Just a quick example showing GraphQL Json structures being parsed into JsonObjects from a raw GraphQL request made by Retrofit and handled via RxJava.
The main question around the use of RxJava at the data layer here is why you wouldn't just use the built-in asynchronous request features of Retrofit2 directly. RxJava is better suited for performing operations on multiple streams of data making the extra boilerplate worth it for those cases.
With GraphQL, you're usually going to get the entire data structure in one shot. The majority of the work there is to map the fields to the views.
You could go really quick and dirty and just parse out JsonObjects from the responses and map the fields directly to the view, or you could use the Gson converter and supply types to be mapped-to before sending off to the view.
Either way, the schema is the source of truth and if it changes the models/bindings/queries have to as well. Since we're writing GQL queries directly in as strings, there's not a lot of boilerplate necessary to get up and running.
The majority of the work on the receiving end would really be the field mapping in almost any scenario. Doing stream work on the resulting data sets may be necessary in some cases for the future, but luckily that's an easy thing to just plug in.
gradle run
Watch terminal output for stringified JSON object containing Account #1 from
the Waldo GraphQL API.