How to formulate quadratic optimization in the json model?

[AncientSwordRage]

JWs comment on this issue #21 implies that this is not possible, but I can't figure out how.

This is my use of this library, https://github.com/AncientSwordRage/ponder

I want to extend my code to allow the model to essentially say: If you have Card X that works with Cards with Attribute Y, it's more valuable the more Cards with attribute Y we have.

I asked here and it looks like this is a linearisable quadratic problem.

But I'm stumped on how to create the right model for this.

Any advice?