pub fn decode_lm_studio_stream( runner: Symbol, model: &str, body: &[u8], include_raw: bool, ) -> Result<Expr>
Decodes LM Studio OpenAI-compatible SSE chunks into a response transcript.