@cogitator-ai/rag 0.1.0 → 0.1.4
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 +2 -2
- package/dist/chunkers/create-chunker.d.ts +3 -0
- package/dist/chunkers/create-chunker.d.ts.map +1 -0
- package/dist/chunkers/create-chunker.js +32 -0
- package/dist/chunkers/create-chunker.js.map +1 -0
- package/dist/chunkers/fixed-chunker.d.ts +12 -0
- package/dist/chunkers/fixed-chunker.d.ts.map +1 -0
- package/dist/chunkers/fixed-chunker.js +40 -0
- package/dist/chunkers/fixed-chunker.js.map +1 -0
- package/dist/chunkers/index.d.ts +5 -0
- package/dist/chunkers/index.d.ts.map +1 -0
- package/dist/chunkers/index.js +5 -0
- package/dist/chunkers/index.js.map +1 -0
- package/dist/chunkers/recursive-chunker.d.ts +18 -0
- package/dist/chunkers/recursive-chunker.d.ts.map +1 -0
- package/dist/chunkers/recursive-chunker.js +161 -0
- package/dist/chunkers/recursive-chunker.js.map +1 -0
- package/dist/chunkers/semantic-chunker.d.ts +19 -0
- package/dist/chunkers/semantic-chunker.d.ts.map +1 -0
- package/dist/chunkers/semantic-chunker.js +176 -0
- package/dist/chunkers/semantic-chunker.js.map +1 -0
- package/dist/index.d.ts +11 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +10 -0
- package/dist/index.js.map +1 -0
- package/dist/loaders/csv-loader.d.ts +16 -0
- package/dist/loaders/csv-loader.d.ts.map +1 -0
- package/dist/loaders/csv-loader.js +57 -0
- package/dist/loaders/csv-loader.js.map +1 -0
- package/dist/loaders/html-loader.d.ts +12 -0
- package/dist/loaders/html-loader.d.ts.map +1 -0
- package/dist/loaders/html-loader.js +40 -0
- package/dist/loaders/html-loader.js.map +1 -0
- package/dist/loaders/index.d.ts +14 -0
- package/dist/loaders/index.d.ts.map +1 -0
- package/dist/loaders/index.js +8 -0
- package/dist/loaders/index.js.map +1 -0
- package/dist/loaders/json-loader.d.ts +16 -0
- package/dist/loaders/json-loader.d.ts.map +1 -0
- package/dist/loaders/json-loader.js +63 -0
- package/dist/loaders/json-loader.js.map +1 -0
- package/dist/loaders/markdown-loader.d.ts +14 -0
- package/dist/loaders/markdown-loader.d.ts.map +1 -0
- package/dist/loaders/markdown-loader.js +65 -0
- package/dist/loaders/markdown-loader.js.map +1 -0
- package/dist/loaders/pdf-loader.d.ts +14 -0
- package/dist/loaders/pdf-loader.d.ts.map +1 -0
- package/dist/loaders/pdf-loader.js +66 -0
- package/dist/loaders/pdf-loader.js.map +1 -0
- package/dist/loaders/text-loader.d.ts +8 -0
- package/dist/loaders/text-loader.d.ts.map +1 -0
- package/dist/loaders/text-loader.js +40 -0
- package/dist/loaders/text-loader.js.map +1 -0
- package/dist/loaders/web-loader.d.ts +13 -0
- package/dist/loaders/web-loader.d.ts.map +1 -0
- package/dist/loaders/web-loader.js +22 -0
- package/dist/loaders/web-loader.js.map +1 -0
- package/dist/rag-builder.d.ts +22 -0
- package/dist/rag-builder.d.ts.map +1 -0
- package/dist/rag-builder.js +74 -0
- package/dist/rag-builder.js.map +1 -0
- package/dist/rag-pipeline.d.ts +28 -0
- package/dist/rag-pipeline.d.ts.map +1 -0
- package/dist/rag-pipeline.js +63 -0
- package/dist/rag-pipeline.js.map +1 -0
- package/dist/rerankers/cohere-reranker.d.ts +12 -0
- package/dist/rerankers/cohere-reranker.d.ts.map +1 -0
- package/dist/rerankers/cohere-reranker.js +41 -0
- package/dist/rerankers/cohere-reranker.js.map +1 -0
- package/dist/rerankers/index.d.ts +3 -0
- package/dist/rerankers/index.d.ts.map +1 -0
- package/dist/rerankers/index.js +3 -0
- package/dist/rerankers/index.js.map +1 -0
- package/dist/rerankers/llm-reranker.d.ts +12 -0
- package/dist/rerankers/llm-reranker.d.ts.map +1 -0
- package/dist/rerankers/llm-reranker.js +56 -0
- package/dist/rerankers/llm-reranker.js.map +1 -0
- package/dist/retrievers/create-retriever.d.ts +33 -0
- package/dist/retrievers/create-retriever.d.ts.map +1 -0
- package/dist/retrievers/create-retriever.js +36 -0
- package/dist/retrievers/create-retriever.js.map +1 -0
- package/dist/retrievers/hybrid-retriever.d.ts +18 -0
- package/dist/retrievers/hybrid-retriever.d.ts.map +1 -0
- package/dist/retrievers/hybrid-retriever.js +44 -0
- package/dist/retrievers/hybrid-retriever.js.map +1 -0
- package/dist/retrievers/index.d.ts +11 -0
- package/dist/retrievers/index.d.ts.map +1 -0
- package/dist/retrievers/index.js +6 -0
- package/dist/retrievers/index.js.map +1 -0
- package/dist/retrievers/mmr-retriever.d.ts +20 -0
- package/dist/retrievers/mmr-retriever.d.ts.map +1 -0
- package/dist/retrievers/mmr-retriever.js +91 -0
