Merge remote-tracking branch 'upstream/master' into future-1

Author: simonjayhawkins

asv_bench/benchmarks/algos/isin.py

@@ -50,8 +50,6 @@ def setup(self, dtype):
 
         elif dtype in ["category[object]", "category[int]"]:
             # Note: sizes are different in this case than others
-            np.random.seed(1234)
-
             n = 5 * 10 ** 5
             sample_size = 100
 
@@ -99,7 +97,6 @@ class IsinAlmostFullWithRandomInt:
     def setup(self, dtype, exponent, title):
         M = 3 * 2 ** (exponent - 2)
         # 0.77-the maximal share of occupied buckets
-        np.random.seed(42)
         self.series = Series(np.random.randint(0, M, M)).astype(dtype)
 
         values = np.random.randint(0, M, M).astype(dtype)
@@ -132,7 +129,6 @@ class IsinWithRandomFloat:
     param_names = ["dtype", "size", "title"]
 
     def setup(self, dtype, size, title):
-        np.random.seed(42)
         self.values = np.random.rand(size)
         self.series = Series(self.values).astype(dtype)
         np.random.shuffle(self.values)
@@ -179,7 +175,6 @@ class IsinWithArange:
 
     def setup(self, dtype, M, offset_factor):
         offset = int(M * offset_factor)
-        np.random.seed(42)
         tmp = Series(np.random.randint(offset, M + offset, 10 ** 6))
         self.series = tmp.astype(dtype)
         self.values = np.arange(M).astype(dtype)
@@ -290,10 +285,8 @@ def setup(self, dtype, MaxNumber, series_type):
             raise NotImplementedError
 
         if series_type == "random_hits":
-            np.random.seed(42)
             array = np.random.randint(0, MaxNumber, N)
         if series_type == "random_misses":
-            np.random.seed(42)
             array = np.random.randint(0, MaxNumber, N) + MaxNumber
         if series_type == "monotone_hits":
             array = np.repeat(np.arange(MaxNumber), N // MaxNumber)
@@ -322,7 +315,6 @@ def setup(self, dtype, series_type):
             raise NotImplementedError
 
         if series_type == "random":
-            np.random.seed(42)
             vals = np.random.randint(0, 10 * N, N)
         if series_type == "monotone":
             vals = np.arange(N)

asv_bench/benchmarks/frame_ctor.py

@@ -67,7 +67,6 @@ class FromDictwithTimestamp:
 
     def setup(self, offset):
         N = 10 ** 3
-        np.random.seed(1234)
         idx = date_range(Timestamp("1/1/1900"), freq=offset, periods=N)
         df = DataFrame(np.random.randn(N, 10), index=idx)
         self.d = df.to_dict()

asv_bench/benchmarks/groupby.py

@@ -507,11 +507,11 @@ def time_frame_agg(self, dtype, method):
         self.df.groupby("key").agg(method)
 
 
-class CumminMax:
+class Cumulative:
     param_names = ["dtype", "method"]
     params = [
         ["float64", "int64", "Float64", "Int64"],
-        ["cummin", "cummax"],
+        ["cummin", "cummax", "cumsum"],
     ]
 
     def setup(self, dtype, method):

asv_bench/benchmarks/hash_functions.py

@@ -67,7 +67,6 @@ class NumericSeriesIndexingShuffled:
 
     def setup(self, index, N):
         vals = np.array(list(range(55)) + [54] + list(range(55, N - 1)))
-        np.random.seed(42)
         np.random.shuffle(vals)
         indices = index(vals)
         self.data = pd.Series(np.arange(N), index=indices)

asv_bench/benchmarks/indexing.py

@@ -368,17 +368,14 @@ def setup(self):
         self.df = DataFrame(index=range(self.N))
 
     def time_insert(self):
-        np.random.seed(1234)
         for i in range(100):
             self.df.insert(0, i, np.random.randn(self.N), allow_duplicates=True)
 
     def time_assign_with_setitem(self):
-        np.random.seed(1234)
         for i in range(100):
             self.df[i] = np.random.randn(self.N)
 
     def time_assign_list_like_with_setitem(self):
-        np.random.seed(1234)
         self.df[list(range(100))] = np.random.randn(self.N, 100)
 
     def time_assign_list_of_columns_concat(self):

asv_bench/benchmarks/series_methods.py

@@ -145,7 +145,6 @@ class Mode:
     param_names = ["N", "dtype"]
 
     def setup(self, N, dtype):
-        np.random.seed(42)
         self.s = Series(np.random.randint(0, N, size=10 * N)).astype(dtype)
 
     def time_mode(self, N, dtype):

ci/code_checks.sh

@@ -77,6 +77,10 @@ if [[ -z "$CHECK" || "$CHECK" == "patterns" ]]; then
     invgrep -R --include="*.rst" -E "[a-zA-Z0-9]\`\`?[a-zA-Z0-9]" doc/source/
     RET=$(($RET + $?)) ; echo $MSG "DONE"
 
