From fac2c89ee634ff572e6c6318c40b1b729a07cbac Mon Sep 17 00:00:00 2001
From: Stephan Hoyer <shoyer@google.com>
Date: Tue, 10 Sep 2024 18:30:29 -0700
Subject: [PATCH] Rename DataTree's "ds" and "data" to "dataset" (#9476)

* Rename DataTree's "ds" and "data" to "dataset"

.ds is kept around as a soft-deprecated alias to facilitate the
transition from xarray-contrib/datatree, though I verified that all
tests pass without it.

* fix data= usage in test_formatting_html.py

* fix formatting test
---
 xarray/backends/zarr.py                |  2 +-
 xarray/core/datatree.py                | 66 +++++++++---------
 xarray/core/datatree_mapping.py        | 11 +--
 xarray/core/formatting.py              |  2 +-
 xarray/tests/test_backends_datatree.py |  4 +-
 xarray/tests/test_datatree.py          | 92 +++++++++++++-------------
 xarray/tests/test_formatting.py        |  8 +--
 xarray/tests/test_formatting_html.py   |  2 +-
 8 files changed, 95 insertions(+), 92 deletions(-)

diff --git a/xarray/backends/zarr.py b/xarray/backends/zarr.py
index 31b367a178b..fd938238ae4 100644
--- a/xarray/backends/zarr.py
+++ b/xarray/backends/zarr.py
@@ -1262,7 +1262,7 @@ def open_datatree(
             ds = open_dataset(
                 filename_or_obj, store=store, group=path_group, engine="zarr", **kwargs
             )
-            new_node: DataTree = DataTree(name=NodePath(path_group).name, data=ds)
+            new_node = DataTree(name=NodePath(path_group).name, dataset=ds)
             tree_root._set_item(
                 path_group,
                 new_node,
diff --git a/xarray/core/datatree.py b/xarray/core/datatree.py
index b12d861624a..d600b18d1d6 100644
--- a/xarray/core/datatree.py
+++ b/xarray/core/datatree.py
@@ -417,7 +417,7 @@ class DataTree(
 
     def __init__(
         self,
-        data: Dataset | None = None,
+        dataset: Dataset | None = None,
         children: Mapping[str, DataTree] | None = None,
         name: str | None = None,
     ):
@@ -430,12 +430,12 @@ def __init__(
 
         Parameters
         ----------
-        data : Dataset, optional
-            Data to store under the .ds attribute of this node.
+        dataset : Dataset, optional
+            Data to store directly at this node.
         children : Mapping[str, DataTree], optional
-            Any child nodes of this node. Default is None.
+            Any child nodes of this node.
         name : str, optional
-            Name for this node of the tree. Default is None.
+            Name for this node of the tree.
 
         Returns
         -------
@@ -449,24 +449,24 @@ def __init__(
             children = {}
 
         super().__init__(name=name)
-        self._set_node_data(_to_new_dataset(data))
+        self._set_node_data(_to_new_dataset(dataset))
 
         # shallow copy to avoid modifying arguments in-place (see GH issue #9196)
         self.children = {name: child.copy() for name, child in children.items()}
 
-    def _set_node_data(self, ds: Dataset):
-        data_vars, coord_vars = _collect_data_and_coord_variables(ds)
+    def _set_node_data(self, dataset: Dataset):
+        data_vars, coord_vars = _collect_data_and_coord_variables(dataset)
         self._data_variables = data_vars
         self._node_coord_variables = coord_vars
-        self._node_dims = ds._dims
-        self._node_indexes = ds._indexes
-        self._encoding = ds._encoding
-        self._attrs = ds._attrs
-        self._close = ds._close
+        self._node_dims = dataset._dims
+        self._node_indexes = dataset._indexes
+        self._encoding = dataset._encoding
+        self._attrs = dataset._attrs
+        self._close = dataset._close
 
     def _pre_attach(self: DataTree, parent: DataTree, name: str) -> None:
         super()._pre_attach(parent, name)
-        if name in parent.ds.variables:
+        if name in parent.dataset.variables:
             raise KeyError(
                 f"parent {parent.name} already contains a variable named {name}"
             )
@@ -534,7 +534,7 @@ def _to_dataset_view(self, rebuild_dims: bool, inherited: bool) -> DatasetView:
         )
 
     @property
-    def ds(self) -> DatasetView:
+    def dataset(self) -> DatasetView:
         """
         An immutable Dataset-like view onto the data in this node.
 
