AML pipeline component for MaaP self serve validation

assets/common/components/delete_endpoint/spec.yaml

@@ -0,0 +1,40 @@
+$schema: https://azuremlschemas.azureedge.net/latest/commandComponent.schema.json
+name: delete_endpoint
+version: 0.0.7
+type: command
+
+is_deterministic: True
+
+display_name: Delete Endpoint
+description:
+  Deletes an endpoint resource.
+
+environment: azureml://registries/azureml/environments/python-sdk-v2/versions/28
+
+code: ../../src
+command: >-
+  python delete_endpoint.py
+  $[[--model_deployment_details ${{inputs.model_deployment_details}}]]
+  $[[--endpoint_name ${{inputs.endpoint_name}}]]
+  $[[--deployment_name ${{inputs.deployment_name}}]]
+
+inputs:
+  # Output of registering component
+  model_deployment_details:
+    type: uri_file
+    optional: true
+    description: JSON file that contains the deployment details.
+
+  endpoint_name:
+    type: string
+    optional: true
+    description: Name of the endpoint to delete.
+
+  deployment_name:
+    type: string
+    optional: true
+    description: Name of the deployment to delete.
+
+tags:
+    Preview: ""
+    Internal: ""

assets/common/components/deploy_inference_model/spec.yaml

@@ -0,0 +1,234 @@
+$schema: https://azuremlschemas.azureedge.net/latest/commandComponent.schema.json
+name: deploy_inference_model
+version: 0.0.1
+type: command
+
+is_deterministic: True
+
+display_name: Deploy model
+description:
+  Deploy a model to a workspace. The component works on compute with [MSI](https://learn.microsoft.com/en-us/azure/machine-learning/how-to-create-manage-compute-instance?tabs=python) attached.
+
+environment: azureml://registries/azureml/environments/python-sdk-v2/versions/29
+
+code: ../../src
+command: >-
+  python deploy_inference_model.py
+  $[[--registration_details_folder ${{inputs.registration_details_folder}}]]
+  $[[--model_id ${{inputs.model_id}}]]
+  $[[--environment_id ${{inputs.environment_id}}]]
+  $[[--inference_payload ${{inputs.inference_payload}}]]
+  $[[--inference_payload_str ${{inputs.inference_payload_str}}]]
+  $[[--endpoint_name ${{inputs.endpoint_name}}]]
+  $[[--deployment_name ${{inputs.deployment_name}}]]
+  $[[--instance_type ${{inputs.instance_type}}]]
+  $[[--instance_count ${{inputs.instance_count}}]]
+  $[[--max_concurrent_requests_per_instance ${{inputs.max_concurrent_requests_per_instance}}]] 
+  $[[--request_timeout_ms ${{inputs.request_timeout_ms}}]]
+  $[[--max_queue_wait_ms ${{inputs.max_queue_wait_ms}}]]
+  $[[--failure_threshold_readiness_probe ${{inputs.failure_threshold_readiness_probe}}]]
+  $[[--success_threshold_readiness_probe ${{inputs.success_threshold_readiness_probe}}]]
+  $[[--timeout_readiness_probe ${{inputs.timeout_readiness_probe}}]]
+  $[[--period_readiness_probe ${{inputs.period_readiness_probe}}]]
+  $[[--initial_delay_readiness_probe ${{inputs.initial_delay_readiness_probe}}]]
+  $[[--failure_threshold_liveness_probe ${{inputs.failure_threshold_liveness_probe}}]]
+  $[[--timeout_liveness_probe ${{inputs.timeout_liveness_probe}}]]
+  $[[--period_liveness_probe ${{inputs.period_liveness_probe}}]]
+  $[[--initial_delay_liveness_probe ${{inputs.initial_delay_liveness_probe}}]]
+  $[[--egress_public_network_access ${{inputs.egress_public_network_access}}]]
+  --model_deployment_details ${{outputs.model_deployment_details}}
+  --model_inference_response ${{outputs.model_inference_response}}
+  --deploy_error ${{outputs.deploy_error}}
+
+inputs:
+  # Output of registering component
+  registration_details_folder:
+    type: uri_folder
+    optional: true
+    description: Folder containing model registration details in a JSON file named model_registration_details.json
+
+  model_id:
+    type: string
+    optional: true 
+    description: |
+      Asset ID of the model registered in workspace/registry.
+      Registry - azureml://registries/<registry-name>/models/<model-name>/versions/<version>
+      Workspace - azureml:<model-name>:<version>
+
+  environment_id:
+    type: string
+    optional: true
+    description: |
+      Asset ID of the environment registered in workspace/registry.
+      Registry - azureml://registries/<registry-name>/environments/<environment-name>/versions/<version>
+      Workspace - azureml:<environment-name>:<version>
+
+  inference_payload:
+    type: uri_file
+    optional: true
+    description: JSON payload which would be used to validate deployment
+
+  inference_payload_str:
+    type: string
+    optional: true
+    description: Serialized JSON payload which would be used to validate deployment
+
+  endpoint_name:
+    type: string
+    optional: true