+    MSG='Check for unnecessary random seeds in asv benchmarks' ; echo $MSG
+    invgrep -R --exclude pandas_vb_common.py -E 'np.random.seed' asv_bench/benchmarks/
+    RET=$(($RET + $?)) ; echo $MSG "DONE"
+
 fi
 
 ### CODE ###

doc/source/whatsnew/v1.3.0.rst

@@ -247,24 +247,24 @@ def group_cumsum(numeric[:, ::1] out,
             for j in range(K):
                 val = values[i, j]
 
+                # For floats, use Kahan summation to reduce floating-point
+                # error (https://en.wikipedia.org/wiki/Kahan_summation_algorithm)
                 if numeric == float32_t or numeric == float64_t:
                     if val == val:
                         y = val - compensation[lab, j]
                         t = accum[lab, j] + y
                         compensation[lab, j] = t - accum[lab, j] - y
                         accum[lab, j] = t
-                        out[i, j] = accum[lab, j]
+                        out[i, j] = t
                     else:
                         out[i, j] = NaN
                         if not skipna:
                             accum[lab, j] = NaN
                             break
                 else:
-                    y = val - compensation[lab, j]
-                    t = accum[lab, j] + y
-                    compensation[lab, j] = t - accum[lab, j] - y
+                    t = val + accum[lab, j]
                     accum[lab, j] = t
-                    out[i, j] = accum[lab, j]
+                    out[i, j] = t
 
 
 @cython.boundscheck(False)

pandas/_libs/groupby.pyx

@@ -15,12 +15,9 @@
 methods that are spread throughout the codebase. This module will make it
 easier to adjust to future upstream changes in the analogous numpy signatures.
 """
-from typing import (
-    Any,
-    Dict,
-    Optional,
-    Union,
-)
+from __future__ import annotations
+
+from typing import Any
 
 from numpy import ndarray
 
@@ -41,7 +38,7 @@ def __init__(
         self,
         defaults,
         fname=None,
-        method: Optional[str] = None,
+        method: str | None = None,
         max_fname_arg_count=None,
     ):
         self.fname = fname
@@ -55,7 +52,7 @@ def __call__(
         kwargs,
         fname=None,
         max_fname_arg_count=None,
-        method: Optional[str] = None,
+        method: str | None = None,
     ) -> None:
         if args or kwargs:
             fname = self.fname if fname is None else fname
@@ -119,7 +116,7 @@ def validate_argmax_with_skipna(skipna, args, kwargs):
     return skipna
 
 
-ARGSORT_DEFAULTS: Dict[str, Optional[Union[int, str]]] = {}
+ARGSORT_DEFAULTS: dict[str, int | str | None] = {}
 ARGSORT_DEFAULTS["axis"] = -1
 ARGSORT_DEFAULTS["kind"] = "quicksort"
 ARGSORT_DEFAULTS["order"] = None
@@ -132,7 +129,7 @@ def validate_argmax_with_skipna(skipna, args, kwargs):
 
 # two different signatures of argsort, this second validation for when the
 # `kind` param is supported
-ARGSORT_DEFAULTS_KIND: Dict[str, Optional[int]] = {}
+ARGSORT_DEFAULTS_KIND: dict[str, int | None] = {}
 ARGSORT_DEFAULTS_KIND["axis"] = -1
 ARGSORT_DEFAULTS_KIND["order"] = None
 validate_argsort_kind = CompatValidator(
@@ -155,7 +152,7 @@ def validate_argsort_with_ascending(ascending, args, kwargs):
     return ascending
 
 
-CLIP_DEFAULTS: Dict[str, Any] = {"out": None}
+CLIP_DEFAULTS: dict[str, Any] = {"out": None}
 validate_clip = CompatValidator(
     CLIP_DEFAULTS, fname="clip", method="both", max_fname_arg_count=3
 )
@@ -176,7 +173,7 @@ def validate_clip_with_axis(axis, args, kwargs):
     return axis
 
