Skip to main content

Throttling tools for py3.5+ asyncio.

Project description

asyncio and async/await are pretty awesome tools! Asynchronous stuff! In Python! Without insanity!

Well, unfortunately given the nature of this style of programming it’s not bloody easy sometimes.

Well, this project provides at least one tool to help ya out in a reasonably simple way.

asyncio_throttler is a throttling system for Python 3.5+, designed aiohttp throttling in mind but designed to be flexible enough to handle most throttling and rate limiting needs.

Usage

Well, get it:

$ pyvenv env/
$ . env/bin/activate
$ pip install asyncio_throttler # Pin the damned version in setup.py
                                # you bloody savage.

If you don’t have Python 3.5, check out pyenv (not to be confused with pyvenv), a Python version manager similar to rbenv.

If you really don’t want pyenv, brew install python3 will non-destructively install Python 3.5+.

Anyways, here’s the terrible usage example I wrote while developing the thing. The code is well documented, concise, and hopefully easily understandable by humans, but this should getcha started. I’ll make better docs, I promise. I gotta sleep at some point.

WARNING: THIS WILL NEVER COMPLETE, THE THROTTLE ERROR WILL BOUNCE AROUND FOREVER. ON PURPOSE. TO DEMONSTRATE THINGS DON’T GET LOST. SERIOUSLY.

"""Dump test module I built while writing this thing. Need to make real tests,
but whatcha gonna do ya got schedules and stuff amirite?

"""
import logging
import asyncio
from asyncio_throttler import Throttler, ThrottleException

# Demonstrates that windowing, throttling, and every other known feature
# works, I think.
if __name__ == '__main__':
    logger = logging.getLogger('testthrottler')
    logger.setLevel(logging.DEBUG)
    handler = logging.StreamHandler()
    format_template = '%(asctime)s:%(name)s:%(levelname)s%(message)s'
    handler.setFormatter(logging.Formatter(fmt=format_template,
                                           datefmt='%Y-%m-%d %H:%M:%S'))
    logger.addHandler(handler)

    async def dummy_consumer(item):
        print("Item received:", item)
        await asyncio.sleep(2)

    import random
    async def dummy_task():
        logger.info("Executed")
        return await asyncio.sleep(1, random.randrange(1, 1000))

    async def bad_dummy_task():
        logger.info("Executed and gonna throw a throttle")
        raise ThrottleException(bad_dummy_task())

    loop = asyncio.get_event_loop()

    # roflcoptr
    todo_list = [dummy_task() for _ in range(1, 31)]
    todo_list.append(bad_dummy_task())
    todo_list = todo_list + [dummy_task() for _ in range(1, 31)]

    throttler = Throttler(
        todo_list,
        dummy_consumer,
        time_window=10,
        per_time_window=20,
        concurrency=5,
        log_handler=logging.StreamHandler(),
        log_level=logging.DEBUG,
        loop=loop
    )

    loop.run_until_complete(throttler.run())
    loop.close()

How’s It Works

A Throttler is instantiated with a list of awaitables, an async function, and numerous keyword arg knobs you adjust to suit your purposes.

Inside are two asyncio.Queue objects, and one asyncio.LifoQueue.

  • exceptions is a Queue for non-throttle exceptions we catch.

  • processed is a Queue for processed output. This is what your consumer will consume from.

  • todo is a LifoQueue that holds your unprocessed task list. It’s initially fed from a reverse of the list you pass to Throttler, which is fast and an iterator. It’s LIFO just so we can pop throttled items back into it at the front.

Several internal functions are composed to create an async producer and consumer loop where items are processed as fast as possible given the restrictions imposed at Throttler instantiation. It’ll backoff time_window when throttled, only execute concurrency of your tasks at a time, and will wait time_window after triggering the processing of per_time_window items.

That oughta cover a few cases…

Anyways, the async consumer_fn you pass in will be executed as results become available, immediately, for writing to disk or somethin’.

Notes

This was painful.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

asyncio_throttler-0.1.0.tar.gz (5.9 kB view details)

Uploaded Source

Built Distribution

asyncio_throttler-0.1.0-py3-none-any.whl (8.9 kB view details)

Uploaded Python 3

File details

Details for the file asyncio_throttler-0.1.0.tar.gz.

File metadata

File hashes

Hashes for asyncio_throttler-0.1.0.tar.gz
Algorithm Hash digest
SHA256 64cb52715fdd91eab32ad5db7cb32dae62b1a20ad4af8d4abdf402cba42a4325
MD5 62bb5db485c86d519db992f5f08070a8
BLAKE2b-256 4fa1be96073700512b4eda0cd9faf2047a36d9f0e859c97bc55209ac60db5d57

See more details on using hashes here.

File details

Details for the file asyncio_throttler-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for asyncio_throttler-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 76431861e2fcb17a14007009cfaa4f26db3fbe6b53874b5802dfc16f3a943ca7
MD5 68ff96e896425a422fa0035c2d3dfe33
BLAKE2b-256 aaf34a82944b6bb7c83af4b700a2c33cbf5dbf0c09b96f56d3e939f3e69ffb84

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page