@cogitator-ai/rag 0.1.10 → 0.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +66 -31
- package/dist/chunkers/create-chunker.js.map +1 -1
- package/dist/chunkers/index.d.ts.map +1 -1
- package/dist/chunkers/index.js.map +1 -1
- package/dist/chunkers/recursive-chunker.d.ts +1 -2
- package/dist/chunkers/recursive-chunker.d.ts.map +1 -1
- package/dist/chunkers/recursive-chunker.js +68 -58
- package/dist/chunkers/recursive-chunker.js.map +1 -1
- package/dist/chunkers/semantic-chunker.d.ts.map +1 -1
- package/dist/chunkers/semantic-chunker.js +4 -1
- package/dist/chunkers/semantic-chunker.js.map +1 -1
- package/dist/index.d.ts +2 -2
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +4 -2
- package/dist/index.js.map +1 -1
- package/dist/loaders/csv-loader.d.ts.map +1 -1
- package/dist/loaders/csv-loader.js +9 -4
- package/dist/loaders/csv-loader.js.map +1 -1
- package/dist/loaders/html-loader.d.ts +4 -0
- package/dist/loaders/html-loader.d.ts.map +1 -1
- package/dist/loaders/html-loader.js +26 -2
- package/dist/loaders/html-loader.js.map +1 -1
- package/dist/loaders/index.d.ts.map +1 -1
- package/dist/loaders/index.js.map +1 -1
- package/dist/loaders/markdown-loader.d.ts.map +1 -1
- package/dist/loaders/markdown-loader.js +7 -3
- package/dist/loaders/markdown-loader.js.map +1 -1
- package/dist/loaders/text-loader.js +1 -1
- package/dist/loaders/text-loader.js.map +1 -1
- package/dist/loaders/web-loader.d.ts +29 -1
- package/dist/loaders/web-loader.d.ts.map +1 -1
- package/dist/loaders/web-loader.js +227 -92
- package/dist/loaders/web-loader.js.map +1 -1
- package/dist/rag-builder.d.ts +10 -1
- package/dist/rag-builder.d.ts.map +1 -1
- package/dist/rag-builder.js +70 -7
- package/dist/rag-builder.js.map +1 -1
- package/dist/rag-pipeline.d.ts +1 -0
- package/dist/rag-pipeline.d.ts.map +1 -1
- package/dist/rag-pipeline.js +55 -19
- package/dist/rag-pipeline.js.map +1 -1
- package/dist/rerankers/index.d.ts.map +1 -1
- package/dist/rerankers/index.js.map +1 -1
- package/dist/retrievers/chunk-indexer.d.ts +19 -0
- package/dist/retrievers/chunk-indexer.d.ts.map +1 -0
- package/dist/retrievers/chunk-indexer.js +8 -0
- package/dist/retrievers/chunk-indexer.js.map +1 -0
- package/dist/retrievers/create-retriever.d.ts +1 -0
- package/dist/retrievers/create-retriever.d.ts.map +1 -1
- package/dist/retrievers/create-retriever.js +1 -0
- package/dist/retrievers/create-retriever.js.map +1 -1
- package/dist/retrievers/hybrid-retriever.d.ts +8 -1
- package/dist/retrievers/hybrid-retriever.d.ts.map +1 -1
- package/dist/retrievers/hybrid-retriever.js +18 -5
- package/dist/retrievers/hybrid-retriever.js.map +1 -1
- package/dist/retrievers/index.d.ts +2 -0
- package/dist/retrievers/index.d.ts.map +1 -1
- package/dist/retrievers/index.js +1 -0
- package/dist/retrievers/index.js.map +1 -1
- package/dist/retrievers/mmr-retriever.d.ts +5 -0
- package/dist/retrievers/mmr-retriever.d.ts.map +1 -1
- package/dist/retrievers/mmr-retriever.js +24 -2
