DynamoDB for Agent

Kern supports using DynamoDB as a storage backend for Agents using the DynamoDb class.

Usage

You need to provide aws_access_key_id and aws_secret_access_key parameters to the DynamoDb class.

1from kern.db.dynamo import DynamoDb
2
3# AWS Credentials
4AWS_ACCESS_KEY_ID = getenv("AWS_ACCESS_KEY_ID")
5AWS_SECRET_ACCESS_KEY = getenv("AWS_SECRET_ACCESS_KEY")
6
7db = DynamoDb(
8    region_name="us-east-1",
9    # aws_access_key_id: AWS access key id
10    aws_access_key_id=AWS_ACCESS_KEY_ID,
11    # aws_secret_access_key: AWS secret access key
12    aws_secret_access_key=AWS_SECRET_ACCESS_KEY,
13)
14
15# Add storage to the Agent
16agent = Agent(db=db)

Params

Parameter	Type	Default	Description
`id`	`Optional[str]`	-	The ID of the database instance. UUID by default.
`db_client`	`None`	-	The DynamoDB client to use.
`region_name`	`optional[str]`	-	AWS region name.
`aws_access_key_id`	`optional[str]`	-	AWS access key ID.
`aws_secret_access_key`	`optional[str]`	-	AWS secret access key.
`session_table`	`optional[str]`	-	The name of the session table.
`memory_table`	`optional[str]`	-	The name of the memory table.
`metrics_table`	`optional[str]`	-	The name of the metrics table.
`eval_table`	`optional[str]`	-	The name of the eval table.
`knowledge_table`	`optional[str]`	-	The name of the knowledge table.
`traces_table`	`optional[str]`	-	The name of the traces table.
`spans_table`	`optional[str]`	-	The name of the spans table.