Anvia

@anvia/weaviate ​

@anvia/weaviate stores precomputed Anvia embeddings in a Weaviate collection and exposes vector search through the shared SDK interface.

Install ​

sh
pnpm add @anvia/weaviate @anvia/core @anvia/openai weaviate-client

The ESM package includes weaviate-client and should use a version compatible with its declared @anvia/core dependency range.

Store and search documents ​

ts
import { retrieveDocuments } from "@anvia/core/vector-store";
import { embedDocuments } from '@anvia/core/embeddings';
import { WeaviateVectorClient } from '@anvia/weaviate';
import { OpenAIClient } from '@anvia/openai';
const openai = new OpenAIClient({
    apiKey: process.env.OPENAI_API_KEY!,
});
const embeddings = openai.embeddingModel({
    modelId: 'text-embedding-3-small'
});
const sourceDocuments = [
    {
        id: 'password-reset',
        text: 'Password reset links expire after 30 minutes.',
    },
];
const { documents } = await embedDocuments({
    model: embeddings,
    documents: sourceDocuments,
    id: (document) => document.id,
    content: (document) => document.text
});
const storeClient = new WeaviateVectorClient({});
const store = storeClient.vectorStore({
    collectionName: 'SupportDocs',
    dimensions: 1536
});
await store.ensure();
await store.upsert({
    documents: documents
});
const results = await retrieveDocuments({
    store: store,
    model: embeddings,
    query: 'How do I reset a password?',
    topK: 5
});

Without an injected client, the HTTP host comes from WEAVIATE_HOST or localhost and the gRPC host from WEAVIATE_GRPC_HOST or localhost; ports come from WEAVIATE_HTTP_PORT and WEAVIATE_GRPC_PORT, defaulting to 8080 and 50051 and validated as integers from 1 to 65535. HTTP and gRPC connections are insecure by default. Inject a configured client for any remote or protected deployment.

Collection ownership ​

ensure() creates a missing collection with no vectorizer, the selected distance, and Anvia's document ID/document properties. The default distance is cosine.

Because Anvia supplies vectors, keep the collection vectorizer disabled. Production infrastructure should own collection creation, replication, vector-index tuning, and property schema; call validate() afterward.

Metadata keys beginning with __anvia_ are reserved. If metadata properties need an explicit Weaviate schema or indexes, provision them before ingestion.

Production patterns ​

  • Inject a client configured for TLS, authentication, endpoint discovery, and lifecycle.
  • Keep the collection distance aligned with the embedding model.
  • Define metadata properties and indexes through schema migrations or deployment automation.
  • Control ingestion batch size around provider and network limits.
  • Keep authorization in the application even when retrieval uses filters.

Reference ​

Built for Anvia.