deps: update mininum support to Python 3.8, Pandas 1.5.0, Numpy 1.20.3

Author: chelsea-lin

.github/sync-repo-settings.yaml

@@ -11,7 +11,6 @@ branchProtectionRules:
     - 'cla/google'
     - 'docs'
     - 'lint'
-    - 'unit (3.8)'
     - 'unit (3.9)'
     - 'unit (3.10)'
     - 'unit (3.11)'

.github/workflows/unittest.yml

@@ -8,7 +8,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10', '3.11', '3.12']
+        python: ['3.9', '3.10', '3.11', '3.12']
     steps:
     - name: Checkout
       uses: actions/checkout@v4
@@ -100,7 +100,7 @@ jobs:
     - name: Setup Python
       uses: actions/setup-python@v5
       with:
-        python-version: "3.8"
+        python-version: "3.9"
     - name: Install coverage
       run: |
         python -m pip install --upgrade setuptools pip wheel

db_dtypes/__init__.py

@@ -18,20 +18,17 @@
 import datetime
 import re
 from typing import Optional, Union
-import warnings
 
 import numpy
-import packaging.version
 import pandas
 import pandas.api.extensions
 from pandas.errors import OutOfBoundsDatetime
 import pyarrow
 import pyarrow.compute
 
 from db_dtypes import core
+from db_dtypes.json import JSONArray, JSONDtype
 from db_dtypes.version import __version__
-from . import _versions_helpers
-
 
 date_dtype_name = "dbdate"
 time_dtype_name = "dbtime"
@@ -47,15 +44,6 @@
 _NP_BOX_DTYPE = "datetime64[us]"
 
 
-# To use JSONArray and JSONDtype, you'll need Pandas 1.5.0 or later. With the removal
-# of Python 3.7 compatibility, the minimum Pandas version will be updated to 1.5.0.
-if packaging.version.Version(pandas.__version__) >= packaging.version.Version("1.5.0"):
-    from db_dtypes.json import JSONArray, JSONDtype
-else:
-    JSONArray = None
-    JSONDtype = None
-
-
 @pandas.api.extensions.register_extension_dtype
 class TimeDtype(core.BaseDatetimeDtype):
     """
@@ -347,33 +335,12 @@ def __sub__(self, other):
         return super().__sub__(other)
 
 
-sys_major, sys_minor, sys_micro = _versions_helpers.extract_runtime_version()
-if sys_major == 3 and sys_minor in (7, 8):
-    warnings.warn(
-        "The python-bigquery library will stop supporting Python 3.7 "
-        "and Python 3.8 in a future major release expected in Q4 2024. "
-        f"Your Python version is {sys_major}.{sys_minor}.{sys_micro}. We "
-        "recommend that you update soon to ensure ongoing support. For "
-        "more details, see: [Google Cloud Client Libraries Supported Python Versions policy](https://cloud.google.com/python/docs/supported-python-versions)",
-        PendingDeprecationWarning,
-    )
-
-
-if not JSONArray or not JSONDtype:
-    __all__ = [
-        "__version__",
-        "DateArray",
-        "DateDtype",
-        "TimeArray",
-        "TimeDtype",
-    ]
-else:
-    __all__ = [
-        "__version__",
-        "DateArray",
-        "DateDtype",
-        "JSONDtype",
-        "JSONArray",
-        "TimeArray",
-        "TimeDtype",
-    ]
+__all__ = [
+    "__version__",
+    "DateArray",
+    "DateDtype",
+    "JSONDtype",
+    "JSONArray",
+    "TimeArray",
+    "TimeDtype",
+]

db_dtypes/_versions_helpers.py

@@ -19,12 +19,11 @@
 the versions in the later versions of pandas.
 """
 
+import operator
 from typing import Any
 
-import numpy
 import packaging.version
 import pandas
-from pandas.api.types import is_integer
 import pandas.compat.numpy.function
 import pandas.core.nanops
 
@@ -35,17 +34,15 @@
 nanany = pandas.core.nanops.nanany
 nanmax = pandas.core.nanops.nanmax
 nanmin = pandas.core.nanops.nanmin
+nanmedian = pandas.core.nanops.nanmedian
 numpy_validate_all = pandas.compat.numpy.function.validate_all
 numpy_validate_any = pandas.compat.numpy.function.validate_any
 numpy_validate_max = pandas.compat.numpy.function.validate_max
 numpy_validate_min = pandas.compat.numpy.function.validate_min
+numpy_validate_median = pandas.compat.numpy.function.validate_median
 
-if pandas_release >= (1, 3):
-    nanmedian = pandas.core.nanops.nanmedian
-    numpy_validate_median = pandas.compat.numpy.function.validate_median
 
-
-def import_default(module_name, force=False, default=None):
+def import_default(module_name, default=None):
     """
     Provide an implementation for a class or function when it can't be imported
 