- package/dist/retrievers/mmr-retriever.js.map +1 -0
- package/dist/retrievers/multi-query-retriever.d.ts +13 -0
- package/dist/retrievers/multi-query-retriever.d.ts.map +1 -0
- package/dist/retrievers/multi-query-retriever.js +29 -0
- package/dist/retrievers/multi-query-retriever.js.map +1 -0
- package/dist/retrievers/similarity-retriever.d.ts +17 -0
- package/dist/retrievers/similarity-retriever.d.ts.map +1 -0
- package/dist/retrievers/similarity-retriever.js +40 -0
- package/dist/retrievers/similarity-retriever.js.map +1 -0
- package/dist/schema.d.ts +70 -0
- package/dist/schema.d.ts.map +1 -0
- package/dist/schema.js +27 -0
- package/dist/schema.js.map +1 -0
- package/dist/tools.d.ts +23 -0
- package/dist/tools.d.ts.map +1 -0
- package/dist/tools.js +45 -0
- package/dist/tools.js.map +1 -0
- package/package.json +1 -6
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"rag-builder.js","sourceRoot":"","sources":["../src/rag-builder.ts"],"names":[],"mappings":"AAUA,OAAO,EAAE,uBAAuB,EAA+B,MAAM,UAAU,CAAC;AAChF,OAAO,EAAE,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAC7C,OAAO,EAAE,aAAa,EAAE,MAAM,2BAA2B,CAAC;AAC1D,OAAO,EAAE,mBAAmB,EAAE,MAAM,mCAAmC,CAAC;AAExE,MAAM,OAAO,kBAAkB;IACrB,MAAM,CAAkB;IACxB,OAAO,CAA0B;IACjC,gBAAgB,CAAoB;IACpC,gBAAgB,CAAoB;IACpC,SAAS,CAAa;IACtB,QAAQ,CAAY;IACpB,WAAW,CAA0B;IAE7C,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,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;YACd,IAAI,mBAAmB,CAAC;gBACtB,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;gBACvC,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;aACxC,CAAC,CAAC;QAEL,OAAO,IAAI,WAAW,CAAC,MAAM,EAAE;YAC7B,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,CAAC,CAAC;IACL,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"}
|
|
@@ -0,0 +1,28 @@
|
|
|
1
|
+
import type { DocumentLoader, Chunker, AsyncChunker, EmbeddingService, EmbeddingAdapter, Retriever, Reranker, RAGPipelineConfig, RetrievalConfig, RetrievalResult } from '@cogitator-ai/types';
|
|
2
|
+
export interface RAGPipelineDeps {
|
|
3
|
+
loader: DocumentLoader;
|
|
4
|
+
chunker: Chunker | AsyncChunker;
|
|
5
|
+
embeddingService: EmbeddingService;
|
|
6
|
+
embeddingAdapter: EmbeddingAdapter;
|
|
7
|
+
retriever: Retriever;
|
|
8
|
+
reranker?: Reranker;
|
|
9
|
+
}
|
|
10
|
+
export declare class RAGPipeline {
|
|
11
|
+
private readonly config;
|
|
12
|
+
private readonly deps;
|
|
13
|
+
private stats;
|
|
14
|
+
constructor(config: RAGPipelineConfig, deps: RAGPipelineDeps);
|
|
15
|
+
ingest(source: string): Promise<{
|
|
16
|
+
documents: number;
|
|
17
|
+
chunks: number;
|
|
18
|
+
}>;
|
|
19
|
+
query(text: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
|
|
20
|
+
getStats(): {
|
|
21
|
+
documentsIngested: number;
|
|
22
|
+
chunksStored: number;
|
|
23
|
+
queriesProcessed: number;
|
|
24
|
+
};
|
|
25
|
+
private chunkDocument;
|
|
26
|
+
private storeChunk;
|
|
27
|
+
}
|
|
28
|
+
//# sourceMappingURL=rag-pipeline.d.ts.map
|
|
@@ -0,0 +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;AAG7B,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;AAED,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;IAKtD,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC;IA8BtE,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAWzF,QAAQ;;;;;YAIM,aAAa;YAIb,UAAU;CAkBzB"}
|
|
@@ -0,0 +1,63 @@
|
|
|
1
|
+
import { RAGPipelineConfigSchema } from './schema';
|
|
2
|
+
export class RAGPipeline {
|
|
3
|
+
config;
|
|
4
|
+
deps;
|
|
5
|
+
stats = { documentsIngested: 0, chunksStored: 0, queriesProcessed: 0 };
|
|
6
|
+
constructor(config, deps) {
|
|
7
|
+
this.config = RAGPipelineConfigSchema.parse(config);
|
|
8
|
+
this.deps = deps;
|
|
9
|
+
}
|
|
10
|
+
async ingest(source) {
|
|
11
|
+
const documents = await this.deps.loader.load(source);
|
|
12
|
+
let totalChunks = 0;
|
|
13
|
+
for (const doc of documents) {
|
|
14
|
+
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`);
|
|
21
|
+
}
|
|
22
|
+
for (let i = 0; i < chunks.length; i++) {
|
|
23
|
+
await this.storeChunk(chunks[i], vectors[i], doc);
|
|
24
|
+
}
|
|
25
|
+
totalChunks += chunks.length;
|
|
26
|
+
}
|
|
27
|
+
this.stats.documentsIngested += documents.length;
|
|
28
|
+
this.stats.chunksStored += totalChunks;
|
|
29
|
+
return { documents: documents.length, chunks: totalChunks };
|
|
30
|
+
}
|
|
31
|
+
async query(text, options) {
|
|
32
|
+
const results = await this.deps.retriever.retrieve(text, options);
