Proposal: Separate protobuf and json parsers

[wi101]

Hey 👋

We're using teleproto in a project to convert to protobuf and we use Spark. We got this error:

ERROR SparkHadoopWriter: Aborting job job_20221214135935241508088406623081_0074.
org.apache.spark.SparkException: Job aborted due to stage failure: Task serialization failed: java.lang.NoClassDefFoundError: org/json4s/JsonAST$JValue
java.lang.NoClassDefFoundError: org/json4s/JsonAST$JValue

Because Spark is using an older json4s version
we had to exclude the "com.thesamet.scalapb" %% "scalapb-json4s to make it work.

• Currently:
  VersionedModelReader has fromJson and fromProto
  and VersionedModelWriter has toJson and toProto

• Suggestion:
  To have a separate module for json and for proto that will be published in separate jars

[saeltz]

Hey Wiem,

thanks for your proposal!

We're actively investigating support for Scala 3 at the moment and don't want to spend too many resources on the Scala 2 version.

An easier alternative might be to declare the dependency on scalapb-json4s as optional in this library.
Do you think this would solve your issue? If so, I'd happily do the change. If not, I'll take another look at modularising the library.

[wi101]

I think the suggested alternative with optional dependency would be helpful :) thank you!