+    description: Name of the endpoint
+
+  deployment_name:
+    type: string
+    optional: true
+    default: default
+    description: Name of the deployment
+
+  instance_type:
+    type: string
+    optional: true
+    enum:
+      - Standard_DS1_v2
+      - Standard_DS2_v2
+      - Standard_DS3_v2
+      - Standard_DS4_v2
+      - Standard_DS5_v2
+      - Standard_F2s_v2
+      - Standard_F4s_v2
+      - Standard_F8s_v2
+      - Standard_F16s_v2
+      - Standard_F32s_v2
+      - Standard_F48s_v2
+      - Standard_F64s_v2
+      - Standard_F72s_v2
+      - Standard_FX24mds
+      - Standard_FX36mds
+      - Standard_FX48mds
+      - Standard_E2s_v3
+      - Standard_E4s_v3
+      - Standard_E8s_v3
+      - Standard_E16s_v3
+      - Standard_E32s_v3
+      - Standard_E48s_v3
+      - Standard_E64s_v3
+      - Standard_NC4as_T4_v3
+      - Standard_NC6s_v2
+      - Standard_NC6s_v3
+      - Standard_NC8as_T4_v3
+      - Standard_NC12s_v2
+      - Standard_NC12s_v3
+      - Standard_NC16as_T4_v3
+      - Standard_NC24s_v2
+      - Standard_NC24s_v3
+      - Standard_NC24rs_v3
+      - Standard_NC24ads_A100_v4
+      - Standard_NC48ads_A100_v4
+      - Standard_NC96ads_A100_v4
+      - Standard_NC64as_T4_v3
+      - Standard_ND40rs_v2
+      - Standard_ND96asr_v4
+      - Standard_ND96amsr_A100_v4
+    default: Standard_NC24s_v3
+    description: Compute instance type to deploy model. Make sure that instance type is available and have enough quota available.
+
+  instance_count:
+    type: integer
+    optional: true
+    default: 1
+    description: Number of instances you want to use for deployment. Make sure instance type have enough quota available.
+
+  max_concurrent_requests_per_instance:
+    type: integer
+    default: 1
+    optional: true
+    description: Maximum concurrent requests to be handled per instance
+
+  request_timeout_ms:
+    type: integer
+    default: 60000
+    optional: true
+    description: Request timeout in ms. Max limit is 90000.
+
+  max_queue_wait_ms:
+    type: integer
+    default: 60000
+    optional: true
+    description: Maximum queue wait time of a request in ms
+
+  failure_threshold_readiness_probe:
+    type: integer
+    default: 10
+    optional: true 
+    description: The number of times system will try after failing the readiness probe
+
+  success_threshold_readiness_probe:
+    type: integer
+    default: 1
+    optional: true 
+    description: The minimum consecutive successes for the readiness probe to be considered successful after having failed
+
+  timeout_readiness_probe:
+    type: integer
+    default: 10
+    optional: true
+    description: The number of seconds after which the readiness probe times out
+
+  period_readiness_probe:
+    type: integer
+    default: 10
+    optional: true
+    description: How often (in seconds) to perform the readiness probe
+
+  initial_delay_readiness_probe:
+    type: integer
+    default: 10
+    optional: true
+    description: The number of seconds after the container has started before the readiness probe is initiated
+
+  failure_threshold_liveness_probe:
+    type: integer
+    default: 30
+    optional: true 
+    description: The number of times system will try after failing the liveness probe
+
+  timeout_liveness_probe:
+    type: integer
+    default: 10
+    optional: true
+    description: The number of seconds after which the liveness probe times out
+
+  period_liveness_probe:
+    type: integer
+    default: 10
+    optional: true 
+    description:  How often (in seconds) to perform the liveness probe
+
+  initial_delay_liveness_probe:
+    type: integer
+    default: 10
+    optional: true
+    description: The number of seconds after the container has started before the liveness probe is initiated
+
+  egress_public_network_access:
+    type: string
+    default: enabled
+    optional: true 
+    enum:
+      - enabled
+      - disabled
+    description: Setting it to disabled secures the deployment by restricting communication between the deployment and the Azure resources used by it
+
+outputs:
+  model_deployment_details:
+    type: uri_file
+    description: Json file to which deployment details will be written
+  model_inference_response:
+    type: uri_file
+    description: JSON file containing inference results
+  deploy_error:
+    type: uri_file
+    description: File containing error messages or stack traces from the validation step.
+
+tags:
+    Preview: ""
+    Internal: ""

assets/common/src/deploy.py

@@ -338,4 +338,4 @@ def main():
 # run script
 if __name__ == "__main__":
     # run main function
-    main()
+    main()