Merge branch 'master' into CLN-libs-2

Author: ShaharNaveh

.github/workflows/ci.yml

@@ -154,6 +154,39 @@ jobs:
         echo "region = BHS" >> $RCLONE_CONFIG_PATH
       if: github.event_name == 'push'
 
-    - name: Sync web
+    - name: Sync web with OVH
       run: rclone sync pandas_web ovh_cloud_pandas_web:dev
       if: github.event_name == 'push'
+
+    - name: Create git repo to upload the built docs to GitHub pages
+      run: |
+        cd pandas_web
+        git init
+        touch .nojekyll
+        echo "dev.pandas.io" > CNAME
+        printf "User-agent: *\nDisallow: /" > robots.txt
+        git add --all .
+        git config user.email "[email protected]"
+        git config user.name "pandas-bot"
+        git commit -m "pandas web and documentation in master"
+      if: github.event_name == 'push'
+
+    # For this task to work, next steps are required:
+    # 1. Generate a pair of private/public keys (i.e. `ssh-keygen -t rsa -b 4096 -C "[email protected]"`)
+    # 2. Go to https://github.com/pandas-dev/pandas/settings/secrets
+    # 3. Click on "Add a new secret"
+    # 4. Name: "github_pagas_ssh_key", Value: <Content of the private ssh key>
+    # 5. The public key needs to be upladed to https://github.com/pandas-dev/pandas-dev.github.io/settings/keys
+    - name: Install GitHub pages ssh deployment key
+      uses: shimataro/ssh-key-action@v2
+      with:
+        key: ${{ secrets.github_pages_ssh_key }}
+        known_hosts: 'github.com,192.30.252.128 ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ=='
+      if: github.event_name == 'push'
+
+    - name: Publish web and docs to GitHub pages
+      run: |
+        cd pandas_web
+        git remote add origin [email protected]:pandas-dev/pandas-dev.github.io.git
+        git push -f origin master || true
+      if: github.event_name == 'push'

asv_bench/benchmarks/algorithms.py

@@ -31,83 +31,62 @@ def time_maybe_convert_objects(self):
 
 class Factorize:
 
-    params = [[True, False], ["int", "uint", "float", "string"]]
-    param_names = ["sort", "dtype"]
-
-    def setup(self, sort, dtype):
-        N = 10 ** 5
-        data = {
-            "int": pd.Int64Index(np.arange(N).repeat(5)),
-            "uint": pd.UInt64Index(np.arange(N).repeat(5)),
-            "float": pd.Float64Index(np.random.randn(N).repeat(5)),
-            "string": tm.makeStringIndex(N).repeat(5),
-        }
-        self.idx = data[dtype]
-
-    def time_factorize(self, sort, dtype):
-        self.idx.factorize(sort=sort)
-
-
-class FactorizeUnique:
-
-    params = [[True, False], ["int", "uint", "float", "string"]]
-    param_names = ["sort", "dtype"]
+    params = [
+        [True, False],
+        [True, False],
+        ["int", "uint", "float", "string", "datetime64[ns]", "datetime64[ns, tz]"],
+    ]
+    param_names = ["unique", "sort", "dtype"]
 
-    def setup(self, sort, dtype):
+    def setup(self, unique, sort, dtype):
         N = 10 ** 5
         data = {
             "int": pd.Int64Index(np.arange(N)),
             "uint": pd.UInt64Index(np.arange(N)),
-            "float": pd.Float64Index(np.arange(N)),
+            "float": pd.Float64Index(np.random.randn(N)),
             "string": tm.makeStringIndex(N),
-        }
-        self.idx = data[dtype]
-        assert self.idx.is_unique
-
-    def time_factorize(self, sort, dtype):
+            "datetime64[ns]": pd.date_range("2011-01-01", freq="H", periods=N),
+            "datetime64[ns, tz]": pd.date_range(
+                "2011-01-01", freq="H", periods=N, tz="Asia/Tokyo"
+            ),
+        }[dtype]
+        if not unique:
+            data = data.repeat(5)
+        self.idx = data
+
+    def time_factorize(self, unique, sort, dtype):
         self.idx.factorize(sort=sort)
 
 
 class Duplicated:
 
