Constructor Detail

• AgentBuilder

  AgentBuilder(String name, String provider, Semver version, String description, List<PromptContributor> promptContributors)

Method Detail

• getActions

   final List<Action> getActions()

• getPromptContributors

   List<PromptContributor> getPromptContributors()

• getName

   final String getName()

• getProvider

   final String getProvider()

• getVersion

   final Semver getVersion()

• getDescription

   final String getDescription()

• action

   final Unit action(Function1<AgentBuilder, Action> block)

  Create an action

• registerPromptContributors

   final Unit registerPromptContributors(PromptContributor promptContributors)

  Require prompt contributors at agent level

• localAgentAction

   final <I extends Any, O extends Any> Unit localAgentAction(Agent agent)

  Add an action is an anonymous agent. This is valuable because the agent will be private and will not pollute the global action space.

• referencedAgentAction

   final <I extends Any, O extends Any> Unit referencedAgentAction(String agentName)

  Add an action that references the agent with a given name

• add

   final Unit add(AgentScope agentScope)

• flow

   final Unit flow(Function1<AgentBuilder, TypedAgentScopeBuilder<?>> block)

  Add a pattern builder, such as an aggregation flow, to the agent. Must be used if you have multiple actions

• transformation

   final <I extends Any, O extends Any> Unit transformation(String name, String description, List<Condition> preConditions, List<String> pre, List<Condition> post, Boolean canRerun, String inputVarName, String outputVarName, Function1<WorldState, Double> cost, Function1<WorldState, Double> value, Set<ToolGroupRequirement> toolGroups, ActionQos qos, Set<String> referencedInputProperties, Transformation<I, O> block)

  Add an action that is a transformation NOT using an LLM.

• promptedTransformer

   final <I extends Any, O extends Any> Unit promptedTransformer(String name, String description, List<Condition> pre, List<Condition> post, String inputVarName, String outputVarName, Function1<WorldState, Double> cost, Set<ToolGroupRequirement> toolGroups, ActionQos qos, Set<String> referencedInputProperties, LlmOptions llm, List<PromptContributor> promptContributors, Boolean canRerun, Collection<Tool> tools, Function1<TransformationActionContext<I, O>, String> prompt)

  Add an action that is a transformation using an LLM.

• goal

   final Unit goal(String name, String description, KClass<?> satisfiedBy, Set<KClass<?>> requires, Set<IoBinding> inputs, List<Condition> pre, Function1<WorldState, Double> value, Export export)

  Add a goal to the agent.

  Parameters:

  name - The name of the goal.

  description - A description of the goal.

  satisfiedBy - A class that satisfies this goal.

  requires - A set of classes that are required to satisfy this goal.

  pre - custom preconditions, in addition to input preconditions

  value - the value of achieving this goal

• condition

   final Unit condition(Function1<AgentBuilder, Condition> block)

• conditionOf

   final AgentBuilder.ConditionDelegateProvider conditionOf(String name, Double cost, Function1<OperationContext, Boolean> block)

• promptedCondition

   final AgentBuilder.ConditionDelegateProvider promptedCondition(String name, Function1<OperationContext, String> prompt, LlmCall llm)

  Declare a condition determined with an LLM. Assign it a name to ensure type safe access val myCondition = condition("custom prompt")

• build

   final Agent build()

  Build the agent