|
|
33
|
+
this.stats.queriesProcessed++;
|
|
34
|
+
if (this.deps.reranker && this.config.reranking?.enabled) {
|
|
35
|
+
return this.deps.reranker.rerank(text, results, this.config.reranking.topN);
|
|
36
|
+
}
|
|
37
|
+
return results;
|
|
38
|
+
}
|
|
39
|
+
getStats() {
|
|
40
|
+
return { ...this.stats };
|
|
41
|
+
}
|
|
42
|
+
async chunkDocument(doc) {
|
|
43
|
+
return this.deps.chunker.chunk(doc.content, doc.id);
|
|
44
|
+
}
|
|
45
|
+
async storeChunk(chunk, vector, doc) {
|
|
46
|
+
const result = await this.deps.embeddingAdapter.addEmbedding({
|
|
47
|
+
sourceId: chunk.id,
|
|
48
|
+
sourceType: 'document',
|
|
49
|
+
vector,
|
|
50
|
+
content: chunk.content,
|
|
51
|
+
metadata: {
|
|
52
|
+
...chunk.metadata,
|
|
53
|
+
documentId: doc.id,
|
|
54
|
+
source: doc.source,
|
|
55
|
+
order: chunk.order,
|
|
56
|
+
},
|
|
57
|
+
});
|
|
58
|
+
if (!result.success) {
|
|
59
|
+
throw new Error(result.error);
|
|
60
|
+
}
|
|
61
|
+
}
|
|
62
|
+
}
|
|
63
|
+
//# sourceMappingURL=rag-pipeline.js.map
|
|
@@ -0,0 +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;QAC1D,IAAI,CAAC,MAAM,GAAG,uBAAuB,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QACpD,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,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACvC,MAAM,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YACpD,CAAC;YAED,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,OAAO,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;QAClE,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"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { Reranker, RetrievalResult } from '@cogitator-ai/types';
|
|
2
|
+
export interface CohereRerankerConfig {
|
|
3
|
+
apiKey: string;
|
|
4
|
+
model?: string;
|
|
5
|
+
}
|
|
6
|
+
export declare class CohereReranker implements Reranker {
|
|
7
|
+
private apiKey;
|
|
8
|
+
private model;
|
|
9
|
+
constructor(config: CohereRerankerConfig);
|
|
10
|
+
rerank(query: string, results: RetrievalResult[], topN?: number): Promise<RetrievalResult[]>;
|
|
11
|
+
}
|
|
12
|
+
//# sourceMappingURL=cohere-reranker.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cohere-reranker.d.ts","sourceRoot":"","sources":["../../src/rerankers/cohere-reranker.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAErE,MAAM,WAAW,oBAAoB;IACnC,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAYD,qBAAa,cAAe,YAAW,QAAQ;IAC7C,OAAO,CAAC,MAAM,CAAS;IACvB,OAAO,CAAC,KAAK,CAAS;gBAEV,MAAM,EAAE,oBAAoB;IAKlC,MAAM,CACV,KAAK,EAAE,MAAM,EACb,OAAO,EAAE,eAAe,EAAE,EAC1B,IAAI,CAAC,EAAE,MAAM,GACZ,OAAO,CAAC,eAAe,EAAE,CAAC;CAmC9B"}
|
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
const DEFAULT_MODEL = 'rerank-v3.5';
|
|
2
|
+
const COHERE_RERANK_URL = 'https://api.cohere.com/v2/rerank';
|
|
3
|
+
export class CohereReranker {
|
|
4
|
+
apiKey;
|
|
5
|
+
model;
|
|
6
|
+
constructor(config) {
|
|
7
|
+
this.apiKey = config.apiKey;
|
|
8
|
+
this.model = config.model ?? DEFAULT_MODEL;
|
|
9
|
+
}
|
|
10
|
+
async rerank(query, results, topN) {
|
|
11
|
+
if (results.length === 0)
|
|
12
|
+
return [];
|
|
13
|
+
const n = topN ?? results.length;
|
|
14
|
+
const response = await fetch(COHERE_RERANK_URL, {
|
|
15
|
+
method: 'POST',
|
|
16
|
+
headers: {
|
|
17
|
+
'Content-Type': 'application/json',
|
|
18
|
+
Authorization: `Bearer ${this.apiKey}`,
|
|
19
|
+
},
|
|
20
|
+
body: JSON.stringify({
|
|
21
|
+
model: this.model,
|
|
22
|
+
query,
|
|
23
|
+
documents: results.map((r) => r.content),
|
|
24
|
+
top_n: n,
|
|
25
|
+
}),
|
|
26
|
+
});
|
|
27
|
+
if (!response.ok) {
|
|
28
|
+
const text = await response.text();
|
|
29
|
+
throw new Error(`Cohere rerank failed (${response.status}): ${text}`);
|
|
30
|
+
}
|
|
31
|
+
const data = (await response.json());
|
|
32
|
+
const mapped = data.results
|
|
33
|
+
.filter(({ index }) => index >= 0 && index < results.length)
|
|
34
|
+
.map(({ index, relevance_score }) => ({
|
|
35
|
+
...results[index],
|
|
36
|
+
score: relevance_score,
|
|
37
|
+
}));
|
|
38
|
+
return mapped.slice(0, n);
|
|
39
|
+
}
|
|
40
|
+
}
|
|
41
|
+
