Add JobTemplates and Flink demo

[ryannedolan]

Summary

• Added JobTemplates
• Added demo flink-template
• Minor planner improvements

Details

JobTemplates, along with TableTemplates, are used by the planner to dynamically assemble a pipeline. The pipeline includes all source tables, the sink table, and any jobs. By "installing" a JobTemplate, you can enable Hoptimator to deploy a physical job (or indeed multiple jobs) as part of a pipeline.

As an example, a basic flink-template is provided, which deploys a FlinkDeployment for each pipeline.

Testing

Integration tests have been updated to account for the job spec.

When we create a materialized view:

0: Hoptimator> create materialized view kafka."test1" as select * from kafka."existing-topic-1";

... we can see that a FlinkDeployment is created:

$ kubectl get flinkdeployments
NAME                   JOB STATUS    LIFECYCLE STATE
kafka-database-test1   RECONCILING   DEPLOYED

The pipeline controller can be seen checking on the status of each source, sink, and job in the pipeline:

Checking status of Pipeline kafka-test1...
Object kafka/KafkaTopic/kafka-database-existing-topic-1 considered ready by default.
KafkaTopic/kafka-database-existing-topic-1 is ready.
Object kafka/KafkaTopic/kafka-database-test1 considered ready by default.
KafkaTopic/kafka-database-test1 is ready.
Pipeline kafka-test1 is NOT ready

[jogrogan]

LGTM