ray/python/ray/actor.py

import copy
import inspect
import logging
import six
import weakref

from abc import ABCMeta, abstractmethod
from collections import namedtuple

from ray.function_manager import FunctionDescriptor
import ray.ray_constants as ray_constants
import ray._raylet
import ray.signature as signature
import ray.worker
from ray import ActorID, ActorClassID

logger = logging.getLogger(__name__)


def method(*args, **kwargs):
    """Annotate an actor method.

    .. code-block:: python

        @ray.remote
        class Foo:
            @ray.method(num_return_vals=2)
            def bar(self):
                return 1, 2

        f = Foo.remote()

        _, _ = f.bar.remote()

    Args:
        num_return_vals: The number of object IDs that should be returned by
            invocations of this actor method.
    """
    assert len(args) == 0
    assert len(kwargs) == 1
    assert "num_return_vals" in kwargs
    num_return_vals = kwargs["num_return_vals"]

    def annotate_method(method):
        method.__ray_num_return_vals__ = num_return_vals
        return method

    return annotate_method


# Create objects to wrap method invocations. This is done so that we can
# invoke methods with actor.method.remote() instead of actor.method().
class ActorMethod:
    """A class used to invoke an actor method.

    Note: This class only keeps a weak ref to the actor, unless it has been
    passed to a remote function. This avoids delays in GC of the actor.

    Attributes:
        _actor: A handle to the actor.
        _method_name: The name of the actor method.
        _num_return_vals: The default number of return values that the method
            invocation should return.
        _decorator: An optional decorator that should be applied to the actor
            method invocation (as opposed to the actor method execution) before
            invoking the method. The decorator must return a function that
            takes in two arguments ("args" and "kwargs"). In most cases, it
            should call the function that was passed into the decorator and
            return the resulting ObjectIDs. For an example, see
            "test_decorated_method" in "python/ray/tests/test_actor.py".
    """

    def __init__(self,
                 actor,
                 method_name,
                 num_return_vals,
                 decorator=None,
                 hardref=False):
        self._actor_ref = weakref.ref(actor)
        self._method_name = method_name
        self._num_return_vals = num_return_vals
        # This is a decorator that is used to wrap the function invocation (as
        # opposed to the function execution). The decorator must return a
        # function that takes in two arguments ("args" and "kwargs"). In most
        # cases, it should call the function that was passed into the decorator
        # and return the resulting ObjectIDs.
        self._decorator = decorator

        # Acquire a hard ref to the actor, this is useful mainly when passing
        # actor method handles to remote functions.
        if hardref:
            self._actor_hard_ref = actor
        else:
            self._actor_hard_ref = None

    def __call__(self, *args, **kwargs):
        raise Exception("Actor methods cannot be called directly. Instead "
                        "of running 'object.{}()', try "
                        "'object.{}.remote()'.".format(self._method_name,
                                                       self._method_name))

    def remote(self, *args, **kwargs):
        return self._remote(args, kwargs)

    def _remote(self, args=None, kwargs=None, num_return_vals=None):
        if num_return_vals is None:
            num_return_vals = self._num_return_vals

        def invocation(args, kwargs):
            actor = self._actor_hard_ref or self._actor_ref()
            if actor is None:
                raise RuntimeError("Lost reference to actor")
            return actor._actor_method_call(
                self._method_name,
                args=args,
                kwargs=kwargs,
                num_return_vals=num_return_vals)

        # Apply the decorator if there is one.
        if self._decorator is not None:
            invocation = self._decorator(invocation)

        return invocation(args, kwargs)

    def __getstate__(self):
        return {
            "actor": self._actor_ref(),
            "method_name": self._method_name,
            "num_return_vals": self._num_return_vals,
            "decorator": self._decorator,
        }

    def __setstate__(self, state):
        self.__init__(
            state["actor"],
            state["method_name"],
            state["num_return_vals"],
            state["decorator"],
            hardref=True)


class ActorClassMetadata:
    """Metadata for an actor class.

