Add support for JSON/camelcase conversion

[malthe]

This adds a new configuration boolean configuration "json.field.names" (defaults to false) which controls whether to convert to JSON/camelcase from protobuf field names (which are underscore-separated by convention) – in toConnectData.

Conversely, going from the internal data representation, if this option is enabled, we convert from JSON/camelcase back to underscore-separated – in fromConnectData.

Note that this option is compatible with the Protobuf Style Guide in the sense that for some languages (e.g. Java), the underscore field naming is translated to camelcase-based methods.

[rayokota]

Thanks for the PR @malthe . Perhaps when going from Protobuf to Connect, it might be better to preserve the original field name as a Connect schema parameter. Then when going from Connect to Protobuf, we can just use the value of such a parameter, if it exists.

[malthe]

@rayokota fwiw, in parallel to this, we have contributed improvements to the ChangeCase transform from https://www.confluent.io/hub/jcustenborder/kafka-connect-transform-common/ such that it works on nested data structures.

This is a more generic approach and the performance will be slightly lower due to having to traverse each data structure.

[cla-assistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}