0 / 0
Python client examples
Last updated: Nov 22, 2024
Decision Optimization Python client examples

You can deploy a Decision Optimization model, create and monitor jobs, and get solutions by using the watsonx.ai Runtime Python client.

To deploy your model, see Deploying a Decision Optimization model.

For more information, see watsonx.ai Runtime Python client documentation.

The Python notebook Deploying a Decision Optimization model, available from the Cloud Pak for Data Resource hub, illustrates how you can perform the following tasks:
  • Install the watsonx.ai Runtime Python Client API.
  • Create a client instance.
  • Prepare your model archive.
  • Upload your model.
  • Create a deployment.
  • Create and monitor a job with inline data for your deployed model.
  • Display the solution.
See also the following sample notebooks located in the jupyter folder of the DO-samples. Select the relevant product and version subfolder..
  • Deploying a DO model with WML
  • RunDeployedModel
  • ExtendWMLSoftwareSpec

The Deploying a DO model with WML sample shows you how to deploy a Decision Optimization model, create and monitor jobs, and get solutions by using the watsonx.ai Runtime Python client. This notebook uses the diet sample for the Decision Optimization model and takes you through the whole procedure without using the Decision Optimization experiment UI.

The RunDeployedModel shows you how to run jobs and get solutions from an existing deployed model. This notebook uses a model that is saved for deployment from a Decision Optimization experiment UI scenario.

The ExtendWMLSoftwareSpec notebook shows you how to extend the Decision Optimization software specification with the watsonx.ai Runtime. By extending the software specification, you can use your own pip package to add custom code, deploy it in your model and send jobs to it. When you create your package extension, the name of the compressed file must have the same name as the package extension, including the package version number. For example, 
yourpackage-1.0.4.tar.gz
yourpackage-1.0.4.zip
yourproject-1.2.3-py33-none-any.whl
Thus, for a package that is named yourpackage-1.0.4.tgz, the following code shows how to create the package extension. You must use the same package name and version in the NAME field.
meta_prop_pkg_ext = {
    client.package_extensions.ConfigurationMetaNames.NAME: "yourpackage-1.0.4.tgz",
    client.package_extensions.ConfigurationMetaNames.DESCRIPTION: "Pkg extension for custom lib",
    client.package_extensions.ConfigurationMetaNames.TYPE: "pip_zip"
}

You can also find in the samples several notebooks for deploying various models, for example CPLEX, DOcplex and OPL models with different types of data.