Replace custom cached_property implementation with functools (NVIDI…

…A#10272)

Replaces our custom `cached_property` with that provided by `functools` since Python 3.8. This uncovered a couple of typing bugs that previously eluded us.

Authors:
  - Ashwin Srinath (https://github.com/shwina)

Approvers:
  - https://github.com/brandon-b-miller
  - Bradley Dice (https://github.com/bdice)

URL: rapidsai/cudf#10272

python/cudf/cudf/core/_base_index.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import pickle
+from functools import cached_property
 from typing import Any, Set
 
 import pandas as pd
@@ -31,7 +32,6 @@
     is_mixed_with_object_dtype,
     numeric_normalize_types,
 )
-from cudf.utils.utils import cached_property
 
 
 class BaseIndex(Serializable):

python/cudf/cudf/core/column_accessor.py

@@ -4,7 +4,7 @@
 
 import itertools
 from collections.abc import MutableMapping
-from functools import reduce
+from functools import cached_property, reduce
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -20,7 +20,6 @@
 
 import cudf
 from cudf.core import column
-from cudf.utils.utils import cached_property
 
 if TYPE_CHECKING:
     from cudf.core.column import ColumnBase
@@ -360,9 +359,9 @@ def select_by_index(self, index: Any) -> ColumnAccessor:
             start, stop, step = index.indices(len(self._data))
             keys = self.names[start:stop:step]
         elif pd.api.types.is_integer(index):
-            keys = [self.names[index]]
+            keys = (self.names[index],)
         else:
-            keys = (self.names[i] for i in index)
+            keys = tuple(self.names[i] for i in index)
         data = {k: self._data[k] for k in keys}
         return self.__class__(
             data, multiindex=self.multiindex, level_names=self.level_names,

python/cudf/cudf/core/frame.py

@@ -90,22 +90,22 @@ def _num_rows(self) -> int:
         return len(self._data.columns[0])
 
     @property
-    def _column_names(self) -> List[Any]:  # TODO: List[str]?
-        return self._data.names
+    def _column_names(self) -> Tuple[Any, ...]:  # TODO: Tuple[str]?
+        return tuple(self._data.names)
 
     @property
-    def _index_names(self) -> List[Any]:  # TODO: List[str]?
+    def _index_names(self) -> Optional[Tuple[Any, ...]]:  # TODO: Tuple[str]?
         # TODO: Temporarily suppressing mypy warnings to avoid introducing bugs
         # by returning an empty list where one is not expected.
         return (
             None  # type: ignore
             if self._index is None
-            else self._index._data.names
+            else tuple(self._index._data.names)
         )
 
     @property
-    def _columns(self) -> List[Any]:  # TODO: List[Column]?
-        return self._data.columns
+    def _columns(self) -> Tuple[Any, ...]:  # TODO: Tuple[Column]?
+        return tuple(self._data.columns)
 
     def serialize(self):
         header = {

python/cudf/cudf/core/groupby/groupby.py

@@ -4,6 +4,7 @@
 import itertools
 import pickle
 import warnings
+from functools import cached_property
 
 import numpy as np
 import pandas as pd
@@ -16,7 +17,7 @@
 from cudf.core.abc import Serializable
 from cudf.core.column.column import arange, as_column
 from cudf.core.multiindex import MultiIndex
-from cudf.utils.utils import GetAttrGetItemMixin, cached_property
+from cudf.utils.utils import GetAttrGetItemMixin
 
 
 # The three functions below return the quantiles [25%, 50%, 75%]

python/cudf/cudf/core/index.py

@@ -1,10 +1,11 @@
-# Copyright (c) 2018-2021, NVIDIA CORPORATION.
+# Copyright (c) 2018-2022, NVIDIA CORPORATION.
 
 from __future__ import annotations
 
 import math
 import pickle
 import warnings
+from functools import cached_property
 from numbers import Number
 from typing import (
     Any,
@@ -54,7 +55,7 @@
 from cudf.core.single_column_frame import SingleColumnFrame
 from cudf.utils.docutils import copy_docstring
 from cudf.utils.dtypes import find_common_type
-from cudf.utils.utils import cached_property, search_range
+from cudf.utils.utils import search_range
 
 T = TypeVar("T", bound="Frame")
 

python/cudf/cudf/core/indexed_frame.py

@@ -6,6 +6,7 @@
 import operator
 import warnings
 from collections import Counter, abc
+from functools import cached_property
 from typing import Callable, Type, TypeVar
 from uuid import uuid4
 
@@ -29,7 +30,6 @@
 from cudf.core.index import Index, RangeIndex, _index_from_columns
 from cudf.core.multiindex import MultiIndex
 from cudf.core.udf.utils import _compile_or_get, _supported_cols_from_frame
-from cudf.utils.utils import cached_property
 
 doc_reset_index_template = """
         Reset the index of the {klass}, or a level of it.

python/cudf/cudf/core/join/join.py

@@ -1,7 +1,7 @@
-# Copyright (c) 2020-2021, NVIDIA CORPORATION.
+# Copyright (c) 2020-2022, NVIDIA CORPORATION.
 from __future__ import annotations
 
-from typing import TYPE_CHECKING, Callable, cast
+from typing import TYPE_CHECKING, Any, Callable, List, cast
 
 import cudf
 from cudf import _lib as libcudf
@@ -320,7 +320,7 @@ def _sort_result(self, result: Frame) -> Frame:
         # same order as given in 'on'. If the indices are used as
         # keys, the index will be sorted. If one index is specified,
         # the key columns on the other side will be used to sort.
-        by = []
+        by: List[Any] = []
         if self._using_left_index and self._using_right_index:
             if result._index is not None:
                 by.extend(result._index._data.columns)

python/cudf/cudf/core/multiindex.py

@@ -1,11 +1,12 @@
-# Copyright (c) 2019-2021, NVIDIA CORPORATION.
+# Copyright (c) 2019-2022, NVIDIA CORPORATION.
 
 from __future__ import annotations
 
 import itertools
 import numbers
 import pickle
 from collections.abc import Sequence
+from functools import cached_property
 from numbers import Integral
 from typing import Any, List, MutableMapping, Optional, Tuple, Union
 
@@ -22,11 +23,7 @@
 from cudf.core._compat import PANDAS_GE_120
 from cudf.core.frame import Frame
 from cudf.core.index import BaseIndex, _lexsorted_equal_range, as_index
-from cudf.utils.utils import (
-    NotIterable,
-    _maybe_indices_to_slice,
-    cached_property,
-)
+from cudf.utils.utils import NotIterable, _maybe_indices_to_slice
 
 
 class MultiIndex(Frame, BaseIndex, NotIterable):

python/cudf/cudf/utils/utils.py

@@ -144,28 +144,6 @@ def set_allocator(
 IS_NEP18_ACTIVE = _is_nep18_active()
 
 
-class cached_property:
-    """
-    Like @property, but only evaluated upon first invocation.
-    To force re-evaluation of a cached_property, simply delete
-    it with `del`.
-    """
-
-    # TODO: Can be replaced with functools.cached_property when we drop support
-    # for Python 3.7.
-
-    def __init__(self, func):
-        self.func = func
-
-    def __get__(self, instance, cls):
-        if instance is None:
-            return self
-        else:
-            value = self.func(instance)
-            object.__setattr__(instance, self.func.__name__, value)
-            return value
-
-
 class GetAttrGetItemMixin:
     """This mixin changes `__getattr__` to attempt a `__getitem__` call.