-    params = [["first", "last", False], ["int", "uint", "float", "string"]]
-    param_names = ["keep", "dtype"]
-
-    def setup(self, keep, dtype):
-        N = 10 ** 5
-        data = {
-            "int": pd.Int64Index(np.arange(N).repeat(5)),
-            "uint": pd.UInt64Index(np.arange(N).repeat(5)),
-            "float": pd.Float64Index(np.random.randn(N).repeat(5)),
-            "string": tm.makeStringIndex(N).repeat(5),
-        }
-        self.idx = data[dtype]
-        # cache is_unique
-        self.idx.is_unique
-
-    def time_duplicated(self, keep, dtype):
-        self.idx.duplicated(keep=keep)
-
-
-class DuplicatedUniqueIndex:
-
-    params = ["int", "uint", "float", "string"]
-    param_names = ["dtype"]
+    params = [
+        [True, False],
+        ["first", "last", False],
+        ["int", "uint", "float", "string", "datetime64[ns]", "datetime64[ns, tz]"],
+    ]
+    param_names = ["unique", "keep", "dtype"]
 
-    def setup(self, dtype):
+    def setup(self, unique, keep, dtype):
         N = 10 ** 5
         data = {
             "int": pd.Int64Index(np.arange(N)),
             "uint": pd.UInt64Index(np.arange(N)),
             "float": pd.Float64Index(np.random.randn(N)),
             "string": tm.makeStringIndex(N),
-        }
-        self.idx = data[dtype]
+            "datetime64[ns]": pd.date_range("2011-01-01", freq="H", periods=N),
+            "datetime64[ns, tz]": pd.date_range(
+                "2011-01-01", freq="H", periods=N, tz="Asia/Tokyo"
+            ),
+        }[dtype]
+        if not unique:
+            data = data.repeat(5)
+        self.idx = data
         # cache is_unique
         self.idx.is_unique
 
-    def time_duplicated_unique(self, dtype):
-        self.idx.duplicated()
+    def time_duplicated(self, unique, keep, dtype):
+        self.idx.duplicated(keep=keep)
 
 
 class Hashing:

asv_bench/benchmarks/binary_ops.py renamed to asv_bench/benchmarks/arithmetic.py

@@ -1,14 +1,23 @@
 import operator
+import warnings
 
 import numpy as np
 
-from pandas import DataFrame, Series, date_range
+import pandas as pd
+from pandas import DataFrame, Series, Timestamp, date_range, to_timedelta
+import pandas._testing as tm
 from pandas.core.algorithms import checked_add_with_arr
 
+from .pandas_vb_common import numeric_dtypes
+
 try:
     import pandas.core.computation.expressions as expr
 except ImportError:
     import pandas.computation.expressions as expr
+try:
+    import pandas.tseries.holiday
+except ImportError:
+    pass
 
 
 class IntFrameWithScalar:
@@ -151,6 +160,110 @@ def time_timestamp_ops_diff_with_shift(self, tz):
         self.s - self.s.shift()
 
 
+class IrregularOps:
+    def setup(self):
+        N = 10 ** 5
+        idx = date_range(start="1/1/2000", periods=N, freq="s")
+        s = Series(np.random.randn(N), index=idx)
+        self.left = s.sample(frac=1)
+        self.right = s.sample(frac=1)
+
+    def time_add(self):
+        self.left + self.right
+
+
+class TimedeltaOps:
+    def setup(self):
+        self.td = to_timedelta(np.arange(1000000))
+        self.ts = Timestamp("2000")
+
+    def time_add_td_ts(self):
+        self.td + self.ts
+
+
+class CategoricalComparisons:
+    params = ["__lt__", "__le__", "__eq__", "__ne__", "__ge__", "__gt__"]
+    param_names = ["op"]
+
+    def setup(self, op):
+        N = 10 ** 5
+        self.cat = pd.Categorical(list("aabbcd") * N, ordered=True)
+
+    def time_categorical_op(self, op):
+        getattr(self.cat, op)("b")
+
+
+class IndexArithmetic:
+
+    params = ["float", "int"]
+    param_names = ["dtype"]
+
+    def setup(self, dtype):
+        N = 10 ** 6
+        indexes = {"int": "makeIntIndex", "float": "makeFloatIndex"}
+        self.index = getattr(tm, indexes[dtype])(N)
+
+    def time_add(self, dtype):
+        self.index + 2
+
+    def time_subtract(self, dtype):
+        self.index - 2
+
+    def time_multiply(self, dtype):
+        self.index * 2
+
+    def time_divide(self, dtype):
+        self.index / 2
+
+    def time_modulo(self, dtype):
+        self.index % 2
+
+
+class NumericInferOps:
+    # from GH 7332
+    params = numeric_dtypes
+    param_names = ["dtype"]
+
+    def setup(self, dtype):
+        N = 5 * 10 ** 5
+        self.df = DataFrame(
+            {"A": np.arange(N).astype(dtype), "B": np.arange(N).astype(dtype)}
+        )
+
+    def time_add(self, dtype):
+        self.df["A"] + self.df["B"]
+
+    def time_subtract(self, dtype):
+        self.df["A"] - self.df["B"]
+
+    def time_multiply(self, dtype):
+        self.df["A"] * self.df["B"]
+
+    def time_divide(self, dtype):
+        self.df["A"] / self.df["B"]
+
+    def time_modulo(self, dtype):
+        self.df["A"] % self.df["B"]
+
+
+class DateInferOps:
+    # from GH 7332
+    def setup_cache(self):
+        N = 5 * 10 ** 5
+        df = DataFrame({"datetime64": np.arange(N).astype("datetime64[ms]")})
+        df["timedelta"] = df["datetime64"] - df["datetime64"]
+        return df
+
+    def time_subtract_datetimes(self, df):
+        df["datetime64"] - df["datetime64"]
+
+    def time_timedelta_plus_datetime(self, df):
+        df["timedelta"] + df["datetime64"]
+
+    def time_add_timedeltas(self, df):
+        df["timedelta"] + df["timedelta"]
+
+
 class AddOverflowScalar:
 
