Astra Vector Store

The Astra Vector Store node integrates DataStax Astra DB, a serverless vector database, into your AI workflow. It allows for upsert operations of embedded data and performs similarity or MMR (Maximal Marginal Relevance) search upon query.

Node Details

• Name: Astra

• Type: Astra

• Version: 2.0

• Category: Vector Stores

Base Classes

• Astra

• VectorStoreRetriever

• BaseRetriever

Parameters

Input Parameters

1. Document (optional, list)

   • Type: Document

   • Description: List of documents to be stored in the vector database.

2. Embeddings

   • Type: Embeddings

   • Description: The embedding model used to convert documents into vector representations.

3. Namespace

   • Type: string

   • Description: The namespace in Astra DB where the data will be stored.

4. Collection

   • Type: string

   • Description: The collection name in Astra DB where the data will be stored.

5. Vector Dimension (optional)

   • Type: number

   • Default: 1536

   • Description: The dimension of the vector embeddings.

6. Similarity Metric (optional)

   • Type: string

   • Options: cosine, euclidean, dot_product

   • Default: cosine

   • Description: The metric used to calculate similarity between vectors.

7. Top K (optional)

   • Type: number

   • Default: 4

   • Description: Number of top results to fetch during retrieval.

8. MMR Parameters (optional)

   • Additional parameters for Maximal Marginal Relevance search.

Credential Parameter

• Connect Credential

  • Type: credential

  • Credential Name: AstraDBApi

  • Description: Authentication credentials for connecting to Astra DB.

Outputs

1. Astra Retriever

   • Type: VectorStoreRetriever

   • Description: A retriever object for querying the Astra vector store.

2. Astra Vector Store

   • Type: AstraDBVectorStore

   • Description: The vector store object for direct interactions with Astra DB.

Functionality

1. Upsert Operation:

   • Allows inserting or updating documents in the Astra DB vector store.

   • Converts documents to vector embeddings before storage.

2. Initialization:

   • Sets up the connection to Astra DB using provided credentials and parameters.

   • Creates or connects to the specified namespace and collection.

3. Retrieval:

   • Supports similarity search and MMR search on the stored vector embeddings.

Use Cases

• Semantic search applications

• Content recommendation systems

• Document similarity analysis

• Any AI application requiring efficient storage and retrieval of vector embeddings

Note

This node requires proper setup of Astra DB and valid credentials for successful operation. Ensure that the vector dimensions and similarity metrics are consistent with your embedding model and use case requirements.