Skip to content

Latest commit

 

History

History
83 lines (60 loc) · 6.26 KB

inference.md

File metadata and controls

83 lines (60 loc) · 6.26 KB

Inference Request APIs

The inference endpoint provides a RESTful interface for triggering a new inference request that is compatible with the ACR DSI API.

Warning

This API is a work in progress and may change between releases.

Note

The inference API is extended based on the draft created by the ACR (American College of Radiology). Please refer to ACR's Platform-Model Communication for AI for more information.

POST /inference

Triggers a new inference job using the specified DICOM dataset from the specified data sources.

Important

For input and output connections that require credentials, ensure that all the connections are secured and encrypted.

Parameters

See the InferenceRequest class definition for examples.

Request Content Type: JSON

Name Type Description
transactionID string (Required) A user-provided transaction ID for correlating an inference request.
priority number The valid range is 0-255. Please refer to Monai.Deploy.InformaticsGateway.Api.Rest.InferenceRequest.Priority for details.
inputMetadata inputMetadata object (Required) The dataset associated with the inference request.
inputResources array of inputResource objects (Required) Data sources from which to retrieve the specified dataset. MONAI Deploy Only: Must include one interface that is type of Algorithm.
outputResources array of inputResource objects (Required) Output destinations where results are exported to

Responses

Response Content Type: JSON - InferenceRequestResponse.

Code Description
200 Inference request received and scheduled for processing.
409 A request with the same transaction ID already exists. The response will be a Problem details object with server error details.
422 Request contains invalid data or is missing required fields. The response will be a Problem details object with server error details.
500 Server error. The response will be a Problem details object with server error details.

GET /inference/status/{transactionId}

Returns the status of an inference request.

Parameters

Name Type Description
transactionId string The transactionId of the original inference request

Responses

Response Content Type: JSON - InferenceStatusResponse.

Code Description
200 Inference request status is available.
404 Inference request not found.
500 Server error. The response will be a Problem details object with server error details.