Update NetCDF/HDF5 page with more recent information.

[wildintellect]

https://guide.cloudnativegeo.org/cloud-optimized-netcdf4-hdf5/

Needs an update based on the latest work by the community. See ESIP talks and discussions for an in depth overview of why and general how https://www.youtube.com/watch?v=hpvVn7CSIaM&t=3183s This needs to be translated into the page for more people to reference what they need to do.

Updates

• It's no longer a mystery how to pack an HDF5 file efficiently with a better than default chunk size. Completely fine to mention that chunk size is still very dependent on the data and use case.
• It's no longer a mystery that when reading a HDF, cloud-optimized or not, there are certain cache settings that should be passed to the HDF library to make i/o more efficient. We should put up some examples demonstrating this @abarciauskas-bgse @chuckwondo @maxrjones
• We should show at a minimum, how to optimize with h5 cli tools and python, and how to pass the correct settings when reading files with tools like fsspec+xarray.
• User h5 cli tools to read metadata and display the relevant metadata, e.g. chunk size.

Some reference material:

• h5repack with atl03 https://gist.github.com/abarciauskas-bgse/8bf4388f8f8989582c807b2451c5cf8c
• Python cloud optimization in ISCE3 , calculating chunk sizes

@betolink do you have some slides, graphics, or code samples you could share?

[betolink]

I can create a folder with some of the Excalidraw illustrations I've used, you can take anything useful from the slides (including the memes 😆)

ESIP slides: https://docs.google.com/presentation/d/1uyzpJUrsBaL-Ks3TJc9sdtyq1LyJtBEhmq1Cjvq8U_o/edit#slide=id.p SciPy poster: https://docs.google.com/presentation/d/15bPnFEGJvcwBTIJr3AIDODSDJF5QRu7L1Esd5_his9Q/edit#slide=id.p