    Attributes:
        modified_class: The original class that was decorated (with some
            additional methods added like __ray_terminate__).
        class_id: The ID of this actor class.
        class_name: The name of this class.
        num_cpus: The default number of CPUs required by the actor creation
            task.
        num_gpus: The default number of GPUs required by the actor creation
            task.
        memory: The heap memory quota for this actor.
        object_store_memory: The object store memory quota for this actor.
        resources: The default resources required by the actor creation task.
        actor_method_cpus: The number of CPUs required by actor method tasks.
        last_export_session_and_job: A pair of the last exported session
            and job to help us to know whether this function was exported.
            This is an imperfect mechanism used to determine if we need to
            export the remote function again. It is imperfect in the sense that
            the actor class definition could be exported multiple times by
            different workers.
        actor_methods: The actor methods.
        method_decorators: Optional decorators that should be applied to the
            method invocation function before invoking the actor methods. These
            can be set by attaching the attribute
            "__ray_invocation_decorator__" to the actor method.
        method_signatures: The signatures of the methods.
        actor_method_names: The names of the actor methods.
        actor_method_num_return_vals: The default number of return values for
            each actor method.
    """

    def __init__(self, modified_class, class_id, max_reconstructions, num_cpus,
                 num_gpus, memory, object_store_memory, resources):
        self.modified_class = modified_class
        self.class_id = class_id
        self.class_name = modified_class.__name__
        self.max_reconstructions = max_reconstructions
        self.num_cpus = num_cpus
        self.num_gpus = num_gpus
        self.memory = memory
        self.object_store_memory = object_store_memory
        self.resources = resources
        self.last_export_session_and_job = None

        self.actor_methods = inspect.getmembers(
            self.modified_class, ray.utils.is_function_or_method)
        self.actor_method_names = [
            method_name for method_name, _ in self.actor_methods
        ]

        constructor_name = "__init__"
        if constructor_name not in self.actor_method_names:
            # Add __init__ if it does not exist.
            # Actor creation will be executed with __init__ together.

            # Assign an __init__ function will avoid many checks later on.
            def __init__(self):
                pass

            self.modified_class.__init__ = __init__
            self.actor_method_names.append(constructor_name)
            self.actor_methods.append((constructor_name, __init__))

        # Extract the signatures of each of the methods. This will be used
        # to catch some errors if the methods are called with inappropriate
        # arguments.
        self.method_decorators = {}
        self.method_signatures = {}
        self.actor_method_num_return_vals = {}
        for method_name, method in self.actor_methods:
            # Whether or not this method requires binding of its first
            # argument. For class and static methods, we do not want to bind
            # the first argument, but we do for instance methods
            is_bound = (ray.utils.is_class_method(method)
                        or ray.utils.is_static_method(self.modified_class,
                                                      method_name))

            # Print a warning message if the method signature is not
            # supported. We don't raise an exception because if the actor
            # inherits from a class that has a method whose signature we
            # don't support, there may not be much the user can do about it.
            self.method_signatures[method_name] = signature.extract_signature(
                method, ignore_first=not is_bound)
            # Set the default number of return values for this method.
            if hasattr(method, "__ray_num_return_vals__"):
                self.actor_method_num_return_vals[method_name] = (
                    method.__ray_num_return_vals__)
            else:
                self.actor_method_num_return_vals[method_name] = (
                    ray_constants.DEFAULT_ACTOR_METHOD_NUM_RETURN_VALS)

            if hasattr(method, "__ray_invocation_decorator__"):
                self.method_decorators[method_name] = (
                    method.__ray_invocation_decorator__)


class ActorClass:
    """An actor class.

    This is a decorated class. It can be used to create actors.

    Attributes:
        __ray_metadata__: Contains metadata for the actor.
    """

    def __init__(cls, name, bases, attr):
        """Prevents users from directly inheriting from an ActorClass.

        This will be called when a class is defined with an ActorClass object
        as one of its base classes. To intentionally construct an ActorClass,
        use the '_ray_from_modified_class' classmethod.