- package/dist/retrievers/mmr-retriever.js.map +1 -1
- package/dist/retrievers/multi-query-retriever.d.ts +12 -1
- package/dist/retrievers/multi-query-retriever.d.ts.map +1 -1
- package/dist/retrievers/multi-query-retriever.js +46 -7
- package/dist/retrievers/multi-query-retriever.js.map +1 -1
- package/dist/retrievers/similarity-retriever.d.ts.map +1 -1
- package/dist/retrievers/similarity-retriever.js +2 -1
- package/dist/retrievers/similarity-retriever.js.map +1 -1
- package/dist/tools.d.ts +11 -2
- package/dist/tools.d.ts.map +1 -1
- package/dist/tools.js +39 -7
- package/dist/tools.js.map +1 -1
- package/package.json +5 -5
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rag-builder.d.ts","sourceRoot":"","sources":["../src/rag-builder.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,cAAc,EACd,OAAO,EACP,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,EAChB,SAAS,EACT,QAAQ,
|
|
1
|
+
{"version":3,"file":"rag-builder.d.ts","sourceRoot":"","sources":["../src/rag-builder.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,cAAc,EACd,OAAO,EACP,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,EAChB,SAAS,EACT,QAAQ,EAER,mBAAmB,EACpB,MAAM,qBAAqB,CAAC;AAC7B,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AACzD,OAAO,EAA2B,KAAK,sBAAsB,EAAE,MAAM,aAAa,CAAC;AACnF,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAOhD,qBAAa,kBAAkB;IAC7B,OAAO,CAAC,MAAM,CAAC,CAAiB;IAChC,OAAO,CAAC,OAAO,CAAC,CAAyB;IACzC,OAAO,CAAC,gBAAgB,CAAC,CAAmB;IAC5C,OAAO,CAAC,gBAAgB,CAAC,CAAmB;IAC5C,OAAO,CAAC,SAAS,CAAC,CAAY;IAC9B,OAAO,CAAC,QAAQ,CAAC,CAAW;IAC5B,OAAO,CAAC,WAAW,CAAC,CAAyB;IAC7C,OAAO,CAAC,YAAY,CAAC,CAAe;IACpC,OAAO,CAAC,aAAa,CAAC,CAAsB;IAC5C,OAAO,CAAC,aAAa,CAAC,CAAuC;IAE7D,UAAU,CAAC,MAAM,EAAE,cAAc,GAAG,IAAI;IAKxC,WAAW,CAAC,OAAO,EAAE,OAAO,GAAG,YAAY,GAAG,IAAI;IAKlD,oBAAoB,CAAC,OAAO,EAAE,gBAAgB,GAAG,IAAI;IAKrD,oBAAoB,CAAC,OAAO,EAAE,gBAAgB,GAAG,IAAI;IAKrD,aAAa,CAAC,SAAS,EAAE,SAAS,GAAG,IAAI;IAKzC,YAAY,CAAC,QAAQ,EAAE,QAAQ,GAAG,IAAI;IAKtC,yFAAyF;IACzF,gBAAgB,CAAC,YAAY,EAAE,YAAY,EAAE,OAAO,CAAC,EAAE,mBAAmB,GAAG,IAAI;IAMjF,mEAAmE;IACnE,iBAAiB,CAAC,WAAW,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,GAAG,IAAI;IAK1E,UAAU,CAAC,MAAM,EAAE,sBAAsB,GAAG,IAAI;IAKhD,KAAK,IAAI,WAAW;IA8BpB,OAAO,CAAC,eAAe;IAmEvB,OAAO,CAAC,aAAa;CAMtB"}
|
package/dist/rag-builder.js
CHANGED
|
@@ -2,6 +2,9 @@ import { RAGPipelineConfigSchema } from './schema.js';
|
|
|
2
2
|
import { RAGPipeline } from './rag-pipeline.js';
|
|
3
3
|
import { createChunker } from './chunkers/create-chunker.js';
|
|
4
4
|
import { SimilarityRetriever } from './retrievers/similarity-retriever.js';
|
|
5
|
+
import { MMRRetriever } from './retrievers/mmr-retriever.js';
|
|
6
|
+
import { HybridRetriever } from './retrievers/hybrid-retriever.js';
|
|
7
|
+
import { MultiQueryRetriever } from './retrievers/multi-query-retriever.js';
|
|
5
8
|
export class RAGPipelineBuilder {
|
|
6
9
|
loader;
|
|
7
10
|
chunker;
|
|
