CLN: assorted cleanups (#33297)

Author: jbrockmendel

pandas/_libs/tslibs/timedeltas.pxd

@@ -1,6 +1,6 @@
 from numpy cimport int64_t
 
 # Exposed for tslib, not intended for outside use.
-cdef int64_t cast_from_unit(object ts, object unit) except? -1
+cdef int64_t cast_from_unit(object ts, str unit) except? -1
 cpdef int64_t delta_to_nanoseconds(delta) except? -1
 cdef convert_to_timedelta64(object ts, object unit)

pandas/_libs/tslibs/timedeltas.pyx

@@ -257,10 +257,15 @@ def array_to_timedelta64(object[:] values, unit='ns', errors='raise'):
     return iresult.base  # .base to access underlying np.ndarray
 
 
-cpdef inline object precision_from_unit(object unit):
+cpdef inline object precision_from_unit(str unit):
     """
     Return a casting of the unit represented to nanoseconds + the precision
     to round the fractional part.
+
+    Notes
+    -----
+    The caller is responsible for ensuring that the default value of "ns"
+    takes the place of None.
     """
     cdef:
         int64_t m
@@ -301,7 +306,7 @@ cpdef inline object precision_from_unit(object unit):
     return m, p
 
 
-cdef inline int64_t cast_from_unit(object ts, object unit) except? -1:
+cdef inline int64_t cast_from_unit(object ts, str unit) except? -1:
     """ return a casting of the unit represented to nanoseconds
         round the fractional part of a float to our precision, p """
     cdef:
@@ -525,15 +530,24 @@ cdef inline timedelta_from_spec(object number, object frac, object unit):
     return cast_from_unit(float(n), unit)
 
 
-cpdef inline object parse_timedelta_unit(object unit):
+cpdef inline str parse_timedelta_unit(object unit):
     """
     Parameters
     ----------
-    unit : an unit string
+    unit : str or None
+
+    Returns
+    -------
+    str
+        Canonical unit string.
+
+    Raises
+    ------
+    ValueError : on non-parseable input
     """
     if unit is None:
-        return 'ns'
-    elif unit == 'M':
+        return "ns"
+    elif unit == "M":
         return unit
     try:
         return timedelta_abbrevs[unit.lower()]
@@ -622,14 +636,14 @@ def _binary_op_method_timedeltalike(op, name):
 # ----------------------------------------------------------------------
 # Timedelta Construction
 
-cdef inline int64_t parse_iso_format_string(object ts) except? -1:
+cdef inline int64_t parse_iso_format_string(str ts) except? -1:
     """
     Extracts and cleanses the appropriate values from a match object with
     groups for each component of an ISO 8601 duration
 
     Parameters
     ----------
-    ts:
+    ts: str
         ISO 8601 Duration formatted string
 
     Returns

pandas/core/generic.py

@@ -4504,6 +4504,8 @@ def _reindex_with_indexers(
                 allow_dups=allow_dups,
                 copy=copy,
             )
+            # If we've made a copy once, no need to make another one
+            copy = False
 
         if copy and new_data is self._mgr:
             new_data = new_data.copy()
@@ -6468,7 +6470,6 @@ def replace(
     ):
         if not (
             is_scalar(to_replace)
-            or isinstance(to_replace, pd.Series)
             or is_re_compilable(to_replace)
             or is_list_like(to_replace)
         ):

pandas/core/indexes/base.py

@@ -1,3 +1,4 @@
+from copy import copy as copy_func
 from datetime import datetime
 import operator
 from textwrap import dedent
@@ -5313,7 +5314,7 @@ def _add_numeric_methods_unary(cls):
         Add in numeric unary methods.
         """
 
-        def _make_evaluate_unary(op, opstr):
+        def _make_evaluate_unary(op, opstr: str_t):
             def _evaluate_numeric_unary(self):
 
                 attrs = self._get_attributes_dict()
@@ -5419,7 +5420,7 @@ def _add_logical_methods(cls):
         """
         )
 
-        def _make_logical_function(name, desc, f):
+        def _make_logical_function(name: str_t, desc: str_t, f):
             @Substitution(outname=name, desc=desc)
             @Appender(_index_shared_docs["index_" + name])
             @Appender(_doc)
@@ -5508,15 +5509,15 @@ def ensure_index_from_sequences(sequences, names=None):
         return MultiIndex.from_arrays(sequences, names=names)
 
 
-def ensure_index(index_like, copy=False):
+def ensure_index(index_like, copy: bool = False):
     """
     Ensure that we have an index from some index-like object.
 
     Parameters
     ----------
-    index : sequence
+    index_like : sequence
         An Index or other sequence
-    copy : bool
+    copy : bool, default False
 
     Returns
     -------
@@ -5567,9 +5568,7 @@ def ensure_index(index_like, copy=False):
         # clean_index_list does the equivalent of copying
         # so only need to do this if not list instance
         if copy:
-            from copy import copy
-
-            index_like = copy(index_like)
+            index_like = copy_func(index_like)
 
     return Index(index_like)
 
@@ -5596,7 +5595,7 @@ def _trim_front(strings):
     return trimmed
 
 
-def _validate_join_method(method):
+def _validate_join_method(method: str):
     if method not in ["left", "right", "inner", "outer"]:
         raise ValueError(f"do not recognize join method {method}")
 

pandas/core/indexing.py

@@ -625,7 +625,7 @@ def _ensure_listlike_indexer(self, key, axis=None):
 
         Parameters
         ----------
-        key : _LocIndexer key or list-like of column labels
+        key : list-like of column labels
             Target labels.
         axis : key axis if known
         """
@@ -636,7 +636,7 @@ def _ensure_listlike_indexer(self, key, axis=None):
             return
 
         if isinstance(key, tuple):
-            # key may be a tuple if key is a _LocIndexer key
+            # key may be a tuple if we are .loc
             # in that case, set key to the column part of key
             key = key[column_axis]
             axis = column_axis
@@ -649,9 +649,7 @@ def _ensure_listlike_indexer(self, key, axis=None):
             and all(is_hashable(k) for k in key)
         ):
             for k in key:
-                try:
-                    self.obj[k]
-                except KeyError:
+                if k not in self.obj:
                     self.obj[k] = np.nan
 
     def __setitem__(self, key, value):

pandas/core/series.py

@@ -415,7 +415,7 @@ def _set_axis(self, axis: int, labels, fastpath: bool = False) -> None:
 
         object.__setattr__(self, "_index", labels)
         if not fastpath:
-            # The ensure_index call aabove ensures we have an Index object
+            # The ensure_index call above ensures we have an Index object
             self._mgr.set_axis(axis, labels)
 
     # ndarray compatibility