@@ -549,11 +549,15 @@ def ds(self) -> DatasetView:
         """
         return self._to_dataset_view(rebuild_dims=True, inherited=True)
 
-    @ds.setter
-    def ds(self, data: Dataset | None = None) -> None:
+    @dataset.setter
+    def dataset(self, data: Dataset | None = None) -> None:
         ds = _to_new_dataset(data)
         self._replace_node(ds)
 
+    # soft-deprecated alias, to facilitate the transition from
+    # xarray-contrib/datatree
+    ds = dataset
+
     def to_dataset(self, inherited: bool = True) -> Dataset:
         """
         Return the data in this node as a new xarray.Dataset object.
@@ -566,7 +570,7 @@ def to_dataset(self, inherited: bool = True) -> Dataset:
 
         See Also
         --------
-        DataTree.ds
+        DataTree.dataset
         """
         coord_vars = self._coord_variables if inherited else self._node_coord_variables
         variables = dict(self._data_variables)
@@ -845,8 +849,8 @@ def get(  # type: ignore[override]
         """
         if key in self.children:
             return self.children[key]
-        elif key in self.ds:
-            return self.ds[key]
+        elif key in self.dataset:
+            return self.dataset[key]
         else:
             return default
 
@@ -1114,7 +1118,7 @@ def from_dict(
         if isinstance(root_data, DataTree):
             obj = root_data.copy()
         elif root_data is None or isinstance(root_data, Dataset):
-            obj = cls(name=name, data=root_data, children=None)
+            obj = cls(name=name, dataset=root_data, children=None)
         else:
             raise TypeError(
                 f'root node data (at "/") must be a Dataset or DataTree, got {type(root_data)}'
@@ -1133,7 +1137,7 @@ def depth(item) -> int:
                 if isinstance(data, DataTree):
                     new_node = data.copy()
                 elif isinstance(data, Dataset) or data is None:
-                    new_node = cls(name=node_name, data=data)
+                    new_node = cls(name=node_name, dataset=data)
                 else:
                     raise TypeError(f"invalid values: {data}")
                 obj._set_item(
@@ -1264,7 +1268,7 @@ def equals(self, other: DataTree, from_root: bool = True) -> bool:
 
         return all(
             [
-                node.ds.equals(other_node.ds)
+                node.dataset.equals(other_node.dataset)
                 for node, other_node in zip(self.subtree, other.subtree, strict=True)
             ]
         )
@@ -1294,7 +1298,7 @@ def identical(self, other: DataTree, from_root=True) -> bool:
             return False
 
         return all(
-            node.ds.identical(other_node.ds)
+            node.dataset.identical(other_node.dataset)
             for node, other_node in zip(self.subtree, other.subtree, strict=True)
         )
 
@@ -1321,7 +1325,7 @@ def filter(self: DataTree, filterfunc: Callable[[DataTree], bool]) -> DataTree:
         map_over_subtree
         """
         filtered_nodes = {
-            node.path: node.ds for node in self.subtree if filterfunc(node)
+            node.path: node.dataset for node in self.subtree if filterfunc(node)
         }
         return DataTree.from_dict(filtered_nodes, name=self.root.name)
 
@@ -1365,7 +1369,7 @@ def match(self, pattern: str) -> DataTree:
             └── Group: /b/B
         """
         matching_nodes = {
-            node.path: node.ds
+            node.path: node.dataset
             for node in self.subtree
             if NodePath(node.path).match(pattern)
         }
@@ -1389,7 +1393,7 @@ def map_over_subtree(
         ----------
         func : callable
             Function to apply to datasets with signature:
-            `func(node.ds, *args, **kwargs) -> Dataset`.
+            `func(node.dataset, *args, **kwargs) -> Dataset`.
 
             Function will not be applied to any nodes without datasets.
         *args : tuple, optional
@@ -1420,7 +1424,7 @@ def map_over_subtree_inplace(
         ----------
         func : callable
             Function to apply to datasets with signature:
-            `func(node.ds, *args, **kwargs) -> Dataset`.
+            `func(node.dataset, *args, **kwargs) -> Dataset`.
 