@@ -10,6 +13,9 @@ export class RAGPipelineBuilder {
|
|
|
10
13
|
retriever;
|
|
11
14
|
reranker;
|
|
12
15
|
configInput;
|
|
16
|
+
hybridSearch;
|
|
17
|
+
hybridWeights;
|
|
18
|
+
queryExpander;
|
|
13
19
|
withLoader(loader) {
|
|
14
20
|
this.loader = loader;
|
|
15
21
|
return this;
|
|
@@ -34,6 +40,17 @@ export class RAGPipelineBuilder {
|
|
|
34
40
|
this.reranker = reranker;
|
|
35
41
|
return this;
|
|
36
42
|
}
|
|
43
|
+
/** HybridSearch used by the `hybrid` strategy (and as the multi-query base when set). */
|
|
44
|
+
withHybridSearch(hybridSearch, weights) {
|
|
45
|
+
this.hybridSearch = hybridSearch;
|
|
46
|
+
this.hybridWeights = weights;
|
|
47
|
+
return this;
|
|
48
|
+
}
|
|
49
|
+
/** Query expansion function used by the `multi-query` strategy. */
|
|
50
|
+
withQueryExpander(expandQuery) {
|
|
51
|
+
this.queryExpander = expandQuery;
|
|
52
|
+
return this;
|
|
53
|
+
}
|
|
37
54
|
withConfig(config) {
|
|
38
55
|
this.configInput = config;
|
|
39
56
|
return this;
|
|
@@ -50,13 +67,7 @@ export class RAGPipelineBuilder {
|
|
|
50
67
|
}
|
|
51
68
|
const config = this.resolveConfig();
|
|
52
69
|
const chunker = this.chunker ?? createChunker(config.chunking, this.embeddingService);
|
|
53
|
-
const retriever = this.retriever ??
|
|
54
|
-
new SimilarityRetriever({
|
|
55
|
-
embeddingAdapter: this.embeddingAdapter,
|
|
56
|
-
embeddingService: this.embeddingService,
|
|
57
|
-
defaultTopK: config.retrieval.topK,
|
|
58
|
-
defaultThreshold: config.retrieval.threshold,
|
|
59
|
-
});
|
|
70
|
+
const retriever = this.retriever ?? this.createRetriever(config, this.embeddingService, this.embeddingAdapter);
|
|
60
71
|
return new RAGPipeline(config, {
|
|
61
72
|
loader: this.loader,
|
|
62
73
|
chunker,
|
|
@@ -66,6 +77,58 @@ export class RAGPipelineBuilder {
|
|
|
66
77
|
reranker: this.reranker,
|
|
67
78
|
}, true);
|
|
68
79
|
}
|
|
80
|
+
createRetriever(config, embeddingService, embeddingAdapter) {
|
|
81
|
+
const { retrieval } = config;
|
|
82
|
+
const vectorRetriever = () => this.hybridSearch
|
|
83
|
+
? new HybridRetriever({
|
|
84
|
+
hybridSearch: this.hybridSearch,
|
|
85
|
+
defaultWeights: this.hybridWeights,
|
|
86
|
+
defaultTopK: retrieval.topK,
|
|
87
|
+
defaultThreshold: retrieval.threshold,
|
|
88
|
+
})
|
|
89
|
+
: new SimilarityRetriever({
|
|
90
|
+
embeddingAdapter,
|
|
91
|
+
embeddingService,
|
|
92
|
+
defaultTopK: retrieval.topK,
|
|
93
|
+
defaultThreshold: retrieval.threshold,
|
|
94
|
+
});
|
|
95
|
+
switch (retrieval.strategy) {
|
|
96
|
+
case 'similarity':
|
|
97
|
+
return new SimilarityRetriever({
|
|
98
|
+
embeddingAdapter,
|
|
99
|
+
embeddingService,
|
|
100
|
+
defaultTopK: retrieval.topK,
|
|
101
|
+
defaultThreshold: retrieval.threshold,
|
|
102
|
+
});
|
|
103
|
+
case 'mmr':
|
|
104
|
+
return new MMRRetriever({
|
|
105
|
+
embeddingAdapter,
|
|
106
|
+
embeddingService,
|
|
107
|
+
defaultLambda: retrieval.mmrLambda,
|
|
108
|
+
defaultTopK: retrieval.topK,
