Support new netcdf4 1.6.0 compression arguments

[markelg]

Is your feature request related to a problem?

When using the netcdf4 engine, I am not able to use the new "compression" argument to choose a compression scheme different from zlib in the encoding.

        if raise_on_invalid:
            invalid = [k for k in encoding if k not in valid_encodings]
            if invalid:
>               raise ValueError(
                    f"unexpected encoding parameters for {backend!r} backend: {invalid!r}. Valid "
                    f"encodings are: {valid_encodings!r}"
                )
E               ValueError: unexpected encoding parameters for 'netCDF4' backend: ['compression']. Valid encodings are: {'fletcher32', 'zlib', 'contiguous', 'dtype', 'least_significant_digit', 'shuffle', 'complevel', '_FillValue', 'chunksizes'}

../../../../netCDF4_.py:279: ValueError

Furthermore, according to the release notes of 1.6.0, zlib argument is to be deprecated:

* add 'compression' kwarg to createVariable to enable new compression
   functionality in netcdf-c 4.9.0.  'None','zlib','szip','zstd','bzip2'
   'blosc_lz','blosc_lz4','blosc_lz4hc','blosc_zlib' and 'blosc_zstd'
   are currently supported. 'blosc_shuffle',
   'szip_mask' and 'szip_pixels_per_block' kwargs also added.
   compression='zlib' is equivalent to (the now deprecated) zlib=True.
   If the environment variable NETCDF_PLUGIN_DIR is set to point to the
   directory with the compression plugin lib__nc* files, then the compression plugins will
   be installed within the package and be automatically available (the binary
   wheels have this).  Otherwise, the environment variable HDF5_PLUGIN_PATH
   needs to be set at runtime  to point to plugins in order to use the new compression
   options.

I am using the last versions

xarray: 2022.6.0
pandas: 1.4.2
numpy: 1.22.4
scipy: None
netCDF4: 1.6.0
pydap: None
h5netcdf: None
h5py: None
Nio: None
zarr: None
cftime: 1.6.0
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: None
dask: 2022.6.0
distributed: 2022.6.0
matplotlib: 3.5.2
cartopy: None
seaborn: None
numbagg: None
fsspec: 2022.5.0
cupy: None
pint: None
sparse: None
flox: None
numpy_groupies: None
setuptools: 62.6.0
pip: 22.1.2
conda: None
pytest: 7.1.2
IPython: 8.4.0
sphinx: 5.1.1

Describe the solution you'd like

Update the netcdf4 backend to support these arguments. Should not be too difficult.

Describe alternatives you've considered

No response

Additional context

I can try to do this myself, it does not look hard.

[dcherian]

I think you just need to add compression here:

xarray/xarray/backends/netCDF4_.py

Lines 236 to 245 in 63d7eb9

	valid_encodings = {
	"zlib",
	"complevel",
	"fletcher32",
	"contiguous",
	"chunksizes",
	"shuffle",
	"_FillValue",
	"dtype",
	}

We'll want a test too in test_backends.py

[markelg]

Thank you @dcherian. I opened a pull request, please take a look. I added "compression" to the list, but it will be removed if a version of netCDF4 below 1.6.0 is detected. It may be a good idea also to check for the latest versions of netcdf-c.