             Function will not be applied to any nodes without datasets,
         *args : tuple, optional
@@ -1433,7 +1437,7 @@ def map_over_subtree_inplace(
 
         for node in self.subtree:
             if node.has_data:
-                node.ds = func(node.ds, *args, **kwargs)
+                node.dataset = func(node.dataset, *args, **kwargs)
 
     def pipe(
         self, func: Callable | tuple[Callable, str], *args: Any, **kwargs: Any
@@ -1499,7 +1503,7 @@ def render(self):
         """Print tree structure, including any data stored at each node."""
         for pre, fill, node in RenderDataTree(self):
             print(f"{pre}DataTree('{self.name}')")
-            for ds_line in repr(node.ds)[1:]:
+            for ds_line in repr(node.dataset)[1:]:
                 print(f"{fill}{ds_line}")
 
     def merge(self, datatree: DataTree) -> DataTree:
@@ -1513,7 +1517,7 @@ def merge_child_nodes(self, *paths, new_path: T_Path) -> DataTree:
     # TODO some kind of .collapse() or .flatten() method to merge a subtree
 
     def to_dataarray(self) -> DataArray:
-        return self.ds.to_dataarray()
+        return self.dataset.to_dataarray()
 
     @property
     def groups(self):
diff --git a/xarray/core/datatree_mapping.py b/xarray/core/datatree_mapping.py
index 1a581629ab8..78abf42601d 100644
--- a/xarray/core/datatree_mapping.py
+++ b/xarray/core/datatree_mapping.py
@@ -99,10 +99,10 @@ def map_over_subtree(func: Callable) -> Callable:
         Function will not be applied to any nodes without datasets.
     *args : tuple, optional
         Positional arguments passed on to `func`. If DataTrees any data-containing nodes will be converted to Datasets
-        via `.ds`.
+        via `.dataset`.
     **kwargs : Any
         Keyword arguments passed on to `func`. If DataTrees any data-containing nodes will be converted to Datasets
-        via `.ds`.
+        via `.dataset`.
 
     Returns
     -------
@@ -160,13 +160,14 @@ def _map_over_subtree(*args, **kwargs) -> DataTree | tuple[DataTree, ...]:
             strict=False,
         ):
             node_args_as_datasetviews = [
-                a.ds if isinstance(a, DataTree) else a for a in all_node_args[:n_args]
+                a.dataset if isinstance(a, DataTree) else a
+                for a in all_node_args[:n_args]
             ]
             node_kwargs_as_datasetviews = dict(
                 zip(
                     [k for k in kwargs_as_tree_length_iterables.keys()],
                     [
-                        v.ds if isinstance(v, DataTree) else v
+                        v.dataset if isinstance(v, DataTree) else v
                         for v in all_node_args[n_args:]
                     ],
                     strict=True,
@@ -183,7 +184,7 @@ def _map_over_subtree(*args, **kwargs) -> DataTree | tuple[DataTree, ...]:
                 )
             elif node_of_first_tree.has_attrs:
                 # propagate attrs
-                results = node_of_first_tree.ds
+                results = node_of_first_tree.dataset
             else:
                 # nothing to propagate so use fastpath to create empty node in new tree
                 results = None
diff --git a/xarray/core/formatting.py b/xarray/core/formatting.py
index d17a72aac27..1cea9a7a28d 100644
--- a/xarray/core/formatting.py
+++ b/xarray/core/formatting.py
@@ -1038,7 +1038,7 @@ def diff_nodewise_summary(a: DataTree, b: DataTree, compat):
 
     summary = []
     for node_a, node_b in zip(a.subtree, b.subtree, strict=True):
-        a_ds, b_ds = node_a.ds, node_b.ds
+        a_ds, b_ds = node_a.dataset, node_b.dataset
 
         if not a_ds._all_compat(b_ds, compat):
             dataset_diff = diff_dataset_repr(a_ds, b_ds, compat_str)
diff --git a/xarray/tests/test_backends_datatree.py b/xarray/tests/test_backends_datatree.py
index e84c77e54ed..d9490385b7d 100644
--- a/xarray/tests/test_backends_datatree.py
+++ b/xarray/tests/test_backends_datatree.py
@@ -56,7 +56,7 @@ def test_netcdf_encoding(self, tmpdir, simple_datatree):
 
