Add Cython Support to Arrow

.github/workflows/continuous_integration.yml

@@ -43,7 +43,7 @@ jobs:
       - name: Install dependencies
         run: |
           pip install -U pip setuptools wheel
-          pip install -U tox tox-gh-actions
+          pip install -U tox tox-gh-actions "Cython>=3.0.0a11"
       - name: Test with tox
         run: tox
       - name: Upload coverage to Codecov

Makefile

@@ -1,4 +1,4 @@
-.PHONY: auto test docs clean
+.PHONY: auto test docs clean compile-cython
 
 auto: build310
 
@@ -15,7 +15,15 @@ build36 build37 build38 build39 build310: clean
 	pip install -U pip setuptools wheel; \
 	pip install -r requirements/requirements-tests.txt; \
 	pip install -r requirements/requirements-docs.txt; \
-	pre-commit install
+	pre-commit install; \
+	python setup.py build_ext --inplace
+
+compile-cython:
+	. venv/bin/activate; \
+	python setup.py build_ext --inplace
+
+clean-cython:
+	rm -f ./arrow/*.c ./arrow/*.so ./arrow/*.pyd
 
 test:
 	rm -f .coverage coverage.xml
@@ -38,14 +46,14 @@ live-docs: clean-docs
 	. venv/bin/activate; \
 	sphinx-autobuild docs docs/_build/html
 
-clean: clean-dist
+clean: clean-dist clean-cython
 	rm -rf venv .pytest_cache ./**/__pycache__
 	rm -f .coverage coverage.xml ./**/*.pyc
 
 clean-dist:
 	rm -rf dist build .egg .eggs arrow.egg-info
 
-build-dist:
+build-dist: compile-cython
 	. venv/bin/activate; \
 	pip install -U pip setuptools twine wheel; \
 	python setup.py sdist bdist_wheel

arrow/arrow.py

@@ -1148,7 +1148,7 @@ def humanize(
         """
 
         locale_name = locale
-        locale = locales.get_locale(locale)
+        locale_cls = locales.get_locale(locale)
 
         if other is None:
             utc = dt_datetime.utcnow().replace(tzinfo=dateutil_tz.tzutc())
@@ -1179,65 +1179,83 @@ def humanize(
         try:
             if granularity == "auto":
                 if diff < 10:
-                    return locale.describe("now", only_distance=only_distance)
+                    return locale_cls.describe("now", only_distance=only_distance)
 
                 if diff < self._SECS_PER_MINUTE:
                     seconds = sign * delta_second
-                    return locale.describe(
+                    return locale_cls.describe(
                         "seconds", seconds, only_distance=only_distance
                     )
 
                 elif diff < self._SECS_PER_MINUTE * 2:
-                    return locale.describe("minute", sign, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "minute", sign, only_distance=only_distance
+                    )
                 elif diff < self._SECS_PER_HOUR:
                     minutes = sign * max(delta_second // self._SECS_PER_MINUTE, 2)
-                    return locale.describe(
+                    return locale_cls.describe(
                         "minutes", minutes, only_distance=only_distance
                     )
 
                 elif diff < self._SECS_PER_HOUR * 2:
-                    return locale.describe("hour", sign, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "hour", sign, only_distance=only_distance
+                    )
                 elif diff < self._SECS_PER_DAY:
                     hours = sign * max(delta_second // self._SECS_PER_HOUR, 2)
-                    return locale.describe("hours", hours, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "hours", hours, only_distance=only_distance
+                    )
                 elif diff < self._SECS_PER_DAY * 2:
-                    return locale.describe("day", sign, only_distance=only_distance)
+                    return locale_cls.describe("day", sign, only_distance=only_distance)
                 elif diff < self._SECS_PER_WEEK:
                     days = sign * max(delta_second // self._SECS_PER_DAY, 2)
-                    return locale.describe("days", days, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "days", days, only_distance=only_distance
+                    )
 
                 elif diff < self._SECS_PER_WEEK * 2:
-                    return locale.describe("week", sign, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "week", sign, only_distance=only_distance
+                    )
                 elif diff < self._SECS_PER_MONTH:
                     weeks = sign * max(delta_second // self._SECS_PER_WEEK, 2)
-                    return locale.describe("weeks", weeks, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "weeks", weeks, only_distance=only_distance
+                    )
 
                 elif diff < self._SECS_PER_MONTH * 2:
-                    return locale.describe("month", sign, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "month", sign, only_distance=only_distance
+                    )
                 elif diff < self._SECS_PER_YEAR:
                     # TODO revisit for humanization during leap years
                     self_months = self._datetime.year * 12 + self._datetime.month
                     other_months = dt.year * 12 + dt.month
 
                     months = sign * max(abs(other_months - self_months), 2)
 
