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IAC/B,CAAC;IAEO,mBAAmB,CAAC,OAA0B;QACpD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA2B,CAAC;QAEhD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,IAAI,MAAM,CAAC,KAAK,GAAG,QAAQ,CAAC,KAAK,EAAE,CAAC;gBAC/C,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YACnC,CAAC;QACH,CAAC;QAED,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAC9D,CAAC;CACF"}
1
+ {"version":3,"file":"multi-query-retriever.js","sourceRoot":"","sources":["../../src/retrievers/multi-query-retriever.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAwC,MAAM,oBAAoB,CAAC;AAU1F,MAAM,aAAa,GAAG,EAAE,CAAC;AAEzB;;;;GAIG;AACH,MAAM,OAAO,mBAAmB;IACb,aAAa,CAAY;IACzB,WAAW,CAAuC;IAClD,WAAW,CAAS;IACpB,iBAAiB,CAAU;IAE5C,YAAY,MAAiC;QAC3C,IAAI,CAAC,aAAa,GAAG,MAAM,CAAC,aAAa,CAAC;QAC1C,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;QACtC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,iBAAiB,GAAG,MAAM,CAAC,iBAAiB,CAAC;IACpD,CAAC;IAED,UAAU,CAAC,KAAmB;QAC5B,IAAI,cAAc,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC;YACvC,IAAI,CAAC,aAAa,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,KAAK,EAAE,OAAO,EAAE,eAAe,CAAC,CAAC;QAE3E,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,UAAU,CACtC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CACzE,CAAC;QAEF,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,CAC9B,CAAC,OAAO,EAAwD,EAAE,CAChE,OAAO,CAAC,MAAM,KAAK,WAAW,CACjC,CAAC;QACF,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC3B,MAAM,YAAY,GAAG,OAAO,CAAC,IAAI,CAC/B,CAAC,OAAO,EAAoC,EAAE,CAAC,OAAO,CAAC,MAAM,KAAK,UAAU,CAC7E,CAAC;YACF,MAAM,YAAY,EAAE,MAAM,YAAY,KAAK;gBACzC,CAAC,CAAC,YAAY,CAAC,MAAM;gBACrB,CAAC,CAAC,IAAI,KAAK,CAAC,+CAA+C,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;QAC/F,CAAC;QAED,MAAM,MAAM,GAAG,IAAI,CAAC,mBAAmB,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;QACvF,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW,CAAC;QAE/C,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IAC/B,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,KAAa,EAAE,YAAqB;QAC9D,IAAI,QAAkB,CAAC;QACvB,IAAI,CAAC;YACH,QAAQ,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC;QAC3C,CAAC;QAAC,MAAM,CAAC;YACP,QAAQ,GAAG,EAAE,CAAC;QAChB,CAAC;QAED,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAC9C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC;QACnC,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,IAAI,OAAO,OAAO,KAAK,QAAQ;gBAAE,SAAS;YAC1C,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;YAC/B,MAAM,GAAG,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;YAClC,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YACxC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACtB,CAAC;QAED,MAAM,GAAG,GAAG,YAAY,IAAI,IAAI,CAAC,iBAAiB,CAAC;QACnD,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;IACvE,CAAC;IAEO,mBAAmB,CAAC,OAA0B;QACpD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA2B,CAAC;QAEhD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAC1C,IAAI,CAAC,QAAQ,IAAI,MAAM,CAAC,KAAK,GAAG,QAAQ,CAAC,KAAK,EAAE,CAAC;gBAC/C,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YACnC,CAAC;QACH,CAAC;QAED,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAC9D,CAAC;CACF"}
@@ -1 +1 @@
1
- {"version":3,"file":"similarity-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/similarity-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,gBAAgB,EAChB,gBAAgB,EAEhB,SAAS,EACT,eAAe,EACf,eAAe,EAChB,MAAM,qBAAqB,CAAC;AAE7B,MAAM,WAAW,yBAAyB;IACxC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAKD,qBAAa,mBAAoB,YAAW,SAAS;IACnD,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAmB;IAC3C,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;IAC7C,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;gBAE9B,MAAM,EAAE,yBAAyB;IAOvC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAiB7F,OAAO,CAAC,iBAAiB;CAa1B"}
1
+ {"version":3,"file":"similarity-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/similarity-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,gBAAgB,EAChB,gBAAgB,EAEhB,SAAS,EACT,eAAe,EACf,eAAe,EAChB,MAAM,qBAAqB,CAAC;AAG7B,MAAM,WAAW,yBAAyB;IACxC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAKD,qBAAa,mBAAoB,YAAW,SAAS;IACnD,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAmB;IAC3C,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;IAC7C,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAS;IACrC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;gBAE9B,MAAM,EAAE,yBAAyB;IAOvC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAiB7F,OAAO,CAAC,iBAAiB;CAa1B"}
