LlamaIndex Vector Store Integration: Actian VectorAI

A LlamaIndex vector store backend powered by Actian VectorAI DB.

Installation

pip install llama-index-vector-stores-actian-vectorai

Requirements

• Python 3.10+
• A running Actian VectorAI DB instance (default endpoint: localhost:6574)

Usage

Basic (synchronous, context manager)

from llama_index.vector_stores.actian_vectorai import ActianVectorAIVectorStore
from llama_index.core.vector_stores.types import VectorStoreQuery

with ActianVectorAIVectorStore() as vector_store:
    # Add nodes
    vector_store.add(nodes)

    # Query
    query = VectorStoreQuery(query_embedding=embedding, similarity_top_k=5)
    result = vector_store.query(query)

    # Delete by source document ID
    vector_store.delete("doc_01")

    # Delete the entire collection
    vector_store.clear()

Async (context manager)

async with ActianVectorAIVectorStore() as vector_store:
    await vector_store.async_add(nodes)

    query = VectorStoreQuery(query_embedding=embedding, similarity_top_k=5)
    result = await vector_store.aquery(query)

    await vector_store.adelete("doc_01")
    await vector_store.aclear()

Manual connection management

vector_store = ActianVectorAIVectorStore()

vector_store.connect()
vector_store.add(nodes)
result = vector_store.query(query)
vector_store.close()

Async equivalent:

await vector_store.aconnect()
await vector_store.async_add(nodes)
result = await vector_store.aquery(query)
await vector_store.aclose()

External client

Pass a pre-configured VectorAIClient (or AsyncVectorAIClient) when you need to share a connection or supply custom client configuration:

from actian_vectorai import VectorAIClient, AsyncVectorAIClient

# Sync
with VectorAIClient("localhost:6574") as client:
    vector_store = ActianVectorAIVectorStore(client=client)
    vector_store.add(nodes)
    result = vector_store.query(query)

# Async
async with AsyncVectorAIClient("localhost:6574") as async_client:
    vector_store = ActianVectorAIVectorStore(async_client=async_client)
    await vector_store.async_add(nodes)
    result = await vector_store.aquery(query)

Constructor Parameters

Custom Vector Configuration

from actian_vectorai import VectorParams, Distance

with ActianVectorAIVectorStore(
    url="localhost:6574",
    collection_name="my_collection",
    dense_vector_name="dense_vector",
    dense_vector_params=VectorParams(size=1536, distance=Distance.Cosine),
) as vector_store:
    vector_store.add(nodes)

Metadata Filtering

Metadata filters can be used with query, delete_nodes, and adelete_nodes.

Supported filter operators

Unsupported operators (ANY, ALL, TEXT_MATCH_INSENSITIVE, CONTAINS) raise NotImplementedError.

Filter conditions

AND, OR, and NOT conditions are supported via FilterCondition.

from llama_index.core.vector_stores.types import (
    FilterCondition,
    FilterOperator,
    MetadataFilter,
    MetadataFilters,
    VectorStoreQuery,
)

# AND condition
query = VectorStoreQuery(
    query_embedding=embedding,
    similarity_top_k=5,
    filters=MetadataFilters(
        filters=[
            MetadataFilter(
                key="category", operator=FilterOperator.EQ, value="ai"
            ),
            MetadataFilter(
                key="score", operator=FilterOperator.GTE, value=0.7
            ),
        ],
        condition=FilterCondition.AND,
    ),
)

# Delete nodes matching a filter
vector_store.delete_nodes(
    filters=MetadataFilters(
        filters=[
            MetadataFilter(
                key="category",
                operator=FilterOperator.IN,
                value=["ai", "energy"],
            ),
        ]
    )
)

Limitations

• Only VectorStoreQueryMode.DEFAULT (dense vector search) is supported.