Figure out productive categories to be auto-detected from source code

[lpmi-13]

So in thinking about how to extract examples of different things from source code for learners to work with (what I did in https://examplar-code.netlify.app, but WAY more fleshed out), saw a list of things from javascript that would be cool to isolate and extract.

And obviously variables are easy to grab via abstract syntax tree parsing, and things like destructuring and rest/spread should be possible to filter out, since the forms are fairly predictable (actually, maybe the AST parsers can already just find these...haven't looked too close yet)...

But how to deal with stuff like state/immutability?

So the idea is basically to just compile on how to deal with these...or even which ones are worth trying to isolate.