ENH: More usage of maybe_unbox_numpy_scalars (#63431)

Author: rhshadrach

pandas/core/frame.py

@@ -90,6 +90,7 @@
     infer_dtype_from_scalar,
     invalidate_string_dtypes,
     maybe_downcast_to_dtype,
+    maybe_unbox_numpy_scalar,
 )
 from pandas.core.dtypes.common import (
     infer_dtype_from_object,
@@ -12824,7 +12825,7 @@ def _get_data() -> DataFrame:
                 df = df.astype(dtype)
                 arr = concat_compat(list(df._iter_column_arrays()))
                 return arr._reduce(name, skipna=skipna, keepdims=False, **kwds)
-            return func(df.values)
+            return maybe_unbox_numpy_scalar(func(df.values))
         elif axis == 1:
             if len(df.index) == 0:
                 # Taking a transpose would result in no columns, losing the dtype.

pandas/core/indexes/base.py

@@ -963,10 +963,10 @@ def __array_ufunc__(self, ufunc: np.ufunc, method: str_t, *inputs, **kwargs):
             return tuple(self.__array_wrap__(x) for x in result)
         elif method == "reduce":
             result = lib.item_from_zerodim(result)
-            return result
+            return maybe_unbox_numpy_scalar(result)
         elif is_scalar(result):
             # e.g. matmul
-            return result
+            return maybe_unbox_numpy_scalar(result)
 
         if result.dtype == np.float16:
             result = result.astype(np.float32)
@@ -6817,7 +6817,7 @@ def _searchsorted_monotonic(self, label, side: Literal["left", "right"] = "left"
             pos = self[::-1].searchsorted(
                 label, side="right" if side == "left" else "left"
             )
-            return len(self) - pos
+            return maybe_unbox_numpy_scalar(len(self) - pos)
 
         raise ValueError("index must be monotonic increasing or decreasing")
 
@@ -7007,6 +7007,8 @@ def slice_locs(
             if start_slice == -1:
                 start_slice -= len(self)
 
+        start_slice = maybe_unbox_numpy_scalar(start_slice)
+        end_slice = maybe_unbox_numpy_scalar(end_slice)
         return start_slice, end_slice
 
     def delete(
@@ -7414,7 +7416,7 @@ def any(self, *args, **kwargs):
             # i.e. EA, call _reduce instead of "any" to get TypeError instead
             #  of AttributeError
             return vals._reduce("any")
-        return np.any(vals)
+        return maybe_unbox_numpy_scalar(np.any(vals))
 
     def all(self, *args, **kwargs):
         """
@@ -7462,7 +7464,7 @@ def all(self, *args, **kwargs):
             # i.e. EA, call _reduce instead of "all" to get TypeError instead
             #  of AttributeError
             return vals._reduce("all")
-        return np.all(vals)
+        return maybe_unbox_numpy_scalar(np.all(vals))
 
     @final
     def _maybe_disable_logical_methods(self, opname: str_t) -> None:

pandas/core/indexes/interval.py

@@ -40,6 +40,7 @@
     infer_dtype_from_scalar,
     maybe_box_datetimelike,
     maybe_downcast_numeric,
+    maybe_unbox_numpy_scalar,
     maybe_upcast_numeric_to_64bit,
 )
 from pandas.core.dtypes.common import (
@@ -812,7 +813,7 @@ def get_loc(self, key) -> int | slice | np.ndarray:
         if matches == 0:
             raise KeyError(key)
         if matches == 1:
-            return mask.argmax()
+            return maybe_unbox_numpy_scalar(mask.argmax())
 
         res = lib.maybe_booleans_to_slice(mask.view("u1"))
         if isinstance(res, slice) and res.stop is None:

