AzureCosmosDBMongoDBVectorStore

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:76

Azure Cosmos DB for MongoDB vCore vector store. To use this, you should have both:

• the mongodb NPM package installed
• a connection string associated with a MongoDB VCore Cluster

You do not need to create a database or collection, it will be created automatically.

You also need an index on the collection, which is by default be created automatically using the createIndex method.

Extends

• BaseVectorStore

Constructors

Constructor

new AzureCosmosDBMongoDBVectorStore(init): AzureCosmosDBMongoDBVectorStore

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:107

Parameters

init

Partial<AzureCosmosDBMongoDBVectorStore> & object & VectorStoreBaseParams

Returns

AzureCosmosDBMongoDBVectorStore

Overrides

BaseVectorStore.constructor

Properties

storesText

storesText: boolean = true

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:77

Overrides

BaseVectorStore.storesText

---

flatMetadata

flatMetadata: boolean = true

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:78

---

dbName

dbName: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:80

---

collectionName

collectionName: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:82

---

indexedMetadataFields

indexedMetadataFields: string[]

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:84

---

mongodbClient

mongodbClient: MongoClient

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:89

The used MongoClient. If not given, a new MongoClient is created based on the MONGODB_URI env variable.

---

indexName

indexName: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:91

---

embeddingKey

embeddingKey: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:93

---

idKey

idKey: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:95

---

textKey

textKey: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:97

---

metadataKey

metadataKey: string

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:99

---

indexOptions

indexOptions: AzureCosmosDBMongoDBIndexOptions

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:101

---

embedModel

embedModel: BaseEmbedding

Defined in: packages/core/src/vector-store/index.ts:128

Inherited from

BaseVectorStore.embedModel

---

isEmbeddingQuery?

optional isEmbeddingQuery: boolean

Defined in: packages/core/src/vector-store/index.ts:130

Inherited from

BaseVectorStore.isEmbeddingQuery

Methods

client()

client(): MongoClient

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:145

Returns

MongoClient

Overrides

BaseVectorStore.client

---

ensureCollection()

ensureCollection(): Promise<Collection<Document>>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:149

Returns

Promise<Collection<Document>>

---

add()

add(nodes): Promise<string[]>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:161

Parameters

nodes

BaseNode<Metadata>[]

Returns

Promise<string[]>

Overrides

BaseVectorStore.add

---

delete()

delete(refDocId, deleteOptions?): Promise<void>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:200

Deletes all nodes from the collection that belong to the given document.

Parameters

refDocId

string

Reference document ID - all nodes with this ref_doc_id will be deleted.

deleteOptions?

object

Additional delete options.

Returns

Promise<void>

A promise that resolves when the documents have been removed.

Overrides

BaseVectorStore.delete

---

query()

query(query, options): Promise<VectorStoreQueryResult>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:210

Parameters

query

VectorStoreQuery

options

AzureCosmosDBMongoDBQueryOptions

Returns

Promise<VectorStoreQueryResult>

Overrides

BaseVectorStore.query

---

createIndex()

createIndex(dimensions, indexType, similarity): Promise<void>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:289

Creates an index on the collection with the specified index name during instance construction.

Setting the numLists parameter correctly is important for achieving good accuracy and performance. Since the vector store uses IVF as the indexing strategy, you should create the index only after you have loaded a large enough sample documents to ensure that the centroids for the respective buckets are faily distributed.

As for the compression, the following options are available:

• “half” - half precision compression for HNSW and IVF indexes
• “pq” - product quantization compression for DiskANN indexes More information on the compression options can be found in the: https://learn.microsoft.com/en-us/azure/cosmos-db/mongodb/vcore/product-quantization

Parameters

dimensions

Number of dimensions for vector similarity. The maximum number of supported dimensions is 2000. If no number is provided, it will be determined automatically by embedding a short text.

number | undefined

indexType

Index Type for Mongo vCore index.

"hnsw" | "ivf" | "diskann"

similarity

AzureCosmosDBMongoDBSimilarityType = AzureCosmosDBMongoDBSimilarityType.COS

Similarity metric to use with the IVF index. Possible options are:

• CosmosDBSimilarityType.COS (cosine distance)
• CosmosDBSimilarityType.L2 (Euclidean distance)
• CosmosDBSimilarityType.IP (inner product)

Returns

Promise<void>

A promise that resolves when the index has been created.

---

checkIndexExists()

checkIndexExists(): Promise<boolean>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:354

Checks if the specified index name during instance construction exists on the collection.

Returns

Promise<boolean>

A promise that resolves to a boolean indicating if the index exists.

---

deleteIndex()

deleteIndex(indexName): Promise<void>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:364

Deletes the index specified during instance construction if it exists.

Parameters

indexName

string

Returns

Promise<void>

A promise that resolves when the index has been deleted.

---

exists()

exists(refDocId): Promise<boolean>

Defined in: packages/providers/storage/azure/src/vectorStore/AzureCosmosDBMongoVectorStore.ts:373

Check if any nodes exist for the given document reference ID. Used for de-duplication during ingestion.

Parameters

refDocId

string

The reference document ID to check

Returns

Promise<boolean>

true if any nodes with this ref_doc_id exist

Overrides

BaseVectorStore.exists