Class LoopPlanner
java.lang.Object
dev.langchain4j.agentic.workflow.impl.LoopPlanner
- All Implemented Interfaces:
Planner
-
Constructor Summary
ConstructorsConstructorDescriptionLoopPlanner(int maxIterations, boolean testExitAtLoopEnd, BiPredicate<AgenticScope, Integer> exitCondition, String exitConditionDescription) -
Method Summary
Modifier and TypeMethodDescription<T extends AgentInstance>
Tas(Class<T> agentInstanceClass, AgentInstance agentInstance) Returns the planner's current execution state as a map of serializable values.firstAction(PlanningContext planningContext) Returns the first action to execute when the planner starts (or resumes).voidinit(InitPlanningContext initPlanningContext) intnextAction(PlanningContext planningContext) Determines the next action to execute based on the result of the previous agent invocation.voidrestoreExecutionState(Map<String, Object> state) Restores the planner's execution state from a previously saved map.booleantopology()Returns the topology of the agentic system managed by this planner.
-
Constructor Details
-
LoopPlanner
public LoopPlanner(int maxIterations, boolean testExitAtLoopEnd, BiPredicate<AgenticScope, Integer> exitCondition, String exitConditionDescription)
-
-
Method Details
-
init
-
firstAction
Description copied from interface:PlannerReturns the first action to execute when the planner starts (or resumes). Defaults to delegating toPlanner.nextAction(PlanningContext).- Specified by:
firstActionin interfacePlanner- Parameters:
planningContext- the current planning context- Returns:
- the first action to execute
-
nextAction
Description copied from interface:PlannerDetermines the next action to execute based on the result of the previous agent invocation.- Specified by:
nextActionin interfacePlanner- Parameters:
planningContext- the current planning context, including the previous agent's result- Returns:
- the next action to execute
-
topology
Description copied from interface:PlannerReturns the topology of the agentic system managed by this planner.- Specified by:
topologyin interfacePlanner- Returns:
- the topology (defaults to
AgenticSystemTopology.SEQUENCE)
-
as
-
maxIterations
public int maxIterations() -
testExitAtLoopEnd
public boolean testExitAtLoopEnd() -
exitCondition
-
executionState
Description copied from interface:PlannerReturns the planner's current execution state as a map of serializable values. This state is persisted to theAgenticScopeafter each agent invocation, enabling the planner to resume from the correct position after a crash.The returned state must be such that, when passed to
Planner.restoreExecutionState(Map)andPlanner.firstAction(PlanningContext)is called, the planner produces the correct resume action.Stateless planners (e.g., parallel, conditional) can use the default empty implementation.
- Specified by:
executionStatein interfacePlanner- Returns:
- a map of state entries to persist, or an empty map if no state needs saving
-
restoreExecutionState
Description copied from interface:PlannerRestores the planner's execution state from a previously saved map. Called by the execution loop beforePlanner.firstAction(PlanningContext)when recovering from a persisted scope.- Specified by:
restoreExecutionStatein interfacePlanner- Parameters:
state- the previously saved execution state
-