tractor/README.rst

22 KiB

tractor

A minimalist actor model built on multiprocessing and trio.

tractor is an attempt to take trionic concurrency concepts and apply them to distributed-multicore Python.

tractor lets you run and spawn Python actors: separate processes which are internally running a trio scheduler and task tree (also known as an async sandwich).

Actors communicate with each other by sending messages over channels, but the details of this in tractor is by default hidden and actors can instead easily invoke remote asynchronous functions using portals.

tractor's tenets non-comprehensively include:

Warning

tractor is in alpha-alpha and is expected to change rapidly! Expect nothing to be set in stone and your ideas about where it should go to be greatly appreciated!

Install

No PyPi release yet!

pip install git+git://github.com/tgoodlet/tractor.git

What's this? Spawning event loops in subprocesses?

Close, but not quite.

The first step to grok tractor is to get the basics of trio down. A great place to start is the trio docs and this blog post by njsmith.

tractor takes much inspiration from pulsar and execnet but attempts to be much more minimal, focus on sophistication of the lower level distributed architecture, and of course does not use asyncio, hence no event loops.

A trynamic first scene

As a first example let's spawn a couple actors and have them run their lines:

import tractor
from functools import partial

_this_module = __name__
the_line = 'Hi my name is {}'


async def hi():
    return the_line.format(tractor.current_actor().name)


async def say_hello(other_actor):
    await trio.sleep(0.4)  # wait for other actor to spawn
    async with tractor.find_actor(other_actor) as portal:
        return await portal.run(_this_module, 'hi')


async def main():
    """Main tractor entry point, the "master" process (for now
    acts as the "director").
    """
    async with tractor.open_nursery() as n:
        print("Alright... Action!")

        donny = await n.start_actor(
            'donny',
            main=partial(say_hello, 'gretchen'),
            rpc_module_paths=[_this_module],
            outlive_main=True
        )
        gretchen = await n.start_actor(
            'gretchen',
            main=partial(say_hello, 'donny'),
            rpc_module_paths=[_this_module],
        )
        print(await gretchen.result())
        print(await donny.result())
        await donny.cancel_actor()
        print("CUTTTT CUUTT CUT!!?! Donny!! You're supposed to say...")


tractor.run(main)

Here, we've spawned two actors, donny and gretchen in separate processes. Each starts up and begins executing their main task defined by an async function, say_hello(). The function instructs each actor to find their partner and say hello by calling their partner's hi() function using a something called a portal. Each actor receives a response and relays that back to the parent actor (in this case our "director").

To gain more insight as to how tractor accomplishes all this please read on!

Actor spawning and causality

tractor tries to take trio's concept of causal task lifetimes to multi-process land. Accordingly tractor's actor nursery behaves similar to the nursery in trio. That is, an ActorNursery created with tractor.open_nursery() waits on spawned sub-actors to complete (or error) in the same causal way trio waits on spawned subtasks. This includes errors from any one sub-actor causing all other actors spawned by the nursery to be cancelled.

To spawn an actor open a nursery block and use the start_actor() method:

def movie_theatre_question():
    """A question asked in a dark theatre, in a tangent
    (errr, I mean different) process.
    """
    return 'have you ever seen a portal?'


async def main():
    """The main ``tractor`` routine.
    """
    async with tractor.open_nursery() as n:
        portal = await n.start_actor(
            'frank',
            # enable the actor to run funcs from this current module
            rpc_module_paths=[__name__],
            outlive_main=True,
        )

        print(await portal.run(__name__, 'movie_theatre_question'))
        # calls the subactor a 2nd time
        print(await portal.run(__name__, 'movie_theatre_question'))

        # the async with will block here indefinitely waiting
        # for our actor "frank" to complete, but since it's an
        # "outlive_main" actor it will never end until cancelled
        await portal.cancel_actor()

Notice the portal instance returned from nursery.start_actor(), we'll get to that shortly.

Spawned actor lifetimes can be configured in one of two ways:

  • the actor terminates when its main task completes (the default if the main kwarg is provided)
  • the actor can be told to outlive_main=True and thus act like an RPC daemon where it runs indefinitely until cancelled

Had we wanted the former in our example it would have been much simpler:

def cellar_door():
    return "Dang that's beautiful"


async def main():
    """The main ``tractor`` routine.
    """
    async with tractor.open_nursery() as n:
        portal = await n.start_actor('some_linguist', main=cellar_door)

    # The ``async with`` will unblock here since the 'some_linguist'
    # actor has completed its main task ``cellar_door``.

    print(await portal.result())

Note that the main task's final result(s) (returned from the provided main function) is always accessed using Portal.result() much like you'd expect from a future.

