Scaffolding for outputting anomaly prediction

[c0c0n3]

Summary

Put in place some basic infra to be able to process incoming shop floor data and output a forecast about possible anomalies.

Intended outcome

On receiving shop floor data form Context Broker, the service writes a prediction back to Context Broker. The prediction is encoded in an NGSI entity. Note we don't need to implement the actual ML logic for this task. All we need is a predict stub function that takes in the NGSI entity sent by Context Broker and outputs a prediction NGSI entity.

How will it work?

Similar to what we've done for roughnator.

[c0c0n3]

done, see code at 601f25e.