CLN: assorted, privatize, easy issues (pandas-dev#45305)

Author: jbrockmendel

LICENSE

@@ -3,7 +3,7 @@ BSD 3-Clause License
 Copyright (c) 2008-2011, AQR Capital Management, LLC, Lambda Foundry, Inc. and PyData Development Team
 All rights reserved.
 
-Copyright (c) 2011-2021, Open source contributors.
+Copyright (c) 2011-2022, Open source contributors.
 
 Redistribution and use in source and binary forms, with or without
 modification, are permitted provided that the following conditions are met:

pandas/_testing/__init__.py

@@ -378,7 +378,11 @@ def makePeriodIndex(k: int = 10, name=None, **kwargs) -> PeriodIndex:
 
 
 def makeMultiIndex(k=10, names=None, **kwargs):
-    return MultiIndex.from_product((("foo", "bar"), (1, 2)), names=names, **kwargs)
+    N = (k // 2) + 1
+    rng = range(N)
+    mi = MultiIndex.from_product([("foo", "bar"), rng], names=names, **kwargs)
+    assert len(mi) >= k  # GH#38795
+    return mi[:k]
 
 
 _names = [

pandas/core/dtypes/cast.py

@@ -200,7 +200,7 @@ def maybe_box_native(value: Scalar) -> Scalar:
     return value
 
 
-def maybe_unbox_datetimelike(value: Scalar, dtype: DtypeObj) -> Scalar:
+def _maybe_unbox_datetimelike(value: Scalar, dtype: DtypeObj) -> Scalar:
     """
     Convert a Timedelta or Timestamp to timedelta64 or datetime64 for setting
     into a numpy array.  Failing to unbox would risk dropping nanoseconds.
@@ -851,7 +851,7 @@ def infer_dtype_from_array(
     return arr.dtype, arr
 
 
-def maybe_infer_dtype_type(element):
+def _maybe_infer_dtype_type(element):
     """
     Try to infer an object's dtype, for use in arithmetic ops.
 
@@ -873,7 +873,7 @@ def maybe_infer_dtype_type(element):
     --------
     >>> from collections import namedtuple
     >>> Foo = namedtuple("Foo", "dtype")
-    >>> maybe_infer_dtype_type(Foo(np.dtype("i8")))
+    >>> _maybe_infer_dtype_type(Foo(np.dtype("i8")))
     dtype('int64')
     """
     tipo = None
@@ -1495,7 +1495,7 @@ def construct_2d_arraylike_from_scalar(
     shape = (length, width)
 
     if dtype.kind in ["m", "M"]:
-        value = maybe_unbox_datetimelike_tz_deprecation(value, dtype)
+        value = _maybe_unbox_datetimelike_tz_deprecation(value, dtype)
     elif dtype == _dtype_obj:
         if isinstance(value, (np.timedelta64, np.datetime64)):
             # calling np.array below would cast to pytimedelta/pydatetime
@@ -1558,17 +1558,17 @@ def construct_1d_arraylike_from_scalar(
             if not isna(value):
                 value = ensure_str(value)
         elif dtype.kind in ["M", "m"]:
-            value = maybe_unbox_datetimelike_tz_deprecation(value, dtype)
+            value = _maybe_unbox_datetimelike_tz_deprecation(value, dtype)
 
         subarr = np.empty(length, dtype=dtype)
         subarr.fill(value)
 
     return subarr
 
 
-def maybe_unbox_datetimelike_tz_deprecation(value: Scalar, dtype: DtypeObj):
+def _maybe_unbox_datetimelike_tz_deprecation(value: Scalar, dtype: DtypeObj):
     """
-    Wrap maybe_unbox_datetimelike with a check for a timezone-aware Timestamp
+    Wrap _maybe_unbox_datetimelike with a check for a timezone-aware Timestamp
     along with a timezone-naive datetime64 dtype, which is deprecated.
     """
     # Caller is responsible for checking dtype.kind in ["m", "M"]
@@ -1578,7 +1578,7 @@ def maybe_unbox_datetimelike_tz_deprecation(value: Scalar, dtype: DtypeObj):
         value = maybe_box_datetimelike(value, dtype)
 
     try:
-        value = maybe_unbox_datetimelike(value, dtype)
+        value = _maybe_unbox_datetimelike(value, dtype)
     except TypeError:
         if (
             isinstance(value, Timestamp)
@@ -1598,7 +1598,7 @@ def maybe_unbox_datetimelike_tz_deprecation(value: Scalar, dtype: DtypeObj):
                 stacklevel=find_stack_level(),
             )
             new_value = value.tz_localize(None)
-            return maybe_unbox_datetimelike(new_value, dtype)
+            return _maybe_unbox_datetimelike(new_value, dtype)
         else:
             raise
     return value
@@ -1748,7 +1748,7 @@ def convert_scalar_for_putitemlike(scalar: Scalar, dtype: np.dtype) -> Scalar:
     """
     if dtype.kind in ["m", "M"]:
         scalar = maybe_box_datetimelike(scalar, dtype)
-        return maybe_unbox_datetimelike(scalar, dtype)
+        return _maybe_unbox_datetimelike(scalar, dtype)
     else:
         _validate_numeric_casting(dtype, scalar)
     return scalar
@@ -1849,11 +1849,9 @@ def np_can_hold_element(dtype: np.dtype, element: Any) -> Any:
     if dtype == _dtype_obj:
         return element
 