        Raises:
            TypeError: Always.
        """
        for base in bases:
            if isinstance(base, ActorClass):
                raise TypeError("Attempted to define subclass '{}' of actor "
                                "class '{}'. Inheriting from actor classes is "
                                "not currently supported. You can instead "
                                "inherit from a non-actor base class and make "
                                "the derived class an actor class (with "
                                "@ray.remote).".format(
                                    name, base.__ray_metadata__.class_name))

        # This shouldn't be reached because one of the base classes must be
        # an actor class if this was meant to be subclassed.
        assert False, ("ActorClass.__init__ should not be called. Please use "
                       "the @ray.remote decorator instead.")

    def __call__(self, *args, **kwargs):
        """Prevents users from directly instantiating an ActorClass.

        This will be called instead of __init__ when 'ActorClass()' is executed
        because an is an object rather than a metaobject. To properly
        instantiated a remote actor, use 'ActorClass.remote()'.

        Raises:
            Exception: Always.
        """
        raise Exception("Actors cannot be instantiated directly. "
                        "Instead of '{}()', use '{}.remote()'.".format(
                            self.__ray_metadata__.class_name,
                            self.__ray_metadata__.class_name))

    @classmethod
    def _ray_from_modified_class(cls, modified_class, class_id,
                                 max_reconstructions, num_cpus, num_gpus,
                                 memory, object_store_memory, resources):
        for attribute in ["remote", "_remote", "_ray_from_modified_class"]:
            if hasattr(modified_class, attribute):
                logger.warning("Creating an actor from class {} overwrites "
                               "attribute {} of that class".format(
                                   modified_class.__name__, attribute))

        # Make sure the actor class we are constructing inherits from the
        # original class so it retains all class properties.
        class DerivedActorClass(cls, modified_class):
            pass

        name = "ActorClass({})".format(modified_class.__name__)
        DerivedActorClass.__module__ = modified_class.__module__
        DerivedActorClass.__name__ = name
        DerivedActorClass.__qualname__ = name
        # Construct the base object.
        self = DerivedActorClass.__new__(DerivedActorClass)

        self.__ray_metadata__ = ActorClassMetadata(
            modified_class, class_id, max_reconstructions, num_cpus, num_gpus,
            memory, object_store_memory, resources)

        return self

    def remote(self, *args, **kwargs):
        """Create an actor.

        Args:
            args: These arguments are forwarded directly to the actor
                constructor.
            kwargs: These arguments are forwarded directly to the actor
                constructor.

        Returns:
            A handle to the newly created actor.
        """
        return self._remote(args=args, kwargs=kwargs)

    def options(self, **options):
        """Convenience method for creating an actor with options.

        Same arguments as Actor._remote(), but returns a wrapped actor class
        that a non-underscore .remote() can be called on.

        Examples:
            # The following two calls are equivalent.
            >>> Actor._remote(num_cpus=4, max_concurrency=8, args=[x, y])
            >>> Actor.options(num_cpus=4, max_concurrency=8).remote(x, y)
        """

        actor_cls = self

        class ActorOptionWrapper:
            def remote(self, *args, **kwargs):
                return actor_cls._remote(args=args, kwargs=kwargs, **options)

        return ActorOptionWrapper()

    def _remote(self,
                args=None,
                kwargs=None,
                num_cpus=None,
                num_gpus=None,
                memory=None,
                object_store_memory=None,
                resources=None,
                is_direct_call=None,
                max_concurrency=None,
                name=None,
                detached=False,
                is_asyncio=False):
        """Create an actor.

        This method allows more flexibility than the remote method because
        resource requirements can be specified and override the defaults in the
        decorator.

