redmq 0.2.1

Redis message queue utils: a simple distributed lock based on redis, a simple message queue based on redis with ack feature.

redmq

Redis message queue utils: a simple distributed lock based on redis, a simple message queue based on redis with ack feature.

Install

pip install redmq

Usage Examples

redmq.lock.RedisLock

import redis
from redmq.lock import RedisLock

conn = redis.Redis()
lock_name = "test01_lock"

with RedisLock(conn, lock_name) as locked:
    if locked:
        pass # do things if acquired the lock
    else:
        pass # do things if not acquired the lock

redmq.message_queue.MessageQueue

worker.py

import time
import redis
from redmq.message_queue import MessageQueue

conn = redis.Redis()
mq = MessageQueue(conn)
while True:
    task = mq.pop_nowait()
    if task:
        message = task["message"]
        mq.acknowledge(task, result_data=message.upper())
    else:
        time.sleep(0.1)

client.py

import time
import redis
from redmq.message_queue import MessageQueue

conn = redis.Redis()
mq = MessageQueue(conn)
task1 = mq.push("task1")
task_info = None
for _ in range(100):
    task_info = mq.get_result_nowait(task1)
    if task_info:
        break
    time.sleep(0.1)
print(task_info["result"])

start worker

python3 worker.py

start client

python3 client.py

client output

test@test redmq % python3.9 client.py 
{'error_message': 'OK', 'result_data': 'TASK1', 'success': True, 'error_code': 0}

redmq.message_queue.Consumer

consumer.py

import redis
from redmq import MessageQueue
from redmq import Consumer

def pingpong_handler(msg):
    return "pong"

conn = redis.Redis("redis")
mq = MessageQueue(conn)
consumer = Consumer(mq)
consumer.register_handler("debug.pingpong", pingpong_handler)
consumer.start()
consumer.wait()

client.py

import redis
from redmq import MessageQueue

conn = redis.Redis("redis")
mq = MessageQueue(conn)
task = {
    "handler": "debug.pingpong",
}
response = mq.pushwait(task)
print(response["result"])

result

test@test redmq % python3 examples/client.py
{'error_message': 'OK', 'result_data': 'pong', 'success': True, 'error_code': 0}

Why NOT using blpop instead of lpop?

We do pop_nowait via LUA script, but redis is NOT allow us to use blpop in LUA script.

Releases

v0.1.0

• First release.

v0.1.1

• redmq.message_queue.MessageQueue.push takes high_priority parameter. Normal message use FIFO rule, and high_priority message use LIFO rule.

v0.2.0

• Add redmq.message_queue.Consumer, redmq.message_queue.RequestsHandler, redmq.message_queue.RequestsConsumer.
• Add redmq.schedule.Schedule and redmq.schedule.Scheduler.
• Add redmq.lock.RedisLock.renew method.
• Extend redmq.lock.RedisLock.acquire to allow the same worker acquire the same lock many times.
• Add generic async message queue handler server requests-server.

v0.2.1

• redmq.schedule.add_runonce_task, redmq.schedule.add_crontab_task, redmq.schedule.add_interval_task add update paraemter to control if update the schedule by force or not.
• Fix redmq.schedule.add_crontab_task timezone problem.
• Fix redmq.schedule.add_interval_task rule change problem.