-    tipo = maybe_infer_dtype_type(element)
+    tipo = _maybe_infer_dtype_type(element)
 
     if dtype.kind in ["i", "u"]:
-        info = np.iinfo(dtype)
-
         if isinstance(element, range):
             if _dtype_can_hold_range(element, dtype):
                 return element
@@ -1863,6 +1861,7 @@ def np_can_hold_element(dtype: np.dtype, element: Any) -> Any:
             # e.g. test_setitem_series_int8 if we have a python int 1
             #  tipo may be np.int32, despite the fact that it will fit
             #  in smaller int dtypes.
+            info = np.iinfo(dtype)
             if info.min <= element <= info.max:
                 return element
             raise ValueError
@@ -1964,7 +1963,7 @@ def np_can_hold_element(dtype: np.dtype, element: Any) -> Any:
 
 def _dtype_can_hold_range(rng: range, dtype: np.dtype) -> bool:
     """
-    maybe_infer_dtype_type infers to int64 (and float64 for very large endpoints),
+    _maybe_infer_dtype_type infers to int64 (and float64 for very large endpoints),
     but in many cases a range can be held by a smaller integer dtype.
     Check if this is one of those cases.
     """

pandas/core/internals/blocks.py

@@ -352,7 +352,7 @@ def getitem_block_columns(
 
         return type(self)(new_values, new_mgr_locs, self.ndim)
 
-    # NB: this cannot be made cache_readonly because in libreduction we pin
+    # NB: this cannot be made cache_readonly because in mgr.set_values we pin
     #  new .values that can have different shape GH#42631
     @property
     def shape(self) -> Shape:

pandas/tests/arrays/interval/test_interval.py

@@ -82,14 +82,15 @@ def test_where_raises(self, other):
             ser.where([True, False, True], other=other)
 
     def test_shift(self):
-        # https://github.com/pandas-dev/pandas/issues/31495
+        # https://github.com/pandas-dev/pandas/issues/31495, GH#22428, GH#31502
         a = IntervalArray.from_breaks([1, 2, 3])
         result = a.shift()
         # int -> float
         expected = IntervalArray.from_tuples([(np.nan, np.nan), (1.0, 2.0)])
         tm.assert_interval_array_equal(result, expected)
 
     def test_shift_datetime(self):
+        # GH#31502, GH#31504
         a = IntervalArray.from_breaks(date_range("2000", periods=4))
         result = a.shift(2)
         expected = a.take([-1, -1, 0], allow_fill=True)

pandas/tests/extension/base/methods.py

@@ -382,6 +382,7 @@ def test_shift_empty_array(self, data, periods):
         self.assert_extension_array_equal(result, expected)
 
     def test_shift_zero_copies(self, data):
+        # GH#31502
         result = data.shift(0)
         assert result is not data
 

pandas/tests/frame/indexing/test_where.py

@@ -758,7 +758,7 @@ def test_where_try_cast_deprecated(frame_or_series):
     "the deprecated path, and also up-cast to int64 instead of int32 "
     "(for now)."
 )
-def test_where_int_downcasting_deprecated(using_array_manager, request):
+def test_where_int_downcasting_deprecated(using_array_manager):
     # GH#44597
     arr = np.arange(6).astype(np.int16).reshape(3, 2)
     df = DataFrame(arr)

pandas/tests/indexing/test_coercion.py

@@ -489,36 +489,20 @@ def test_where_series_bool(self, fill_val, exp_dtype):
         ],
         ids=["datetime64", "datetime64tz"],
     )
-    def test_where_series_datetime64(self, fill_val, exp_dtype):
-        klass = pd.Series
-
-        obj = klass(pd.date_range("2011-01-01", periods=4, freq="D")._with_freq(None))
-        assert obj.dtype == "datetime64[ns]"
-        self._run_test(obj, fill_val, klass, exp_dtype)
+    def test_where_datetime64(self, index_or_series, fill_val, exp_dtype):
+        klass = index_or_series
 