        Args:
            args: The arguments to forward to the actor constructor.
            kwargs: The keyword arguments to forward to the actor constructor.
            num_cpus: The number of CPUs required by the actor creation task.
            num_gpus: The number of GPUs required by the actor creation task.
            memory: Restrict the heap memory usage of this actor.
            object_store_memory: Restrict the object store memory used by
                this actor when creating objects.
            resources: The custom resources required by the actor creation
                task.
            is_direct_call: Use direct actor calls.
            max_concurrency: The max number of concurrent calls to allow for
                this actor. This only works with direct actor calls. The max
                concurrency defaults to 1 for threaded execution, and 1000 for
                asyncio execution. Note that the execution order is not
                guaranteed when max_concurrency > 1.
            name: The globally unique name for the actor.
            detached: Whether the actor should be kept alive after driver
                exits.
            is_asyncio: Turn on async actor calls. This only works with direct
                actor calls.

        Returns:
            A handle to the newly created actor.
        """
        if args is None:
            args = []
        if kwargs is None:
            kwargs = {}
        if is_direct_call is None:
            is_direct_call = ray_constants.direct_call_enabled()
        if max_concurrency is None:
            if is_asyncio:
                max_concurrency = 1000
            else:
                max_concurrency = 1

        if max_concurrency > 1 and not is_direct_call:
            raise ValueError(
                "setting max_concurrency requires is_direct_call=True")
        if max_concurrency < 1:
            raise ValueError("max_concurrency must be >= 1")

        if is_asyncio and not is_direct_call:
            raise ValueError(
                "Setting is_asyncio requires is_direct_call=True.")

        worker = ray.worker.get_global_worker()
        if worker.mode is None:
            raise Exception("Actors cannot be created before ray.init() "
                            "has been called.")

        meta = self.__ray_metadata__

        if detached and name is None:
            raise Exception("Detached actors must be named. "
                            "Please use Actor._remote(name='some_name') "
                            "to associate the name.")

        # Check whether the name is already taken.
        if name is not None:
            try:
                ray.experimental.get_actor(name)
            except ValueError:  # name is not taken, expected.
                pass
            else:
                raise ValueError(
                    "The name {name} is already taken. Please use "
                    "a different name or get existing actor using "
                    "ray.experimental.get_actor('{name}')".format(name=name))

        # Set the actor's default resources if not already set. First three
        # conditions are to check that no resources were specified in the
        # decorator. Last three conditions are to check that no resources were
        # specified when _remote() was called.
        if (meta.num_cpus is None and meta.num_gpus is None
                and meta.resources is None and num_cpus is None
                and num_gpus is None and resources is None):
            # In the default case, actors acquire no resources for
            # their lifetime, and actor methods will require 1 CPU.
            cpus_to_use = ray_constants.DEFAULT_ACTOR_CREATION_CPU_SIMPLE
            actor_method_cpu = ray_constants.DEFAULT_ACTOR_METHOD_CPU_SIMPLE
        else:
            # If any resources are specified (here or in decorator), then
            # all resources are acquired for the actor's lifetime and no
            # resources are associated with methods.
            cpus_to_use = (ray_constants.DEFAULT_ACTOR_CREATION_CPU_SPECIFIED
                           if meta.num_cpus is None else meta.num_cpus)
            actor_method_cpu = ray_constants.DEFAULT_ACTOR_METHOD_CPU_SPECIFIED

        function_name = "__init__"
        function_descriptor = FunctionDescriptor(
            meta.modified_class.__module__, function_name,
            meta.modified_class.__name__)

        # Do not export the actor class or the actor if run in LOCAL_MODE
        # Instead, instantiate the actor locally and add it to the worker's
        # dictionary
        if worker.mode == ray.LOCAL_MODE:
            actor_id = ActorID.from_random()
            worker.actors[actor_id] = meta.modified_class(
                *copy.deepcopy(args), **copy.deepcopy(kwargs))
        else:
            # Export the actor.
            if (meta.last_export_session_and_job !=
                    worker.current_session_and_job):
                # If this actor class was not exported in this session and job,
                # we need to export this function again, because current GCS
                # doesn't have it.
                meta.last_export_session_and_job = (
                    worker.current_session_and_job)
                worker.function_actor_manager.export_actor_class(
                    meta.modified_class, meta.actor_method_names)

            resources = ray.utils.resources_from_resource_arguments(
                cpus_to_use, meta.num_gpus, meta.memory,
                meta.object_store_memory, meta.resources, num_cpus, num_gpus,
                memory, object_store_memory, resources)

