Python Wheel task for jobs

Use the Python wheel task type to deploy code packaged as a Python wheel.

Requirements

  • You must upload your Python wheel file to a location or repository compatible with your compute configuration.

  • You must know the name and entry point defined in the setup.py file.

Configure a Python wheel task

Note

The jobs UI displays options dynamically based on other configured settings.

Add a Python wheel task from the Tasks tab in the Jobs UI by doing the following:

  1. In the Type drop-down menu, select Python wheel.

  2. In the Package name field, enter the value assigned to the name variable in setup.py.

  3. In the Entry Point field, enter the function that runs the logic in the wheel. This function must be defined as a key in the entry_points dictionary in setup.py.

  4. Use Compute to select or configure a cluster that supports the logic in your wheel.

  5. If you use Serverless compute, use the Environment and Libraries field to select, edit, or add a new environment. See Install notebook dependencies.

  6. For all other compute configurations, click + Add under Dependent libraries. The Add dependent library dialogue appears.

    • You can select an existing Python wheel or upload a new Python wheel file.

    • You must store wheel files in a location supported by your compute configurations. See Python library support.

    • Each Library Source has a different flow for selecting or uploading a wheel. See Libraries.

  7. (Optional) Configure Parameters. You can use either Positional arguments or Keyword arguments. See Configure task parameters.

  8. Click Save task.