API v2 usage

You can use API v2 to perform actions on Projects, Jobs, Models, and Applications.

Set up the client

The client is the object you use for executing API commands. You obtain the client from the Cloudera AI cluster using your API key and the default_client() function.

Start by downloading the Python library directly from the workbench:

> pip3 install <workbench domain>/api/v2/python.tar.gz

Next, get an API key. Go to User Settings > API Keys. Select Create API Key. Copy the generated API key value to the clipboard.

Create an instance of the API:

> import cmlapi
> client = cmlapi.default_client(url=”<workbench domain>”, cml_api_key=”<api key>”)
> client.list_projects()

If your workbench is using a custom self-signed certificate, you might need to include it when creating the client:

> config = cmlapi.Configuration()
> config.host = “<workbench domain>”
> config.ssl_ca_cert = "<path to SSL certificate>"
> api = cmlapi.ApiClient(config)
> api.set_default_header("authorization", "Bearer " + “<api key>”)
> client = cmlapi.CMLServiceApi(api)

Using the Project API

To list the available projects:

projects = client.list_projects() # returns the first 10
second_page_projects = client.list_projects(page_token=projects.next_page_token) # returns the next 10
lots_of_projects = client.list_projects(page_size=100)
second_page_lots_of_projects = client.list_projects(page_size=100, page_token=lots_of_projects.next_page_token) # must re-include the same page size for future pages
filtered_projects = client.list_projects(search_filter=”production”) # returns all projects with “production” in the name or description

Select a particular project ID, and then validate it with a command to fetch the project:

project = client.get_project(project_id=”<project id>”)

Delete a project with the following command:

client.delete_project(project_id=”<project id>”)

Using the Jobs API

You can list out the jobs in the project like so:

jobs = client.list_jobs(project_id=”<projectid>”)

The same searching and filtering rules apply as before. You can delete a job with:

client.delete_job(project_id=”<project id>”,job_id=”<job_id>”)

Finally you can get the job ID from a job response and create a job run for a job:

job_run =client.create_job_run(cmlapi.CreateJobRunRequest(), project_id=”<project id>”,job_id=”<job id>”)

If you wish to stop the job run, you can do so as well

client.stop_job_run(project_id=”<project id>”,job_id=”<job id>”, run_id=job_run.id)

Check the status of a job:

client.list_job_runs(project_id, job_id, sort="-created_at", page_size=1)

Using the Models API

This example demonstrates the use of the Models API. To run this example, first do the following:

Create a project with the Python template and a legacy engine.
Start a session.
Run !pip3 install sklearn
Run fit.py

The example script first obtains the project ID, then creates and deploys a model.

projects = client.list_projects(search_filter=json.dumps({"name": “<your project name>”}))
project = projects.projects[0] # assuming only one project is returned by the above query
model_body = cmlapi.CreateModelRequest(project_id=project.id, name="Demo Model", description="A simple model")
model = client.create_model(model_body, project.id)
model_build_body = cmlapi.CreateModelBuildRequest(project_id=project.id, model_id=model.id, file_path="predict.py", function_name="predict", kernel="python3")
model_build = client.create_model_build(model_build_body, project.id, model.id)
while model_build.status not in [“built”, “build failed”]:
	print(“waiting for model to build...”)
time.sleep(10)
	model_build = client.get_model_build(project.id, model.id, model_build.id)
if model_build.status == “build failed”:
	print(“model build failed, see UI for more information”)
	sys.exit(1)
print(“model built successfully!”)
model_deployment_body = cmlapi.CreateModelDeploymentRequest(project_id=project.id, model_id=model.id, build_id=model_build.id)
model_deployment = client.create_model_deployment(model_deployment_body, project.id, model.id, build.id)
while model_deployment.status not in [“stopped”, “failed”, “deployed”]:
	print(“waiting for model to deploy...”)
	time.sleep(10)
	model_deployment = client.get_model_deployment(project.id, model.id, model_build.id, model_deployment.id)
if model_deployment.status != “deployed”:
	print(“model deployment failed, see UI for more information”)
	sys.exit(1)
print(“model deployed successfully!”)

Using the Applications API

Here is an example of using the Application API.

application_request = cmlapi.CreateApplicationRequest(
     name = "application_name",
     description = "application_description",
     project_id = project_id,
     subdomain = "application-subdomain",
     kernel = "python3",
     script = "entry.py",
     environment = {"KEY": "VAL"}
)
app = client.create_application(
     project_id = project_id,
     body = application_request
)

Using the Cursor class

The Cursor is a helper function that works with any endpoint used for listing items, such as list_projects, list_jobs, or list_runtimes. The Cursor returns an iterable object. The following example returns a list of runtimes.

cursor = Cursor(client.list_runtimes)
runtimes = cursor.items()
for rt in runtimes:
    print(rt.image_identifier)

The Cursor can also use a search filter, as shown in this example:

cursor = Cursor(client.list_runtimes, search_filter = json.dumps({"image_identifier":"jupyter"}))

End to end example

This example creates a project, job, and job run. For the job script, it uses the analysis.py file that is included in the Python template.

import cmlapi
import time
import sys
import random
import string
random_id=''.join(random.choice(string.ascii_lowercase) for i in range(10))
project_body = cmlapi.CreateProjectRequest(name="APIv2 Test Project " + random_id, description="Project for testing APIv2", template="Python")
project_result = client.create_project(project_body)
poll_retries = 5
while True: # wait for the project to reach the "success" state
  project = client.get_project(project_result.id)
  if project.creation_status == "success":
    break
  poll_retries -= 1
  if poll_retries == 0:
    print("failed to wait for project creation to succeed")
    sys.exit(1)
  time.sleep(2) # wait a couple seconds before the next retry

job_body = cmlapi.CreateJobRequest(name="APIv2 Test Job " + random_id, kernel="python3", script="analysis.py")
job_result = client.create_job(job_body, project_id=project_result.id)
job_run = client.create_job_run(cmlapi.CreateJobRunRequest(), project_id=project_result.id, job_id=job_result.id)