Fstd2nc

* Fixed crop_to_smallest_grid option, which was broken in .to_xarray()
after fused records were implemented.
* Added doc string in .from_fstpy().
* Improved support for mesh grids (grtyp='M'). Removed extra
dimensions from mesh_face_nodes variable that shouldn't be there,
and got rid of warning message about decoding '' record.
* Fixed interp option so it now works in dask clusters.
* Better error message when an empty list of files is passed in.

* Added vertical bounds information to --bounds option.
* Removed "coordinates" from internal encoding of xarray DataArray
objects. This makes it easier for the user to remove it from the
attributes in xarray to_netcdf routine.
* Added CCC file interface to the project (work in progress).
* Refactored internal _decode routines to be class methods, which
means the full Buffer object no longer needs to be included in
dask graphs (greatly reducing the size of dask graphs).
* Automatically "fuse" records together in xarray interface, which
increases the chunk sizes in the dask graphs (making operations work
more efficiently in dask).

* Fix compatibility issue with older versions of numpy, triggered
when using gridhacks options.

* Fix handling of accented characters from vardicts and metadata
files.
* Removed 'axis' attribute from 2D lat/lon fields, since they are not
truly axes in that case.
* New Python helper function (fstd2nc.extra.get_crs) for getting
projection information from an xarray dataset.
* Turned off annoying librmn messages when using this module from
Python.
* Some support for creating a new Buffer object by instantiating from
an existing Buffer, allowing more specific filtering options.
* Fix filtering for etikets that contain spaces.
* Detect and transform netCDF variable names that contain a '/'.
This can happen, for instance if there are multiple typvars and one
of them contains a '/' character.
* Some support for complex data (datyp=8) from Python interface.
Note: not supported for writing to netCDF, just xarray interface.
* New to_fstd() method to write back to FSTD format instead of netCDF.
* Use only "current" definitions for --opdict and --vardict options,
ignore the obsolete ones.

* Detect newer surface aggregate fields from GEM 5.1 / 5.2 (assign
"sfctype" categories on those fields).
* Detect SWE (from CalDAS) as a surface aggregate field as well.
* New command-line option --sfc-agg-vars to allow the user to specify
additional fields that should be treated as surface-type fields.
* Fix pickling of Buffer objects.
* Fix compatibility issue when using .to_xarray() on Python 3.10.
* Add support for pathlib.Path objects.
* When invoking this module from a Python script, re-route warning and
error messages through the warnings.warn and Exception interfaces
respectively. This will make it easier for the calling script to
filter the warnings and implement exception handling.
* Traduire en (mauvais) français.

* Allow fstd2nc to be "pip install"ed from an environment that doesn't
yet have numpy installed. Installation no longer fails with a weird
error message.
* Fix buggy behaviour when there's a mask field with values > 1.
E.g., values of "2" sometimes represent partially submerged points.
Now treat such points the same as mask value "1" (keep grid point).
* Handle edge case where one of multiple FST input files does not
contain any records.
* Fix --exclude option to properly remove leadtime/reftime again. These
stopped being excludable after some recent code updates in previous
release.
* Fix various bugs in --fstd-compat option.