A streamProcessor for creation of a Custom Text Classification Model

Needs to prepare the data, store it in an S3 location and call Sagemaker with the correct Docker image.

Docker image is selected based on the model definition created by the user.

Returs the reference for the model in SageMaker, and several metainfo regarding the training : Accuracy, etc
The sagemaker reference should be used later to create an endpoint for use by the enrichment pipeline