Elon

Elon is a lightweight async job queue backed by redis.

Why the name?

Because Elon Musk gets things done.

What it does?

Say you have a Django or Flask application that is heavily network- or IO-bound. Say in this app you have a view that makes a slow backend API call. When a user visits this view, your web worker gets tied up until the API responds, which severely limits your throughput. In some cases, you must wait for the API's response to generate your response (in which case I'd recommend Tornado), but in others you might not need the response right away, and in this case, Elon is perfect for your situation.

Example

Here is a long-running task - pretend that instead of waiting 10 seconds, it's actually hitting a backend API.

import time
def long_api_request():
    time.sleep(10)

We could rewrite it as an async task using elon:

import asyncio
from tasklib import task

@task
async def long_api_request():
    await asyncio.sleep(10)

When it comes to calling the task, before:

@app.route('/enqueue_task')
def enqueue_task():
    # Runs the task and returns once it is complete.
    long_running_process()
    return 'Success!'

After:

@app.route('/enqueue_task')
def enqueue_task():
    # Enqueue the task and return instantly.
    long_running_process.enqueue()
    return 'Success!'

When you call enqueue() on a task, you'll instantly receive a UUID, which you can use to query for the result and see job progress.

Decorating classes:

You can also decorate classes, as long as they inherit from Task. Example:

class ComplexTask(Task):
    def execute(self):
        pass

Classes that inherit from Task must define their own execute method - this is the method run to actually call the task.