CLN: avoid internals, some misc cleanups (#33083)

Author: jbrockmendel

pandas/core/groupby/grouper.py

@@ -561,7 +561,8 @@ def get_grouper(
     # if the actual grouper should be obj[key]
     def is_in_axis(key) -> bool:
         if not _is_label_like(key):
-            items = obj._data.items
+            # items -> .columns for DataFrame, .index for Series
+            items = obj.axes[-1]
             try:
                 items.get_loc(key)
             except (KeyError, TypeError, InvalidIndexError):

pandas/core/indexes/period.py

@@ -149,6 +149,7 @@ class PeriodIndex(DatetimeIndexOpsMixin, Int64Index):
     _infer_as_myclass = True
 
     _data: PeriodArray
+    freq: DateOffset
 
     _engine_type = libindex.PeriodEngine
     _supports_partial_string_indexing = True

pandas/core/indexing.py

@@ -1,4 +1,4 @@
-from typing import Hashable, List, Tuple, Union
+from typing import TYPE_CHECKING, Hashable, List, Tuple, Union
 
 import numpy as np
 
@@ -29,6 +29,9 @@
 )
 from pandas.core.indexes.api import Index, InvalidIndexError
 
+if TYPE_CHECKING:
+    from pandas import DataFrame  # noqa:F401
+
 # "null slice"
 _NS = slice(None, None)
 
@@ -2108,7 +2111,7 @@ def _tuplify(ndim: int, loc: Hashable) -> Tuple[Union[Hashable, slice], ...]:
     return tuple(_tup)
 
 
-def convert_to_index_sliceable(obj, key):
+def convert_to_index_sliceable(obj: "DataFrame", key):
     """
     If we are index sliceable, then return my slicer, otherwise return None.
     """
@@ -2119,7 +2122,7 @@ def convert_to_index_sliceable(obj, key):
     elif isinstance(key, str):
 
         # we are an actual column
-        if key in obj._data.items:
+        if key in obj.columns:
             return None
 
         # We might have a datetimelike string that we can translate to a

pandas/core/series.py

@@ -74,7 +74,7 @@
     sanitize_array,
 )
 from pandas.core.generic import NDFrame
-from pandas.core.indexers import maybe_convert_indices, unpack_1tuple
+from pandas.core.indexers import unpack_1tuple
 from pandas.core.indexes.accessors import CombinedDatetimelikeProperties
 from pandas.core.indexes.api import (
     Float64Index,
@@ -435,7 +435,8 @@ def dtypes(self) -> DtypeObj:
         """
         Return the dtype object of the underlying data.
         """
-        return self._data.dtype
+        # DataFrame compatibility
+        return self.dtype
 
     @property
     def name(self) -> Label:
@@ -828,15 +829,7 @@ def take(self, indices, axis=0, is_copy=None, **kwargs) -> "Series":
 
         indices = ensure_platform_int(indices)
         new_index = self.index.take(indices)
-
-        if is_categorical_dtype(self):
-            # https://github.com/pandas-dev/pandas/issues/20664
-            # TODO: remove when the default Categorical.take behavior changes
-            indices = maybe_convert_indices(indices, len(self._get_axis(axis)))
-            kwargs = {"allow_fill": False}
-        else:
-            kwargs = {}
-        new_values = self._values.take(indices, **kwargs)
+        new_values = self._values.take(indices)
 
         return self._constructor(
             new_values, index=new_index, fastpath=True

pandas/io/formats/info.py

@@ -265,7 +265,8 @@ def _sizeof_fmt(num, size_qualifier):
         else:
             _verbose_repr()
 
-    counts = data._data.get_dtype_counts()
+    # groupby dtype.name to collect e.g. Categorical columns
+    counts = data.dtypes.value_counts().groupby(lambda x: x.name).sum()
     dtypes = [f"{k[0]}({k[1]:d})" for k in sorted(counts.items())]
     lines.append(f"dtypes: {', '.join(dtypes)}")