 
-CUM_FUNC_DEFAULTS: Dict[str, Any] = {}
+CUM_FUNC_DEFAULTS: dict[str, Any] = {}
 CUM_FUNC_DEFAULTS["dtype"] = None
 CUM_FUNC_DEFAULTS["out"] = None
 validate_cum_func = CompatValidator(
@@ -201,7 +198,7 @@ def validate_cum_func_with_skipna(skipna, args, kwargs, name):
     return skipna
 
 
-ALLANY_DEFAULTS: Dict[str, Optional[bool]] = {}
+ALLANY_DEFAULTS: dict[str, bool | None] = {}
 ALLANY_DEFAULTS["dtype"] = None
 ALLANY_DEFAULTS["out"] = None
 ALLANY_DEFAULTS["keepdims"] = False
@@ -224,28 +221,28 @@ def validate_cum_func_with_skipna(skipna, args, kwargs, name):
     MINMAX_DEFAULTS, fname="max", method="both", max_fname_arg_count=1
 )
 
-RESHAPE_DEFAULTS: Dict[str, str] = {"order": "C"}
+RESHAPE_DEFAULTS: dict[str, str] = {"order": "C"}
 validate_reshape = CompatValidator(
     RESHAPE_DEFAULTS, fname="reshape", method="both", max_fname_arg_count=1
 )
 
-REPEAT_DEFAULTS: Dict[str, Any] = {"axis": None}
+REPEAT_DEFAULTS: dict[str, Any] = {"axis": None}
 validate_repeat = CompatValidator(
     REPEAT_DEFAULTS, fname="repeat", method="both", max_fname_arg_count=1
 )
 
-ROUND_DEFAULTS: Dict[str, Any] = {"out": None}
+ROUND_DEFAULTS: dict[str, Any] = {"out": None}
 validate_round = CompatValidator(
     ROUND_DEFAULTS, fname="round", method="both", max_fname_arg_count=1
 )
 
-SORT_DEFAULTS: Dict[str, Optional[Union[int, str]]] = {}
+SORT_DEFAULTS: dict[str, int | str | None] = {}
 SORT_DEFAULTS["axis"] = -1
 SORT_DEFAULTS["kind"] = "quicksort"
 SORT_DEFAULTS["order"] = None
 validate_sort = CompatValidator(SORT_DEFAULTS, fname="sort", method="kwargs")
 
-STAT_FUNC_DEFAULTS: Dict[str, Optional[Any]] = {}
+STAT_FUNC_DEFAULTS: dict[str, Any | None] = {}
 STAT_FUNC_DEFAULTS["dtype"] = None
 STAT_FUNC_DEFAULTS["out"] = None
 
@@ -279,13 +276,13 @@ def validate_cum_func_with_skipna(skipna, args, kwargs, name):
     MEDIAN_DEFAULTS, fname="median", method="both", max_fname_arg_count=1
 )
 
-STAT_DDOF_FUNC_DEFAULTS: Dict[str, Optional[bool]] = {}
+STAT_DDOF_FUNC_DEFAULTS: dict[str, bool | None] = {}
 STAT_DDOF_FUNC_DEFAULTS["dtype"] = None
 STAT_DDOF_FUNC_DEFAULTS["out"] = None
 STAT_DDOF_FUNC_DEFAULTS["keepdims"] = False
 validate_stat_ddof_func = CompatValidator(STAT_DDOF_FUNC_DEFAULTS, method="kwargs")
 
-TAKE_DEFAULTS: Dict[str, Optional[str]] = {}
+TAKE_DEFAULTS: dict[str, str | None] = {}
 TAKE_DEFAULTS["out"] = None
 TAKE_DEFAULTS["mode"] = "raise"
 validate_take = CompatValidator(TAKE_DEFAULTS, fname="take", method="kwargs")
@@ -392,7 +389,7 @@ def validate_resampler_func(method: str, args, kwargs) -> None:
             raise TypeError("too many arguments passed in")
 
 
-def validate_minmax_axis(axis: Optional[int], ndim: int = 1) -> None:
+def validate_minmax_axis(axis: int | None, ndim: int = 1) -> None:
     """
     Ensure that the axis argument passed to min, max, argmin, or argmax is zero
     or None, as otherwise it will be incorrectly ignored.