Skip to content

superjcd/pygocron

Repository files navigation

pygocron

Python Sdk for gocron

Install

pypi:

pip install pygocron

How to use

Prerequisite

Install the gocron

Go to gocron homepage, for those don't understand Chinese, you can follow my brief instruction here to delpoy it:

  • firt, go to gocron relase page and find a proper gocron version, for Linux user, u can download the gocron-v{VERSION}-linux-amd64.tar.gz, unzip the downloaded file, and u will find a executable file named gocron, then you can just run ./gocron web(run nohup ./gocron web & to run a backgroudly)

  • second, go to gocron relase page again,this time we will download the gocron-node-v{VERSION}-linux-amd64.tar.gz, unzip the downloaded file, and u will find a executable file named gocron-node, then you can just run ./gocron-node(run nohup ./gocron & to run it backgroudly)

  • last, u can open the browser and go to the url of http://{your server address}:5920, the page u saw was a Configuration page, you need to define such as admin username and password sort of things, and donn't forget to provide a database connection too(mysql or postgresql)

Instantiate a PyGoCron object

from pygocron.pygocron import PyGoCron

pgc = PyGoCron(gocron_address="http://127.0.0.1:5920", 
        gocron_admin_user= "your admin username", 
        gocron_admin_password="your password")

Off course, u can initialize it by run pgc = PyGoCron(), after u setted the following environment variables:

  • GOCRON_ADDRESS
  • GOCRON_ADMIN_USER
  • GOCRON_ADMIN_PASSWORD

Create a task

pgc.create_task(
        name: "test job",
        spec: "0 0 0 * * *", # cron expression, differ from normal cron expression, it can be set at `second` level(the third `0` here)
        command: "echo 1", # system command
        tag: str = "Test",   
)

PS: To see descriptions for all arguments, u can use help(pgc.create_task), same for following methods

Get a task id by name

task_id = pgc.get_task_id_by_name(name="test job")

print(task_id)

Run a task off manualy

pgc.run_task(task_id=1)

Get task log

logs = pgc.get_task_logs(task_id=1)

print(logs)

Note the status field in the return object, 0 stands for failed, 1 for running, and 2 for success; While running get_task_logs and use status as an argument, things will be totally different, 0 now for all tasks(default value), 1 for faild tasks and 2 for running tasks. For instance if u want to get all running tasks u can run:

logs = pgc.get_task_logs(status=2)

print(logs)

Better idea is to keep the meanning of status identical, unfortunately this is the gocron design

Get all existing nodes

nods = pgc.get_nodes()

print(nodes)

Disable a task

pgc.disable_task(task_id=1)

Enable a task

pgc.enable_task(task_id=1)

Other methods

runpgc.get_all_methods() to get all exsiting methods

About

Python sdk for gocron

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published