//# sourceMappingURL=cohere-reranker.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"cohere-reranker.js","sourceRoot":"","sources":["../../src/rerankers/cohere-reranker.ts"],"names":[],"mappings":"AAcA,MAAM,aAAa,GAAG,aAAa,CAAC;AACpC,MAAM,iBAAiB,GAAG,kCAAkC,CAAC;AAE7D,MAAM,OAAO,cAAc;IACjB,MAAM,CAAS;IACf,KAAK,CAAS;IAEtB,YAAY,MAA4B;QACtC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;QAC5B,IAAI,CAAC,KAAK,GAAG,MAAM,CAAC,KAAK,IAAI,aAAa,CAAC;IAC7C,CAAC;IAED,KAAK,CAAC,MAAM,CACV,KAAa,EACb,OAA0B,EAC1B,IAAa;QAEb,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEpC,MAAM,CAAC,GAAG,IAAI,IAAI,OAAO,CAAC,MAAM,CAAC;QAEjC,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,iBAAiB,EAAE;YAC9C,MAAM,EAAE,MAAM;YACd,OAAO,EAAE;gBACP,cAAc,EAAE,kBAAkB;gBAClC,aAAa,EAAE,UAAU,IAAI,CAAC,MAAM,EAAE;aACvC;YACD,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC;gBACnB,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,KAAK;gBACL,SAAS,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC;gBACxC,KAAK,EAAE,CAAC;aACT,CAAC;SACH,CAAC,CAAC;QAEH,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;YACjB,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAC;YACnC,MAAM,IAAI,KAAK,CAAC,yBAAyB,QAAQ,CAAC,MAAM,MAAM,IAAI,EAAE,CAAC,CAAC;QACxE,CAAC;QAED,MAAM,IAAI,GAAG,CAAC,MAAM,QAAQ,CAAC,IAAI,EAAE,CAAyB,CAAC;QAE7D,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO;aACxB,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,IAAI,CAAC,IAAI,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC;aAC3D,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,eAAe,EAAE,EAAE,EAAE,CAAC,CAAC;YACpC,GAAG,OAAO,CAAC,KAAK,CAAC;YACjB,KAAK,EAAE,eAAe;SACvB,CAAC,CAAC,CAAC;QAEN,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IAC5B,CAAC;CACF"}
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { Reranker, RetrievalResult } from '@cogitator-ai/types';
|
|
2
|
+
export interface LLMRerankerConfig {
|
|
3
|
+
generateFn: (prompt: string) => Promise<string>;
|
|
4
|
+
}
|
|
5
|
+
export declare class LLMReranker implements Reranker {
|
|
6
|
+
private generateFn;
|
|
7
|
+
constructor(config: LLMRerankerConfig);
|
|
8
|
+
rerank(query: string, results: RetrievalResult[], topN?: number): Promise<RetrievalResult[]>;
|
|
9
|
+
private buildPrompt;
|
|
10
|
+
private parseScores;
|
|
11
|
+
}
|
|
12
|
+
//# sourceMappingURL=llm-reranker.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"llm-reranker.d.ts","sourceRoot":"","sources":["../../src/rerankers/llm-reranker.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAErE,MAAM,WAAW,iBAAiB;IAChC,UAAU,EAAE,CAAC,MAAM,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,CAAC,CAAC;CACjD;AAED,qBAAa,WAAY,YAAW,QAAQ;IAC1C,OAAO,CAAC,UAAU,CAAsC;gBAE5C,MAAM,EAAE,iBAAiB;IAI/B,MAAM,CACV,KAAK,EAAE,MAAM,EACb,OAAO,EAAE,eAAe,EAAE,EAC1B,IAAI,CAAC,EAAE,MAAM,GACZ,OAAO,CAAC,eAAe,EAAE,CAAC;IA0B7B,OAAO,CAAC,WAAW;IAcnB,OAAO,CAAC,WAAW;CAiBpB"}
|
|
@@ -0,0 +1,56 @@
|
|
|
1
|
+
export class LLMReranker {
|
|
2
|
+
generateFn;
|
|
3
|
+
constructor(config) {
|
|
4
|
+
this.generateFn = config.generateFn;
|
|
5
|
+
}
|
|
6
|
+
async rerank(query, results, topN) {
|
|
7
|
+
if (results.length === 0)
|
|
8
|
+
return [];
|
|
9
|
+
try {
|
|
10
|
+
const prompt = this.buildPrompt(query, results);
|
|
11
|
+
const response = await this.generateFn(prompt);
|
|
12
|
+
const scores = this.parseScores(response, results.length);
|
|
13
|
+
const scored = results.map((result, i) => {
|
|
14
|
+
const entry = scores.find((s) => s.index === i);
|
|
15
|
+
return { result, llmScore: entry?.score ?? 0 };
|
|
16
|
+
});
|
|
17
|
+
scored.sort((a, b) => b.llmScore - a.llmScore);
|
|
18
|
+
const reranked = scored.map(({ result, llmScore }) => ({
|
|
19
|
+
...result,
|
|
20
|
+
score: Math.max(0, Math.min(1, llmScore / 10)),
|
|
21
|
+
}));
|
|
22
|
+
return topN ? reranked.slice(0, topN) : reranked;
|
|
23
|
+
}
|
|
24
|
+
catch {
|
|
25
|
+
const fallback = [...results];
|
|
26
|
+
return topN ? fallback.slice(0, topN) : fallback;
|
|
27
|
+
}
|
|
28
|
+
}
|
|
29
|
+
buildPrompt(query, results) {
|
|
30
|
+
const docs = results.map((r, i) => `[${i}] ${r.content}`).join('\n\n');
|
|
31
|
+
return [
|
|
32
|
+
"Score each document's relevance to the query on a scale of 0-10.",
|
|
33
|
+
'Return ONLY a JSON array: [{ "index": number, "score": number }]',
|
|
34
|
+
'',
|
|
35
|
+
`Query: ${query}`,
|
|
36
|
+
'',
|
|
37
|
+
'Documents:',
|
|
38
|
+
docs,
|
|
39
|
+
].join('\n');
|
|
40
|
+
}
|
|
41
|
+
parseScores(response, count) {
|
|
42
|
+
const jsonMatch = /\[[\s\S]*?\]/.exec(response);
|
|
43
|
+
if (!jsonMatch)
|
|
44
|
+
throw new Error('No JSON array found in response');
|
|
45
|
+
const parsed = JSON.parse(jsonMatch[0]);
|
|
46
|
+
if (!Array.isArray(parsed))
|
|
47
|
+
