Weaviate

Python

FirstBatch integrates Weaviate through the Weaviate Class.

Example run:

from firstbatch import Weaviate
import weaviate

auth_config = weaviate.AuthApiKey(api_key=os.environ["WEAVIATE_API_KEY"])
client = weaviate.Client(
    url=os.environ["WEAVIATE_URL"],
    auth_client_secret=auth_config,
)
index_name = "default"
embedding_size = int(os.environ["EMBEDDING_SIZE"])

config = Config(batch_size=20, verbose=True)
personalized = FirstBatch(api_key=os.environ["FIRSTBATCH_API_KEY"], config=config)
personalized.add_vdb("weaviate_db", Weaviate(client=client, index_name=index_name, embedding_size=embedding_size))

Distance metric can explicitly be provided. If not, default value is COSINE_SIM

class DistanceMetric(Enum):
    COSINE_SIM = "cosine_sim"
    EUCLIDEAN_DIST = "euclidean_dist"
    DOT_PRODUCT = "dot_product"

from firstbatch import DistanceMetric

Weaviate(client=client, index_name=index_name,
distance_metric=DistanceMetric.EUCLIDEAN_DIST)

TypeScript

FirstBatch integrates Weaviate through the Weaviate class.

First, we create a VectorStore instance:

import weaviate, {ApiKey} from 'weaviate-ts-client';
import {Weaviate} from 'firstbatch';

// create a Weaviate client
const client = weaviate.client({
    scheme: 'your-scheme',
    host: 'your-host',
    apiKey: new ApiKey('api-key'),
});

// instantiate the vector store
const vs = new Weaviate(client);

Then, we can use the Add a VectorDB method to register our client to the SDK:

await personalized.addVdb('my_db', vs);

The Weaviate vector store accepts optional arguments:

• className name of the Weaviate schema

• outputFields fields to output

• embeddingSize vector embedding dimension

• historyField the name of the field for filtering that avoids previously retrieved data

• distanceMetrics the distance metric