Having true JSON Schema[0] is being considered, but on the back-burner due to the sheer number of parsers to build schemas for. Also, it is more difficult to accurately define the schema for a small subset of parsers since their command output are so variable.
I’ve been trying to build something like this but simply don’t have the free time currently.
The plan: adapt the parser VM from lpeg (or similar, there’s a paper I’ve been reading on an Earley parser VM) into a command line app that takes a grammar + text input (or stdin) and spits out json to a file (or stdout). Probably not as general purpose as this one but also wouldn’t need a pull request to add a new format.
All the pieces are there but without the free time…
I was actually curious if there was any demand for such a thing, I just want it to parse my payroll statements because this billion dollar company can only manage crappy pdfs and, well, it’s an interesting problem.
—edit—
Oh, output schema. Totally different than what I’m going on about.