@anvia/fastembed 0.1.4 → 0.2.1
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 -1
- package/dist/index.d.ts +1 -1
- package/dist/index.js.map +1 -1
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -19,7 +19,8 @@ pnpm --filter @anvia/fastembed build
|
|
|
19
19
|
## Usage
|
|
20
20
|
|
|
21
21
|
```ts
|
|
22
|
-
import { embedDocuments
|
|
22
|
+
import { embedDocuments } from "@anvia/core/embeddings";
|
|
23
|
+
import { InMemoryVectorStore } from "@anvia/core/vector-store";
|
|
23
24
|
import { createFastEmbedEmbeddingModel } from "@anvia/fastembed";
|
|
24
25
|
|
|
25
26
|
const embeddingModel = await createFastEmbedEmbeddingModel();
|
package/dist/index.d.ts
CHANGED
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import { EmbeddingModel, Embedding } from '@anvia/core';
|
|
1
|
+
import { EmbeddingModel, Embedding } from '@anvia/core/embeddings';
|
|
2
2
|
import { EmbeddingModel as EmbeddingModel$1, ExecutionProvider } from 'fastembed';
|
|
3
3
|
|
|
4
4
|
declare const DEFAULT_FASTEMBED_EMBEDDING_MODEL = FastEmbedModel.BGESmallENV15;
|
package/dist/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["../src/index.ts"],"sourcesContent":["import type { Embedding, EmbeddingModel } from \"@anvia/core\";\nimport type { ExecutionProvider } from \"fastembed\";\nimport { EmbeddingModel as FastEmbedModel, FlagEmbedding } from \"fastembed\";\n\nexport const DEFAULT_FASTEMBED_EMBEDDING_MODEL = FastEmbedModel.BGESmallENV15;\n\nexport type FastEmbedEmbeddingModelName = `${Exclude<FastEmbedModel, FastEmbedModel.CUSTOM>}`;\n\nexport type FastEmbedRuntime = {\n embed(texts: string[], batchSize?: number): AsyncIterable<unknown>;\n};\n\nexport type FastEmbedEmbeddingModelOptions = {\n model?: FastEmbedEmbeddingModelName | undefined;\n maxBatchSize?: number | undefined;\n initOptions?:\n | {\n executionProviders?: ExecutionProvider[] | undefined;\n maxLength?: number | undefined;\n cacheDir?: string | undefined;\n showDownloadProgress?: boolean | undefined;\n modelName?: string | undefined;\n }\n | undefined;\n};\n\nexport class FastEmbedEmbeddingModel implements EmbeddingModel {\n readonly model: string;\n readonly maxBatchSize: number;\n\n constructor(\n private readonly runtime: FastEmbedRuntime,\n options: FastEmbedEmbeddingModelOptions = {},\n ) {\n this.model = options.model ?? DEFAULT_FASTEMBED_EMBEDDING_MODEL;\n this.maxBatchSize = Math.max(1, Math.trunc(options.maxBatchSize ?? 256));\n }\n\n static async create(\n options: FastEmbedEmbeddingModelOptions = {},\n ): Promise<FastEmbedEmbeddingModel> {\n const model = options.model ?? DEFAULT_FASTEMBED_EMBEDDING_MODEL;\n const runtime = await FlagEmbedding.init({\n ...