Implement zoneinfo support and make pytz optional (#940)

CHANGES.rst

@@ -1,6 +1,13 @@
 Babel Changelog
 ===============
 
+Unreleased
+----------
+
+* Use `zoneinfo` timezone resolving on python 3.9+, while keeping pytz support
+  for lower versions
+
+
 Version 2.11.0
 --------------
 

babel/dates.py

@@ -19,20 +19,24 @@
 
 import re
 import warnings
+from typing import TYPE_CHECKING, SupportsInt
+
+try:
+    import pytz
+except ModuleNotFoundError:
+    pytz = None
+    import zoneinfo
+
 from bisect import bisect_right
 from collections.abc import Iterable
 from datetime import date, datetime, time, timedelta, tzinfo
-from typing import TYPE_CHECKING, SupportsInt
-
-import pytz as _pytz
 
+from babel import localtime
 from babel.core import Locale, default_locale, get_global
 from babel.localedata import LocaleDataDict
-from babel.util import LOCALTZ, UTC
 
 if TYPE_CHECKING:
     from typing_extensions import Literal, TypeAlias
-
     _Instant: TypeAlias = date | time | float | None
     _PredefinedTimeFormat: TypeAlias = Literal['full', 'long', 'medium', 'short']
     _Context: TypeAlias = Literal['format', 'stand-alone']
@@ -48,6 +52,12 @@
 NO_INHERITANCE_MARKER = u'\u2205\u2205\u2205'
 
 
+if pytz:
+    UTC = pytz.utc
+else:
+    UTC = zoneinfo.ZoneInfo('UTC')
+LOCALTZ = localtime.LOCALTZ
+
 LC_TIME = default_locale('LC_TIME')
 
 # Aliases for use in scopes where the modules are shadowed by local variables
@@ -56,6 +66,24 @@
 time_ = time
 
 
+def _localize(tz: tzinfo, dt: datetime) -> datetime:
+    # Support localizing with both pytz and zoneinfo tzinfos
+    # nothing to do
+    if dt.tzinfo is tz:
+        return dt
+
+    if hasattr(tz, 'localize'):  # pytz
+        return tz.localize(dt)
+
+    if dt.tzinfo is None:
+        # convert naive to localized
+        return dt.replace(tzinfo=tz)
+
+    # convert timezones
+    return dt.astimezone(tz)
+
+
+
 def _get_dt_and_tzinfo(dt_or_tzinfo: _DtOrTzinfo) -> tuple[datetime_ | None, tzinfo]:
     """
     Parse a `dt_or_tzinfo` value into a datetime and a tzinfo.
@@ -150,15 +178,15 @@ def _ensure_datetime_tzinfo(datetime: datetime_, tzinfo: tzinfo | None = None) -
 
     If a tzinfo is passed in, the datetime is normalized to that timezone.
 
-    >>> _ensure_datetime_tzinfo(datetime(2015, 1, 1)).tzinfo.zone
+    >>> _get_tz_name(_ensure_datetime_tzinfo(datetime(2015, 1, 1)))
     'UTC'
 
     >>> tz = get_timezone("Europe/Stockholm")
     >>> _ensure_datetime_tzinfo(datetime(2015, 1, 1, 13, 15, tzinfo=UTC), tzinfo=tz).hour
     14
 
     :param datetime: Datetime to augment.
-    :param tzinfo: Optional tznfo.
+    :param tzinfo: optional tzinfo
     :return: datetime with tzinfo
     :rtype: datetime
     """
@@ -184,8 +212,10 @@ def _get_time(time: time | datetime | None, tzinfo: tzinfo | None = None) -> tim
         time = datetime.utcnow()
     elif isinstance(time, (int, float)):
         time = datetime.utcfromtimestamp(time)
+
     if time.tzinfo is None:
         time = time.replace(tzinfo=UTC)
+
     if isinstance(time, datetime):
         if tzinfo is not None:
             time = time.astimezone(tzinfo)
@@ -197,28 +227,40 @@ def _get_time(time: time | datetime | None, tzinfo: tzinfo | None = None) -> tim
     return time
 
 
-def get_timezone(zone: str | _pytz.BaseTzInfo | None = None) -> _pytz.BaseTzInfo:
+def get_timezone(zone: str | tzinfo | None = None) -> tzinfo:
     """Looks up a timezone by name and returns it.  The timezone object
-    returned comes from ``pytz`` and corresponds to the `tzinfo` interface and
-    can be used with all of the functions of Babel that operate with dates.
+    returned comes from ``pytz`` or ``zoneinfo``, whichever is available.
+    It corresponds to the `tzinfo` interface and can be used with all of
+    the functions of Babel that operate with dates.
 
     If a timezone is not known a :exc:`LookupError` is raised.  If `zone`
     is ``None`` a local zone object is returned.
 
     :param zone: the name of the timezone to look up.  If a timezone object
-                 itself is passed in, mit's returned unchanged.
+                 itself is passed in, it's returned unchanged.
     """
     if zone is None:
         return LOCALTZ
     if not isinstance(zone, str):
         return zone
-    try:
-        return _pytz.timezone(zone)
-    except _pytz.UnknownTimeZoneError:
-        raise LookupError(f"Unknown timezone {zone}")
 
+    exc = None
+    if pytz:
+        try:
+            return pytz.timezone(zone)
+        except pytz.UnknownTimeZoneError as exc:
+            pass
+    else:
+        assert zoneinfo
+        try:
+            return zoneinfo.ZoneInfo(zone)
+        except zoneinfo.ZoneInfoNotFoundError as exc:
+            pass
+
+    raise LookupError(f"Unknown timezone {zone}") from exc
 
