KernelFunctionSelectionStrategy Class

Determines agent selection based on the evaluation of a Kernel Function.

Note: This class is marked as 'experimental' and may change in the future.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Constructor

KernelFunctionSelectionStrategy(*, has_selected: bool = False, initial_agent: Agent | None = None, agent_variable_name: str | None = '_agent_', history_variable_name: str | None = '_history_', arguments: KernelArguments | None = None, function: KernelFunction, kernel: Kernel, result_parser: Callable[[...], str] = None, history_reducer: ChatHistoryReducer | None = None)

Keyword-Only Parameters

Name	Description
has_selected Required
initial_agent Required
agent_variable_name	Default value: _agent_
history_variable_name	Default value: _history_
arguments Required
function Required
kernel Required
result_parser Required
history_reducer Required

Methods

select_agent

Select the next agent to interact with.

select_agent

Select the next agent to interact with.

async select_agent(agents: list[Agent], history: list[ChatMessageContent]) -> Agent

Parameters

Name	Description
agents Required	The list of agents to select from.
history Required	The history of messages in the conversation.

Returns

Type	Description
	The next agent to interact with.

Exceptions

Type	Description
AgentExecutionException	If the strategy fails to execute the function or select the next agent

Attributes

DEFAULT_AGENT_VARIABLE_NAME

DEFAULT_AGENT_VARIABLE_NAME: ClassVar[str] = '_agent_'

DEFAULT_HISTORY_VARIABLE_NAME

DEFAULT_HISTORY_VARIABLE_NAME: ClassVar[str] = '_history_'

agent_variable_name

agent_variable_name: str | None

arguments

arguments: KernelArguments | None

function

function: KernelFunction

history_reducer

history_reducer: ChatHistoryReducer | None

history_variable_name

history_variable_name: str | None

is_experimental

is_experimental = True

kernel

kernel: Kernel

result_parser

result_parser: Callable[[...], str]

stage_status

stage_status = 'experimental'