(options.initOptions ?? {}),\n model,\n } as never);\n\n return new FastEmbedEmbeddingModel(runtime, { ...options, model });\n }\n\n async embedTexts(texts: string[]): Promise<Embedding[]> {\n if (texts.length === 0) {\n return [];\n }\n\n const vectors: number[][] = [];\n for await (const batch of this.runtime.embed(texts, this.maxBatchSize)) {\n vectors.push(...parseBatch(batch, vectors.length));\n }\n\n if (vectors.length !== texts.length) {\n throw new Error(\n `FastEmbed embedding model returned ${vectors.length} embeddings for ${texts.length} texts`,\n );\n }\n\n return texts.map((document, index) => ({\n document,\n vector: vectors[index] as number[],\n }));\n }\n}\n\nexport function createFastEmbedEmbeddingModel(\n options: FastEmbedEmbeddingModelOptions = {},\n): Promise<FastEmbedEmbeddingModel> {\n return FastEmbedEmbeddingModel.create(options);\n}\n\nfunction parseBatch(batch: unknown, offset: number): number[][] {\n if (!Array.isArray(batch)) {\n throw new Error(`FastEmbed embedding model returned an invalid batch at offset ${offset}`);\n }\n\n return batch.map((vector, index) => {\n const values = vectorToArray(vector);\n if (values === undefined) {\n throw new Error(\n `FastEmbed embedding model returned an invalid vector at index ${offset + index}`,\n );\n }\n return values;\n });\n}\n\nfunction vectorToArray(vector: unknown): number[] | undefined {\n if (Array.isArray(vector) && vector.every((item) => typeof item === \"number\")) {\n return vector;\n }\n\n if (ArrayBuffer.isView(vector) && !(vector instanceof DataView)) {\n return Array.from(vector as unknown as ArrayLike<number>);\n }\n\n return undefined;\n}\n"],"mappings":";AAEA,SAAS,kBAAkB,gBAAgB,qBAAqB;AAEzD,IAAM,oCAAoC,eAAe;AAsBzD,IAAM,0BAAN,MAAM,yBAAkD;AAAA,EAI7D,YACmB,SACjB,UAA0C,CAAC,GAC3C;AAFiB;AAGjB,SAAK,QAAQ,QAAQ,SAAS;AAC9B,SAAK,eAAe,KAAK,IAAI,GAAG,KAAK,MAAM,QAAQ,gBAAgB,GAAG,CAAC;AAAA,EACzE;AAAA,EALmB;AAAA,EAJV;AAAA,EACA;AAAA,EAUT,aAAa,OACX,UAA0C,CAAC,GACT;AAClC,UAAM,QAAQ,QAAQ,SAAS;AAC/B,UAAM,UAAU,MAAM,cAAc,KAAK;AAAA,MACvC,GAAI,QAAQ,eAAe,CAAC;AAAA,MAC5B;AAAA,IACF,CAAU;AAEV,WAAO,IAAI,yBAAwB,SAAS,EAAE,GAAG,SAAS,MAAM,CAAC;AAAA,EACnE;AAAA,EAEA,MAAM,WAAW,OAAuC;AACtD,QAAI,MAAM,WAAW,GAAG;AACtB,aAAO,CAAC;AAAA,IACV;AAEA,UAAM,UAAsB,CAAC;AAC7B,qBAAiB,SAAS,KAAK,QAAQ,MAAM,OAAO,KAAK,YAAY,GAAG;AACtE,cAAQ,KAAK,GAAG,WAAW,OAAO,QAAQ,MAAM,CAAC;AAAA,IACnD;AAEA,QAAI,QAAQ,WAAW,MAAM,QAAQ;AACnC,YAAM,IAAI;AAAA,QACR,sCAAsC,QAAQ,MAAM,mBAAmB,MAAM,MAAM;AAAA,MACrF;AAAA,IACF;AAEA,WAAO,MAAM,IAAI,CAAC,UAAU,WAAW;AAAA,MACrC;AAAA,MACA,QAAQ,QAAQ,KAAK;AAAA,IACvB,EAAE;AAAA,EACJ;AACF;AAEO,SAAS,8BACd,UAA0C,CAAC,GACT;AAClC,SAAO,wBAAwB,OAAO,OAAO;AAC/C;AAEA,SAAS,WAAW,OAAgB,QAA4B;AAC9D,MAAI,CAAC,MAAM,QAAQ,KAAK,GAAG;AACzB,UAAM,IAAI,MAAM,iEAAiE,MAAM,EAAE;AAAA,EAC3F;AAEA,SAAO,MAAM,IAAI,CAAC,QAAQ,UAAU;AAClC,UAAM,SAAS,cAAc,MAAM;AACnC,QAAI,WAAW,QAAW;AACxB,YAAM,IAAI;AAAA,QACR,iEAAiE,SAAS,KAAK;AAAA,MACjF;AAAA,IACF;AACA,WAAO;AAAA,EACT,CAAC;AACH;AAEA,SAAS,cAAc,QAAuC;AAC5D,MAAI,MAAM,QAAQ,MAAM,KAAK,OAAO,MAAM,CAAC,SAAS,OAAO,SAAS,QAAQ,GAAG;AAC7E,WAAO;AAAA,EACT;AAEA,MAAI,YAAY,OAAO,MAAM,KAAK,EAAE,kBAAkB,WAAW;AAC/D,WAAO,MAAM,KAAK,MAAsC;AAAA,EAC1D;AAEA,SAAO;AACT;","names":[]}
