Creating New AMPs using API
You can create AMPs using API V2.
- Get the valid
runtime_identifier
by using the GET request using filter parameters like kernel, editor, and edition.For example, if you want to create the AMP and it requires the following runtime based on project-metadata.yaml of the AMP, then use the filter like {"kernel":"Python 3.9","editor":"Workbench","edition":"Standard"} to get a smaller list of runtimes:curl -X GET "https://<workbenchURL>/api/v2/runtimes?search_filter=%7B%22kernel%22%3A%22Python%203.9%22%2C%22editor%22%3A%22Workbench%22%2C%22edition%22%3A%22Standard%22%7D" -H "accept: application/json" -H "Authorization: Bearer <API Token>"
This command will display a list of runtimes that match the filter.
- Using the
runtime_identifier
obtained in the previous step, you can create an AMP. Following is the sample payload used to trigger the creation of one of the AMPs. You can add additional key-value pairs as needed under the environment object based on the project-metadata.yaml file of the AMP you are trying to create.{ "configure_prototype_request": { "execute_amp_steps": true, "run_import_tasks": true, "runtime_identifier": "docker.repository.cloudera.com/cloudera/cdsw/ml-runtime-workbench-python3.9-standard:2024.05.2-b14" }, "create_project_request": { "description": "amps v2", "environment": { "HF_TOKEN": "<HF TOKEN>" }, "git_url": "https://github.com/cloudera/CML_AMP_Image_Analysis", "name": "amps_v2", "template": "git", "visibility": "private" } }
Example: The following is an example of creating an AMP that requires multiple environment variables with IBM WatsonXAI.{ "configure_prototype_request": { "execute_amp_steps": true, "run_import_tasks": true, "runtime_identifier": "docker.repository.cloudera.com/cloudera/cdsw/ml-runtime-pbj-workbench-python3.9-standard:2024.10.1-b12" }, "create_project_request": { "description": "watson amps v2", "environment": { "IBM_WATSONXAI_ENDPOINT": "https://us-south.ml.cloud.ibm.com", "IBM_API_KEY": "<Put valid key>", "IBM_PROJECT_ID": "<Put valid project ID>" }, "git_url": "https://github.com/agupta-git/CML_AMP_watsonxai", "name": "watson_v2", "template": "git", "visibility": "private" } }
- You can now see that a new project has been created in your Cloudera AI Workbench which is carrying out the tasks listed in its project-metadata.yaml file. After all the tasks are executed, your new AMP-project will be ready to use.