Refactor SonarCloud showcase test coverage to show meaningful results

[mpeddada1]

We have SonarCloud set up to track showcase coverage for both unit tests and integration tests. Integration test coverage can help with determining how much of end-to-end behavior we are testing and also provide more relevant information as showcase ITs are handwritten. On the other hand, the unit tests are generated.

Tracking unit test coverage can sometimes add noise to the PRs. This is particularly relevant when we are upgrading grpc which results in the re-generation of classes that are generated by the gRPC proto compiler (example: #1989). This makes it a little difficult to determine whether the coverage result is actionable or not.

Proposed changes after discussion with @burkedavison to ensure that test coverage results are more meaningful:

• Remove unit test coverage.
• Refactor integration test coverage to only include the gapic-showcase/ folder. This will allow us to focus on coverage for code generated by gapic-generator-java.

[JoeWang1127]

Refactor integration test coverage to only include the gapic-showcase/ folder.

Hi @mpeddada1, do you mean exclude api-common and gax-java when analyzing showcase integration test coverage?

[burkedavison]

@JoeWang1127 : Showcase integration test coverage currently includes class coverage in:

• <root>/showcase/gapic-showcase
• <root>/showcase/grpc-gapic-showcase-v1beta1
• <root>/showcase/proto-gapic-showcase-v1beta1

This should be modified to only include the classes generated by gapic-generator-java. These are found exclusively in <root>/showcase/gapic-showcase/src/main.