-    @pytest.mark.parametrize(
-        "fill_val",
-        [
-            pd.Timestamp("2012-01-01"),
-            pd.Timestamp("2012-01-01").to_datetime64(),
-            pd.Timestamp("2012-01-01").to_pydatetime(),
-        ],
-    )
-    def test_where_index_datetime(self, fill_val):
-        exp_dtype = "datetime64[ns]"
-        klass = pd.Index
         obj = klass(pd.date_range("2011-01-01", periods=4, freq="D")._with_freq(None))
         assert obj.dtype == "datetime64[ns]"
-        self._run_test(obj, fill_val, klass, exp_dtype)
-
-    def test_where_index_datetime64tz(self):
-        klass = pd.Index
 
-        fill_val = pd.Timestamp("2012-01-01", tz="US/Eastern")
-        exp_dtype = object
-        obj = klass(pd.date_range("2011-01-01", periods=4, freq="D")._with_freq(None))
-        assert obj.dtype == "datetime64[ns]"
-        self._run_test(obj, fill_val, klass, exp_dtype)
+        fv = fill_val
+        # do the check with each of the available datetime scalars
+        if exp_dtype == "datetime64[ns]":
+            for scalar in [fv, fv.to_pydatetime(), fv.to_datetime64()]:
+                self._run_test(obj, scalar, klass, exp_dtype)
+        else:
+            for scalar in [fv, fv.to_pydatetime()]:
+                self._run_test(obj, fill_val, klass, exp_dtype)
 
     @pytest.mark.xfail(reason="Test not implemented")
     def test_where_index_complex128(self):
@@ -640,11 +624,16 @@ def test_fillna_float64(self, index_or_series, fill_val, fill_dtype):
             (True, object),
         ],
     )
-    def test_fillna_series_complex128(self, fill_val, fill_dtype):
-        obj = pd.Series([1 + 1j, np.nan, 3 + 3j, 4 + 4j])
+    def test_fillna_complex128(self, index_or_series, fill_val, fill_dtype, request):
+        klass = index_or_series
+        if klass is pd.Index:
+            mark = pytest.mark.xfail(reason="No Index[complex]")
+            request.node.add_marker(mark)
+
+        obj = klass([1 + 1j, np.nan, 3 + 3j, 4 + 4j], dtype=np.complex128)
         assert obj.dtype == np.complex128
 
-        exp = pd.Series([1 + 1j, fill_val, 3 + 3j, 4 + 4j])
+        exp = klass([1 + 1j, fill_val, 3 + 3j, 4 + 4j], dtype=fill_dtype)
         self._assert_fillna_conversion(obj, fill_val, exp, fill_dtype)
 
     @pytest.mark.parametrize(

pandas/tests/series/indexing/test_indexing.py

@@ -321,7 +321,7 @@ def test_frozenset_index():
     assert s[idx1] == 3
 
 
-class TestDepreactedIndexers:
+class TestDeprecatedIndexers:
     @pytest.mark.parametrize("key", [{1}, {1: 1}])
     def test_getitem_dict_and_set_deprecated(self, key):
         # GH#42825

pandas/tests/series/methods/test_drop.py

@@ -21,8 +21,8 @@ def test_drop_unique_and_non_unique_index(
     data, index, axis, drop_labels, expected_data, expected_index
 ):
 
-    s = Series(data=data, index=index)
-    result = s.drop(drop_labels, axis=axis)
+    ser = Series(data=data, index=index)
+    result = ser.drop(drop_labels, axis=axis)
     expected = Series(data=expected_data, index=expected_index)
     tm.assert_series_equal(result, expected)
 
@@ -45,17 +45,17 @@ def test_drop_exception_raised(data, index, drop_labels, axis, error_type, error
 
 def test_drop_with_ignore_errors():
     # errors='ignore'
-    s = Series(range(3), index=list("abc"))
-    result = s.drop("bc", errors="ignore")
-    tm.assert_series_equal(result, s)
-    result = s.drop(["a", "d"], errors="ignore")
-    expected = s.iloc[1:]
+    ser = Series(range(3), index=list("abc"))
+    result = ser.drop("bc", errors="ignore")
+    tm.assert_series_equal(result, ser)
+    result = ser.drop(["a", "d"], errors="ignore")
+    expected = ser.iloc[1:]
     tm.assert_series_equal(result, expected)
 
     # GH 8522
-    s = Series([2, 3], index=[True, False])
-    assert s.index.is_object()
-    result = s.drop(True)
+    ser = Series([2, 3], index=[True, False])
+    assert ser.index.is_object()
+    result = ser.drop(True)
     expected = Series([3], index=[False])
     tm.assert_series_equal(result, expected)
 

pandas/tests/util/test_make_objects.py

@@ -0,0 +1,15 @@
+"""
+Tests for tm.makeFoo functions.
+"""
+
+
+import numpy as np
+
+import pandas._testing as tm
+
+
+def test_make_multiindex_respects_k():
+    # GH#38795 respect 'k' arg
+    N = np.random.randint(0, 100)
+    mi = tm.makeMultiIndex(k=N)
+    assert len(mi) == N