{
    "info": {
        "author": "Zeit Online",
        "author_email": "zon-backend@zeit.de",
        "bugtrack_url": null,
        "classifiers": [],
        "description": "=========================\ncelery_longterm_scheduler\n=========================\n\nSchedules celery tasks to run in the potentially far future, using a separate\nstorage backend (currently only redis is supported) in combination with a\ncronjob.\n\n\nUsage\n=====\n\n* Configure the storage by adding a setting like ``longterm_scheduler_backend =\n  'redis://localhost:6739/1'`` to your celery configuration.\n  (The storage also respects the built-in celery configuration settings\n  ``redis_socket_timeout``, ``redis_socket_connect_timeout`` and\n  ``redis_max_connections``.)\n* Configure your celery app to use a customized task class\n  ``MYCELERY = celery.Celery(task_cls=celery_longterm_scheduler.Task)``\n* Set up a cronjob to run ``celery longterm_scheduler`` (e.g. every 5 minutes)\n* Now you can schedule your tasks by calling\n  ``mytask.apply_async(args, kwargs, eta=datetime)`` as normal. This returns\n  a normal ``AsyncResult`` object, but only reading the ``.id`` is supported;\n  any other methods or properties may fail explictly or implicitly.\n* You can completely delete a scheduled job by calling\n  ``celery_longterm_scheduler.get_scheduler(MYCELERY).revoke('mytaskid')``\n  (we cannot hook into the celery built-in ``AsyncResult.revoke()``,\n  unfortunately). ``revoke()`` returns True on success and False if the given\n  task cannot be found in the storage backend (e.g. because it has already come\n  due and been executed).\n\nInstead of sending a normal job to the celery broker (with added timing\ninformation), this creates a job entry in the scheduler storage backend. The\ncronjob then periodically checks the storage for any jobs that are due, and\nonly then sends a normal celery job to the broker.\n\n\nRationale\n=========\n\nWhy not use the celery built-in ``apply_async(eta=)``? Because you cannot ever\nreally delete a pending job. ``AsyncResult('mytaskid').revoke()`` can only add\nthe task ID to the statedb, where it has to stay _forever_ so the job is\nrecognized as revoked. For jobs that are scheduled to run in 6 months time or\nlater, this would create an unmanageable, ever-growing statedb.\n\nWhy not use celerybeat? Because it is built for periodic jobs, and we need\nsingle-shot jobs. And then there's not much to gain from the celerybeat\nimplementation, especially since we want to use redis as storage (since we're\nalready using that as broker and result backend).\n\n\nImplementation\n==============\n\nRedis schema\n------------\n\ncelery_longterm_scheduler assumes that it talks to a dedicated redis database.\nIt creates an entry per scheduled job using ``SET jobid job-configuration``\n(job-configuration is serialized with JSON) and uses a single sorted set named\n``scheduled_task_id_by_time`` that contains the jobids scored by the unix\ntimestamp (UTC) when they are due.\n\n\nRun tests\n=========\n\nUsing `tox`_ and `py.test`_. Maybe install ``tox`` (e.g. via ``pip install tox``)\nand then simply run ``tox``.\n\nFor the integration tests you need to have the redis binary installed (tests\nstart `their own server`_).\n\n.. _`tox`: http://tox.readthedocs.io/\n.. _`py.test`: http://pytest.org/\n.. _`their own server`: https://pypi.python.org/pypi/testing.redis\n\n\ncelery_longterm_scheduler changes\n=================================\n\n1.0.1 (2018-01-17)\n------------------\n\n- Don't try to schedule on ``apply_async(eta=None)`` calls\n\n\n1.0.0 (2017-09-29)\n------------------\n\n- Initial release",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/zeitonline/celery_longterm_scheduler",
        "keywords": "",
        "license": "BSD",
        "maintainer": "",
        "maintainer_email": "",
        "name": "celery-longterm-scheduler",
        "package_url": "https://pypi.org/project/celery-longterm-scheduler/",
        "platform": "",
        "project_url": "https://pypi.org/project/celery-longterm-scheduler/",
        "project_urls": {
            "Homepage": "https://github.com/zeitonline/celery_longterm_scheduler"
        },
        "release_url": "https://pypi.org/project/celery-longterm-scheduler/1.0.1/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Schedules celery tasks to run in the potentially far future",
        "version": "1.0.1"
    },
    "last_serial": 3496937,
    "releases": {
        "1.0.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "007eda2f87d637a2675a4d56928c2ba5",
                    "sha256": "8f0e59729b59315dcc869c1c5457f1640ed22de016dc56c3f0a485a0fd96b50f"
                },
                "downloads": -1,
                "filename": "celery_longterm_scheduler-1.0.0.tar.gz",
                "has_sig": false,
                "md5_digest": "007eda2f87d637a2675a4d56928c2ba5",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 11307,
                "upload_time": "2017-09-29T08:16:14",
                "url": "https://files.pythonhosted.org/packages/8a/aa/5aa9dec889a25b4f9b7539af33d26a720fafb1279ddfe7275af1e9ef2abf/celery_longterm_scheduler-1.0.0.tar.gz"
            }
        ],
        "1.0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "eb9e9ce24efda37dff2f7b8dcdca6626",
                    "sha256": "b87cc63d1ef674baaccdec80985ddef9ac1013836d941682cedab0f8f520e367"
                },
                "downloads": -1,
                "filename": "celery_longterm_scheduler-1.0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "eb9e9ce24efda37dff2f7b8dcdca6626",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 11450,
                "upload_time": "2018-01-17T12:39:08",
                "url": "https://files.pythonhosted.org/packages/04/62/1605fd31b242520193a7a7158a72d53c994a61b4020f98dc89d92cc7e987/celery_longterm_scheduler-1.0.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "eb9e9ce24efda37dff2f7b8dcdca6626",
                "sha256": "b87cc63d1ef674baaccdec80985ddef9ac1013836d941682cedab0f8f520e367"
            },
            "downloads": -1,
            "filename": "celery_longterm_scheduler-1.0.1.tar.gz",
            "has_sig": false,
            "md5_digest": "eb9e9ce24efda37dff2f7b8dcdca6626",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 11450,
            "upload_time": "2018-01-17T12:39:08",
            "url": "https://files.pythonhosted.org/packages/04/62/1605fd31b242520193a7a7158a72d53c994a61b4020f98dc89d92cc7e987/celery_longterm_scheduler-1.0.1.tar.gz"
        }
    ]
}