-                    return locale.describe(
+                    return locale_cls.describe(
                         "months", months, only_distance=only_distance
                     )
 
                 elif diff < self._SECS_PER_YEAR * 2:
-                    return locale.describe("year", sign, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "year", sign, only_distance=only_distance
+                    )
                 else:
                     years = sign * max(delta_second // self._SECS_PER_YEAR, 2)
-                    return locale.describe("years", years, only_distance=only_distance)
+                    return locale_cls.describe(
+                        "years", years, only_distance=only_distance
+                    )
 
             elif isinstance(granularity, str):
                 granularity = cast(TimeFrameLiteral, granularity)  # type: ignore[assignment]
 
                 if granularity == "second":
                     delta = sign * float(delta_second)
                     if abs(delta) < 2:
-                        return locale.describe("now", only_distance=only_distance)
+                        return locale_cls.describe("now", only_distance=only_distance)
                 elif granularity == "minute":
                     delta = sign * delta_second / self._SECS_PER_MINUTE
                 elif granularity == "hour":
@@ -1260,7 +1278,9 @@ def humanize(
 
                 if trunc(abs(delta)) != 1:
                     granularity += "s"  # type: ignore[assignment]
-                return locale.describe(granularity, delta, only_distance=only_distance)
+                return locale_cls.describe(
+                    granularity, delta, only_distance=only_distance
+                )
 
             else:
 
@@ -1304,7 +1324,9 @@ def gather_timeframes(_delta: float, _frame: TimeFrameLiteral) -> float:
                         "Please select between 'second', 'minute', 'hour', 'day', 'week', 'month', 'quarter' or 'year'."
                     )
 
-                return locale.describe_multi(timeframes, only_distance=only_distance)
+                return locale_cls.describe_multi(
+                    timeframes, only_distance=only_distance
+                )
 
         except KeyError as e:
             raise ValueError(
@@ -1410,9 +1432,7 @@ def dehumanize(self, input_string: str, locale: str = "en_us") -> "Arrow":
 
                 # Add change value to the correct unit (incorporates the plurality that exists within timeframe i.e second v.s seconds)
                 time_unit_to_change = str(unit)
-                time_unit_to_change += (
-                    "s" if (str(time_unit_to_change)[-1] != "s") else ""
-                )
+                time_unit_to_change += "s" if time_unit_to_change[-1] != "s" else ""
                 time_object_info[time_unit_to_change] = change_value
                 unit_visited[time_unit_to_change] = True
 
@@ -1663,7 +1683,8 @@ def isocalendar(self) -> Tuple[int, int, int]:
 
         """
 
-        return self._datetime.isocalendar()
+        cal = tuple(self._datetime.isocalendar())
+        return (cal[0], cal[1], cal[2])
 
     def isoformat(self, sep: str = "T", timespec: str = "auto") -> str:
         """Returns an ISO 8601 formatted representation of the date and time.

requirements/requirements.txt

@@ -1,2 +1,3 @@
+Cython>=3.0.0a11
 python-dateutil>=2.7.0
 typing_extensions; python_version < '3.8'

setup.py

@@ -1,13 +1,22 @@
 # mypy: ignore-errors
 from pathlib import Path
 
-from setuptools import setup
+from Cython.Build import build_ext, cythonize
+from setuptools import Extension, setup
 
 readme = Path("README.rst").read_text(encoding="utf-8")
 version = Path("arrow/_version.py").read_text(encoding="utf-8")
 about = {}
 exec(version, about)
 
+extensions = [
+    Extension(
+        "*",
+        ["arrow/*.py"],
+        define_macros=[("CYTHON_TRACE", "1")],
+    )
+]
+
 setup(
     name="arrow",
     version=about["__version__"],
@@ -46,4 +55,8 @@
         "Bug Reports": "https://github.com/arrow-py/arrow/issues",
         "Documentation": "https://arrow.readthedocs.io",
     },
+    ext_modules=cythonize(
+        extensions, language_level="3", compiler_directives={"linetrace": True}
+    ),
+    cmdclass={"build_ext": build_ext},
 )

tests/test_locales.py

@@ -77,13 +77,10 @@ def test_get_locale_by_class_name(self, mocker):
         mock_locale_cls = mocker.Mock()
         mock_locale_obj = mock_locale_cls.return_value = mocker.Mock()
 
-        globals_fn = mocker.Mock()
-        globals_fn.return_value = {"NonExistentLocale": mock_locale_cls}
-
         with pytest.raises(ValueError):
             arrow.locales.get_locale_by_class_name("NonExistentLocale")
 
-        mocker.patch.object(locales, "globals", globals_fn)
+        mocker.patch.object(locales, "NonExistentLocale", mock_locale_cls, create=True)
         result = arrow.locales.get_locale_by_class_name("NonExistentLocale")
 
         mock_locale_cls.assert_called_once_with()

tox.ini

@@ -47,3 +47,6 @@ include_trailing_comma = true
 [flake8]
 per-file-ignores = arrow/__init__.py:F401,tests/*:ANN001,ANN201
 ignore = E203,E501,W503,ANN101,ANN102,ANN401
+
+[coverage:run]
+plugins = Cython.Coverage