Package com.codedstream.otterstream.inference.model

Class InferenceResult

java.lang.Object

com.codedstream.otterstream.inference.model.InferenceResult

public class InferenceResult
extends Object

Container for ML inference results including predictions and metadata.

Represents the output from an inference operation, containing the model's predictions, timing information, and success status.

Result Types:

• Success: Contains predictions in outputs map
• Failure: Contains error message and empty outputs

Usage Examples:

Creating Success Result:

 Map<String, Object> outputs = Map.of(
     "probability", 0.92,
     "class", "fraudulent",
     "confidence", 0.88
 );
 InferenceResult result = new InferenceResult(outputs, 45, "fraud-detector");
 

Creating Failure Result:

 InferenceResult result = new InferenceResult(
     "fraud-detector",
     "Model execution timeout",
     5000
 );
 

Consuming Results:

 if (result.isSuccess()) {
     double prob = result.getOutput("probability");
     String prediction = result.getOutput("class");
     System.out.println("Prediction: " + prediction + " (" + prob + ")");
 } else {
     log.error("Inference failed: {}", result.getErrorMessage());
 }
 

Since:

1.0.0

Author:

Nestor Martourez, Sr Software and Data Streaming Engineer @ CodedStreams

Constructor Summary

Constructors

Constructor	Description
InferenceResult(String modelId, String errorMessage, long inferenceTimeMs)	Creates a failed inference result.
InferenceResult(Map<String,Object> outputs, long inferenceTimeMs, String modelId)	Creates a successful inference result.

Method Summary

Modifier and Type	Method	Description
String	getErrorMessage()	Gets error message for failed inferences.
long	getInferenceTimeMs()	Gets inference duration.
String	getModelId()	Gets model identifier.
<T> T	getOutput(String key)	Gets a specific output value by name.
Map<String,Object>	getOutputs()	Gets all output predictions.
boolean	isSuccess()	Checks if inference was successful.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

InferenceResult

public InferenceResult(Map<String,Object> outputs,
                       long inferenceTimeMs,
                       String modelId)

Creates a successful inference result.

Parameters:

outputs - map of output name to predicted value

inferenceTimeMs - inference duration in milliseconds

modelId - identifier of the model that made the prediction

InferenceResult

public InferenceResult(String modelId,
                       String errorMessage,
                       long inferenceTimeMs)

Creates a failed inference result.

Parameters:

modelId - identifier of the model that failed

errorMessage - description of the failure

inferenceTimeMs - time spent before failure in milliseconds

Method Detail

getOutputs

public Map<String,Object> getOutputs()

Gets all output predictions.

Returns:

immutable map of output name to value

getInferenceTimeMs

public long getInferenceTimeMs()

Gets inference duration.

Returns:

duration in milliseconds

getModelId

public String getModelId()

Gets model identifier.

Returns:

model ID

isSuccess

public boolean isSuccess()

Checks if inference was successful.

Returns:

true if successful, false if failed

getErrorMessage

public String getErrorMessage()

Gets error message for failed inferences.

Returns:

error message, or null if successful

getOutput

public <T> T getOutput(String key)

Gets a specific output value by name.

Convenience method with type casting.

Type Parameters:

T - expected type of the output

Parameters:

key - output name

Returns:

output value cast to type T, or null if not found

Throws:

ClassCastException - if type T doesn't match actual type