For a given float WMO, a new class to easily open netcdf dataset for local and remote GDAC

[gmaze]

This PR shall provides a new float store to delegate to argopy all Argo netcdf files access/read methods, for local and remote GDAC content.

This new data store, ArgoFloat, aims to provide methods for third party libraries, or operator and expert workflow, to access local, or remote, Argo GDAC netcdf files without the burden of transfer protocol and GDAC paths handling.

Online API:

from argopy import ArgoFloat
ds = ArgoFloat(6903091).open_dataset('prof')

More detailed Expected API:

from argopy import ArgoFloat
WMO = 6903091

af = ArgoFloat(WMO)  # Use argopy 'gdac' option by default
af = ArgoFloat(WMO, host='/home/ref-argo/gdac')  # Use your local GDAC copy
af = ArgoFloat(WMO, host='https')  # Shortcut for https://data-argo.ifremer.fr
af = ArgoFloat(WMO, host='ftp')    # shortcut for ftp://ftp.ifremer.fr/ifremer/argo
af = ArgoFloat(WMO, host='s3')     # Shortcut for s3://argo-gdac-sandbox/pub

# Load any netcdf files from this float:
ds = af.open_dataset('meta') # load <WMO>_meta.nc
ds = af.open_dataset('prof') # load <WMO>_prof.nc
ds = af.open_dataset('tech') # load <WMO>_tech.nc
ds = af.open_dataset('Rtraj') # load <WMO>_Rtraj.nc

# List all available datasets for this float:
af.list_dataset()

Once merged, #385 and #423 will provide lazy methods for this class.

This new class shall provides much more attributes and methods. The API design is laid out in here: https://gist.github.com/gmaze/6924fc26405f42aa36fd5a22a64657ea

Online documentation can be found here:
https://argopy--429.org.readthedocs.build/en/429/generated/argopy.ArgoFloat.html