Add stream computing platform

[stingerpk]

Since data coming into Kafka topics from divolte needs to be joined and/or augmented, there is a need for stream computing before data is pushed to druid. Since druid is configured to read from Kafka topics, the transformations are going to be Kafka --> Kafka. Therefore it makes sense to use Kafka Streams.

• Add Confluent schema registry and REST proxy
• Write code to deserialise and serialise Avro
• Write code for a simple lookup and data augmentation to solve 6 digit postal code match for US and Canada #1