            # If the actor methods require CPU resources, then set the required
            # placement resources. If actor_placement_resources is empty, then
            # the required placement resources will be the same as resources.
            actor_placement_resources = {}
            assert actor_method_cpu in [0, 1]
            if actor_method_cpu == 1:
                actor_placement_resources = resources.copy()
                actor_placement_resources["CPU"] += 1
            function_signature = meta.method_signatures[function_name]
            creation_args = signature.flatten_args(function_signature, args,
                                                   kwargs)
            actor_id = worker.core_worker.create_actor(
                function_descriptor.get_function_descriptor_list(),
                creation_args, meta.max_reconstructions, resources,
                actor_placement_resources, is_direct_call, max_concurrency,
                detached, is_asyncio)

        actor_handle = ActorHandle(
            actor_id,
            meta.modified_class.__module__,
            meta.class_name,
            meta.actor_method_names,
            meta.method_decorators,
            meta.method_signatures,
            meta.actor_method_num_return_vals,
            actor_method_cpu,
            worker.current_session_and_job,
            original_handle=True)

        if name is not None:
            ray.experimental.register_actor(name, actor_handle)

        return actor_handle


class ActorHandle:
    """A handle to an actor.

    The fields in this class are prefixed with _ray_ to hide them from the user
    and to avoid collision with actor method names.

    An ActorHandle can be created in three ways. First, by calling .remote() on
    an ActorClass. Second, by passing an actor handle into a task (forking the
    ActorHandle). Third, by directly serializing the ActorHandle (e.g., with
    cloudpickle).

    Attributes:
        _ray_actor_id: Actor ID.
        _ray_module_name: The module name of this actor.
        _ray_actor_method_names: The names of the actor methods.
        _ray_method_decorators: Optional decorators for the function
            invocation. This can be used to change the behavior on the
            invocation side, whereas a regular decorator can be used to change
            the behavior on the execution side.
        _ray_method_signatures: The signatures of the actor methods.
        _ray_method_num_return_vals: The default number of return values for
            each method.
        _ray_class_name: The name of the actor class.
        _ray_actor_method_cpus: The number of CPUs required by actor methods.
        _ray_original_handle: True if this is the original actor handle for a
            given actor. If this is true, then the actor will be destroyed when
            this handle goes out of scope.
    """

    def __init__(self,
                 actor_id,
                 module_name,
                 class_name,
                 actor_method_names,
                 method_decorators,
                 method_signatures,
                 method_num_return_vals,
                 actor_method_cpus,
                 session_and_job,
                 original_handle=False):
        self._ray_actor_id = actor_id
        self._ray_module_name = module_name
        self._ray_original_handle = original_handle
        self._ray_actor_method_names = actor_method_names
        self._ray_method_decorators = method_decorators
        self._ray_method_signatures = method_signatures
        self._ray_method_num_return_vals = method_num_return_vals
        self._ray_class_name = class_name
        self._ray_actor_method_cpus = actor_method_cpus
        self._ray_session_and_job = session_and_job
        self._ray_function_descriptor_lists = {
            method_name: FunctionDescriptor(
                self._ray_module_name, method_name,
                self._ray_class_name).get_function_descriptor_list()
            for method_name in self._ray_method_signatures.keys()
        }

        for method_name in actor_method_names:
            method = ActorMethod(
                self,
                method_name,
                self._ray_method_num_return_vals[method_name],
                decorator=self._ray_method_decorators.get(method_name))
            setattr(self, method_name, method)

    def _actor_method_call(self,
                           method_name,
                           args=None,
                           kwargs=None,
                           num_return_vals=None):
        """Method execution stub for an actor handle.

        This is the function that executes when
        `actor.method_name.remote(*args, **kwargs)` is called. Instead of
        executing locally, the method is packaged as a task and scheduled
        to the remote actor instance.

