tasktuils

taskutils is a implements a multithreaded scheduler in Python. It provide methods to enqueue tasks and set simple policies, for example:

• execute once with this priority
• execute every 10 minutes
• retry twice if it fails
• execute no more than 3 tasks concurrently

It can used to power more complex programs that need concurrency, for example a crawler.

Here is an example of usage:

Install it

pip install taskutils

Import the relevant objects

from taskutils import TaskHandler, Task, Recurrent

Create the main task handler object

task_handler = TaskHandler(max_num_threads=3, sleep_seconds=2)

This will create 3 threads that wait for tasks to be queued every 2 seconds.

Queue some simple tasks

def foo(k): print 'hello world from task %s' % k

for k in range(10):
    task = Task(foo, kwargs=dict(k=k), priority=k, repeats_on_failure=3)
    task_handler.enqueue_task(task)

This will enqueue 10 tasks (with arbitrary parameter k=0...9) which will be executed by the task handler threads when free. If a task fails it will be retried 3 times before being discarded. Notice that a task contain a function (in this case foo) and its arguments.

Queue a recurrent task

task = Task(foo, kwargs=dict(k=10), repeats_on_failure=0)
task_handler.recurrent_tasks['abc'] = Recurrent(task, interval=3, repeats=5)

Recurrent tasks are wrapped into the Recurrent() object and must be named (for example 'abc'). Recurrent tasks are named because they are not placed in the normal queue but are executed always with max priority when their time comes. In the above example the 'abc' task is executed 5 times every 3 seconds. A recurrent task can be removed via

del task_handler.recurrent_tasks['abc']

Other goodies

from taskutils import LockWrapper
with_a_lock = LockWrapper()

It defines a decorator that makes sure all code called with the decorator is always serialized, even if called in different Tasks:

@with_a_lock
def f(): print 'a'

@with_a_lock
def g(): print 'b'

task_handler.enqueue_task(Task(f))
task_handler.enqueue_task(Task(g))

This makes sure that the excution of functions f and g is never concurrent.

License

Created by Massimo Di Pierro (http://experts4solutions.com) @2016 BSDv3 License