Pystac v1.0 compatability

[gadomski]

Feedback appreciated, especially on the StacIO stuff, where I did my best to just port over the old behavior to The New Way™.

• IO is broken until Remove duplicate read_text without orjson pystac#348 lands in the next beta release
• [ ] Pystac 1.0 resolves parent collections, so some landset tests are broken because https://landsatlook.usgs.gov/data/collection02/landsat-c2l2-sr.json is not publicly accessible.
• [ ] Other landsat tests break because landsat schema is no longer accessible: https://landsat.usgs.gov/stac/landsat-extension/schema.json.
• Aster and CGLS_LC100 both do not validate against the eo:bands schema. CGLS_LC100 has "Change-Confidence-layer" for a common name, and aster has 'yellow/green'.

cc @volaya since you said you might be working on this effort as well.

[gadomski]

Ok, I've fixed up stuff to use extensions correctly (hopefully). There's some issues with a couple subpackages, some of which are due to STAC v1 validation, some of which are landsat infrastructure issues. I've cataloged them as best as I can in this PR's description so we can try to axe them down.

For the aster+cgls_lc100, @lossyrob can you advise on what we should do about the band names?

The landsat stuff is trickier -- relying on someone's else's schemas and collections as a part of our test suite seems fragile.

[lossyrob]

For Aster, it looks like the center wavelength falls in the "green" category from this table, so I'd go with that....

For CGLS_LC100 I'd remove the common_name

[gadomski]

Updated for the v0.2 reorganization (makes this PR a lot smaller).

[gadomski]

@duckontheweb mind giving this one a once-over? It's much simpler now w/o the dataset packages.

[duckontheweb]

Looks good to me, just a couple of minor comments!

[duckontheweb]

👍