atq

atq is a pure-Python asynchronous task queue built to work with asyncio.
It is designed to run costly functions outside main event loop using
distributed workers.

atq requires Python 3.5+ and is distributed under BSD license.

Usage

First, you should start workers on the servers you plan to use for task execution

… code-block ::

atqserver --host --port --worker

where

• <hostname> is a hostname of the server
• <port-number> is a port that server will listen on
• <num-workers> is a number of worker processes

Please note that code of tasks should be accessible by atq, so it’s advised
to run atq server from your project root directory in more complex
situations.

Then you will need to create a client using hostnames and ports of initialized
servers:

… code-block:: python

import atq

q = atq.Q([
(“localhost”, 12345),
])

Finally you can use atq in your code:

… code-block:: python

import atq
import asyncio
import requests
from collections import ChainMap
from collections import Counter

URLS = [
    ...
]

q = atq.Q([
    ("localhost", 12345),
])

def top_words(url, n):
    """Returns top n words from text specified by url."""
    text = requests.get(url).text.split()
    return {url: Counter(text).most_common(n)}

async def get_top_words(urls, n):
    """Returns top n words in documents specified by URLs."""
    tops_in_url = await asyncio.gather(
        *[q.q(top_words, url, n) for url in urls])
    return ChainMap(*tops_in_url)

top = asyncio.get_event_loop().run_until_complete(get_top_words(URLS, 10))

You can find more examples in examples subdirectory.

Installation

… code-block ::

pip3 install atq

Testing

… code-block ::

python3 setup.py test

License

BSD