Class: planner_action

GOAP action class implementing both action base and action planner. Captures execution as action and plans its internal logic with separate logics graph.

Examples: combat planner, anomaly planner, items looting planner, alife planner, state planner.

Source

C++ class planner_action : action_planner,action_base

Custom Constructor

planner_action

Extends

• action_planner

Constructors

Constructor

new planner_action(): planner_action;

Default constructor.

Returns

planner_action

Overrides

action_planner.constructor

Constructor

new planner_action(object?: Nillable<game_object>): planner_action;

Parameters

Parameter	Type	Description
object?	Nillable<game_object>	Target game object to work with, null is OK since correct object will be injected on setup.

Returns

planner_action

Overrides

action_planner.constructor

Constructor

new planner_action(object: Nillable<game_object>, name: string): planner_action;

Parameters

Parameter	Type	Description
object	Nillable<game_object>	Target game object to work with, null is OK since correct object will be injected on setup.
name	string	Name of the action, used for debug purposes mainly.

Returns

planner_action

Overrides

action_planner.constructor

Methods

action()

action(id: number): action_base;

Get action instance by unique id.

Parameters

Parameter	Type	Description
id	number	Unique identifier of the action to get.

Returns

action_base

Registered action instance.

Inherited from

action_planner.action

---

actual()

actual(): boolean;

Returns

boolean

Whether state of current planner is actual.

Inherited from

action_planner.actual

---

add_action()

add_action(id: number, action: action_base): void;

Add generic action by id for planner execution.

Parameters

Parameter	Type	Description
id	number	Unique identifier of new action.
action	action_base	Action implementation containing preconditions, logics, effects and other meta infos.

Returns

void

Remarks

Do not add actions from evaluator/action callbacks running inside update; the engine asserts on planner graph changes during solving.

Inherited from

action_planner.add_action

---

add_effect()

add_effect(property: world_property): void;

Add action effect. Describes what target world state is expected to be if action is completed.

Parameters

Parameter	Type	Description
property	world_property	World state property describing a pair of evaluator ID and value.

Returns

void

Remarks

Do not mutate effects while the owner planner is updating; the engine asserts on graph changes during solving.

---

add_evaluator()

add_evaluator(id: number, evaluator: property_evaluator): void;

Add evaluator instance for current action planner.

Parameters

Parameter	Type	Description
id	number	Unique identifier of the evaluator.
evaluator	property_evaluator	Instance of evaluator linked to the id.

Returns

void

Remarks

Do not add evaluators from evaluator/action callbacks running inside update; the engine asserts on planner graph changes during solving.

Inherited from

action_planner.add_evaluator

---

add_precondition()

add_precondition(property: world_property): void;

Add action execution precondition. When building logics graph, action will be considered blocked by some preconditions.

Parameters

Parameter	Type	Description
property	world_property	World state property describing a pair of evaluator ID and value.

Returns

void

Remarks

Do not mutate preconditions while the owner planner is updating; the engine asserts on graph changes during solving.

---

clear()

clear(): void;

Clear state of current action planner.

Returns

void

Inherited from

action_planner.clear

---

current_action()

current_action(): action_base;

Get currently active action being executed.

Returns

action_base

Current action instance reference.

Remarks

Requires the planner to be initialized, usually after a successful update.

Inherited from

action_planner.current_action

---

current_action_id()

current_action_id(): number;

Get currently active action identifier.

Returns

number

Unique identifier of current action.

Throws

If the planner is not initialized yet.

Inherited from

action_planner.current_action_id

---

evaluator()

evaluator(id: number): property_evaluator;

Get evaluator instance by id.

Parameters

Parameter	Type	Description
id	number	Unique identifier of the evaluator to get.

Returns

property_evaluator

Registered property evaluator.

Inherited from

action_planner.evaluator

---

execute()

execute(): void;

Lifecycle method. Execution tick of the action, called from object logics update cycle when current action is active.

Returns

void

Remarks

Called by the owning outer planner after initialize. Direct calls bypass planner state.