|
|
109
|
+
defaultThreshold: retrieval.threshold,
|
|
110
|
+
});
|
|
111
|
+
case 'hybrid':
|
|
112
|
+
if (!this.hybridSearch) {
|
|
113
|
+
throw new Error('retrieval strategy "hybrid" requires withHybridSearch() (or a custom withRetriever())');
|
|
114
|
+
}
|
|
115
|
+
return vectorRetriever();
|
|
116
|
+
case 'multi-query':
|
|
117
|
+
if (!this.queryExpander) {
|
|
118
|
+
throw new Error('retrieval strategy "multi-query" requires withQueryExpander() (or a custom withRetriever())');
|
|
119
|
+
}
|
|
120
|
+
return new MultiQueryRetriever({
|
|
121
|
+
baseRetriever: vectorRetriever(),
|
|
122
|
+
expandQuery: this.queryExpander,
|
|
123
|
+
defaultTopK: retrieval.topK,
|
|
124
|
+
defaultMaxQueries: retrieval.multiQueryCount,
|
|
125
|
+
});
|
|
126
|
+
default: {
|
|
127
|
+
const exhaustive = retrieval.strategy;
|
|
128
|
+
throw new Error(`Unknown retrieval strategy: ${String(exhaustive)}`);
|
|
129
|
+
}
|
|
130
|
+
}
|
|
131
|
+
}
|
|
69
132
|
resolveConfig() {
|
|
70
133
|
if (!this.configInput) {
|
|
71
134
|
throw new Error('config is required — call withConfig() before build()');
|
package/dist/rag-builder.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rag-builder.js","sourceRoot":"","sources":["../src/rag-builder.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"rag-builder.js","sourceRoot":"","sources":["../src/rag-builder.ts"],"names":[],"mappings":"AAYA,OAAO,EAAE,uBAAuB,EAA+B,MAAM,aAAa,CAAC;AACnF,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,aAAa,EAAE,MAAM,8BAA8B,CAAC;AAC7D,OAAO,EAAE,mBAAmB,EAAE,MAAM,sCAAsC,CAAC;AAC3E,OAAO,EAAE,YAAY,EAAE,MAAM,+BAA+B,CAAC;AAC7D,OAAO,EAAE,eAAe,EAAE,MAAM,kCAAkC,CAAC;AACnE,OAAO,EAAE,mBAAmB,EAAE,MAAM,uCAAuC,CAAC;AAE5E,MAAM,OAAO,kBAAkB;IACrB,MAAM,CAAkB;IACxB,OAAO,CAA0B;IACjC,gBAAgB,CAAoB;IACpC,gBAAgB,CAAoB;IACpC,SAAS,CAAa;IACtB,QAAQ,CAAY;IACpB,WAAW,CAA0B;IACrC,YAAY,CAAgB;IAC5B,aAAa,CAAuB;IACpC,aAAa,CAAwC;IAE7D,UAAU,CAAC,MAAsB;QAC/B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,OAAO,IAAI,CAAC;IACd,CAAC;IAED,WAAW,CAAC,OAA+B;QACzC,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;QACvB,OAAO,IAAI,CAAC;IACd,CAAC;IAED,oBAAoB,CAAC,OAAyB;QAC5C,IAAI,CAAC,gBAAgB,GAAG,OAAO,CAAC;QAChC,OAAO,IAAI,CAAC;IACd,CAAC;IAED,oBAAoB,CAAC,OAAyB;QAC5C,IAAI,CAAC,gBAAgB,GAAG,OAAO,CAAC;QAChC,OAAO,IAAI,CAAC;IACd,CAAC;IAED,aAAa,CAAC,SAAoB;QAChC,IAAI,CAAC,SAAS,GAAG,SAAS,CAAC;QAC3B,OAAO,IAAI,CAAC;IACd,CAAC;IAED,YAAY,CAAC,QAAkB;QAC7B,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;QACzB,OAAO,IAAI,CAAC;IACd,CAAC;IAED,yFAAyF;IACzF,gBAAgB,CAAC,YAA0B,EAAE,OAA6B;QACxE,IAAI,CAAC,YAAY,GAAG,YAAY,CAAC;QACjC,IAAI,CAAC,aAAa,GAAG,OAAO,CAAC;QAC7B,OAAO,IAAI,CAAC;IACd,CAAC;IAED,mEAAmE;IACnE,iBAAiB,CAAC,WAAiD;QACjE,IAAI,CAAC,aAAa,GAAG,WAAW,CAAC;QACjC,OAAO,IAAI,CAAC;IACd,CAAC;IAED,UAAU,CAAC,MAA8B;QACvC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC;QAC1B,OAAO,IAAI,CAAC;IACd,CAAC;IAED,KAAK;QACH,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjB,MAAM,IAAI,KAAK,CAAC,uDAAuD,CAAC,CAAC;QAC3E,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC3B,MAAM,IAAI,KAAK,CAAC,2EAA2E,CAAC,CAAC;QAC/F,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC3B,MAAM,IAAI,KAAK,CAAC,2EAA2E,CAAC,CAAC;QAC/F,CAAC;QAED,MAAM,MAAM,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC;QACpC,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,IAAI,aAAa,CAAC,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,gBAAgB,CAAC,CAAC;QACtF,MAAM,SAAS,GACb,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,eAAe,CAAC,MAAM,EAAE,IAAI,CAAC,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,CAAC,CAAC;QAE/F,OAAO,IAAI,WAAW,CACpB,MAAM,EACN;YACE,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,OAAO;YACP,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;YACvC,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;YACvC,SAAS;YACT,QAAQ,EAAE,IAAI,CAAC,QAAQ;SACxB,EACD,IAAI,CACL,CAAC;IACJ,CAAC;IAEO,eAAe,CACrB,MAAyB,EACzB,gBAAkC,EAClC,gBAAkC;QAElC,MAAM,EAAE,SAAS,EAAE,GAAG,MAAM,CAAC;QAC7B,MAAM,eAAe,GAAG,GAAc,EAAE,CACtC,IAAI,CAAC,YAAY;YACf,CAAC,CAAC,IAAI,eAAe,CAAC;gBAClB,YAAY,EAAE,IAAI,CAAC,YAAY;gBAC/B,cAAc,EAAE,IAAI,CAAC,aAAa;gBAClC,WAAW,EAAE,SAAS,CAAC,IAAI;gBAC3B,gBAAgB,EAAE,SAAS,CAAC,SAAS;aACtC,CAAC;YACJ,CAAC,CAAC,IAAI,mBAAmB,CAAC;gBACtB,gBAAgB;gBAChB,gBAAgB;gBAChB,WAAW,EAAE,SAAS,CAAC,IAAI;gBAC3B,gBAAgB,EAAE,SAAS,CAAC,SAAS;aACtC,CAAC,CAAC;QAET,QAAQ,SAAS,CAAC,QAAQ,EAAE,CAAC;YAC3B,KAAK,YAAY;gBACf,OAAO,IAAI,mBAAmB,CAAC;oBAC7B,gBAAgB;oBAChB,gBAAgB;oBAChB,WAAW,EAAE,SAAS,CAAC,IAAI;oBAC3B,gBAAgB,EAAE,SAAS,CAAC,SAAS;iBACtC,CAAC,CAAC;YAEL,KAAK,KAAK;gBACR,OAAO,IAAI,YAAY,CAAC;oBACtB,gBAAgB;oBAChB,gBAAgB;oBAChB,aAAa,EAAE,SAAS,CAAC,SAAS;oBAClC,WAAW,EAAE,SAAS,CAAC,IAAI;oBAC3B,gBAAgB,EAAE,SAAS,CAAC,SAAS;iBACtC,CAAC,CAAC;YAEL,KAAK,QAAQ;gBACX,IAAI,CAAC,IAAI,CAAC,YAAY,EAAE,CAAC;oBACvB,MAAM,IAAI,KAAK,CACb,uFAAuF,CACxF,CAAC;gBACJ,CAAC;gBACD,OAAO,eAAe,EAAE,CAAC;YAE3B,KAAK,aAAa;gBAChB,IAAI,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC;oBACxB,MAAM,IAAI,KAAK,CACb,6FAA6F,CAC9F,CAAC;gBACJ,CAAC;gBACD,OAAO,IAAI,mBAAmB,CAAC;oBAC7B,aAAa,EAAE,eAAe,EAAE;oBAChC,WAAW,EAAE,IAAI,CAAC,aAAa;oBAC/B,WAAW,EAAE,SAAS,CAAC,IAAI;oBAC3B,iBAAiB,EAAE,SAAS,CAAC,eAAe;iBAC7C,CAAC,CAAC;YAEL,OAAO,CAAC,CAAC,CAAC;gBACR,MAAM,UAAU,GAAU,SAAS,CAAC,QAAQ,CAAC;gBAC7C,MAAM,IAAI,KAAK,CAAC,+BAA+B,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;YACvE,CAAC;QACH,CAAC;IACH,CAAC;IAEO,aAAa;QACnB,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;YACtB,MAAM,IAAI,KAAK,CAAC,uDAAuD,CAAC,CAAC;QAC3E,CAAC;QACD,OAAO,uBAAuB,CAAC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;IACzD,CAAC;CACF"}
|
package/dist/rag-pipeline.d.ts
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rag-pipeline.d.ts","sourceRoot":"","sources":["../src/rag-pipeline.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,cAAc,EACd,OAAO,EACP,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,EAChB,SAAS,EACT,QAAQ,EACR,iBAAiB,EACjB,eAAe,EACf,eAAe,EAGhB,MAAM,qBAAqB,CAAC;
