Rewrite DataTree.to_netcdf and support netCDF4 in-memory

This PR includes a handful of significant changes:

1. It refactors the internal structure of `DataTree.to_netcdf()` and
   `DataTree.to_zarr()` to use lower level interfaces, rather than
   calling `Dataset` methods. This allows for properly supporting
   `compute=False` (and likely various other improvements).
2. Reading and writing in-memory data with netCDF4-python is now
   supported, including DataTree.
3. The `engine` argument in `DataTree.to_netcdf()` is now set
   consistently with `Dataset.to_netcdf()`, preferring `netcdf4` to
   `h5netcdf`.
3. Calling `Dataset.to_netcdf()` without a target now always returns a
   `memoryview` object, *including* in the case where `engine='scipy'`
   is used (which currently returns `bytes`). This is a breaking change,
   rather than merely issuing a warning as is done in #10571. I believe
   it probably makes sense to do as a this breaking change because (1)
   it offers significant performance benefits, (2) the default behavior
   without specifying an engine will already change (because `netcdf4`
   is preferred to the `scipy` backend) and (3) restoring previous
   behavior is easy (by wrapping the memoryview with `bytes()`).

mypy

Author: shoyer

xarray/__init__.py

@@ -4,6 +4,7 @@
 from xarray.backends.api import (
     load_dataarray,
     load_dataset,
+    load_datatree,
     open_dataarray,
     open_dataset,
     open_datatree,
@@ -96,6 +97,7 @@
     "infer_freq",
     "load_dataarray",
     "load_dataset",
+    "load_datatree",
     "map_blocks",
     "map_over_datasets",
     "merge",