I want to do as much coding in Elixir as possible, even when extending my Rails app. Does anyone have any good interop stories?

The Rails app is the front-end presenting the results of an ETL process. (Publishing laws online.) I’ve been using the strategy of slowly moving functions away from it when I can and introducing JSON or HTTP interfaces. I’m wondering if there’s something in-process and tighter and faster that’s possible for calling Elixir code.

Another idea is to create a CLI elixir app which performs parsing and formatting functions — and I can shell out to it from my Rake import tasks. That’s a little sub-optimal both in speed and coding, but time isn’t a big priority in these tasks. I’m thinking I could make a Ruby wrapper around the Elixir CLI invocation to check the Unix exit code and convert to exceptions…