Agent Bricks

These dataclasses are used in the SDK to represent API requests and responses for services in the databricks.sdk.service.agentbricks module.

class databricks.sdk.service.agentbricks.CustomLlm(name: 'str', instructions: 'str', agent_artifact_path: 'Optional[str]' = None, creation_time: 'Optional[str]' = None, creator: 'Optional[str]' = None, datasets: 'Optional[List[Dataset]]' = None, endpoint_name: 'Optional[str]' = None, guidelines: 'Optional[List[str]]' = None, id: 'Optional[str]' = None, optimization_state: 'Optional[State]' = None)

name: str

Name of the custom LLM

instructions: str

Instructions for the custom LLM to follow

agent_artifact_path: str | None = None

creation_time: str | None = None

Creation timestamp of the custom LLM

creator: str | None = None

Creator of the custom LLM

datasets: List[Dataset] | None = None

Datasets used for training and evaluating the model, not for inference

endpoint_name: str | None = None

Name of the endpoint that will be used to serve the custom LLM

guidelines: List[str] | None = None

Guidelines for the custom LLM to adhere to

id: str | None = None

optimization_state: State | None = None

If optimization is kicked off, tracks the state of the custom LLM

as_dict() → dict

Serializes the CustomLlm into a dictionary suitable for use as a JSON request body.

as_shallow_dict() → dict

Serializes the CustomLlm into a shallow dictionary of its immediate attributes.

classmethod from_dict(d: Dict[str, Any]) → CustomLlm

Deserializes the CustomLlm from a dictionary.

class databricks.sdk.service.agentbricks.Dataset(table: 'Table')

table: Table

as_dict() → dict

Serializes the Dataset into a dictionary suitable for use as a JSON request body.

as_shallow_dict() → dict

Serializes the Dataset into a shallow dictionary of its immediate attributes.

classmethod from_dict(d: Dict[str, Any]) → Dataset

Deserializes the Dataset from a dictionary.

class databricks.sdk.service.agentbricks.State

States of Custom LLM optimization lifecycle.

CANCELLED = "CANCELLED"

COMPLETED = "COMPLETED"

CREATED = "CREATED"

FAILED = "FAILED"

PENDING = "PENDING"

RUNNING = "RUNNING"

class databricks.sdk.service.agentbricks.Table(table_path: 'str', request_col: 'str', response_col: 'Optional[str]' = None)

table_path: str

Full UC table path in catalog.schema.table_name format

request_col: str

Name of the request column

response_col: str | None = None

Optional: Name of the response column if the data is labeled

as_dict() → dict

Serializes the Table into a dictionary suitable for use as a JSON request body.

as_shallow_dict() → dict

Serializes the Table into a shallow dictionary of its immediate attributes.

classmethod from_dict(d: Dict[str, Any]) → Table

Deserializes the Table from a dictionary.