FEAT-#1911: support cat methods

modin/backends/pandas/query_compiler.py

@@ -179,6 +179,8 @@ def default_to_pandas(self, pandas_op, *args, **kwargs):
 
         result = pandas_op(self.to_pandas(), *args, **kwargs)
         if isinstance(result, pandas.Series):
+            if result.name is None:
+                result.name = "__reduced__"
             result = result.to_frame()
         if isinstance(result, pandas.DataFrame):
             return self.from_pandas(result, type(self._modin_frame))
@@ -2240,3 +2242,9 @@ def sort_columns_by_row_values(self, rows, ascending=True, **kwargs):
             by=rows, axis=1, ascending=ascending, kind=kind, na_position=na_position,
         ).columns
         return self.reindex(1, new_columns)
+
+    # Cat operations
+    def cat_codes(self):
+        return self.default_to_pandas(lambda df: df[df.columns[0]].cat.codes)
+
+    # END Cat operations

modin/pandas/series.py

@@ -1627,7 +1627,7 @@ def axes(self):
 
     @property
     def cat(self):
-        return self._default_to_pandas(pandas.Series.cat)
+        return CategoryMethods(self)
 
     @property
     def dt(self):
@@ -2240,3 +2240,76 @@ def _default_to_pandas(self, op, *args, **kwargs):
         return self._series._default_to_pandas(
             lambda series: op(series.str, *args, **kwargs)
         )
+
+
+class CategoryMethods(object):
+    def __init__(self, series):
+        self._series = series
+        self._query_compiler = series._query_compiler
+
+    @property
+    def categories(self):
+        return self._series._default_to_pandas(pandas.Series.cat).categories
+
+    @categories.setter
+    def categories(self, categories):
+        def set_categories(series, categories):
+            series.cat.categories = categories
+
+        self._series._default_to_pandas(set_categories, categories=categories)
+
+    @property
+    def ordered(self):
+        return self._series._default_to_pandas(pandas.Series.cat).ordered
+
+    @property
+    def codes(self):
+        return Series(query_compiler=self._query_compiler.cat_codes())
+
+    def rename_categories(self, new_categories, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.rename_categories, new_categories, inplace=inplace
+        )
+
+    def reorder_categories(self, new_categories, ordered=None, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.reorder_categories,
+            new_categories,
+            ordered=ordered,
+            inplace=inplace,
+        )
+
+    def add_categories(self, new_categories, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.add_categories, new_categories, inplace=inplace
+        )
+
+    def remove_categories(self, removals, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.remove_categories, removals, inplace=inplace
+        )
+
+    def remove_unused_categories(self, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.remove_unused_categories, inplace=inplace
+        )
+
+    def set_categories(self, new_categories, ordered=None, rename=False, inplace=False):
+        return self._default_to_pandas(
+            pandas.Series.cat.set_categories,
+            new_categories,
+            ordered=ordered,
+            rename=rename,
+            inplace=inplace,
+        )
+
+    def as_ordered(self, inplace=False):
+        return self._default_to_pandas(pandas.Series.cat.as_ordered, inplace=inplace)
+
+    def as_unordered(self, inplace=False):
+        return self._default_to_pandas(pandas.Series.cat.as_unordered, inplace=inplace)
+
+    def _default_to_pandas(self, op, *args, **kwargs):
+        return self._series._default_to_pandas(
+            lambda series: op(series.cat, *args, **kwargs)
+        )

modin/pandas/test/test_series.py

@@ -57,6 +57,8 @@
     eval_general,
     test_data_small_values,
     test_data_small_keys,
+    test_data_categorical_values,
+    test_data_categorical_keys,
 )
 
 pd.DEFAULT_NPARTITIONS = 4
@@ -4100,3 +4102,141 @@ def test_hasattr_sparse(data):
     else:
         modin_result = hasattr(modin_series, "sparse")
         assert modin_result == pandas_result
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+def test_cat_categories(data):
+    modin_series, pandas_series = create_test_series(data.copy())
+    df_equals(modin_series.cat.categories, pandas_series.cat.categories)
+    pandas_series.cat.categories = list("qwert")
+    modin_series.cat.categories = list("qwert")
+    df_equals(modin_series, pandas_series)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+def test_cat_ordered(data):
+    modin_series, pandas_series = create_test_series(data.copy())
+    assert modin_series.cat.ordered == pandas_series.cat.ordered
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+def test_cat_codes(data):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.codes
+    modin_result = modin_series.cat.codes
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_rename_categories(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.rename_categories(list("qwert"), inplace=inplace)
+    modin_result = modin_series.cat.rename_categories(list("qwert"), inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("ordered", bool_arg_values, ids=bool_arg_keys)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_reorder_categories(data, ordered, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.reorder_categories(
+        list("tades"), ordered=ordered, inplace=inplace
+    )
+    modin_result = modin_series.cat.reorder_categories(
+        list("tades"), ordered=ordered, inplace=inplace
+    )
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_add_categories(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.add_categories(list("qw"), inplace=inplace)
+    modin_result = modin_series.cat.add_categories(list("qw"), inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_remove_categories(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.remove_categories(list("at"), inplace=inplace)
+    modin_result = modin_series.cat.remove_categories(list("at"), inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_remove_unused_categories(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_series[1] = np.nan
+    pandas_result = pandas_series.cat.remove_unused_categories(inplace=inplace)
+    modin_series[1] = np.nan
+    modin_result = modin_series.cat.remove_unused_categories(inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("ordered", bool_arg_values, ids=bool_arg_keys)
+@pytest.mark.parametrize("rename", [True, False])
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_set_categories(data, ordered, rename, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.set_categories(
+        list("qwert"), ordered=ordered, rename=rename, inplace=inplace
+    )
+    modin_result = modin_series.cat.set_categories(
+        list("qwert"), ordered=ordered, rename=rename, inplace=inplace
+    )
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_as_ordered(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.as_ordered(inplace=inplace)
+    modin_result = modin_series.cat.as_ordered(inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)
+
+
+@pytest.mark.parametrize(
+    "data", test_data_categorical_values, ids=test_data_categorical_keys
+)
+@pytest.mark.parametrize("inplace", [True, False])
+def test_cat_as_unordered(data, inplace):
+    modin_series, pandas_series = create_test_series(data.copy())
+    pandas_result = pandas_series.cat.as_unordered(inplace=inplace)
+    modin_result = modin_series.cat.as_unordered(inplace=inplace)
+    df_equals(modin_series, pandas_series)
+    df_equals(modin_result, pandas_result)

modin/pandas/test/utils.py

@@ -173,6 +173,14 @@
 test_data_with_duplicates_values = list(test_data_with_duplicates.values())
 test_data_with_duplicates_keys = list(test_data_with_duplicates.keys())
 
+test_data_categorical = {
+    "ordered": pandas.Categorical(list("testdata"), ordered=True),
+    "unordered": pandas.Categorical(list("testdata"), ordered=False),
+}
+
+test_data_categorical_values = list(test_data_categorical.values())
+test_data_categorical_keys = list(test_data_categorical.keys())
+
 numeric_dfs = [
     "empty_data",
     "columns_only",