[Parser] Decouple code from Swagger-Core

I would like to investigate if it is possible to replace usage of

<dependency>
    <groupId>io.swagger.core.v3</groupId>
    <artifactId>swagger-models</artifactId>
    <version>2.0.6</version>
</dependency>

With the interfaces defined in Eclipse MicroProfile:

<dependency>
    <groupId>org.eclipse.microprofile.openapi</groupId>
    <artifactId>microprofile-openapi-api</artifactId>
    <version>2.0-MR1</version>
</dependency>

To model existing OpenAPI specifications, the Eclipse MicroProfile is using a model that is really similar to what Swagger-Core is providing.
But it follows the JavaEE principle: there is a clear API with interfaces and different vendors can implement them.

With the release of the EMPOA project, we have a way to continue to use Swagger-Parser as we did before. EMPOA Swagger-Core is a bridge that implements the Eclipse MicroProfile interface by delegating to the Swagger-Core objects.

With this approach, the output of Swagger-Parser can be consumed as valid Eclipse MicroProfile implementation:

io.swagger.v3.oas.models.OpenAPI swaggerOpenAPI = parserResult.getOpenAPI();
OpenAPI openAPI = SwAdapter.toOpenAPI(swaggerOpenAPI);

Advantage of this approach: decoupling!

The EMPOA project could implement a second wrapper for the model used by KaiZen-OpenAPI-Parser.
This way we would have an alternative for corner cases that are not well supported by Swagger-Parser.

Start by tracing Swagger-Parser's parserResult.getOpenAPI() output and its use of the swagger-models dependency. Compare that model with the MicroProfile OpenAPI interfaces and the EMPOA Swagger-Core bridge. Done means determining whether the direct Swagger-Core dependency can be replaced while preserving consumption through the MicroProfile model.