A parser is trivial. It's XML and you have a schema.
What you want is a compiler (e.g., into a different document format) or an interpreter (e.g., for running a search or a spell checker).
That's a task that's massively complicated because you cannot give an LLM the semantic definition of the XML and your target (both typically are under documented and under specified). Without that information, the LLM would almost certainly generate an incomplete or broken implementation.
What you want is a compiler (e.g., into a different document format) or an interpreter (e.g., for running a search or a spell checker).
That's a task that's massively complicated because you cannot give an LLM the semantic definition of the XML and your target (both typically are under documented and under specified). Without that information, the LLM would almost certainly generate an incomplete or broken implementation.