|
|
1
|
+
{"version":3,"file":"rag-pipeline.d.ts","sourceRoot":"","sources":["../src/rag-pipeline.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,cAAc,EACd,OAAO,EACP,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,EAChB,SAAS,EACT,QAAQ,EACR,iBAAiB,EACjB,eAAe,EACf,eAAe,EAGhB,MAAM,qBAAqB,CAAC;AAI7B,MAAM,WAAW,eAAe;IAC9B,MAAM,EAAE,cAAc,CAAC;IACvB,OAAO,EAAE,OAAO,GAAG,YAAY,CAAC;IAChC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,gBAAgB,EAAE,gBAAgB,CAAC;IACnC,SAAS,EAAE,SAAS,CAAC;IACrB,QAAQ,CAAC,EAAE,QAAQ,CAAC;CACrB;AASD,qBAAa,WAAW;IACtB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAoB;IAC3C,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAkB;IACvC,OAAO,CAAC,KAAK,CAAkE;gBAEnE,MAAM,EAAE,iBAAiB,EAAE,IAAI,EAAE,eAAe,EAAE,aAAa,UAAQ;IAK7E,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC;IA2BtE,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAezF,QAAQ;;;;;YAIM,aAAa;YAKb,WAAW;YAiBX,UAAU;CAkCzB"}
|
package/dist/rag-pipeline.js
CHANGED
|
@@ -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
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
|
|
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
|
-
|
|
23
|
-
|
|
32
|
+
ingestedDocuments++;
|
|
33
|
+
this.stats.documentsIngested++;
|
|
24
34
|
}
|
|
25
|
-
|
|
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 = {
|
|
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
|
-
|
|
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
|
package/dist/rag-pipeline.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rag-pipeline.js","sourceRoot":"","sources":["../src/rag-pipeline.ts"],"names":[],"mappings":"AAcA,OAAO,EAAE,uBAAuB,EAAE,MAAM,
|
|
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,
|
|
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,
|
|
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;
|
|
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"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"create-retriever.js","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,mBAAmB,EAAE,MAAM,
|
|
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
|
-
|
|
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;
|
|
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
|
|
29
|
-
const
|
|
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
|
|
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
|
-
...
|
|
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":"
|
|
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,
|
|
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"}
|
package/dist/retrievers/index.js
CHANGED
|
@@ -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,
|
|
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;
|
|
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
|
|
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":"
|
|
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
|
-
|
|
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
|