|
|
1
|
+
{"version":3,"sources":["../src/index.ts"],"sourcesContent":["import type { Embedding, EmbeddingModel } from \"@anvia/core/embeddings\";\nimport type { ExecutionProvider } from \"fastembed\";\nimport { EmbeddingModel as FastEmbedModel, FlagEmbedding } from \"fastembed\";\n\nexport const DEFAULT_FASTEMBED_EMBEDDING_MODEL = FastEmbedModel.BGESmallENV15;\n\nexport type FastEmbedEmbeddingModelName = `${Exclude<FastEmbedModel, FastEmbedModel.CUSTOM>}`;\n\nexport type FastEmbedRuntime = {\n embed(texts: string[], batchSize?: number): AsyncIterable<unknown>;\n};\n\nexport type FastEmbedEmbeddingModelOptions = {\n model?: FastEmbedEmbeddingModelName | undefined;\n maxBatchSize?: number | undefined;\n initOptions?:\n | {\n executionProviders?: ExecutionProvider[] | undefined;\n maxLength?: number | undefined;\n cacheDir?: string | undefined;\n showDownloadProgress?: boolean | undefined;\n modelName?: string | undefined;\n }\n | undefined;\n};\n\nexport class FastEmbedEmbeddingModel implements EmbeddingModel {\n readonly model: string;\n readonly maxBatchSize: number;\n\n constructor(\n private readonly runtime: FastEmbedRuntime,\n options: FastEmbedEmbeddingModelOptions = {},\n ) {\n this.model = options.model ?? DEFAULT_FASTEMBED_EMBEDDING_MODEL;\n this.maxBatchSize = Math.max(1, Math.trunc(options.maxBatchSize ?? 256));\n }\n\n static async create(\n options: FastEmbedEmbeddingModelOptions = {},\n ): Promise<FastEmbedEmbeddingModel> {\n const model = options.model ?? DEFAULT_FASTEMBED_EMBEDDING_MODEL;\n const runtime = await FlagEmbedding.init({\n ...(options.initOptions ?? {}),\n model,\n } as never);\n\n return new FastEmbedEmbeddingModel(runtime, { ...options, model });\n }\n\n async embedTexts(texts: string[]): Promise<Embedding[]> {\n if (texts.length === 0) {\n return [];\n }\n\n const vectors: number[][] = [];\n for await (const batch of this.runtime.embed(texts, this.maxBatchSize)) {\n vectors.push(...parseBatch(batch, vectors.length));\n }\n\n if (vectors.length !== texts.length) {\n throw new Error(\n `FastEmbed embedding model returned ${vectors.length} embeddings for ${texts.length} texts`,\n );\n }\n\n return texts.map((document, index) => ({\n document,\n vector: vectors[index] as number[],\n }));\n }\n}\n\nexport function createFastEmbedEmbeddingModel(\n options: FastEmbedEmbeddingModelOptions = {},\n): Promise<FastEmbedEmbeddingModel> {\n return FastEmbedEmbeddingModel.create(options);\n}\n\nfunction parseBatch(batch: