feat(hiopy): add kwargs to hiopy-configure add-variable

The arguments are forwrded to `create_dataset`, this yields more
flexibility for creating the arrays like passing a compressor or storage_transformers.

apps/hiopy/configure.py

@@ -59,25 +59,23 @@ def add_variable(
     chunk_shape=None,
     frac_mask=None,
     yac_name=None,
+    **kwargs,
 ):
     for g in _collect_groups(dataset):
         ntime = g.time.shape[0]
         ncells = healpy.nside2npix(g.crs.attrs["healpix_nside"])
         if zaxis is None:
             shape = (ntime, ncells)
-            _chunk_shape = np.minimum(chunk_shape, shape) if chunk_shape is not None else None
-            v = g.create_dataset(
-                name, shape=shape, dtype=np.float32, fill_value=np.nan, chunks=_chunk_shape
-            )
-            v.attrs["_ARRAY_DIMENSIONS"] = ("time", "cell")
+            dimensions = ("time", "cell")
         else:
             nheight = g[zaxis].shape[0]
             shape = (ntime, nheight, ncells)
-            _chunk_shape = np.minimum(chunk_shape, shape) if chunk_shape is not None else None
-            v = g.create_dataset(
-                name, shape=shape, dtype=np.float32, fill_value=np.nan, chunks=_chunk_shape
-            )
-            v.attrs["_ARRAY_DIMENSIONS"] = ("time", zaxis, "cell")
+            dimensions = ("time", zaxis, "cell")
+        _chunk_shape = np.minimum(chunk_shape, shape) if chunk_shape is not None else None
+        v = g.create_dataset(
+            name, shape=shape, dtype=np.float32, fill_value=np.nan, chunks=_chunk_shape, **kwargs
+        )
+        v.attrs["_ARRAY_DIMENSIONS"] = dimensions
         if time_method is not None:
             v.attrs["hiopy::time_method"] = time_method
         if yac_name is not None: