Revamp at STAC sprint

I still think intake-stac serves an interesting role. Intake is domain-agnostic and python-specific whereas STAC is geo-specific and language-agnostic. So it is totally reasonable that someone might be more comfortable accessing all their data geo and non-geo via intake.

This repo has been a little neglected over the past year or so and the ecosystem has come a long way. There is a STAC sprint next week and it might be a good opportunity to tidy things up. These are the things I am thinking of:

- Triage open issues
- Triage open PRs
- Bring linters/actions/deps uptodate
- Check the geopandas read logic vs what's in https://github.com/TomAugspurger/staccontainers
- Remove dependency on intake-xarray
- Switch the xarray read to just use xpystac (which delegates stacking to stackstac or odc-stac)

Just wanted to put this up in case it's not a good time for some reason (maybe because of https://github.com/intake/intake/pull/737 @martindurant)