@@ -56,10 +53,7 @@ def import_default(module_name, force=False, default=None):
     """
 
     if default is None:
-        return lambda func_or_class: import_default(module_name, force, func_or_class)
-
-    if force:
-        return default
+        return lambda func_or_class: import_default(module_name, func_or_class)
 
     name = default.__name__
     module = __import__(module_name, {}, {}, [name])
@@ -73,85 +67,11 @@ def import_default(module_name, force=False, default=None):
 # 'datetime.date'
 @import_default("pandas.core.arraylike")
 class OpsMixin:
-    def _cmp_method(self, other, op):  # pragma: NO COVER
-        return NotImplemented
+    pass
 
 
 # TODO: use public API once pandas 1.5 / 2.x is released.
 # See: https://github.com/pandas-dev/pandas/pull/45544
-@import_default("pandas.core.arrays._mixins", pandas_release < (1, 3))
+@import_default("pandas.core.arrays._mixins")
 class NDArrayBackedExtensionArray(pandas.core.arrays.base.ExtensionArray):
-    def __init__(self, values, dtype):
-        assert isinstance(values, numpy.ndarray)
-        self._ndarray = values
-        self._dtype = dtype
-
-    @classmethod
-    def _from_backing_data(cls, data):
-        return cls(data, data.dtype)
-
-    def __getitem__(self, index):
-        value = self._ndarray[index]
-        if is_integer(index):
-            return self._box_func(value)
-        return self.__class__(value, self._dtype)
-
-    def __setitem__(self, index, value):
-        self._ndarray[index] = self._validate_setitem_value(value)
-
-    def __len__(self):
-        return len(self._ndarray)
-
-    @property
-    def shape(self):
-        return self._ndarray.shape
-
-    @property
-    def ndim(self) -> int:
-        return self._ndarray.ndim
-
-    @property
-    def size(self) -> int:
-        return self._ndarray.size
-
-    @property
-    def nbytes(self) -> int:
-        return self._ndarray.nbytes
-
-    def copy(self):
-        return self[:]
-
-    def repeat(self, n):
-        return self.__class__(self._ndarray.repeat(n), self._dtype)
-
-    def take(
-        self,
-        indices,
-        *,
-        allow_fill: bool = False,
-        fill_value: Any = None,
-        axis: int = 0,
-    ):
-        from pandas.core.algorithms import take
-
-        if allow_fill:
-            fill_value = self._validate_scalar(fill_value)
-
-        new_data = take(
-            self._ndarray,
-            indices,
-            allow_fill=allow_fill,
-            fill_value=fill_value,
-            axis=axis,
-        )
-        return self._from_backing_data(new_data)
-
-    @classmethod
-    def _concat_same_type(cls, to_concat, axis=0):
-        dtypes = {str(x.dtype) for x in to_concat}
-        if len(dtypes) != 1:
-            raise ValueError("to_concat must have the same dtype (tz)", dtypes)
-
-        new_values = [x._ndarray for x in to_concat]
-        new_values = numpy.concatenate(new_values, axis=axis)
-        return to_concat[0]._from_backing_data(new_values)  # type: ignore[arg-type]
+    pass

db_dtypes/pandas_backports.py

@@ -30,10 +30,10 @@
 release_status = "Development Status :: 5 - Production/Stable"
 
 dependencies = [
-    "packaging >= 17.0",
-    "pandas >= 0.24.2",
+    "packaging>=17.0",
+    "pandas>=1.5.0",
     "pyarrow>=3.0.0",
-    "numpy >= 1.16.6",
+    "numpy>=1.20.3",
 ]
 
 package_root = os.path.abspath(os.path.dirname(__file__))
@@ -63,8 +63,6 @@ def readme():
         "License :: OSI Approved :: Apache Software License",
         "Programming Language :: Python",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.7",
-        "Programming Language :: Python :: 3.8",
         "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
         "Programming Language :: Python :: 3.11",
@@ -74,6 +72,6 @@ def readme():
     ],
     platforms="Posix; MacOS X; Windows",
     install_requires=dependencies,
-    python_requires=">=3.7",
+    python_requires=">=3.9",
     tests_require=["pytest"],
 )

setup.py

@@ -1,3 +1,3 @@
 # Make sure we test with pandas 1.5.3. The Python version isn't that relevant.
 pandas==1.5.3
-numpy==1.24.0
+numpy==1.20.3

testing/constraints-3.7.txt

@@ -23,7 +23,6 @@
 import pytest
 
 import db_dtypes
-from db_dtypes import pandas_backports
 
 VALUE_PARSING_TEST_CASES = [
     # Min/Max values for pandas.Timestamp.
@@ -258,10 +257,6 @@ def test_date_min_2d():
     )
 
 
-@pytest.mark.skipif(
-    not hasattr(pandas_backports, "numpy_validate_median"),
-    reason="median not available with this version of pandas",
-)
 @pytest.mark.parametrize(
     "values, expected",
     [
@@ -284,10 +279,6 @@ def test_date_median(values, expected):
     assert series.median() == expected
 
 
-@pytest.mark.skipif(
-    not hasattr(pandas_backports, "numpy_validate_median"),
-    reason="median not available with this version of pandas",
-)
 def test_date_median_2d():
     input_array = db_dtypes.DateArray(
         numpy.array(