agents-core/ai.koog.agents.core.dsl.builder

Package-level declarations

Types

annotation class AIAgentBuilderDslMarker

Indicates that a class or method is part of a DSL for building AI agents.

class AIAgentEdgeBuilder<IncomingOutput, OutgoingInput, CompatibleOutput : OutgoingInput> : BaseBuilder<AIAgentEdge<IncomingOutput, OutgoingInput>>

A builder class for constructing an AIAgentEdge instance, which represents a directed edge connecting two nodes in a graph of an AI agent's processing pipeline. This edge defines the flow of data from a source node to a target node, enabling transformation or filtering of the source node's output before passing it to the target node.

AIAgentEdgeBuilderIntermediate

expect class AIAgentEdgeBuilderIntermediate<IncomingOutput, IntermediateOutput, OutgoingInput> : AIAgentEdgeBuilderIntermediateBase<IncomingOutput, IntermediateOutput, OutgoingInput>

Represents an intermediate stage in the construction of a directed edge between two nodes in an AI agent strategy graph. This class provides mechanisms to define conditions and transformations that dictate how data flows and is processed between the nodes.

actual class AIAgentEdgeBuilderIntermediate<IncomingOutput, IntermediateOutput, OutgoingInput> : AIAgentEdgeBuilderIntermediateBase<IncomingOutput, IntermediateOutput, OutgoingInput>

actual open class AIAgentEdgeBuilderIntermediate<IncomingOutput, IntermediateOutput, OutgoingInput> : AIAgentEdgeBuilderIntermediateBase<IncomingOutput, IntermediateOutput, OutgoingInput>

AIAgentEdgeBuilderIntermediateBase

abstract class AIAgentEdgeBuilderIntermediateBase<IncomingOutput, IntermediateOutput, OutgoingInput>

Base class for AIAgentEdgeBuilderIntermediate with default implementations.

AIAgentGraphStrategyBuilder

class AIAgentGraphStrategyBuilder<TInput, TOutput>(name: String, inputType: TypeToken, outputType: TypeToken, toolSelectionStrategy: ToolSelectionStrategy) : AIAgentSubgraphBuilderBase<TInput, TOutput> , BaseBuilder<AIAgentGraphStrategy<TInput, TOutput>>

A builder class responsible for constructing an instance of AIAgentGraphStrategy. The AIAgentGraphStrategyBuilder serves as a specific configuration for creating AI agent strategies with a defined start and finish node, along with a designated tool selection strategy.

AIAgentNodeDelegate

open class AIAgentNodeDelegate<Input, Output>(val name: String?, val inputType: TypeToken, val outputType: TypeToken, val execute: suspend AIAgentGraphContextBase.(Input) -> Output)

A delegate for creating and managing an instance of AIAgentNodeBase.

AIAgentParallelNodesMergeContext

class AIAgentParallelNodesMergeContext<Input, Output>(underlyingContextBase: AIAgentGraphContextBase, val results: List<ParallelResult<Input, Output>>) : AIAgentGraphContextBase

Context for merging parallel node execution results.

AIAgentSubgraphBuilder

class AIAgentSubgraphBuilder<Input, Output>(val name: String? = null, inputType: TypeToken, outputType: TypeToken, toolSelectionStrategy: ToolSelectionStrategy, llmModel: LLModel?, llmParams: LLMParams?, responseProcessor: ResponseProcessor? = null) : AIAgentSubgraphBuilderBase<Input, Output> , BaseBuilder<AIAgentSubgraphDelegate<Input, Output>>

Builder class for creating AI agent subgraphs with a defined tool selection strategy.

AIAgentSubgraphBuilderBase

abstract class AIAgentSubgraphBuilderBase<Input, Output>

Abstract base class for building AI agent subgraphs.

AIAgentSubgraphDelegate

open class AIAgentSubgraphDelegate<Input, Output>

A delegate that provides dynamic access to an instance of AIAgentSubgraph. This class facilitates constructing and associating a subgraph with specific start and finish nodes and a defined tool selection strategy upon access.

BaseBuilder

interface BaseBuilder<T>

A generic base interface for builders that enables the construction of an object of type T. This interface is typically used in DSL patterns, providing a fluent and structured way to build complex objects.

ContextualCondition

fun interface ContextualCondition<Output>

ContextualCondition is a functional interface designed to evaluate a condition based on an output value and the given AI agent graph context.

