Xarray and Sliderule IO Improvements

[scottyhq]

• Closes Create Altimeter Point Plotting Function #21, Update COP-30 Search to Pull DEM #22, Update WorldCover Search to Return LULC Raster #23
• Tests added
• Tests passing
• Full type hint coverage
• Changes are documented in docs/releases.rst
• New functions/methods are listed in api.rst
• New functionality has documentation

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[scottyhq]

@Jack-Hayes I'm going to merge this since it has useful functions for AGU prep. Note this takes largely takes care of what you started in #30 but still would love your help with tests, docs, and any features you think are still missing (like the polygon validation).

[Jack-Hayes]

Gotcha! I'm hoping to implement that test and an example notebook for the docs once I make the loading faster with dask if that sounds good to you

[dshean]

Great to explore dask, but it should be optional.

Hoping we can offer options to directly load overviews (instead of full-res raster) from COGs, or load coarser raster products.

[scottyhq]

Great to explore dask, but it should be optional.

Yes, could certainly revisit choice to use dask by default here (but can override by passing chunks=None)

coincident/src/coincident/io/xarray.py

Line 65 in 7ce2087

ds = odc.stac.load(items, bands=bands, geopolygon=aoi, chunks={}, **kwargs)

I'm not sure if it's better or not to include the kitchen sink of dependencies in this project or to keep a lot of the I/O and visualization dependencies as optional... Including everything is the current approach but may get complicated (for example if we add PDAL for custom 3DEP workflows, which pins older arrow & hdf versions which i'd like to avoid b/c of search relying on recent geoparquet/arrow)...

coincident/pyproject.toml

Lines 70 to 74 in 7ce2087

	]
	# all = [
	# "matplotlib-base>=3.9.2,<4",
	# "sliderule>=4.7.1,<5"
	# ]

Hoping we can offer options to directly load overviews (instead of full-res raster) from COGs, or load coarser raster products.

With odc-stac, reading from appropriate overviews should happen automatically by setting the resolution= keyword, regardless of dask usage (for example https://coincident.readthedocs.io/en/latest/examples/contextual_data.html#load-raster-datasets-with-xarray)

More discussion of reading overviews including confirming they are used via logging is here gjoseph92/stackstac#196