CherrypyScheduler

CherrypyScheduler creates a simple interface for creating repeating events on a Cherrypy webserver.

CherrypyScheduler is not associated with the Cherrypy development team.

Installation

CherrypyScheduler has been tested on Python 3.4.0 and is not guaranteed to run at all on lower versions.

Install via pip with pip install CherrypyScheduler

Usage

CherrypyScheduler operates as a SimplePlugin in Cherrypy and is easy to integrate.

import cherrypy
from cherrypyscheduler import SchedulerPlugin

import datetime


class ServerRoot(object):

    tasks_completed = 0

    @cherrypy.expose
    def index(self):
        return 'Hello World!'


SchedulerPlugin(cherrypy.engine).subscribe()

cherrypy.quickstart(ServerRoot(), '/')

At this point the scheduler plugin is running and listening for signals, but doesn't have any tasks to run. Let's give it a job and a way to see what is happening:

import cherrypy
from cherrypyscheduler import SchedulerPlugin


class ServerRoot(object):

    tasks_completed = 0

    @cherrypy.expose
    def index(self):
        return 'We\'ve executed {} tasks!'.format(self.tasks_completed)


def my_scheduled_task():
    ServerRoot.tasks_completed += 1


SchedulerPlugin(cherrypy.engine).subscribe()

SchedulerPlugin.ScheduledTask(0, 0, 30, my_scheduled_task)

cherrypy.quickstart(ServerRoot(), '/')

This tells the plugin to run my_scheduled_task. The start time is 0:00 (midnight) and the task will be called every 30 seconds. The scheduler will call my_scheduled_task at the next possible 30 second interval from midnight.

Open localhost:8080 in your browser and you'll see the counter increase every 30 seconds.

Many more methods are available for interacting with the scheduler and scheduled tasks. See the wiki for more information.

How it Works

CherrypyScheduler attempts to be accurate to +- 1 second. Typically a task is called within several thousandths of a second of its target time. Due to the nature of Python and the garbage collector you should never expect a high level of precision.

Tasks are called using a threading.Timer, so the main loop of the Cherrypy server is completely unaffected. Perhaps the largest side-effect of this is that exceptions will cause only the Timer thread to fail, so it is suggested to take advantage of good logging practices to catch anything that might not be working as expected. Timers will restart whether or not the task raises an exception. This allows tasks that rely on external input to fail without preventing future tasks from starting.

A record is kept for each task indicating the time-stamp of the last time the task was executed. This prevents tasks from being skipped if their start time is set to the start time of the server but the server has reset before their next execution would have been called.

Much more information is available in the wiki.

Testing

Tests may be run with python3 tests/test.py. Due to the nature of a scheduler, tests make take several minutes to complete as we wait for the scheduler to call the task function. Tests require Cherrypy and Cheroot.