     params = [1, -1, 0]
@@ -188,4 +301,68 @@ def time_add_overflow_both_arg_nan(self):
         )
 
 
+hcal = pd.tseries.holiday.USFederalHolidayCalendar()
+# These offsets currently raise a NotImplimentedError with .apply_index()
+non_apply = [
+    pd.offsets.Day(),
+    pd.offsets.BYearEnd(),
+    pd.offsets.BYearBegin(),
+    pd.offsets.BQuarterEnd(),
+    pd.offsets.BQuarterBegin(),
+    pd.offsets.BMonthEnd(),
+    pd.offsets.BMonthBegin(),
+    pd.offsets.CustomBusinessDay(),
+    pd.offsets.CustomBusinessDay(calendar=hcal),
+    pd.offsets.CustomBusinessMonthBegin(calendar=hcal),
+    pd.offsets.CustomBusinessMonthEnd(calendar=hcal),
+    pd.offsets.CustomBusinessMonthEnd(calendar=hcal),
+]
+other_offsets = [
+    pd.offsets.YearEnd(),
+    pd.offsets.YearBegin(),
+    pd.offsets.QuarterEnd(),
+    pd.offsets.QuarterBegin(),
+    pd.offsets.MonthEnd(),
+    pd.offsets.MonthBegin(),
+    pd.offsets.DateOffset(months=2, days=2),
+    pd.offsets.BusinessDay(),
+    pd.offsets.SemiMonthEnd(),
+    pd.offsets.SemiMonthBegin(),
+]
+offsets = non_apply + other_offsets
+
+
+class OffsetArrayArithmetic:
+
+    params = offsets
+    param_names = ["offset"]
+
+    def setup(self, offset):
+        N = 10000
+        rng = pd.date_range(start="1/1/2000", periods=N, freq="T")
+        self.rng = rng
+        self.ser = pd.Series(rng)
+
+    def time_add_series_offset(self, offset):
+        with warnings.catch_warnings(record=True):
+            self.ser + offset
+
+    def time_add_dti_offset(self, offset):
+        with warnings.catch_warnings(record=True):
+            self.rng + offset
+
+
+class ApplyIndex:
+    params = other_offsets
+    param_names = ["offset"]
+
+    def setup(self, offset):
+        N = 10000
+        rng = pd.date_range(start="1/1/2000", periods=N, freq="T")
+        self.rng = rng
+
+    def time_apply_index(self, offset):
+        offset.apply_index(self.rng)
+
+
 from .pandas_vb_common import setup  # noqa: F401 isort:skip

asv_bench/benchmarks/categoricals.py

@@ -63,18 +63,6 @@ def time_existing_series(self):
         pd.Categorical(self.series)
 
 
-class CategoricalOps:
-    params = ["__lt__", "__le__", "__eq__", "__ne__", "__ge__", "__gt__"]
-    param_names = ["op"]
-
-    def setup(self, op):
-        N = 10 ** 5
-        self.cat = pd.Categorical(list("aabbcd") * N, ordered=True)
-
-    def time_categorical_op(self, op):
-        getattr(self.cat, op)("b")
-
-
 class Concat:
     def setup(self):
         N = 10 ** 5
@@ -270,9 +258,6 @@ def setup(self):
     def time_get_loc(self):
         self.index.get_loc(self.category)
 
-    def time_shape(self):
-        self.index.shape
-
     def time_shallow_copy(self):
         self.index._shallow_copy()