creek-service / creek-kafka / #4432

97%

push

github

Pre work for JSON serde (#429)

* Pre work for JSON serde

- Add TypeOverrides interface as a way of users to provide overrides for specific types. This is more flexible than having specific methods to override specific types. This also allows the `TopicClient` to be made private, as only the test helps should be overriding this.

- Switch `KafkaTopicClient` to be per-cluster, as this is simplifies the code.

- Change how `KafkaClientsExtensionProvider` initialises, moving the logic for creating serde into the Topic resource handlers new `prepare` method.  This ensures we don't type to create serde until any external resources, e.g. schema, have been created.

- Move initialisation of serde resources out of `TopicClient` and into `TopicResourceHandler`.

- Add `ClustersSerdeProviders` as a way for serde providers to be initialised once per cluster, allowing for shared state per cluster, e.g. a schema registry client.  This is handled in `KafkaSerdeProvider` by having an `initialize` method that returns a new `SerdeProvider` type.

- Dropped `TearDownTestListener` as system tests now ensure extensions are closed.

151 of 156 new or added lines in 16 files covered. (96.79%)

1951 of 1985 relevant lines covered (98.29%)

0.98 hits per line