Using the machine learning endpoint as an adopted application

These deployed endpoints are accessible in API Gateway to:
  • Get a list of registered endpoints
  • Get the schema of an endpoint
  • Retrieve a prediction for a given endpoint.
  • Retrieve a detailed prediction for a given endpoint

This table shows the process of calling the endpoints through API Gateway:

Endpoint Coleman/v1/endpoints Coleman/v1/endpoints/{endpoint name}/schema Coleman/v1/endpoints/{endpoint name}/prediction Coleman/v1/endpoints/{endpoint name}/detailed-prediction
Method GET GET POST POST
Type List Schema Prediction Detailed prediction
Description Get a list of endpoints.

Retrieve a list of registered endpoints.

 Get the schema for a given endpoint.

Retrieve the endpoint schema based on the endpoint name. Use the schema to create the input for the prediction method.

 Retrieve a prediction for a given endpoint.

Provide input data for a given endpoint and retrieve prediction based on the input values.

 Retrieve a detailed prediction for a given endpoint.

Provide input data for a given endpoint and retrieve prediction based on the input values.

Input data is expected in key-value pairs according to the endpoint schema definition. The response is returned in the same format.
Input Status: Active endpoint name endpoint name

Input values for prediction put into the JSON. For example:

{
    data: [
        [value1, value2, value3, value4],
        [value1, value2, value3, value4],
        ...
    ]
}

Bulk & single upload of predictions schema

 endpoint name

Input values for prediction put into an array of JSONs.

For example:

[
    {
        "prediction1_input_key": "value",
        "prediction1_input_key2": "value2"
    },
    {
        "prediction2_input_key": "value",
        "prediction2_input_key2": "value2"
    }
]
bulk & single upload of predictions schema
Output
  • endpoint Id custom
  • name
  • status
  • failure code
  • created on
  • created by
  • last updated on
  • last updated by
  • quest
  • quest id
 JSON schema for selected endpoint name predictions input data with predictions, everything in the key-value pair format. Output of the example from above:
[
    {
        "prediction1_input_key": "value",
        "prediction1_input_key2": "value2",
        "prediction1_output_key": "prediction_value"
    },
    {
        "prediction2_input_key": "value",
        "prediction2_input_key2": "value2",
        "prediction2_output_key": "prediction_value"
    }
]