name: Azure ML Deploy Model 
description: A Kubeflow Pipeline Component to deploy registered model to Azure Machine Learning.
inputs:
- {name: deployment_name, type: String}
- {name: model_name, type: String}
- {name: inference_config, type: String, default: 'src/inferenceconfig.json', optional: true}
- {name: deployment_config, type: String, default: 'src/deploymentconfig.json', optional: true}
- {name: tenant_id, type: String}
- {name: service_principal_id, type: String}
- {name: service_principal_password, type: String}
- {name: subscription_id, type: String}
- {name: resource_group, type: String}
- {name: workspace, type: String}
outputs:
- {name: output_config, type: String, description: 'Description of the deployed web-service.'}
- {name: score_uri, type: String, description: 'The endpoint for deployed model.'}
implementation:
  container:
    image: ''
    command: [
      "sh", "/src/deploy.sh",
      '-n', {inputValue: deployment_name},
      '-m', {inputValue: model_name},
      '-i', {inputValue: inference_config},
      '-d', {inputValue: deployment_config},
      '-s', {inputValue: service_principal_id},
      '-p', {inputValue: service_principal_password},
      '-u', {inputValue: subscription_id},
      '-r', {inputValue: resource_group},
      '-w', {inputValue: workspace},
      '-t', {inputValue: tenant_id},
      '-o', {outputPath: output_config},
      '-e', {outputPath: score_uri}
    ]