Inference Request APIs

The Clara DICOM Adapter supports the following RESTful APIs on (default) port 5000.

POST /api/inference

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

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.

Parameters

Please see the InferenceRequest class definition for examples.

Request Content Type: JSON

Name	Type	Description
transactionID	string	Required. User provided transaction ID for correlating an inference request.
priority	number	Valid range 0-255. Please refer to Nvidia.Clara.DicomAdapter.API.Rest.InferenceRequest.Priority for details.
inputMetadata	inputMetadata object	Required. Specifies the dataset associated with the inference request.
inputResources	array of inputResource objects	Required. Data sources where the specified dataset to be retrieved. Clara 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.
422	Request contains invalid data or is missing required fields.
500	Server error.

---

GET /api/inference/status/{id}

Retrieves status of an inference request.

Parameters

The transaction ID or the Clara Job ID must be provided as part of the request URI.

Responses

Response Content Type: JSON - InferenceStatusResponse.

Code	Description
200	Inference request status is available.
404	Inference request cannot be found.
500	Server error.