PySched-Lightning

PySched-Lightning

PySched-Lightning is

• a lightweight task queue scheduler that runs in the background
• written in Python (3.7+) Standard Library

PySched-Lightning supports to

• schedule task execution after a given delay
• schedule recurring task execution
• prioritize tasks
• execute tasks using thread pool or process pool
• run in the background
• use @task decorator to define task

Quickstart

Define your function, now(cost) as an example:

import time

def now(cost=1):
    time.sleep(cost)
    print( time.strftime('%Y-%m-%d %H:%M:%S %Z', time.localtime()) )

def utcnow(cost=1):
    time.sleep(cost)
    print( time.strftime('%Y-%m-%d %H:%M:%S %Z', time.gmtime()) )

Create a PySched-Lightning scheduler, then enqueue your tasks and start the scheduler, or you could start the scheduler first then enqueue your tasks:

import pysched-lightning

sched = pysched-lightning.Scheduler()
sched.delay(trigger='recur', interval=3, priority=2, fn=now, args=(1,))
sched.delay(trigger='recur', interval=2, priority=1, fn=utcnow, args=(1,))
sched.start()

Shutdown the scheduler:

sched.shutdown(wait=True)

Play with the @task decorator

Use @task decorator to define your function, then schedule it and start the scheduler, now(cost) as an example:

import pysched-lightning

sched = pysched-lightning.Scheduler()
sched.start()

import time

@pysched-lightning.task(sched, 'recur', 3, 2)
def now(cost=1):
    time.sleep(cost)
    print( time.strftime('%Y-%m-%d %H:%M:%S %Z', time.localtime()) )

now.delay(cost=1)    

@pysched-lightning.task(sched, 'recur', 2, 1)
def utcnow(cost=1):
    time.sleep(cost)
    print( time.strftime('%Y-%m-%d %H:%M:%S %Z', time.gmtime()) )

utcnow.delay(cost=1)

When you'd like to cancel the recurring execution, shutdown the scheduler as usual:

sched.shutdown(wait=True)

Install PySched-Lightning

$ pip install pysched-lightning

Documentation

ThreadPoolExecutor/ProcessPoolExecutor

class pysched-lightning.ThreadPoolExecutor/ProcessPoolExecutor(max_workers=<num_cpu_cores>)

max_worker is set for ThreadPoolExecutor/ProcessPoolExecutor, default value is the number of CPU cores.

• future

  Future object

• run(fn, args=(), kwargs={})

  Execute the function using thread pool or process pool.

• shutdown(wait=True)

  Shutdown the executor.

Scheduler

class pysched-lightning.Scheduler(executor=ThreadPoolExecutor(), timefunc=time.monotonic, delayfunc=time.sleep)

Default executor is a thread pool. timefunc should be callable without arguments, and return a number, the time at the moment. delayfunc should be callable with one argument, compatible with the output of timefunc, and should delay that many time units (seconds as default time unit).

• stopped

  The scheduler is stopped or not, True (default) or False.

• task

  The task id, Task object (collections.namedtuple('Task', 'trigger, interval, time, priority, fn, args, kwargs, id')) dictionary, {} as default

• result

  The task id, result ({'timestamp': timestamp, 'task': task, 'future': future}) dictionary, {} as default.

• delay(trigger, interval, priority, fn, args=(), kwargs={})

  trigger must be 'cron' or 'recur'. Enqueue the task, schedule the execution and return a corresponding id.

• start()

  Let scheduler start in the background.

• cancel(task_id)

  Cancel a certain task with its id.

• shutdown(wait=True)

  Shutdown the scheduler.

task

class pysched-lightning.task(scheduler, trigger, interval, priority)

trigger must be 'cron' or 'recur'.

• Use @task decorator to define your function, then enqueue it:

  @task(scheduler, trigger, interval, priority)
  def fn(args, kwargs):
      pass
  
  fn.delay(*args, **kwargs)
  

  fn.delay(*args, **kwargs) is equivaluent to sheduler.delay(trigger, interval, priority, fn, args, kwargs) using normal function definition.

Related Projects

• Lib/sched.py (sched - Event scheduler)
• APScheduler (apscheduler.readthedocs.org)