general structure-based selectors or matchers

[yshavit]

Do I want selectors or matchers that work off of the general structure of the doc, without reference to specific items? For example, # [0] for "the first section", or @parent for the parent element?

My starting position has been no; that jq already does this very well, so I should just let people use --json and then have jq take it from there.

But link references complicate this a bit. Let's say I have the following doc:

# Section one

My [first link][1].

# Section two

My [second link][2].

[1]: https://example.com/1
[2]: https://example.com/2

In this case, picking just the first section means that the output would not contain the [2]: https://example.com/2 link. If I just piped it to jq '{section: .items[0], links: .links}', then I'd get just the first section, but both links. At the jq level, it's not very easy to filter away the links that aren't needed anymore.

This is a low-priority enhancement for now, and I'd probably want to get a couple use cases before I go down this route, to make sure I get the right abstraction.