unknown, offset: number): number[][] {\n if (!Array.isArray(batch)) {\n throw new Error(`FastEmbed embedding model returned an invalid batch at offset ${offset}`);\n }\n\n return batch.map((vector, index) => {\n const values = vectorToArray(vector);\n if (values === undefined) {\n throw new Error(\n `FastEmbed embedding model returned an invalid vector at index ${offset + index}`,\n );\n }\n return values;\n });\n}\n\nfunction vectorToArray(vector: unknown): number[] | undefined {\n if (Array.isArray(vector) && vector.every((item) => typeof item === \"number\")) {\n return vector;\n }\n\n if (ArrayBuffer.isView(vector) && !(vector instanceof DataView)) {\n return Array.from(vector as unknown as ArrayLike<number>);\n }\n\n return undefined;\n}\n"],"mappings":";AAEA,SAAS,kBAAkB,gBAAgB,qBAAqB;AAEzD,IAAM,oCAAoC,eAAe;AAsBzD,IAAM,0BAAN,MAAM,yBAAkD;AAAA,EAI7D,YACmB,SACjB,UAA0C,CAAC,GAC3C;AAFiB;AAGjB,SAAK,QAAQ,QAAQ,SAAS;AAC9B,SAAK,eAAe,KAAK,IAAI,GAAG,KAAK,MAAM,QAAQ,gBAAgB,GAAG,CAAC;AAAA,EACzE;AAAA,EALmB;AAAA,EAJV;AAAA,EACA;AAAA,EAUT,aAAa,OACX,UAA0C,CAAC,GACT;AAClC,UAAM,QAAQ,QAAQ,SAAS;AAC/B,UAAM,UAAU,MAAM,cAAc,KAAK;AAAA,MACvC,GAAI,QAAQ,eAAe,CAAC;AAAA,MAC5B;AAAA,IACF,CAAU;AAEV,WAAO,IAAI,yBAAwB,SAAS,EAAE,GAAG,SAAS,MAAM,CAAC;AAAA,EACnE;AAAA,EAEA,MAAM,WAAW,OAAuC;AACtD,QAAI,MAAM,WAAW,GAAG;AACtB,aAAO,CAAC;AAAA,IACV;AAEA,UAAM,UAAsB,CAAC;AAC7B,qBAAiB,SAAS,KAAK,QAAQ,MAAM,OAAO,KAAK,YAAY,GAAG;AACtE,cAAQ,KAAK,GAAG,WAAW,OAAO,QAAQ,MAAM,CAAC;AAAA,IACnD;AAEA,QAAI,QAAQ,WAAW,MAAM,QAAQ;AACnC,YAAM,IAAI;AAAA,QACR,sCAAsC,QAAQ,MAAM,mBAAmB,MAAM,MAAM;AAAA,MACrF;AAAA,IACF;AAEA,WAAO,MAAM,IAAI,CAAC,UAAU,WAAW;AAAA,MACrC;AAAA,MACA,QAAQ,QAAQ,KAAK;AAAA,IACvB,EAAE;AAAA,EACJ;AACF;AAEO,SAAS,8BACd,UAA0C,CAAC,GACT;AAClC,SAAO,wBAAwB,OAAO,OAAO;AAC/C;AAEA,SAAS,WAAW,OAAgB,QAA4B;AAC9D,MAAI,CAAC,MAAM,QAAQ,KAAK,GAAG;AACzB,UAAM,IAAI,MAAM,iEAAiE,MAAM,EAAE;AAAA,EAC3F;AAEA,SAAO,MAAM,IAAI,CAAC,QAAQ,UAAU;AAClC,UAAM,SAAS,cAAc,MAAM;AACnC,QAAI,WAAW,QAAW;AACxB,YAAM,IAAI;AAAA,QACR,iEAAiE,SAAS,KAAK;AAAA,MACjF;AAAA,IACF;AACA,WAAO;AAAA,EACT,CAAC;AACH;AAEA,SAAS,cAAc,QAAuC;AAC5D,MAAI,MAAM,QAAQ,MAAM,KAAK,OAAO,MAAM,CAAC,SAAS,OAAO,SAAS,QAAQ,GAAG;AAC7E,WAAO;AAAA,EACT;AAEA,MAAI,YAAY,OAAO,MAAM,KAAK,EAAE,kBAAkB,WAAW;AAC/D,WAAO,MAAM,KAAK,MAAsC;AAAA,EAC1D;AAEA,SAAO;AACT;","names":[]}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@anvia/fastembed",
|
|
3
|
-
"version": "0.1
|
|
3
|
+
"version": "0.2.1",
|
|
4
4
|
"description": "FastEmbed embedding model adapter for Anvia.",
|
|
5
5
|
"author": "anvia",
|
|
6
6
|
"maintainer": "Indra Zulfi",
|
|
@@ -27,7 +27,7 @@
|
|
|
27
27
|
},
|
|
28
28
|
"dependencies": {
|
|
29
29
|
"fastembed": "^2.1.0",
|
|
30
|
-
"@anvia/core": "0.
|
|
30
|
+
"@anvia/core": "0.4.1"
|
|
31
31
|
},
|
|
32
32
|
"devDependencies": {
|
|
33
33
|
"@types/node": "^24.9.1",
|