Astra DB

DataStax Astra DB is a serverless vector-capable database built on Cassandra and made conveniently available through an easy-to-use JSON API.

AstraDBStore and AstraDBByteStore need the astrapy package to be installed:

%pip install --upgrade --quiet  astrapy

The Store takes the following parameters:

• api_endpoint: Astra DB API endpoint. Looks like https://01234567-89ab-cdef-0123-456789abcdef-us-east1.apps.astra.datastax.com
• token: Astra DB token. Looks like AstraCS:6gBhNmsk135....
• collection_name : Astra DB collection name
• namespace: (Optional) Astra DB namespace

AstraDBStore

The AstraDBStore is an implementation of BaseStore that stores everything in your DataStax Astra DB instance. The store keys must be strings and will be mapped to the _id field of the Astra DB document. The store values can be any object that can be serialized by json.dumps. In the database, entries will have the form:

{
  "_id": "<key>",
  "value": <value>
}

from langchain_community.storage import AstraDBStore

from getpass import getpass

ASTRA_DB_API_ENDPOINT = input("ASTRA_DB_API_ENDPOINT = ")
ASTRA_DB_APPLICATION_TOKEN = getpass("ASTRA_DB_APPLICATION_TOKEN = ")

store = AstraDBStore(
    api_endpoint=ASTRA_DB_API_ENDPOINT,
    token=ASTRA_DB_APPLICATION_TOKEN,
    collection_name="my_store",
)

store.mset([("k1", "v1"), ("k2", [0.1, 0.2, 0.3])])
print(store.mget(["k1", "k2"]))

['v1', [0.1, 0.2, 0.3]]

Usage with CacheBackedEmbeddings

You may use the AstraDBStore in conjunction with a CacheBackedEmbeddings to cache the result of embeddings computations. Note that AstraDBStore stores the embeddings as a list of floats without converting them first to bytes so we don’t use fromByteStore there.

from langchain.embeddings import CacheBackedEmbeddings, OpenAIEmbeddings

embeddings = CacheBackedEmbeddings(
    underlying_embeddings=OpenAIEmbeddings(), document_embedding_store=store
)

AstraDBByteStore

The AstraDBByteStore is an implementation of ByteStore that stores everything in your DataStax Astra DB instance. The store keys must be strings and will be mapped to the _id field of the Astra DB document. The store bytes values are converted to base64 strings for storage into Astra DB. In the database, entries will have the form:

{
  "_id": "<key>",
  "value": "bytes encoded in base 64"
}

from langchain_community.storage import AstraDBByteStore

from getpass import getpass

ASTRA_DB_API_ENDPOINT = input("ASTRA_DB_API_ENDPOINT = ")
ASTRA_DB_APPLICATION_TOKEN = getpass("ASTRA_DB_APPLICATION_TOKEN = ")

store = AstraDBByteStore(
    api_endpoint=ASTRA_DB_API_ENDPOINT,
    token=ASTRA_DB_APPLICATION_TOKEN,
    collection_name="my_store",
)

store.mset([("k1", b"v1"), ("k2", b"v2")])
print(store.mget(["k1", "k2"]))

[b'v1', b'v2']