---

finalize()

finalize(): void;

Lifecycle method called once on action execution stop. Means that action is finished / preconditions are not met anymore.

Returns

void

Remarks

Called by the owning outer planner. Direct calls are only useful when driving the planner action manually in tests or tools.

---

initialize()

initialize(): void;

Lifecycle method called once on action execution start. Means that lifecycle of the action begun.

Returns

void

Remarks

Called by the owning outer planner. Direct calls are only useful when driving the planner action manually in tests or tools.

---

initialized()

initialized(): boolean;

Returns

boolean

Whether object action planner is already initialized.

Inherited from

action_planner.initialized

---

remove_action()

remove_action(id: number): void;

Remove action by unique id.

Parameters

Parameter	Type	Description
id	number	Unique identifier of the action to remove.

Returns

void

Remarks

Do not remove actions from evaluator/action callbacks running inside update; the engine asserts on planner graph changes during solving.

Inherited from

action_planner.remove_action

---

remove_effect()

remove_effect(id: number): void;

Remove action effect. Action will be not considered as property changing for id anymore.

Parameters

Parameter	Type	Description
id	number	World state property id.

Returns

void

Remarks

Do not mutate effects while the owner planner is updating; the engine asserts on graph changes during solving.

---

remove_evaluator()

remove_evaluator(id: number): void;

Remove evaluator instance from current action planner.

Parameters

Parameter	Type	Description
id	number	Unique identifier of the evaluator for removal.

Returns

void

Remarks

Do not remove evaluators from evaluator/action callbacks running inside update; the engine asserts on planner graph changes during solving.

Inherited from

action_planner.remove_evaluator

---

remove_precondition()

remove_precondition(id: number): void;

Remove precondition for action. When building logics graph, action will not be considered blocked by evaluator id states.

Parameters

Parameter	Type	Description
id	number	World state property id.

Returns

void

Remarks

Do not mutate preconditions while the owner planner is updating; the engine asserts on graph changes during solving.

---

set_goal_world_state()

set_goal_world_state(state: world_state): void;

Set target world state to try to reach with all the graph logics. All graphs will be built from current state to goal world state with the shortest possible path.

Parameters

Parameter	Type	Description
state	world_state	Target world state to reach with planner.

Returns

void

Inherited from

action_planner.set_goal_world_state

---

set_weight()

set_weight(weight: number): void;

Set weight of planner action switch.

Parameters

Parameter	Type	Description
weight	number	Weight value to express how prioritized action is.

Returns

void

---

setup()

setup(object: game_object): void;

Setup planner for game object.

Parameters

Parameter	Type	Description
object	game_object	Client game object to setup planner for.

Returns

void

Throws

If object is missing.

Inherited from

action_planner.setup

---

show()

show(prefix: string): void;

Debug method. With mixed / debug build allows investigation of evaluators and actions matches. Helps to debug custom actions and actions pre-conditions with state printing in log files.

Note: Available only in mixed / debug engine builds, not for direct usage from lua.

Parameters

Parameter	Type	Description
prefix	string	String prefix to display current action state in logs.

Returns

void

Inherited from

action_planner.show

---

update()

update(): void;

Lifecycle method to handle generic game loop updates.

Returns

void

Throws

If the planner cannot build a non-empty action sequence to the target world state.

Inherited from

action_planner.update

---

weight()

weight(first: world_state, second: world_state): number;

Comparator to check weight of switching one state to another.

Parameters

Parameter	Type	Description
first	world_state	First state.
second	world_state	Second state.

Returns

number

Transition weight.

Properties

__name

readonly __name: string;

LuaBind instance constructor name.

Inherited from

action_planner.__name

---

object

readonly object: game_object;

Game object that is handled by current planner instance.

Inherited from

action_planner.object

---

storage

readonly storage: property_storage;

Container reference with state of planner preconditions.

Inherited from

action_planner.storage

---

__name

readonly static __name: string;

LuaBind class constructor name.

Inherited from

action_planner.__name