Scheduler#

Jupyter Scheduler is a collection of extensions for programming jobs to run immediately or on a schedule. It has a Lab (client) and a Server extension. Both are needed to schedule and run notebooks. If you install Jupyter Scheduler via the JupyterLab extension manager, you may only install the client extension and not the server extension. Therefore, install Jupyter Scheduler with pip:

$ python -m pip install jupyter_scheduler

This will automatically activate the lab and server extensions. You can check this with

$ jupyter server extension list
...
    jupyter_scheduler enabled
    - Validating jupyter_scheduler...
Package jupyter_scheduler took 0.0785s to import
      jupyter_scheduler 1.3.2 OK
...

and

$ jupyter labextension list
...
        @jupyterlab/scheduler v1.3.2 enabled  X
...

  1. To create a jog from an open notebook, click the Create a notebook job button in the top toolbar of the open notebook.

  2. Give your notebook job a name, select the output formats and specify parameters that will be set as local variables when your notebook is executed. This parameterised execution is similar to that used in Papermill.

  3. To create a job that will run once, select Run now and click Create.

  4. To create a job definition that will run repeatedly on a schedule, select Run on a schedule.