         # add compression
         comp = dict(zlib=True, complevel=9)
-        enc = {"/set2": {var: comp for var in original_dt["/set2"].ds.data_vars}}
+        enc = {"/set2": {var: comp for var in original_dt["/set2"].dataset.data_vars}}
 
         original_dt.to_netcdf(filepath, encoding=enc, engine=self.engine)
         roundtrip_dt = open_datatree(filepath, engine=self.engine)
@@ -246,7 +246,7 @@ def test_zarr_encoding(self, tmpdir, simple_datatree):
         original_dt = simple_datatree
 
         comp = {"compressor": zarr.Blosc(cname="zstd", clevel=3, shuffle=2)}
-        enc = {"/set2": {var: comp for var in original_dt["/set2"].ds.data_vars}}
+        enc = {"/set2": {var: comp for var in original_dt["/set2"].dataset.data_vars}}
         original_dt.to_zarr(filepath, encoding=enc)
         roundtrip_dt = open_datatree(filepath, engine="zarr")
 
diff --git a/xarray/tests/test_datatree.py b/xarray/tests/test_datatree.py
index 353f1fab708..e6eb4c8dd22 100644
--- a/xarray/tests/test_datatree.py
+++ b/xarray/tests/test_datatree.py
@@ -36,11 +36,11 @@ def test_bad_names(self):
 
     def test_data_arg(self):
         ds = xr.Dataset({"foo": 42})
-        tree: DataTree = DataTree(data=ds)
+        tree: DataTree = DataTree(dataset=ds)
         assert_identical(tree.to_dataset(), ds)
 
         with pytest.raises(TypeError):
-            DataTree(data=xr.DataArray(42, name="foo"))  # type: ignore[arg-type]
+            DataTree(dataset=xr.DataArray(42, name="foo"))  # type: ignore[arg-type]
 
 
 class TestFamilyTree:
@@ -141,38 +141,38 @@ def test_relative_paths(self):
 class TestStoreDatasets:
     def test_create_with_data(self):
         dat = xr.Dataset({"a": 0})
-        john = DataTree(name="john", data=dat)
+        john = DataTree(name="john", dataset=dat)
 
         assert_identical(john.to_dataset(), dat)
 
         with pytest.raises(TypeError):
-            DataTree(name="mary", data="junk")  # type: ignore[arg-type]
+            DataTree(name="mary", dataset="junk")  # type: ignore[arg-type]
 
     def test_set_data(self):
         john = DataTree(name="john")
         dat = xr.Dataset({"a": 0})
-        john.ds = dat  # type: ignore[assignment]
+        john.dataset = dat  # type: ignore[assignment]
 
         assert_identical(john.to_dataset(), dat)
 
         with pytest.raises(TypeError):
-            john.ds = "junk"  # type: ignore[assignment]
+            john.dataset = "junk"  # type: ignore[assignment]
 
     def test_has_data(self):
-        john = DataTree(name="john", data=xr.Dataset({"a": 0}))
+        john = DataTree(name="john", dataset=xr.Dataset({"a": 0}))
         assert john.has_data
 
-        john_no_data = DataTree(name="john", data=None)
+        john_no_data = DataTree(name="john", dataset=None)
         assert not john_no_data.has_data
 
     def test_is_hollow(self):
-        john = DataTree(data=xr.Dataset({"a": 0}))
+        john = DataTree(dataset=xr.Dataset({"a": 0}))
         assert john.is_hollow
 
         eve = DataTree(children={"john": john})
         assert eve.is_hollow
 
-        eve.ds = xr.Dataset({"a": 1})  # type: ignore[assignment]
+        eve.dataset = xr.Dataset({"a": 1})  # type: ignore[assignment]
         assert not eve.is_hollow
 
 
@@ -197,7 +197,7 @@ def test_parent_already_has_variable_with_childs_name(self):
             DataTree.from_dict({"/": xr.Dataset({"a": [0], "b": 1}), "/a": None})
 
     def test_parent_already_has_variable_with_childs_name_update(self):
-        dt = DataTree(data=xr.Dataset({"a": [0], "b": 1}))
+        dt = DataTree(dataset=xr.Dataset({"a": [0], "b": 1}))
         with pytest.raises(ValueError, match="already contains a variable named a"):
             dt.update({"a": DataTree()})
 
@@ -209,13 +209,13 @@ def test_assign_when_already_child_with_variables_name(self):
         )
 
         with pytest.raises(ValueError, match="node already contains a variable"):
-            dt.ds = xr.Dataset({"a": 0})  # type: ignore[assignment]
+            dt.dataset = xr.Dataset({"a": 0})  # type: ignore[assignment]
 
