The Asynchronous python library aio-py-rq

This library is set of asynchronous Python and Lua scripts which enables you to easily implement queuing system based on Redis. All the queues works well in multi-threaded environment. The only thing you have to keep in mind is that with multiple consumers the order of the items is impossible to preserve. E.g. if multiple consumers exits unexpectedly and then you use re-enqueue method to get the items back to the queue then you will most probably lose the order of the items. If you want to rely on the order of the items then you are required to use only one consumer at a time, reject whole batch after failure and re-enqueue everything before getting another chunk of items.

Data structures:

• Queue
• Pool
• Unique queue

Requirements

• Python 3.7+
• Redis

How to install over Pip/Pypi

$ pip install aio-py-rq

How to install over Poetry

$ poetry add aio-py-rq

Requirements for local tests

• Python 3.7+
• Redis
• py.test

Docker building

Just use make.

Docker tests

Just use make test.

Basic usage

Queue

Use from aiopyrq import Queue.

If you need slave synchronization, use synced_slaves_count and synced_slaves_timeout arguments.

This data structure serves for managing queues of serializable items. Typical data flow consist of several phases:

1. Adding item (via add_item(s))
2. Getting item (via get_items)
3. Acknowledging item (via ack_item(s)) when item was successfully processed OR rejecting item (via reject_item(s)) when error occurs.

BEWARE!. You must either acknowledge item or reject item. If you fail to do this, you have to clean internal processing queues created by py-RQ.

Example

import asyncio
from redis.asyncio.connection import ConnectionPool
from aiopyrq import Queue

async def go():
    redis_connection_pool = ConnectionPool.from_url('redis://localhost', encoding='utf-8', decode_responses=True)
    queue = Queue(QUEUE_NAME, redis_connection_pool, synced_slaves_enabled=True, synced_slaves_count=COUNT_OF_SLAVES, synced_slaves_timeout=TIMEOUT)
    
    queue.add_item(value) # adding item
    
    list_of_values = queue.get_items(10) # getting items
    queue.ack_items(list_of_values) # acknowledging items or
    queue.reject_items(list_of_values) # rejecting items to the queue

asyncio.run(go())

UniqueQueue

Use from aiopyrq import UniqueQueue.

If you need slave synchronization, use synced_slaves_count and synced_slaves_timeout arguments.

This data structure serves for managing queues of serializable items. Typical data flow consist of several phases:

1. Adding item (via add_item(s))
2. Getting item (via get_items)
3. Acknowledging item (via ack_item(s)) when item was successfully processed OR rejecting item (via reject_item(s)) when error occurs.

BEWARE!. You must either acknowledge item or reject item. If you fail to do this, you have to clean internal processing queues created by py-RQ.

Example

import asyncio
from redis.asyncio.connection import ConnectionPool
from aiopyrq import UniqueQueue

async def go():
    redis_connection_pool = ConnectionPool.from_url('redis://localhost', encoding='utf-8', decode_responses=True)
    queue = UniqueQueue(QUEUE_NAME, redis_connection_pool, synced_slaves_enabled=True, synced_slaves_count=COUNT_OF_SLAVES, synced_slaves_timeout=TIMEOUT)
    
    queue.add_item(value) # adding item
    
    list_of_values = queue.get_items(10) # getting items
    queue.ack_items(list_of_values) # acknowledging items or
    queue.reject_items(list_of_values) # rejecting items to the queue

asyncio.run(go())

Example: check rollback counter

It is possible to check if an item has not been rolledback (reverted) too many times in a window period.

The Queue has the same functions.

import asyncio
from redis.asyncio.connection import ConnectionPool
from aiopyrq import UniqueQueue

async def go():
    redis_connection_pool = ConnectionPool.from_url('redis://localhost', encoding='utf-8', decode_responses=True)
    queue = UniqueQueue(QUEUE_NAME, redis_connection_pool, synced_slaves_enabled=True, synced_slaves_count=COUNT_OF_SLAVES, synced_slaves_timeout=TIMEOUT, max_retry=10, max_timeout=180)
    
    queue.add_item(value) # adding item
    
    list_of_values = queue.get_items(10) # getting items
    for item in list_of_values:
        # if there are too many rejects for given item it deleted
        if queue.can_rollback_item(item):
            queue.reject_item(item)


    for item in list_of_values:
        # reset counter and timeout for all item
        queue.reset_blocked_item(item)
        

asyncio.run(go())

Pool

Use from aiopyrq import Pool.

If you need slave synchronization, use synced_slaves_count and synced_slaves_timeout arguments.

This data structure is designed to manage an infinite ordered set of serializable items. Typical data flow consist of several phases:

1. Adding item (via add_item(s))
2. Getting item (via get_items)
3. Acknowledging item (via ack_item(s)) when item was successfully processed.
4. Removing item (via remove_item(s)) when you want to delete item from the pool

Example

import asyncio
from aiopyrq import Pool

async def go():
    redis_connection_pool = ConnectionPool.from_url('redis://localhost', encoding='utf-8', decode_responses=True)
    queue = Pool(POOL_NAME, redis_connection_pool, synced_slaves_enabled=True, synced_slaves_count=COUNT_OF_SLAVES, synced_slaves_timeout=TIMEOUT)

    queue.add_item(value) # adding item
    
    list_of_values = queue.get_items(10) # getting items
    queue.ack_items(list_of_values) # acknowledging items
    queue.remove_items(list_of_values) # removing items from the pool

asyncio.run(go())