        Args:
            method_name: The name of the actor method to execute.
            args: A list of arguments for the actor method.
            kwargs: A dictionary of keyword arguments for the actor method.
            num_return_vals (int): The number of return values for the method.

        Returns:
            object_ids: A list of object IDs returned by the remote actor
                method.
        """
        worker = ray.worker.get_global_worker()

        args = args or []
        kwargs = kwargs or {}
        function_signature = self._ray_method_signatures[method_name]

        if not args and not kwargs and not function_signature:
            list_args = []
        else:
            list_args = signature.flatten_args(function_signature, args,
                                               kwargs)
        if worker.mode == ray.LOCAL_MODE:
            function = getattr(worker.actors[self._actor_id], method_name)
            object_ids = worker.local_mode_manager.execute(
                function, method_name, args, kwargs, num_return_vals)
        else:
            object_ids = worker.core_worker.submit_actor_task(
                self._ray_actor_id,
                self._ray_function_descriptor_lists[method_name], list_args,
                num_return_vals, self._ray_actor_method_cpus)

        if len(object_ids) == 1:
            object_ids = object_ids[0]
        elif len(object_ids) == 0:
            object_ids = None

        return object_ids

    # Make tab completion work.
    def __dir__(self):
        return self._ray_actor_method_names

    def __repr__(self):
        return "Actor({}, {})".format(self._ray_class_name,
                                      self._actor_id.hex())

    def __del__(self):
        """Terminate the worker that is running this actor."""
        # TODO(swang): Also clean up forked actor handles.
        # Kill the worker if this is the original actor handle, created
        # with Class.remote(). TODO(rkn): Even without passing handles around,
        # this is not the right policy. the actor should be alive as long as
        # there are ANY handles in scope in the process that created the actor,
        # not just the first one.
        worker = ray.worker.get_global_worker()
        exported_in_current_session_and_job = (
            self._ray_session_and_job == worker.current_session_and_job)
        if (worker.mode == ray.worker.SCRIPT_MODE
                and not exported_in_current_session_and_job):
            # If the worker is a driver and driver id has changed because
            # Ray was shut down re-initialized, the actor is already cleaned up
            # and we don't need to send `__ray_terminate__` again.
            logger.warning(
                "Actor is garbage collected in the wrong driver." +
                " Actor id = %s, class name = %s.", self._ray_actor_id,
                self._ray_class_name)
            return
        if worker.connected and self._ray_original_handle:
            # Note: in py2 the weakref is destroyed prior to calling __del__
            # so we need to set the hardref here briefly
            try:
                self.__ray_terminate__._actor_hard_ref = self
                self.__ray_terminate__.remote()
            finally:
                self.__ray_terminate__._actor_hard_ref = None

    def __ray_kill__(self):
        """Kill the actor that this actor handle refers to immediately.

        This will cause any outstanding tasks submitted to the actor to fail
        and the actor to exit in the same way as if it crashed. In general,
        you should prefer to just delete the actor handle and let it clean up
        gracefull.

        Returns:
            None.
        """
        worker = ray.worker.get_global_worker()
        worker.core_worker.kill_actor(self._ray_actor_id)

    @property
    def _actor_id(self):
        return self._ray_actor_id

    def _serialization_helper(self, ray_forking):
        """This is defined in order to make pickling work.

        Args:
            ray_forking: True if this is being called because Ray is forking
                the actor handle and false if it is being called by pickling.

        Returns:
            A dictionary of the information needed to reconstruct the object.
        """
        worker = ray.worker.get_global_worker()
        worker.check_connected()
        state = {
            # Local mode just uses the actor ID.
            "core_handle": worker.core_worker.serialize_actor_handle(
                self._ray_actor_id)
            if hasattr(worker, "core_worker") else self._ray_actor_id,
            "module_name": self._ray_module_name,
            "class_name": self._ray_class_name,
            "actor_method_names": self._ray_actor_method_names,
            "method_decorators": self._ray_method_decorators,
            "method_signatures": self._ray_method_signatures,
            "method_num_return_vals": self._ray_method_num_return_vals,
            "actor_method_cpus": self._ray_actor_method_cpus
        }

        return state

    def _deserialization_helper(self, state, ray_forking):
        """This is defined in order to make pickling work.