The rpc_module_paths kwarg above is a list of module path strings that will be loaded and made accessible for execution in the remote actor through a call to Portal.run(). For now this is a simple mechanism to restrict the functionality of the remote (daemonized) actor and uses Python's module system to limit the allowed remote function namespace(s).

tractor is opinionated about the underlying threading model used for each actor. Since Python has a GIL and an actor model by definition shares no state, there is no reason to use anything other then a multiprocessing Process for execution. This makes tractor programs able leverage not only multi-core hardware but also distribute over many hardware hosts (each actor can talk to all others with ease over standard network protocols).

Eventually tractor plans to support different supervision strategies like erlang.

Transparent function calling using portals

tractor introdces the concept of a portal which is an API borrowed from trio. A portal may seems similar to the idea of a RPC future except a portal allows invoking remote async functions and generators and intermittently blocking to receive responses. This allows for fully async-native IPC between actors.

When you invoke another actor's routines using a portal it looks as though it was called locally in the current actor. So when you see a call to await portal.run() what you get back is what you'd expect to if you'd called the function directly in-process. This approach avoids the need to add any special RPC proxy objects to the library by instead just relying on the built-in (async) function calling semantics and protocols of Python.

Depending on the function type Portal.run() tries to correctly interface exactly like a local version of the remote built-in Python function type. Currently async functions, generators, and regular functions are supported. Inspiration for this API comes from the way execnet does remote function execution but without the client code (necessarily) having to worry about the underlying channels system or shipping code over the network.

This portal approach turns out to be paricularly exciting with the introduction of asynchronous generators in Python 3.6! It means that actors can compose nicely in a data processing pipeline.

Say you wanted to spawn two actors which each pulling data feeds from two different sources (and wanted this work spread across 2 cpus). You also want to aggregate these feeds, do some processing on them and then deliver the final result stream to a client (or in this case parent) actor and print the results to your screen:

import time
import trio
import tractor


# this is the first 2 actors, streamer_1 and streamer_2
async def stream_data(seed):
    for i in range(seed):
        yield i
        await trio.sleep(0)  # trigger scheduler


# this is the third actor; the aggregator
async def aggregate(seed):
    """Ensure that the two streams we receive match but only stream
    a single set of values to the parent.
    """
    async with tractor.open_nursery() as nursery:
        portals = []
        for i in range(1, 3):
            # fork point
            portal = await nursery.start_actor(
                name=f'streamer_{i}',
                rpc_module_paths=[__name__],
                outlive_main=True,  # daemonize these actors
            )

            portals.append(portal)

        q = trio.Queue(500)

        async def push_to_q(portal):
            async for value in await portal.run(
                __name__, 'stream_data', seed=seed
            ):
                # leverage trio's built-in backpressure
                await q.put(value)

            await q.put(None)
            print(f"FINISHED ITERATING {portal.channel.uid}")

        # spawn 2 trio tasks to collect streams and push to a local queue
        async with trio.open_nursery() as n:
            for portal in portals:
                n.start_soon(push_to_q, portal)

            unique_vals = set()
            async for value in q:
                if value not in unique_vals:
                    unique_vals.add(value)
                    # yield upwards to the spawning parent actor
                    yield value

                    if value is None:
                        break

                assert value in unique_vals

            print("FINISHED ITERATING in aggregator")

        await nursery.cancel()
        print("WAITING on `ActorNursery` to finish")
    print("AGGREGATOR COMPLETE!")


# this is the main actor and *arbiter*
async def main():
    # a nursery which spawns "actors"
    async with tractor.open_nursery() as nursery:

        seed = int(1e3)
        import time
        pre_start = time.time()

        portal = await nursery.start_actor(
            name='aggregator',
            # executed in the actor's "main task" immediately
            main=partial(aggregate, seed),
        )

        start = time.time()
        # the portal call returns exactly what you'd expect
        # as if the remote "main" function was called locally
        result_stream = []
        async for value in await portal.result():
            result_stream.append(value)

        print(f"STREAM TIME = {time.time() - start}")
        print(f"STREAM + SPAWN TIME = {time.time() - pre_start}")
        assert result_stream == list(range(seed)) + [None]
        return result_stream


final_stream = tractor.run(main, arbiter_addr=('127.0.0.1', 1616))

Here there's four actors running in separate processes (using all the cores on you machine). Two are streaming in stream_data(), one is aggregating values from those two in aggregate() and shipping the single stream of unique values up the parent actor (the 'MainProcess' as multiprocessing calls it) which is running main().

There has also been some discussion about adding support for reactive programming primitives and native support for asyncitertools like libs -so keep an eye out for that!