-        dt.ds = xr.Dataset()  # type: ignore[assignment]
+        dt.dataset = xr.Dataset()  # type: ignore[assignment]
 
         new_ds = dt.to_dataset().assign(a=xr.DataArray(0))
         with pytest.raises(ValueError, match="node already contains a variable"):
-            dt.ds = new_ds  # type: ignore[assignment]
+            dt.dataset = new_ds  # type: ignore[assignment]
 
 
 class TestGet: ...
@@ -238,7 +238,7 @@ def test_getitem_self(self):
 
     def test_getitem_single_data_variable(self):
         data = xr.Dataset({"temp": [0, 50]})
-        results = DataTree(name="results", data=data)
+        results = DataTree(name="results", dataset=data)
         assert_identical(results["temp"], data["temp"])
 
     def test_getitem_single_data_variable_from_node(self):
@@ -257,14 +257,14 @@ def test_getitem_nonexistent_node(self):
 
     def test_getitem_nonexistent_variable(self):
         data = xr.Dataset({"temp": [0, 50]})
-        results = DataTree(name="results", data=data)
+        results = DataTree(name="results", dataset=data)
         with pytest.raises(KeyError):
             results["pressure"]
 
     @pytest.mark.xfail(reason="Should be deprecated in favour of .subset")
     def test_getitem_multiple_data_variables(self):
         data = xr.Dataset({"temp": [0, 50], "p": [5, 8, 7]})
-        results = DataTree(name="results", data=data)
+        results = DataTree(name="results", dataset=data)
         assert_identical(results[["temp", "p"]], data[["temp", "p"]])  # type: ignore[index]
 
     @pytest.mark.xfail(
@@ -272,7 +272,7 @@ def test_getitem_multiple_data_variables(self):
     )
     def test_getitem_dict_like_selection_access_to_dataset(self):
         data = xr.Dataset({"temp": [0, 50]})
-        results = DataTree(name="results", data=data)
+        results = DataTree(name="results", dataset=data)
         assert_identical(results[{"temp": 1}], data[{"temp": 1}])  # type: ignore[index]
 
 
@@ -440,7 +440,7 @@ def test_setitem_unnamed_child_node_becomes_named(self):
 
     def test_setitem_new_grandchild_node(self):
         john = DataTree.from_dict({"/Mary/Rose": DataTree()})
-        new_rose = DataTree(data=xr.Dataset({"x": 0}))
+        new_rose = DataTree(dataset=xr.Dataset({"x": 0}))
         john["Mary/Rose"] = new_rose
 
         grafted_rose = john["Mary/Rose"]
@@ -466,7 +466,7 @@ def test_setitem_overwrite_data_in_node_with_none(self):
         john["mary"] = DataTree()
         assert_identical(john["mary"].to_dataset(), xr.Dataset())
 
-        john.ds = xr.Dataset()  # type: ignore[assignment]
+        john.dataset = xr.Dataset()  # type: ignore[assignment]
         with pytest.raises(ValueError, match="has no name"):
             john["."] = DataTree()
 
@@ -518,20 +518,20 @@ def test_setitem_coerce_to_dataarray(self):
     def test_setitem_add_new_variable_to_empty_node(self):
         results = DataTree(name="results")
         results["pressure"] = xr.DataArray(data=[2, 3])
-        assert "pressure" in results.ds
+        assert "pressure" in results.dataset
         results["temp"] = xr.Variable(data=[10, 11], dims=["x"])
-        assert "temp" in results.ds
+        assert "temp" in results.dataset
 
         # What if there is a path to traverse first?
         results_with_path = DataTree(name="results")
         results_with_path["highres/pressure"] = xr.DataArray(data=[2, 3])
-        assert "pressure" in results_with_path["highres"].ds
+        assert "pressure" in results_with_path["highres"].dataset
         results_with_path["highres/temp"] = xr.Variable(data=[10, 11], dims=["x"])
-        assert "temp" in results_with_path["highres"].ds
+        assert "temp" in results_with_path["highres"].dataset
 
     def test_setitem_dataarray_replace_existing_node(self):
         t = xr.Dataset({"temp": [0, 50]})
-        results = DataTree(name="results", data=t)
+        results = DataTree(name="results", dataset=t)
         p = xr.DataArray(data=[2, 3])
         results["pressure"] = p
         expected = t.assign(pressure=p)
@@ -620,7 +620,7 @@ def test_full(self, simple_datatree):
         ]
 
     def test_datatree_values(self):
-        dat1 = DataTree(data=xr.Dataset({"a": 1}))
+        dat1 = DataTree(dataset=xr.Dataset({"a": 1}))
         expected = DataTree()
         expected["a"] = dat1
 
