agents-core/ai.koog.agents.core.agent.entity/SubgraphWithTaskBuilder

SubgraphWithTaskBuilder

class SubgraphWithTaskBuilder<Input : Any, Output : Any>(name: String?, toolSelectionStrategy: ToolSelectionStrategy = ToolSelectionStrategy.ALL, llmModel: LLModel? = null, llmParams: LLMParams? = null, responseProcessor: ResponseProcessor? = null, inputClass: ERROR CLASS: Symbol not found for Class<Input>, outputOption: OutputOption<Output>, defineTask: ContextualAction<Input, String>, runMode: ToolCalls = ToolCalls.SEQUENTIAL, assistantResponseRepeatMax: Int? = null) : TypedAIAgentSubgraphBuilderBase<Input, Output, SubgraphWithTaskBuilder<Input, Output>> (source)

A builder class for creating an AI agent subgraph that incorporates task definition as part of its configuration. This builder allows customizing the construction of a subgraph while defining how tasks are specified and executed within the subgraph.

The class is designed for Java interoperability and simplifies the process of building subgraphs with task-specific logic, including specifying input/output types, tool selection

Constructors

constructor(name: String?, toolSelectionStrategy: ToolSelectionStrategy = ToolSelectionStrategy.ALL, llmModel: LLModel? = null, llmParams: LLMParams? = null, responseProcessor: ResponseProcessor? = null, inputClass: ERROR CLASS: Symbol not found for Class<Input>, outputOption: OutputOption<Output>, defineTask: ContextualAction<Input, String>, runMode: ToolCalls = ToolCalls.SEQUENTIAL, assistantResponseRepeatMax: Int? = null)

Functions

assistantResponseRepeatMax

fun assistantResponseRepeatMax(assistantResponseRepeatMax: Int): SubgraphWithTaskBuilder<Input, Output>

Sets the maximum number of times the assistant's response can be repeated.

build

fun build(): AIAgentSubgraph<Input, Output>

Builds and returns an instance of AIAgentSubgraph configured with the specified parameters.

limitedTools

fun limitedTools(vararg toolSets: ERROR CLASS: Symbol not found for ToolSet): SubgraphWithTaskBuilder<Input, Output>

Configures the builder with a selection of tools defined by the provided tool sets. The tools will be extracted from each ToolSet and applied to the builder's tool selection strategy.

fun limitedTools(tools: List<Tool<*, *>>): SubgraphWithTaskBuilder<Input, Output>

Configures the builder to use a specific list of tools for the AI agent's subgraph.

runMode

fun runMode(runMode: ToolCalls): SubgraphWithTaskBuilder<Input, Output>

Configures the run mode for

usingLLM

fun usingLLM(llmModel: LLModel): SubgraphWithTaskBuilder<Input, Output>

Sets the specified Large Language Model (LLM) for the agent subgraph builder.

withInput

fun <Input : Any> withInput(outputClass: ERROR CLASS: Symbol not found for Class<Input>): AIAgentSubgraphBuilderWithInput<Input, *>

Configures the builder with the specified input type and returns a new instance of AIAgentSubgraphBuilderWithInput, allowing further configuration for the specified input type.

withLLMParams

fun withLLMParams(llmParams: LLMParams): SubgraphWithTaskBuilder<Input, Output>

Sets the parameters for the Language Learning Model (LLM) in the current builder.

withResponseProcessor

fun withResponseProcessor(responseProcessor: ResponseProcessor): SubgraphWithTaskBuilder<Input, Output>

Sets the specified response processor to handle and modify LLM responses.

withToolSelectionStrategy

fun withToolSelectionStrategy(strategy: ToolSelectionStrategy): SubgraphWithTaskBuilder<Input, Output>

Sets the tool selection strategy for the subgraph builder and returns the updated builder instance.