Fix type issues from pandas stubs (pydata#10128)

* Fix type issues from pandas stubs

for transparency, claude code also helped here, but I did more auditing & fixing than the idxmax PR (sorry to externalize that to reviewers...)

* .

* Don't skip tests when on a `mypy` branch

For some reason, my local was skipping tests when on a branch named `mypy`, or at least something was setting a `mypy` keyword on the test items.

This changes the check check whether the test has the `mypy` marker set, which is more precise than just looking at the keywords.

Author: max-sixty

ci/requirements/environment.yml

@@ -56,6 +56,7 @@ dependencies:
   - types-pytz
   - types-PyYAML
   - types-setuptools
+  - types-openpyxl
   - typing_extensions
   - zarr
   - pip:

pyproject.toml

@@ -55,6 +55,7 @@ types = [
   "types-pexpect",
   "types-psutil",
   "types-pycurl",
+  "types-openpyxl",
   "types-python-dateutil",
   "types-pytz",
   "types-setuptools",

xarray/core/coordinates.py

@@ -177,13 +177,13 @@ def to_index(self, ordered_dims: Sequence[Hashable] | None = None) -> pd.Index:
 
                 # compute the cartesian product
                 code_list += [
-                    np.tile(np.repeat(code, repeat_counts[i]), tile_counts[i])
+                    np.tile(np.repeat(code, repeat_counts[i]), tile_counts[i]).tolist()
                     for code in codes
                 ]
                 level_list += levels
                 names += index.names
 
-        return pd.MultiIndex(level_list, code_list, names=names)
+        return pd.MultiIndex(levels=level_list, codes=code_list, names=names)
 
 
 class Coordinates(AbstractCoordinates):

xarray/core/groupby.py

@@ -527,7 +527,7 @@ def factorize(self) -> EncodedGroups:
         _flatcodes = where(mask.data, -1, _flatcodes)
 
         full_index = pd.MultiIndex.from_product(
-            (grouper.full_index.values for grouper in groupers),
+            list(grouper.full_index.values for grouper in groupers),
             names=tuple(grouper.name for grouper in groupers),
         )
         # This will be unused when grouping by dask arrays, so skip..

xarray/core/indexes.py

@@ -1046,9 +1046,11 @@ def stack(
                 *[lev.factorize() for lev in level_indexes], strict=True
             )
             labels_mesh = np.meshgrid(*split_labels, indexing="ij")
-            labels = [x.ravel() for x in labels_mesh]
+            labels = [x.ravel().tolist() for x in labels_mesh]
 
-            index = pd.MultiIndex(levels, labels, sortorder=0, names=variables.keys())
+            index = pd.MultiIndex(
+                levels=levels, codes=labels, sortorder=0, names=variables.keys()
+            )
         level_coords_dtype = {k: var.dtype for k, var in variables.items()}
 
         return cls(index, dim, level_coords_dtype=level_coords_dtype)
@@ -1120,7 +1122,8 @@ def from_variables_maybe_expand(
             levels.append(cat.categories)
             level_variables[name] = var
 
-        index = pd.MultiIndex(levels, codes, names=names)
+        codes_as_lists = [list(x) for x in codes]
+        index = pd.MultiIndex(levels=levels, codes=codes_as_lists, names=names)
         level_coords_dtype = {k: var.dtype for k, var in level_variables.items()}
         obj = cls(index, dim, level_coords_dtype=level_coords_dtype)
         index_vars = obj.create_variables(level_variables)

xarray/tests/test_dataset.py

@@ -2923,7 +2923,7 @@ def test_drop_indexes(self) -> None:
         assert_identical(actual, ds)
 
         # test index corrupted
-        midx = pd.MultiIndex.from_tuples([([1, 2]), ([3, 4])], names=["a", "b"])
+        midx = pd.MultiIndex.from_tuples([(1, 2), (3, 4)], names=["a", "b"])
         midx_coords = Coordinates.from_pandas_multiindex(midx, "x")
         ds = Dataset(coords=midx_coords)
 
@@ -3219,7 +3219,7 @@ def test_rename_dimension_coord_warnings(self) -> None:
             ds.rename(x="x")
 
     def test_rename_multiindex(self) -> None:
-        midx = pd.MultiIndex.from_tuples([([1, 2]), ([3, 4])], names=["a", "b"])
+        midx = pd.MultiIndex.from_tuples([(1, 2), (3, 4)], names=["a", "b"])
         midx_coords = Coordinates.from_pandas_multiindex(midx, "x")
         original = Dataset({}, midx_coords)