@@ -675,11 +675,11 @@ def test_array_values(self):
 class TestDatasetView:
     def test_view_contents(self):
         ds = create_test_data()
-        dt = DataTree(data=ds)
+        dt = DataTree(dataset=ds)
         assert ds.identical(
-            dt.ds
+            dt.dataset
         )  # this only works because Dataset.identical doesn't check types
-        assert isinstance(dt.ds, xr.Dataset)
+        assert isinstance(dt.dataset, xr.Dataset)
 
     def test_immutability(self):
         # See issue https://github.com/xarray-contrib/datatree/issues/38
@@ -694,28 +694,28 @@ def test_immutability(self):
         with pytest.raises(
             AttributeError, match="Mutation of the DatasetView is not allowed"
         ):
-            dt.ds["a"] = xr.DataArray(0)
+            dt.dataset["a"] = xr.DataArray(0)
 
         with pytest.raises(
             AttributeError, match="Mutation of the DatasetView is not allowed"
         ):
-            dt.ds.update({"a": 0})
+            dt.dataset.update({"a": 0})
 
         # TODO are there any other ways you can normally modify state (in-place)?
         # (not attribute-like assignment because that doesn't work on Dataset anyway)
 
     def test_methods(self):
         ds = create_test_data()
-        dt = DataTree(data=ds)
-        assert ds.mean().identical(dt.ds.mean())
-        assert isinstance(dt.ds.mean(), xr.Dataset)
+        dt = DataTree(dataset=ds)
+        assert ds.mean().identical(dt.dataset.mean())
+        assert isinstance(dt.dataset.mean(), xr.Dataset)
 
     def test_arithmetic(self, create_test_datatree):
         dt = create_test_datatree()
         expected = create_test_datatree(modify=lambda ds: 10.0 * ds)[
             "set1"
         ].to_dataset()
-        result = 10.0 * dt["set1"].ds
+        result = 10.0 * dt["set1"].dataset
         assert result.identical(expected)
 
     def test_init_via_type(self):
@@ -727,12 +727,12 @@ def test_init_via_type(self):
             dims=["x", "y", "time"],
             coords={"area": (["x", "y"], np.random.rand(3, 4))},
         ).to_dataset(name="data")
-        dt = DataTree(data=a)
+        dt = DataTree(dataset=a)
 
         def weighted_mean(ds):
             return ds.weighted(ds.area).mean(["x", "y"])
 
-        weighted_mean(dt.ds)
+        weighted_mean(dt.dataset)
 
 
 class TestAccess:
@@ -961,7 +961,7 @@ def test_inherited_dims(self):
         assert dt.b.sizes == {"x": 2, "y": 1}
         assert dt.c.sizes == {"x": 2, "y": 3}
         # dataset objects created from nodes should not
-        assert dt.b.ds.sizes == {"y": 1}
+        assert dt.b.dataset.sizes == {"y": 1}
         assert dt.b.to_dataset(inherited=True).sizes == {"y": 1}
         assert dt.b.to_dataset(inherited=False).sizes == {"y": 1}
 
@@ -1021,10 +1021,10 @@ def test_inconsistent_dims(self):
         with pytest.raises(ValueError, match=expected_msg):
             dt["/b/c"] = xr.DataArray([3.0], dims=["x"])
 
-        b = DataTree(data=xr.Dataset({"c": (("x",), [3.0])}))
+        b = DataTree(dataset=xr.Dataset({"c": (("x",), [3.0])}))
         with pytest.raises(ValueError, match=expected_msg):
             DataTree(
-                data=xr.Dataset({"a": (("x",), [1.0, 2.0])}),
+                dataset=xr.Dataset({"a": (("x",), [1.0, 2.0])}),
                 children={"b": b},
             )
 