        Args:
            state: The serialized state of the actor handle.
            ray_forking: True if this is being called because Ray is forking
                the actor handle and false if it is being called by pickling.
        """
        worker = ray.worker.get_global_worker()
        worker.check_connected()

        self.__init__(
            # TODO(swang): Accessing the worker's current task ID is not
            # thread-safe.
            # Local mode just uses the actor ID.
            worker.core_worker.deserialize_and_register_actor_handle(
                state["core_handle"])
            if hasattr(worker, "core_worker") else state["core_handle"],
            state["module_name"],
            state["class_name"],
            state["actor_method_names"],
            state["method_decorators"],
            state["method_signatures"],
            state["method_num_return_vals"],
            state["actor_method_cpus"],
            worker.current_session_and_job)

    def __getstate__(self):
        """This code path is used by pickling but not by Ray forking."""
        return self._serialization_helper(False)

    def __setstate__(self, state):
        """This code path is used by pickling but not by Ray forking."""
        return self._deserialization_helper(state, False)


def make_actor(cls, num_cpus, num_gpus, memory, object_store_memory, resources,
               max_reconstructions):
    # Give an error if cls is an old-style class.
    if not issubclass(cls, object):
        raise TypeError(
            "The @ray.remote decorator cannot be applied to old-style "
            "classes. In Python 2, you must declare the class with "
            "'class ClassName(object):' instead of 'class ClassName:'.")

    if issubclass(cls, Checkpointable) and inspect.isabstract(cls):
        raise TypeError(
            "A checkpointable actor class should implement all abstract "
            "methods in the `Checkpointable` interface.")

    if max_reconstructions is None:
        max_reconstructions = 0

    if not (ray_constants.NO_RECONSTRUCTION <= max_reconstructions <=
            ray_constants.INFINITE_RECONSTRUCTION):
        raise Exception("max_reconstructions must be in range [%d, %d]." %
                        (ray_constants.NO_RECONSTRUCTION,
                         ray_constants.INFINITE_RECONSTRUCTION))

    # Modify the class to have an additional method that will be used for
    # terminating the worker.
    class Class(cls):
        def __ray_terminate__(self):
            worker = ray.worker.get_global_worker()
            if worker.mode != ray.LOCAL_MODE:
                ray.actor.exit_actor()

        def __ray_checkpoint__(self):
            """Save a checkpoint.

            This task saves the current state of the actor, the current task
            frontier according to the raylet, and the checkpoint index
            (number of tasks executed so far).
            """
            worker = ray.worker.global_worker
            if not isinstance(self, ray.actor.Checkpointable):
                raise Exception(
                    "__ray_checkpoint__.remote() may only be called on actors "
                    "that implement ray.actor.Checkpointable")
            return worker._save_actor_checkpoint()

    Class.__module__ = cls.__module__
    Class.__name__ = cls.__name__

    return ActorClass._ray_from_modified_class(
        Class, ActorClassID.from_random(), max_reconstructions, num_cpus,
        num_gpus, memory, object_store_memory, resources)


def exit_actor():
    """Intentionally exit the current actor.

    This function is used to disconnect an actor and exit the worker.

    Raises:
        Exception: An exception is raised if this is a driver or this
            worker is not an actor.
    """
    worker = ray.worker.global_worker
    if worker.mode == ray.WORKER_MODE and not worker.actor_id.is_nil():
        # Intentionally disconnect the core worker from the raylet so the
        # raylet won't push an error message to the driver.
        ray.disconnect()
        # Disconnect global state from GCS.
        ray.state.state.disconnect()
        # Set a flag to indicate this is an intentional actor exit. This
        # reduces log verbosity.
        exit = SystemExit(0)
        exit.is_ray_terminate = True
        raise exit
        assert False, "This process should have terminated."
    else:
        raise Exception("exit_actor called on a non-actor worker.")


ray.worker.global_worker.make_actor = make_actor

CheckpointContext = namedtuple(
    "CheckpointContext",
    [
        # Actor's ID.
        "actor_id",
        # Number of tasks executed since last checkpoint.
        "num_tasks_since_last_checkpoint",
        # Time elapsed since last checkpoint, in milliseconds.
        "time_elapsed_ms_since_last_checkpoint",
    ],
)
"""A namedtuple that contains information about actor's last checkpoint."""

