In v3.1 I believe some customers have created "proxy" semantic parsers which can switch to other language parsers when they find a language change token, etc.
This specific feature is something we started working on for v4.0 and have it somewhat working although we plan to focus on it more after the public beta is out shortly. We had it at a point where it would automatically transition from one semantic parser for a language to another with no work on your end to achieve that. Then it would combine the semantic parse results from each language transition (which were ASTs) into one large AST. We still have some other things left to do to finalize that feature but it's really cool so far.