throw new Error('Response is not an array');
|
|
48
|
+
return parsed.filter((item) => typeof item === 'object' &&
|
|
49
|
+
item !== null &&
|
|
50
|
+
typeof item.index === 'number' &&
|
|
51
|
+
typeof item.score === 'number' &&
|
|
52
|
+
item.index >= 0 &&
|
|
53
|
+
item.index < count);
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
//# sourceMappingURL=llm-reranker.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"llm-reranker.js","sourceRoot":"","sources":["../../src/rerankers/llm-reranker.ts"],"names":[],"mappings":"AAMA,MAAM,OAAO,WAAW;IACd,UAAU,CAAsC;IAExD,YAAY,MAAyB;QACnC,IAAI,CAAC,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC;IACtC,CAAC;IAED,KAAK,CAAC,MAAM,CACV,KAAa,EACb,OAA0B,EAC1B,IAAa;QAEb,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEpC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;YAChD,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;YAC/C,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,CAAC,QAAQ,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;YAE1D,MAAM,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE;gBACvC,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC;gBAChD,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;YACjD,CAAC,CAAC,CAAC;YAEH,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;YAC/C,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAC;gBACrD,GAAG,MAAM;gBACT,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,GAAG,EAAE,CAAC,CAAC;aAC/C,CAAC,CAAC,CAAC;YAEJ,OAAO,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;QACnD,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,QAAQ,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC;YAC9B,OAAO,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;QACnD,CAAC;IACH,CAAC;IAEO,WAAW,CAAC,KAAa,EAAE,OAA0B;QAC3D,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAEvE,OAAO;YACL,kEAAkE;YAClE,kEAAkE;YAClE,EAAE;YACF,UAAU,KAAK,EAAE;YACjB,EAAE;YACF,YAAY;YACZ,IAAI;SACL,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACf,CAAC;IAEO,WAAW,CAAC,QAAgB,EAAE,KAAa;QACjD,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAChD,IAAI,CAAC,SAAS;YAAE,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QAEnE,MAAM,MAAM,GAAY,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;QACjD,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,0BAA0B,CAAC,CAAC;QAExE,OAAO,MAAM,CAAC,MAAM,CAClB,CAAC,IAAI,EAA4C,EAAE,CACjD,OAAO,IAAI,KAAK,QAAQ;YACxB,IAAI,KAAK,IAAI;YACb,OAAQ,IAAgC,CAAC,KAAK,KAAK,QAAQ;YAC3D,OAAQ,IAAgC,CAAC,KAAK,KAAK,QAAQ;YAC1D,IAA0B,CAAC,KAAK,IAAI,CAAC;YACrC,IAA0B,CAAC,KAAK,GAAG,KAAK,CAC5C,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
import type { Retriever, EmbeddingAdapter, EmbeddingService, HybridSearchWeights } from '@cogitator-ai/types';
|
|
2
|
+
import type { HybridSearch } from '@cogitator-ai/memory';
|
|
3
|
+
interface BaseRetrieverDeps {
|
|
4
|
+
embeddingAdapter: EmbeddingAdapter;
|
|
5
|
+
embeddingService: EmbeddingService;
|
|
6
|
+
}
|
|
7
|
+
interface SimilarityDeps extends BaseRetrieverDeps {
|
|
8
|
+
strategy: 'similarity';
|
|
9
|
+
topK?: number;
|
|
10
|
+
threshold?: number;
|
|
11
|
+
}
|
|
12
|
+
interface MMRDeps extends BaseRetrieverDeps {
|
|
13
|
+
strategy: 'mmr';
|
|
14
|
+
lambda?: number;
|
|
15
|
+
topK?: number;
|
|
16
|
+
threshold?: number;
|
|
17
|
+
}
|
|
18
|
+
interface HybridDeps {
|
|
19
|
+
strategy: 'hybrid';
|
|
20
|
+
hybridSearch: HybridSearch;
|
|
21
|
+
weights?: HybridSearchWeights;
|
|
22
|
+
topK?: number;
|
|
23
|
+
threshold?: number;
|
|
24
|
+
}
|
|
25
|
+
interface MultiQueryDeps {
|
|
26
|
+
strategy: 'multi-query';
|
|
27
|
+
baseRetriever: Retriever;
|
|
28
|
+
expandQuery: (query: string) => Promise<string[]>;
|
|
29
|
+
}
|
|
30
|
+
export type CreateRetrieverConfig = SimilarityDeps | MMRDeps | HybridDeps | MultiQueryDeps;
|
|
31
|
+
export declare function createRetriever(config: CreateRetrieverConfig): Retriever;
|
|
32
|
+
export {};
|
|
33
|
+
//# sourceMappingURL=create-retriever.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"create-retriever.d.ts","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,gBAAgB,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,qBAAqB,CAAC;AAC9G,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;CACnD;AAED,MAAM,MAAM,qBAAqB,GAAG,cAAc,GAAG,OAAO,GAAG,UAAU,GAAG,cAAc,CAAC;AAE3F,wBAAgB,eAAe,CAAC,MAAM,EAAE,qBAAqB,GAAG,SAAS,CAiCxE"}
|
|
@@ -0,0 +1,36 @@
|
|
|
1
|
+
import { SimilarityRetriever } from './similarity-retriever';