@@ -1054,14 +1054,14 @@ def test_inconsistent_child_indexes(self):
             )
 
         dt = DataTree()
-        dt.ds = xr.Dataset(coords={"x": [1.0]})  # type: ignore[assignment]
+        dt.dataset = xr.Dataset(coords={"x": [1.0]})  # type: ignore[assignment]
         dt["/b"] = DataTree()
         with pytest.raises(ValueError, match=expected_msg):
-            dt["/b"].ds = xr.Dataset(coords={"x": [2.0]})
+            dt["/b"].dataset = xr.Dataset(coords={"x": [2.0]})
 
         b = DataTree(xr.Dataset(coords={"x": [2.0]}))
         with pytest.raises(ValueError, match=expected_msg):
-            DataTree(data=xr.Dataset(coords={"x": [1.0]}), children={"b": b})
+            DataTree(dataset=xr.Dataset(coords={"x": [1.0]}), children={"b": b})
 
     def test_inconsistent_grandchild_indexes(self):
         expected_msg = _exact_match(
@@ -1089,15 +1089,15 @@ def test_inconsistent_grandchild_indexes(self):
             )
 
         dt = DataTree()
-        dt.ds = xr.Dataset(coords={"x": [1.0]})  # type: ignore[assignment]
+        dt.dataset = xr.Dataset(coords={"x": [1.0]})  # type: ignore[assignment]
         dt["/b/c"] = DataTree()
         with pytest.raises(ValueError, match=expected_msg):
-            dt["/b/c"].ds = xr.Dataset(coords={"x": [2.0]})
+            dt["/b/c"].dataset = xr.Dataset(coords={"x": [2.0]})
 
         c = DataTree(xr.Dataset(coords={"x": [2.0]}))
         b = DataTree(children={"c": c})
         with pytest.raises(ValueError, match=expected_msg):
-            DataTree(data=xr.Dataset(coords={"x": [1.0]}), children={"b": b})
+            DataTree(dataset=xr.Dataset(coords={"x": [1.0]}), children={"b": b})
 
     def test_inconsistent_grandchild_dims(self):
         expected_msg = _exact_match(
diff --git a/xarray/tests/test_formatting.py b/xarray/tests/test_formatting.py
index 6066eeacfee..2e48b3a4219 100644
--- a/xarray/tests/test_formatting.py
+++ b/xarray/tests/test_formatting.py
@@ -626,7 +626,7 @@ def test_datatree_print_empty_node(self):
 
     def test_datatree_print_empty_node_with_attrs(self):
         dat = xr.Dataset(attrs={"note": "has attrs"})
-        dt: DataTree = DataTree(name="root", data=dat)
+        dt: DataTree = DataTree(name="root", dataset=dat)
         printout = str(dt)
         assert printout == dedent(
             """\
@@ -638,7 +638,7 @@ def test_datatree_print_empty_node_with_attrs(self):
 
     def test_datatree_print_node_with_data(self):
         dat = xr.Dataset({"a": [0, 2]})
-        dt: DataTree = DataTree(name="root", data=dat)
+        dt: DataTree = DataTree(name="root", dataset=dat)
         printout = str(dt)
         expected = [
             "<xarray.DataTree 'root'>",
@@ -646,8 +646,6 @@ def test_datatree_print_node_with_data(self):
             "Dimensions",
             "Coordinates",
             "a",
-            "Data variables",
-            "*empty*",
         ]
         for expected_line, printed_line in zip(
             expected, printout.splitlines(), strict=True
@@ -666,7 +664,7 @@ def test_datatree_printout_nested_node(self):
 
     def test_datatree_repr_of_node_with_data(self):
         dat = xr.Dataset({"a": [0, 2]})
-        dt: DataTree = DataTree(name="root", data=dat)
+        dt: DataTree = DataTree(name="root", dataset=dat)
         assert "Coordinates" in repr(dt)
 
     def test_diff_datatree_repr_structure(self):
diff --git a/xarray/tests/test_formatting_html.py b/xarray/tests/test_formatting_html.py
index ada7f75b21b..e1f51e38882 100644
--- a/xarray/tests/test_formatting_html.py
+++ b/xarray/tests/test_formatting_html.py
@@ -220,7 +220,7 @@ def childfree_tree_factory(self):
 
         def _childfree_tree_factory():
             return DataTree(
-                data=xr.Dataset({"z": ("y", [randint(1, 100) for _ in range(3)])})
+                dataset=xr.Dataset({"z": ("y", [randint(1, 100) for _ in range(3)])})
             )
 
         return _childfree_tree_factory