ContextualTransformation

fun interface ContextualTransformation<Output, NewOutput>

A functional interface representing a transformation operation that processes an input output of type Output in the context of AIAgentGraphContextBase and produces a transformed result of type NewOutput.

EdgeTransformationDslMarker

@DslMarker

@Target(allowedTargets = [AnnotationTarget.CLASS, AnnotationTarget.FUNCTION])

annotation class EdgeTransformationDslMarker

Marks a function as a transformation specific to edges within the AI agent's DSL to ensure its proper highlighting.

ParallelNodeExecutionResult

data class ParallelNodeExecutionResult<Output>(val output: Output, val context: AIAgentContext)

Represents the result of a parallel node execution, containing both the output value and the execution context.

ParallelResult

data class ParallelResult<Input, Output>(val nodeName: String, val nodeInput: Input, val nodeResult: ParallelNodeExecutionResult<Output>)

Represents the completed result of a parallel node execution.

SimpleCondition

fun interface SimpleCondition<Output>

Represents a simple, stateless functional interface that evaluates a condition on a given output. This is particularly tailored for Java compatibility, enabling usage in contexts where Java code interacts with Kotlin.

SimpleTransformation

fun interface SimpleTransformation<Output, NewOutput>

A functional interface representing a transformation operation from a given input type to a new output type.

Functions

forwardTo

infix fun <IncomingOutput, OutgoingInput> AIAgentNodeBase<*, IncomingOutput>.~~forwardTo~~(otherNode: AIAgentNodeBase<OutgoingInput, *>): AIAgentEdgeBuilderIntermediate<IncomingOutput, IncomingOutput, OutgoingInput>

Creates a directed edge from this AIAgentNodeBase to another AIAgentNodeBase, allowing data to flow from the output of the current node to the input of the specified node.

node

inline fun <Input, Output> node(name: String? = null, noinline execute: suspend AIAgentGraphContextBase.(input: Input) -> Output): AIAgentNodeDelegate<Input, Output>

fun <Input, Output> node(name: String? = null, inputType: TypeToken, outputType: TypeToken, execute: suspend AIAgentGraphContextBase.(input: Input) -> Output): AIAgentNodeDelegate<Input, Output>

Defines a new node in the agent's stage, representing a unit of execution that takes an input and produces an output.

parallel

fun <Input, Output> parallel(vararg nodes: AIAgentNodeBase<Input, Output>, dispatcher: CoroutineDispatcher = Dispatchers.Default, name: String? = null, merge: suspend AIAgentParallelNodesMergeContext<Input, Output>.() -> ParallelNodeExecutionResult<Output>): AIAgentNodeDelegate<Input, Output>

Creates a node that executes multiple nodes in parallel.

strategy

inline fun <Input, Output> strategy(name: String, toolSelectionStrategy: ToolSelectionStrategy = ToolSelectionStrategy.ALL, init: AIAgentGraphStrategyBuilder<Input, Output>.() -> Unit): AIAgentGraphStrategy<Input, Output>

Builds a local AI agent that processes user input through a sequence of stages.

subgraph

inline fun <Input, Output> subgraph(name: String? = null, toolSelectionStrategy: ToolSelectionStrategy = ToolSelectionStrategy.ALL, llmModel: LLModel? = null, llmParams: LLMParams? = null, responseProcessor: ResponseProcessor? = null, define: AIAgentSubgraphBuilderBase<Input, Output>.() -> Unit): AIAgentSubgraphDelegate<Input, Output>

fun <Input : Any, Output : Any> subgraph(name: String? = null, inputType: TypeToken, outputType: TypeToken, toolSelectionStrategy: ToolSelectionStrategy = ToolSelectionStrategy.ALL, llmModel: LLModel? = null, llmParams: LLMParams? = null, responseProcessor: ResponseProcessor? = null, define: AIAgentSubgraphBuilderBase<Input, Output>.() -> Unit): AIAgentSubgraphDelegate<Input, Output>

Creates a subgraph with a specified tool selection strategy.

inline fun <Input, Output> subgraph(name: String? = null, tools: List<Tool<*, *>>, llmModel: LLModel? = null, llmParams: LLMParams? = null, responseProcessor: ResponseProcessor? = null, define: AIAgentSubgraphBuilderBase<Input, Output>.() -> Unit): AIAgentSubgraphDelegate<Input, Output>

Creates a subgraph with specified tools.