Cancellation

tractor supports trio's cancellation system verbatim:

import trio
import tractor
from itertools import repeat


async def stream_forever():
    for i in repeat("I can see these little future bubble things"):
        yield i
        await trio.sleep(0.01)


async def main():
    # stream for at most 1 second
    with trio.move_on_after(1) as cancel_scope:
        async with tractor.open_nursery() as n:
            portal = await n.start_actor(
                f'donny',
                rpc_module_paths=[__name__],
                outlive_main=True
            )
            async for letter in await portal.run(__name__, 'stream_forever'):
                print(letter)

    assert cancel_scope.cancelled_caught
    assert n.cancelled

tractor.run(main)

Cancelling a nursery block cancels all actors spawned by it.

Remote error propagation

Any task invoked in a remote actor should ship any error(s) back to the calling actor where it is raised and expected to be dealt with. This way remote actor's are never cancelled unless explicitly asked or there's a bug in tractor itself.

async def assert_err():
    assert 0

async def main():
    async with tractor.open_nursery() as n:
        real_actors = []
        for i in range(3):
            real_actors.append(await n.start_actor(
                f'actor_{i}',
                rpc_module_paths=[__name__],
                outlive_main=True
            ))

        # start one actor that will fail immediately
        await n.start_actor('extra', main=assert_err)

    # should error here with a ``RemoteActorError`` containing
    # an ``AssertionError`` and all the other actors have been cancelled

try:
    # also raises
    tractor.run(main)
except tractor.RemoteActorError:
    print("Look Maa that actor failed hard, hehhh!")

You'll notice the nursery cancellation conducts a one-cancels-all supervisory strategy exactly like trio. The plan is to add more erlang strategies in the near future by allowing nurseries to accept a Supervisor type.

Shared task state

Although tractor uses a shared-nothing architecture between processes you can of course share state within an actor. trio tasks spawned via multiple RPC calls to an actor can access global data using the per actor statespace dictionary:

statespace = {'doggy': 10}


def check_statespace():
    # Remember this runs in a new process so no changes
    # will propagate back to the parent actor
    assert tractor.current_actor().statespace == statespace


async def main():
    async with tractor.open_nursery() as n:
        await n.start_actor(
            'checker', main=check_statespace,
            statespace=statespace
        )

How do actors find each other (a poor man's service discovery)?

Though it will be built out much more in the near future, tractor currently keeps track of actors by (name: str, id: str) using a special actor called the arbiter. Currently the arbiter must exist on a host (or it will be created if one can't be found) and keeps a simple dict of actor names to sockets for discovery by other actors. Obviously this can be made more sophisticated (help me with it!) but for now it does the trick.

To find the arbiter from the current actor use the get_arbiter() function and to find an actor's socket address by name use the find_actor() function:

import tractor


async def main(service_name):

    async with tractor.get_arbiter() as portal:
        print(f"Arbiter is listening on {portal.channel}")

    async with tractor.find_actor(service_name) as sockaddr:
        print(f"my_service is found at {my_service}")


tractor.run(main, service_name)

The name value you should pass to find_actor() is the one you passed as the first argument to either tractor.run() or ActorNursery.start_actor().

Using Channel directly (undocumented)

You can use the Channel api if necessary by simply defining a chan and cid kwarg in your async function definition. tractor will treat such async functions like async generators on the calling side (for now anyway) such that you can push stream values a little more granularly if you find yielding values to be restrictive. I am purposely not documenting this feature with code because I'm not yet sure yet how it should be used correctly. If you'd like more details please feel free to ask me on the trio gitter channel.

Running actors standalone (without spawning)

You don't have to spawn any actors using open_nursery() if you just want to run a single actor that connects to an existing cluster. All the comms and arbiter registration stuff still works. This can somtimes turn out being handy when debugging mult-process apps when you need to hop into a debugger. You just need to pass the existing arbiter's socket address you'd like to connect to:

tractor.run(main, arbiter_addr=('192.168.0.10', 1616))

Enabling logging

Considering how complicated distributed software can become it helps to know what exactly it's doing (even at the lowest levels). Luckily tractor has tons of logging throughout the core. tractor isn't opinionated on how you use this information and users are expected to consume log messages in whichever way is appropriate for the system at hand. That being said, when hacking on tractor there is a prettified console formatter which you can enable to see what the heck is going on. Just put the following somewhere in your code:

from tractor.log import get_console_log
log = get_console_log('trace')

What the future holds

Stuff I'd like to see tractor do one day:

  • erlang-like supervisors
  • native support for zeromq as a channel transport
  • native gossip protocol support for service discovery and arbiter election
  • a distributed log ledger for tracking cluster behaviour
  • a slick multi-process aware debugger much like in celery but with better pdb++ support

If you're interested in tackling any of these please do shout about it on the trio gitter channel!