Real-time inference

If you remember, earlier in this chapter, we added a placeholder in our decoding lambda code. That was specifically placed to allow us to leverage the model endpoint as real-time data is processed. The following set of code can now be added to reference our endpoint and evaluate incoming data:

### run against the model to calculate an anomaly score.
    data = {"current":str(event["current"])}
    payload = data['current']
    print("payload: ", type(payload), payload)
    runtime= boto3.client('runtime.SageMaker')
    response = runtime.invoke_endpoint(EndpointName="current-rcf-2022-09-26-06-56", ContentType='text/csv', Accept='application/json', Body=payload)
    print(response)
    result = json.loads(response['Body'].read().decode...