Support for Kinesis streams for SPARK computations

[XxChauhanxX]

Child of #1089

As a Product owner I want to make Tempus application in Native AWS, since most of the customers prefers and trusts the Native AWS architecture and applications.

ACs:

1. Need to start supporting Kinesis streaming for SPARK computations.

[pal-abhishek]

Possible approach to do the same:-

-> Existing computations will be refactored to implement three steps of spark computation (load, transform and extract).
-> The loaders will be plug-able (kafka or kinesis) depending upon the selected value on UI.
-> Descriptor will have parameters for both kafka and kinesis.
-> The form loading of the UI will have some changes depending on the selected value (kafka or kinesis). We need to the values associated to the specific stream.
-> The contract of loader and transform can be defined in a separate module in Tempus-Extensions. The contract will be used to implement loader or transformer (current computations). The module can be repackaged in Uber jar for uploading in Tempus.
-> Combining the three implementation spark Pipeline can be created Ex - Pipeline(KafkaLoader, drillBitCompulationTranformer, MqttPublisher).
-> These will be done by driver, only the transformer will use the parallel processing across nodes.