Add WW3 Stations Spec reader

[mpiannucci]

This PR adds support for reading WW3 Station Spectral Wave Files. These files are distributed with GFS Wave output by NOAA. An example location of these files is here: https://noaa-gfs-bdp-pds.s3.amazonaws.com/index.html#gfs.20230820/18/wave/station/bulls.t18z/ . They are also available from NOMADS and GCP.

Because these files are available from cloud providers, the reader accepts a file like object instead of a filename to parse out the dataset.

Example usage with fsspec:

import wavespectra as ws
import fsspec

fs = fsspec.filesystem("s3", anon=True)
with fs.open('s3://noaa-gfs-bdp-pds/gfs.20230820/18/wave/station/bulls.t18z/gfswave.44097.spec') as file:
    ds = ws.read_ww3_station(file)

ds.spec

[rafa-guedes]

This looks great @mpiannucci thanks for this PR!

Minor comments before merging this, shall we also support files specified by strings rather than a file object in case those are available locally? I think it would be nice in fact to set this up more broadly for other file types as well as you suggested in #91. Also, could you please run black in the new code you have added, we use it to ensure all the code reads the same.

[mpiannucci]

Thanks for the feedback!

I will make those changes and then take a stab at making the changes to support fsspec more broadly.

[mpiannucci]

I took care of the low hanging fruit, but the methods that expect file globs for ascii data are a decent bit harder

[rafa-guedes]

I get an IndexError rather than a ValueError when trying to open a file-like object with xr.open_mfdataset - do you get ValueError in wavespectra?

[mpiannucci]

This gives me a ValueError, saying that it does not know which engine to use because it is just a file like object passed in.

import xarray as xr

with open('./sample_files/swanfile.nc', 'rb') as file:
    xr.open_mfdataset(file, chunks={}, combine="by_coords")

If I specify the engine, that seems to give me a ValueError as well:

with open('./sample_files/swanfile.nc', 'rb') as file:
    xr.open_mfdataset(file, chunks={}, combine="by_coords", engine="netcdf4")

gives:

ValueError: can only read bytes or file-like objects with engine='scipy' or 'h5netcdf'

Switching to h5netcdf also throws a ValueError for me:

ValueError: can't open netCDF4/HDF5 as bytes try passing a path or file-like object

Which doesnt totally make sense, but throwing an error does. I am unable to get an IndexError, can you show me the code to reproduce the IndexError? I'm also happy to just catch any exception, but better to be explicit

[mpiannucci]

I am not sure if the error is messed up between versions of xarray or netcdf, but for now I just changed it to Exception. If that unacceptable, please let me know.

Otherwise I think this is as far as I want to refactor for this PR, it doesnt remove any existing functionality, but adds flexibility to many of the readers.