|
|
2
|
+
import { MMRRetriever } from './mmr-retriever';
|
|
3
|
+
import { HybridRetriever } from './hybrid-retriever';
|
|
4
|
+
import { MultiQueryRetriever } from './multi-query-retriever';
|
|
5
|
+
export function createRetriever(config) {
|
|
6
|
+
switch (config.strategy) {
|
|
7
|
+
case 'similarity':
|
|
8
|
+
return new SimilarityRetriever({
|
|
9
|
+
embeddingAdapter: config.embeddingAdapter,
|
|
10
|
+
embeddingService: config.embeddingService,
|
|
11
|
+
defaultTopK: config.topK,
|
|
12
|
+
defaultThreshold: config.threshold,
|
|
13
|
+
});
|
|
14
|
+
case 'mmr':
|
|
15
|
+
return new MMRRetriever({
|
|
16
|
+
embeddingAdapter: config.embeddingAdapter,
|
|
17
|
+
embeddingService: config.embeddingService,
|
|
18
|
+
defaultLambda: config.lambda,
|
|
19
|
+
defaultTopK: config.topK,
|
|
20
|
+
defaultThreshold: config.threshold,
|
|
21
|
+
});
|
|
22
|
+
case 'hybrid':
|
|
23
|
+
return new HybridRetriever({
|
|
24
|
+
hybridSearch: config.hybridSearch,
|
|
25
|
+
defaultWeights: config.weights,
|
|
26
|
+
defaultTopK: config.topK,
|
|
27
|
+
defaultThreshold: config.threshold,
|
|
28
|
+
});
|
|
29
|
+
case 'multi-query':
|
|
30
|
+
return new MultiQueryRetriever({
|
|
31
|
+
baseRetriever: config.baseRetriever,
|
|
32
|
+
expandQuery: config.expandQuery,
|
|
33
|
+
});
|
|
34
|
+
}
|
|
35
|
+
}
|
|
36
|
+
//# sourceMappingURL=create-retriever.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"create-retriever.js","sourceRoot":"","sources":["../../src/retrievers/create-retriever.ts"],"names":[],"mappings":"AAEA,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;AAoC9D,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;aAChC,CAAC,CAAC;IACP,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
import type { Retriever, RetrievalConfig, RetrievalResult, HybridSearchWeights } from '@cogitator-ai/types';
|
|
2
|
+
import type { HybridSearch } from '@cogitator-ai/memory';
|
|
3
|
+
export interface HybridRetrieverConfig {
|
|
4
|
+
hybridSearch: HybridSearch;
|
|
5
|
+
defaultWeights?: HybridSearchWeights;
|
|
6
|
+
defaultTopK?: number;
|
|
7
|
+
defaultThreshold?: number;
|
|
8
|
+
}
|
|
9
|
+
export declare class HybridRetriever implements Retriever {
|
|
10
|
+
private readonly hybridSearch;
|
|
11
|
+
private readonly defaultWeights?;
|
|
12
|
+
private readonly defaultTopK;
|
|
13
|
+
private readonly defaultThreshold;
|
|
14
|
+
constructor(config: HybridRetrieverConfig);
|
|
15
|
+
retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
|
|
16
|
+
private toRetrievalResult;
|
|
17
|
+
}
|
|
18
|
+
//# sourceMappingURL=hybrid-retriever.d.ts.map
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +1,44 @@
|
|
|
1
|
+
const DEFAULT_TOP_K = 10;
|
|
2
|
+
const DEFAULT_THRESHOLD = 0.0;
|
|
3
|
+
export class HybridRetriever {
|
|
4
|
+
hybridSearch;
|
|
5
|
+
defaultWeights;
|
|
6
|
+
defaultTopK;
|
|
7
|
+
defaultThreshold;
|
|
8
|
+
constructor(config) {
|
|
9
|
+
this.hybridSearch = config.hybridSearch;
|
|
10
|
+
this.defaultWeights = config.defaultWeights;
|
|
11
|
+
this.defaultTopK = config.defaultTopK ?? DEFAULT_TOP_K;
|
|
12
|
+
this.defaultThreshold = config.defaultThreshold ?? DEFAULT_THRESHOLD;
|
|
13
|
+
}
|
|
14
|
+
async retrieve(query, options) {
|
|
15
|
+
const result = await this.hybridSearch.search({
|
|
16
|
+
query,
|
|
17
|
+
strategy: 'hybrid',
|
|
18
|
+
weights: this.defaultWeights,
|
|
19
|
+
limit: options?.topK ?? this.defaultTopK,
|
|
20
|
+
threshold: options?.threshold ?? this.defaultThreshold,
|
|
21
|
+
});
|
|
22
|
+
if (!result.success) {
|
|
23
|
+
throw new Error(result.error);
|
|
24
|
+
}
|
|
25
|
+
return result.data.map((entry) => this.toRetrievalResult(entry));
|
|
26
|
+
}
|
|
27
|
+
toRetrievalResult(entry) {
|
|
28
|
+
const rawDocId = entry.metadata?.documentId;
|
|
29
|
+
const documentId = typeof rawDocId === 'string' ? rawDocId : entry.sourceId;
|
|
30
|
+
return {
|
|
31
|
+
chunkId: entry.sourceId,
|
|
32
|
+
documentId,
|
|
33
|
+
content: entry.content,
|
|
34
|
+
score: entry.score,
|
|
35
|
+
source: entry.sourceType,
|
|
36
|
+
metadata: {
|
|
37
|
+
...entry.metadata,
|
|
38
|
+
vectorScore: entry.vectorScore,
|
|
39
|
+
keywordScore: entry.keywordScore,
|
|
40
|
+
},
|
|
41
|
+
};
|
|
42
|
+
}
|
|
43
|
+
}
|
|
44
|
+
//# sourceMappingURL=hybrid-retriever.js.map
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