@@ -1,3 +1,4 @@
1
+ import { resultSource } from './chunk-indexer.js';
1
2
  const DEFAULT_TOP_K = 10;
2
3
  const DEFAULT_THRESHOLD = 0.0;
3
4
  export class SimilarityRetriever {
@@ -32,7 +33,7 @@ export class SimilarityRetriever {
32
33
  documentId,
33
34
  content: entry.content,
34
35
  score: entry.score,
35
- source: entry.sourceType,
36
+ source: resultSource(entry.metadata, entry.sourceType),
36
37
  metadata: entry.metadata,
37
38
  };
38
39
  }
@@ -1 +1 @@
1
- {"version":3,"file":"similarity-retriever.js","sourceRoot":"","sources":["../../src/retrievers/similarity-retriever.ts"],"names":[],"mappings":"AAgBA,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAE9B,MAAM,OAAO,mBAAmB;IACb,OAAO,CAAmB;IAC1B,SAAS,CAAmB;IAC5B,WAAW,CAAS;IACpB,gBAAgB,CAAS;IAE1C,YAAY,MAAiC;QAC3C,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACvC,IAAI,CAAC,SAAS,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACzC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAEjD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;YACvC,MAAM;YACN,KAAK;YACL,KAAK,EAAE,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW;YACxC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB;SACvD,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IACnE,CAAC;IAEO,iBAAiB,CAAC,KAAoC;QAC5D,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;QAC5C,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC;QAE5E,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,QAAQ;YACvB,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,KAAK,CAAC,UAAU;YACxB,QAAQ,EAAE,KAAK,CAAC,QAAQ;SACzB,CAAC;IACJ,CAAC;CACF"}
1
+ {"version":3,"file":"similarity-retriever.js","sourceRoot":"","sources":["../../src/retrievers/similarity-retriever.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AASlD,MAAM,aAAa,GAAG,EAAE,CAAC;AACzB,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAE9B,MAAM,OAAO,mBAAmB;IACb,OAAO,CAAmB;IAC1B,SAAS,CAAmB;IAC5B,WAAW,CAAS;IACpB,gBAAgB,CAAS;IAE1C,YAAY,MAAiC;QAC3C,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACvC,IAAI,CAAC,SAAS,GAAG,MAAM,CAAC,gBAAgB,CAAC;QACzC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,IAAI,aAAa,CAAC;QACvD,IAAI,CAAC,gBAAgB,GAAG,MAAM,CAAC,gBAAgB,IAAI,iBAAiB,CAAC;IACvE,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAkC;QAC9D,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAEjD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;YACvC,MAAM;YACN,KAAK;YACL,KAAK,EAAE,OAAO,EAAE,IAAI,IAAI,IAAI,CAAC,WAAW;YACxC,SAAS,EAAE,OAAO,EAAE,SAAS,IAAI,IAAI,CAAC,gBAAgB;SACvD,CAAC,CAAC;QAEH,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,iBAAiB,CAAC,KAAK,CAAC,CAAC,CAAC;IACnE,CAAC;IAEO,iBAAiB,CAAC,KAAoC;QAC5D,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;QAC5C,MAAM,UAAU,GAAG,OAAO,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC;QAE5E,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,QAAQ;YACvB,UAAU;YACV,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,YAAY,CAAC,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,UAAU,CAAC;YACtD,QAAQ,EAAE,KAAK,CAAC,QAAQ;SACzB,CAAC;IACJ,CAAC;CACF"}
package/dist/tools.d.ts CHANGED
@@ -17,7 +17,16 @@ declare const IngestParamsSchema: z.ZodObject<{
17
17
  }, z.core.$strip>;
18
18
  type IngestParams = z.infer<typeof IngestParamsSchema>;
19
19
  export declare function createSearchTool(pipeline: RAGPipeline): RAGTool<SearchParams>;
20
- export declare function createIngestTool(pipeline: RAGPipeline): RAGTool<IngestParams>;
21
- export declare function ragTools(pipeline: RAGPipeline): [RAGTool<SearchParams>, RAGTool<IngestParams>];
20
+ export interface IngestToolOptions {
21
+ /**
22
+ * Restrict local sources to these directories (symlinks are resolved). When set, any
23
+ * path outside them is rejected. Strongly recommended when the tool is exposed to an LLM.
24
+ */
25
+ allowedRoots?: string[];
26
+ /** Allow http(s) URLs as sources. Defaults to `true`. */
27
+ allowUrls?: boolean;
28
+ }
29
+ export declare function createIngestTool(pipeline: RAGPipeline, options?: IngestToolOptions): RAGTool<IngestParams>;
30
+ export declare function ragTools(pipeline: RAGPipeline, ingestOptions?: IngestToolOptions): [RAGTool<SearchParams>, RAGTool<IngestParams>];
22
31
  export {};
23
32
  //# sourceMappingURL=tools.d.ts.map