Think about why you're wanting to do this. Is there a material benefit you'll get from splitting your codebase like so? Enough to overcome the fact that you've split your codebase into two very different languages?
When you're going between languages like this, you either need some kind of communication protocol (e.g. JRPC, TCP, or maybe something home cooked) or you need a stable ABI to allow the programs to talk to each other directly.
My point is, you probably don't want to do what you're trying to do. Unless you have a really good reason, pick one language and stick with it.