export { SimilarityRetriever } from './similarity-retriever';
|
|
2
|
+
export type { SimilarityRetrieverConfig } from './similarity-retriever';
|
|
3
|
+
export { MMRRetriever } from './mmr-retriever';
|
|
4
|
+
export type { MMRRetrieverConfig } from './mmr-retriever';
|
|
5
|
+
export { HybridRetriever } from './hybrid-retriever';
|
|
6
|
+
export type { HybridRetrieverConfig } from './hybrid-retriever';
|
|
7
|
+
export { MultiQueryRetriever } from './multi-query-retriever';
|
|
8
|
+
export type { MultiQueryRetrieverConfig } from './multi-query-retriever';
|
|
9
|
+
export { createRetriever } from './create-retriever';
|
|
10
|
+
export type { CreateRetrieverConfig } from './create-retriever';
|
|
11
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +1,6 @@
|
|
|
1
|
+
export { SimilarityRetriever } from './similarity-retriever';
|
|
2
|
+
export { MMRRetriever } from './mmr-retriever';
|
|
3
|
+
export { HybridRetriever } from './hybrid-retriever';
|
|
4
|
+
export { MultiQueryRetriever } from './multi-query-retriever';
|
|
5
|
+
export { createRetriever } from './create-retriever';
|
|
6
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +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"}
|
|
@@ -0,0 +1,20 @@
|
|
|
1
|
+
import type { EmbeddingAdapter, EmbeddingService, Retriever, RetrievalConfig, RetrievalResult } from '@cogitator-ai/types';
|
|
2
|
+
export interface MMRRetrieverConfig {
|
|
3
|
+
embeddingAdapter: EmbeddingAdapter;
|
|
4
|
+
embeddingService: EmbeddingService;
|
|
5
|
+
defaultLambda?: number;
|
|
6
|
+
defaultTopK?: number;
|
|
7
|
+
defaultThreshold?: number;
|
|
8
|
+
}
|
|
9
|
+
export declare class MMRRetriever implements Retriever {
|
|
10
|
+
private readonly adapter;
|
|
11
|
+
private readonly embedding;
|
|
12
|
+
private readonly defaultLambda;
|
|
13
|
+
private readonly defaultTopK;
|
|
14
|
+
private readonly defaultThreshold;
|
|
15
|
+
constructor(config: MMRRetrieverConfig);
|
|
16
|
+
retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
|
|
17
|
+
private selectMMR;
|
|
18
|
+
private toRetrievalResult;
|
|
19
|
+
}
|
|
20
|
+
//# sourceMappingURL=mmr-retriever.d.ts.map
|
|
@@ -0,0 +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;AAuBD,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;IAuB7F,OAAO,CAAC,SAAS;IAsCjB,OAAO,CAAC,iBAAiB;CAa1B"}
|
|
@@ -0,0 +1,91 @@
|
|
|
1
|
+
const DEFAULT_TOP_K = 10;
|
|
2
|
+
const DEFAULT_THRESHOLD = 0.0;
|
|
3
|
+
const DEFAULT_LAMBDA = 0.7;
|
|
4
|
+
const CANDIDATE_MULTIPLIER = 3;
|
|
5
|
+
function cosineSimilarity(a, b) {
|
|
6
|
+
const len = Math.min(a.length, b.length);
|
|
7
|
+
let dot = 0;
|
|
8
|
+
let normA = 0;
|
|
9
|
+
let normB = 0;
|
|
10
|
+
for (let i = 0; i < len; i++) {
|
|
11
|
+
dot += a[i] * b[i];
|
|
12
|
+
normA += a[i] * a[i];
|
|
13
|
+
normB += b[i] * b[i];
|
|
14
|
+
}
|
|
15
|
+
const denom = Math.sqrt(normA) * Math.sqrt(normB);
|
|
16
|
+
return denom === 0 ? 0 : dot / denom;
|
|
17
|
+
}
|
|
18
|
+
export class MMRRetriever {
|
|
19
|
+
adapter;
|
|
20
|
+
embedding;
|
|
21
|
+
defaultLambda;
|
|
22
|
+
defaultTopK;
|
|
23
|
+
defaultThreshold;
|
|
24
|
+
constructor(config) {
|
|
25
|
+
this.adapter = config.embeddingAdapter;
|
|
26
|
+
this.embedding = config.embeddingService;
|
|
27
|
+
this.defaultLambda = config.defaultLambda ?? DEFAULT_LAMBDA;
|
|
28
|
+
this.defaultTopK = config.defaultTopK ?? DEFAULT_TOP_K;
|
|
29
|
+
this.defaultThreshold = config.defaultThreshold ?? DEFAULT_THRESHOLD;
|
|
30
|
+
}
|
|
31
|
+
async retrieve(query, options) {
|
|
32
|
+
const topK = options?.topK ?? this.defaultTopK;
|
|
33
|
+
const threshold = options?.threshold ?? this.defaultThreshold;
|
|
34
|
+
const queryVector = await this.embedding.embed(query);
|
|
35
|
+
const result = await this.adapter.search({
|
|
36
|
+
vector: queryVector,
|
|
37
|
+
query,
|
|
38
|
+
limit: topK * CANDIDATE_MULTIPLIER,
|
|
39
|
+
threshold,
|
|
40
|
+
});
|
|
41
|
+
if (!result.success) {
|
|
42
|
+
throw new Error(result.error);
|
|
43
|
+
}
|
|
44
|
+
const candidates = result.data;
|
|
45
|
+
if (candidates.length === 0)
|
|
46
|
+
return [];
|
|
47
|
+
const selected = this.selectMMR(candidates, topK);
|
|
48
|
+
return selected.map((entry) => this.toRetrievalResult(entry));
|
|
49
|
+
}
|
|
50
|
+
selectMMR(candidates, topK) {
|
|
51
|
+
const lambda = this.defaultLambda;
|
|
52
|
+
const remaining = new Set(candidates.map((_, i) => i));
|
|
53
|
+