pandas/core/indexes/multi.py

@@ -58,7 +58,10 @@
 )
 from pandas.util._exceptions import find_stack_level
 
-from pandas.core.dtypes.cast import coerce_indexer_dtype
+from pandas.core.dtypes.cast import (
+    coerce_indexer_dtype,
+    maybe_unbox_numpy_scalar,
+)
 from pandas.core.dtypes.common import (
     ensure_int64,
     ensure_platform_int,
@@ -3132,7 +3135,9 @@ def get_slice_bound(
         """
         if not isinstance(label, tuple):
             label = (label,)
-        return self._partial_tup_index(label, side=side)
+        result = self._partial_tup_index(label, side=side)
+        result = maybe_unbox_numpy_scalar(result)
+        return result
 
     def slice_locs(self, start=None, end=None, step=None) -> tuple[int, int]:
         """
@@ -3719,7 +3724,7 @@ def convert_indexer(start, stop, step, indexer=indexer, codes=level_codes):
             if start == end:
                 # The label is present in self.levels[level] but unused:
                 raise KeyError(key)
-            return slice(start, end)
+            return slice(maybe_unbox_numpy_scalar(start), maybe_unbox_numpy_scalar(end))
 
     def get_locs(self, seq) -> npt.NDArray[np.intp]:
         """

pandas/core/series.py

@@ -2667,7 +2667,7 @@ def quantile(
             return self._constructor(result, index=idx, name=self.name)
         else:
             # scalar
-            return result.iloc[0]
+            return maybe_unbox_numpy_scalar(result.iloc[0])
 
     def corr(
         self,
@@ -2754,9 +2754,11 @@ def corr(
         other_values = other.to_numpy(dtype=float, na_value=np.nan, copy=False)
 
         if method in ["pearson", "spearman", "kendall"] or callable(method):
-            return nanops.nancorr(
+            result = nanops.nancorr(
                 this_values, other_values, method=method, min_periods=min_periods
             )
+            result = maybe_unbox_numpy_scalar(result)
+            return result
 
         raise ValueError(
             "method must be either 'pearson', "
@@ -2808,9 +2810,11 @@ def cov(
             return np.nan
         this_values = this.to_numpy(dtype=float, na_value=np.nan, copy=False)
         other_values = other.to_numpy(dtype=float, na_value=np.nan, copy=False)
-        return nanops.nancov(
+        result = nanops.nancov(
             this_values, other_values, min_periods=min_periods, ddof=ddof
         )
+        result = maybe_unbox_numpy_scalar(result)
+        return result
 
     @doc(
         klass="Series",
@@ -3023,11 +3027,12 @@ def dot(self, other: AnyArrayLike | DataFrame) -> Series | np.ndarray:
                 np.dot(lvals, rvals), index=other.columns, copy=False, dtype=common_type
             ).__finalize__(self, method="dot")
         elif isinstance(other, Series):
-            return np.dot(lvals, rvals)
+            result = np.dot(lvals, rvals)
         elif isinstance(rvals, np.ndarray):
-            return np.dot(lvals, rvals)
+            result = np.dot(lvals, rvals)
         else:  # pragma: no cover
             raise TypeError(f"unsupported type: {type(other)}")
+        return maybe_unbox_numpy_scalar(result)
 
     def __matmul__(self, other):
         """
@@ -5701,7 +5706,7 @@ def pop(self, item: Hashable) -> Any:
         2    3
         dtype: int64
         """
-        return super().pop(item=item)
+        return maybe_unbox_numpy_scalar(super().pop(item=item))
 
     def info(
         self,

pandas/tests/series/test_ufunc.py

@@ -433,10 +433,13 @@ def test_np_matmul():
 
 
 @pytest.mark.parametrize("box", [pd.Index, pd.Series])
-def test_np_matmul_1D(box):
+def test_np_matmul_1D(box, using_python_scalars):
     result = np.matmul(box([1, 2]), box([2, 3]))
     assert result == 8
-    assert isinstance(result, np.int64)
+    if using_python_scalars:
+        assert type(result) == int, type(result)
+    else:
+        assert type(result) == np.int64, type(result)
 
 
 def test_array_ufuncs_for_many_arguments():