aio-beanstalk is a asyncio client for beanstalk
Status: Low level protocol is fully implemented but not well tested. Higher-level interface is in prototype stage.
Basically it looks like:
.. code-block:: python
import asyncio from aiobeanstalk.proto import Client
@asyncio.coroutine def main(): client = yield from Client.connect('localhost', 11300) yield from client.send_command('use', 'mytesttube') res = yield from client.send_command('put', 100, # priority 0, # delay 10, # TTR = time to run body=b'task body') if isinstance(res, Exception): raise res print("Job queued with id", res.job_id) client.close()
if name == 'main': asyncio.get_event_loop().run_until_complete(main())
The low-level interface is intentionally has no methods for each command, and does not raise exceptions. Higher-level interface fill-in the gaps, and provides usual consumer-producer abstractions (At low level both the process which sends the tasks and worker that processes them share same protocol. It's how beanstalk is designed).
Client methods:
connect(host, port)
A classmethod coroutine that returns Client
object that is connected
to beanstalkd server on the specified host and port.
send_command(cmd, *args, body=None)
A coroutine that sends command to the beanstalkd server and waits for the
reply. You must not put length of the body in args
as it will be
added automatially.
Reply is returned as object from ``aiobeanstalk.packets`` or
``aiobeanstalk.exceptions``. Exceptions are returned rather than raised,
so you must always check result. But, ``EOFError`` may be *raised* in the
case connection is closed before receiving a reply.
It's *safe* to call it from many coroutines simultaneously, requests will
be pipelined. However, it's probably useless to call any command
simultaneously with ``reserve``, because the latter blocks on server.
close()
Closes socket.
TBD