const selected = [];
|
|
54
|
+
while (selected.length < topK && remaining.size > 0) {
|
|
55
|
+
let bestIdx = -1;
|
|
56
|
+
let bestScore = -Infinity;
|
|
57
|
+
for (const idx of remaining) {
|
|
58
|
+
const relevance = candidates[idx].score;
|
|
59
|
+
let maxSelectedSim = 0;
|
|
60
|
+
for (const sel of selected) {
|
|
61
|
+
const sim = cosineSimilarity(candidates[idx].vector, sel.vector);
|
|
62
|
+
if (sim > maxSelectedSim)
|
|
63
|
+
maxSelectedSim = sim;
|
|
64
|
+
}
|
|
65
|
+
const mmrScore = lambda * relevance - (1 - lambda) * maxSelectedSim;
|
|
66
|
+
if (mmrScore > bestScore) {
|
|
67
|
+
bestScore = mmrScore;
|
|
68
|
+
bestIdx = idx;
|
|
69
|
+
}
|
|
70
|
+
}
|
|
71
|
+
if (bestIdx === -1)
|
|
72
|
+
break;
|
|
73
|
+
remaining.delete(bestIdx);
|
|
74
|
+
selected.push({ ...candidates[bestIdx], mmrScore: bestScore });
|
|
75
|
+
}
|
|
76
|
+
return selected;
|
|
77
|
+
}
|
|
78
|
+
toRetrievalResult(entry) {
|
|
79
|
+
const rawDocId = entry.metadata?.documentId;
|
|
80
|
+
const documentId = typeof rawDocId === 'string' ? rawDocId : entry.sourceId;
|
|
81
|
+
return {
|
|
82
|
+
chunkId: entry.sourceId,
|
|
83
|
+
documentId,
|
|
84
|
+
content: entry.content,
|
|
85
|
+
score: entry.mmrScore,
|
|
86
|
+
source: entry.sourceType,
|
|
87
|
+
metadata: entry.metadata,
|
|
88
|
+
};
|
|
89
|
+
}
|
|
90
|
+
}
|
|
91
|
+
//# sourceMappingURL=mmr-retriever.js.map
|
|
@@ -0,0 +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,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC;IACzC,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,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7B,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,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAClD,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;QAEZ,MAAM,MAAM,GAAG,IAAI,CAAC,aAAa,CAAC;QAClC,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,CAAC;gBACvB,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;gBAED,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,KAAK,CAAC,QAAQ;SACzB,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
import type { Retriever, RetrievalConfig, RetrievalResult } from '@cogitator-ai/types';
|
|
2
|
+
export interface MultiQueryRetrieverConfig {
|
|
3
|
+
baseRetriever: Retriever;
|
|
4
|
+
expandQuery: (query: string) => Promise<string[]>;
|
|
5
|
+
}
|
|
6
|
+
export declare class MultiQueryRetriever implements Retriever {
|
|
7
|
+
private readonly baseRetriever;
|
|
8
|
+
private readonly expandQuery;
|
|
9
|
+
constructor(config: MultiQueryRetrieverConfig);
|
|
10
|
+
retrieve(query: string, options?: Partial<RetrievalConfig>): Promise<RetrievalResult[]>;
|
|
11
|
+
private deduplicateAndMerge;
|
|
12
|
+
}
|
|
13
|
+
//# sourceMappingURL=multi-query-retriever.d.ts.map
|
|
@@ -0,0 +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;CACnD;AAID,qBAAa,mBAAoB,YAAW,SAAS;IACnD,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAY;IAC1C,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAuC;gBAEvD,MAAM,EAAE,yBAAyB;IAKvC,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,eAAe,CAAC,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC;IAe7F,OAAO,CAAC,mBAAmB;CAY5B"}
|
|
@@ -0,0 +1,29 @@
|
|
|
1
|
+
const DEFAULT_TOP_K = 10;
|
|
2
|
+
export class MultiQueryRetriever {
|
|
3
|
+
baseRetriever;
|
|
4
|
+
expandQuery;
|
|
5
|
+
constructor(config) {
|
|
6
|
+
this.baseRetriever = config.baseRetriever;
|
|
7
|
+
this.expandQuery = config.expandQuery;
|
|
8
|
+
}
|
|
9
|
+
async retrieve(query, options) {
|
|
10
|
+
const variants = await this.expandQuery(query);
|
|
11
|
+
if (variants.length === 0)
|
|
12
|
+
return [];
|
|
13
|
+
const allResults = await Promise.all(variants.map((variant) => this.baseRetriever.retrieve(variant, options)));
|
|
14
|
+
const merged = this.deduplicateAndMerge(allResults.flat());
|
|
15
|
+
const topK = options?.topK ?? DEFAULT_TOP_K;
|
|
16
|
+
return merged.slice(0, topK);
|
|
17
|
+
}
|
|
18
|
+
deduplicateAndMerge(results) {
|
|
19
|
+
const best = new Map();
|
|
20
|
+
for (const result of results) {
|
|
21
|
+
const existing = best.get(result.chunkId);
|
|
22
|
+
if (!existing || result.score > existing.score) {
|
|
23
|
+
best.set(result.chunkId, result);
|
|
24
|
+
}
|
|
25
|
+
}
|
|
26
|
+
return [...best.values()].sort((a, b) => b.score - a.score);
|
|
27
|
+
}
|
|
28
|
+
}
|
|
29
|
+
//# sourceMappingURL=multi-query-retriever.js.map
|