Checkpoint = namedtuple(
    "Checkpoint",
    [
        # ID of this checkpoint.
        "checkpoint_id",
        # The timestamp at which this checkpoint was saved,
        # represented as milliseconds elapsed since Unix epoch.
        "timestamp",
    ],
)
"""A namedtuple that represents a checkpoint."""


class Checkpointable(six.with_metaclass(ABCMeta, object)):
    """An interface that indicates an actor can be checkpointed."""

    @abstractmethod
    def should_checkpoint(self, checkpoint_context):
        """Whether this actor needs to be checkpointed.

        This method will be called after every task. You should implement this
        callback to decide whether this actor needs to be checkpointed at this
        time, based on the checkpoint context, or any other factors.

        Args:
            checkpoint_context: A namedtuple that contains info about last
                checkpoint.

        Returns:
            A boolean value that indicates whether this actor needs to be
            checkpointed.
        """
        pass

    @abstractmethod
    def save_checkpoint(self, actor_id, checkpoint_id):
        """Save a checkpoint to persistent storage.

        If `should_checkpoint` returns true, this method will be called. You
        should implement this callback to save actor's checkpoint and the given
        checkpoint id to persistent storage.

        Args:
            actor_id: Actor's ID.
            checkpoint_id: ID of this checkpoint. You should save it together
                with actor's checkpoint data. And it will be used by the
                `load_checkpoint` method.
        Returns:
            None.
        """
        pass

    @abstractmethod
    def load_checkpoint(self, actor_id, available_checkpoints):
        """Load actor's previous checkpoint, and restore actor's state.

        This method will be called when an actor is reconstructed, after
        actor's constructor.
        If the actor needs to restore from previous checkpoint, this function
        should restore actor's state and return the checkpoint ID. Otherwise,
        it should do nothing and return None.
        Note, this method must return one of the checkpoint IDs in the
        `available_checkpoints` list, or None. Otherwise, an exception will be
        raised.

        Args:
            actor_id: Actor's ID.
            available_checkpoints: A list of `Checkpoint` namedtuples that
                contains all available checkpoint IDs and their timestamps,
                sorted by timestamp in descending order.
        Returns:
            The ID of the checkpoint from which the actor was resumed, or None
            if the actor should restart from the beginning.
        """
        pass

    @abstractmethod
    def checkpoint_expired(self, actor_id, checkpoint_id):
        """Delete an expired checkpoint.

        This method will be called when an checkpoint is expired. You should
        implement this method to delete your application checkpoint data.
        Note, the maximum number of checkpoints kept in the backend can be
        configured at `RayConfig.num_actor_checkpoints_to_keep`.

        Args:
            actor_id: ID of the actor.
            checkpoint_id: ID of the checkpoint that has expired.
        Returns:
            None.
        """
        pass


def get_checkpoints_for_actor(actor_id):
    """Get the available checkpoints for the given actor ID, return a list
    sorted by checkpoint timestamp in descending order.
    """
    checkpoint_info = ray.state.state.actor_checkpoint_info(actor_id)
    if checkpoint_info is None:
        return []
    checkpoints = [
        Checkpoint(checkpoint_id, timestamp) for checkpoint_id, timestamp in
        zip(checkpoint_info["CheckpointIds"], checkpoint_info["Timestamps"])
    ]
    return sorted(
        checkpoints,
        key=lambda checkpoint: checkpoint.timestamp,
        reverse=True,
    )