-def get_next_timezone_transition(zone: _pytz.BaseTzInfo | None = None, dt: _Instant = None) -> TimezoneTransition:
+
+def get_next_timezone_transition(zone: tzinfo | None = None, dt: _Instant = None) -> TimezoneTransition:
     """Given a timezone it will return a :class:`TimezoneTransition` object
     that holds the information about the next timezone transition that's going
     to happen.  For instance this can be used to detect when the next DST
@@ -474,7 +516,7 @@ def get_timezone_gmt(datetime: _Instant = None, width: Literal['long', 'short',
     >>> get_timezone_gmt(dt, locale='en', width='iso8601_short')
     u'+00'
     >>> tz = get_timezone('America/Los_Angeles')
-    >>> dt = tz.localize(datetime(2007, 4, 1, 15, 30))
+    >>> dt = _localize(tz, datetime(2007, 4, 1, 15, 30))
     >>> get_timezone_gmt(dt, locale='en')
     u'GMT-07:00'
     >>> get_timezone_gmt(dt, 'short', locale='en')
@@ -608,7 +650,7 @@ def get_timezone_name(dt_or_tzinfo: _DtOrTzinfo = None, width: Literal['long', '
     u'PST'
 
     If this function gets passed only a `tzinfo` object and no concrete
-    `datetime`,  the returned display name is indenpendent of daylight savings
+    `datetime`,  the returned display name is independent of daylight savings
     time. This can be used for example for selecting timezones, or to set the
     time of events that recur across DST changes:
 
@@ -755,12 +797,11 @@ def format_datetime(datetime: _Instant = None, format: _PredefinedTimeFormat | s
     >>> format_datetime(dt, locale='en_US')
     u'Apr 1, 2007, 3:30:00 PM'
 
-    For any pattern requiring the display of the time-zone, the third-party
-    ``pytz`` package is needed to explicitly specify the time-zone:
+    For any pattern requiring the display of the timezone:
 
     >>> format_datetime(dt, 'full', tzinfo=get_timezone('Europe/Paris'),
     ...                 locale='fr_FR')
-    u'dimanche 1 avril 2007 \xe0 17:30:00 heure d\u2019\xe9t\xe9 d\u2019Europe centrale'
+    'dimanche 1 avril 2007 à 17:30:00 heure d’été d’Europe centrale'
     >>> format_datetime(dt, "yyyy.MM.dd G 'at' HH:mm:ss zzz",
     ...                 tzinfo=get_timezone('US/Eastern'), locale='en')
     u'2007.04.01 AD at 11:30:00 EDT'
@@ -806,9 +847,9 @@ def format_time(time: time | datetime | float | None = None, format: _Predefined
 
     >>> t = datetime(2007, 4, 1, 15, 30)
     >>> tzinfo = get_timezone('Europe/Paris')
-    >>> t = tzinfo.localize(t)
+    >>> t = _localize(tzinfo, t)
     >>> format_time(t, format='full', tzinfo=tzinfo, locale='fr_FR')
-    u'15:30:00 heure d\u2019\xe9t\xe9 d\u2019Europe centrale'
+    '15:30:00 heure d’été d’Europe centrale'
     >>> format_time(t, "hh 'o''clock' a, zzzz", tzinfo=get_timezone('US/Eastern'),
     ...             locale='en')
     u"09 o'clock AM, Eastern Daylight Time"
@@ -841,12 +882,17 @@ def format_time(time: time | datetime | float | None = None, format: _Predefined
     :param tzinfo: the time-zone to apply to the time for display
     :param locale: a `Locale` object or a locale identifier
     """
+
+    # get reference date for if we need to find the right timezone variant
+    # in the pattern
+    ref_date = time.date() if isinstance(time, datetime) else None
+
     time = _get_time(time, tzinfo)
 
     locale = Locale.parse(locale)
     if format in ('full', 'long', 'medium', 'short'):
         format = get_time_format(format, locale=locale)
-    return parse_pattern(format).apply(time, locale)
+    return parse_pattern(format).apply(time, locale, reference_date=ref_date)
 
 
 def format_skeleton(skeleton: str, datetime: _Instant = None, tzinfo: tzinfo | None = None,
@@ -1124,7 +1170,7 @@ def format_interval(start: _Instant, end: _Instant, skeleton: str | None = None,
     return _format_fallback_interval(start, end, skeleton, tzinfo, locale)
 
 
-def get_period_id(time: _Instant, tzinfo: _pytz.BaseTzInfo | None = None, type: Literal['selection'] | None = None,
+def get_period_id(time: _Instant, tzinfo: tzinfo | None = None, type: Literal['selection'] | None = None,
                   locale: Locale | str | None = LC_TIME) -> str:
     """
     Get the day period ID for a given time.
@@ -1327,18 +1373,29 @@ def __mod__(self, other: DateTimeFormat) -> str:
             return NotImplemented
         return self.format % other
 
-    def apply(self, datetime: date | time, locale: Locale | str | None) -> str:
-        return self % DateTimeFormat(datetime, locale)
+    def apply(
+        self,
+        datetime: date | time,
+        locale: Locale | str | None,
+        reference_date: date | None = None
+    ) -> str:
+        return self % DateTimeFormat(datetime, locale, reference_date)
 
 
 class DateTimeFormat:
 
-    def __init__(self, value: date | time, locale: Locale | str):
+    def __init__(
+        self,
+        value: date | time,
+        locale: Locale | str,
+        reference_date: date | None = None
+    ):
         assert isinstance(value, (date, datetime, time))
         if isinstance(value, (datetime, time)) and value.tzinfo is None:
             value = value.replace(tzinfo=UTC)
         self.value = value
         self.locale = Locale.parse(locale)
+        self.reference_date = reference_date
 
     def __getitem__(self, name: str) -> str:
         char = name[0]
@@ -1558,46 +1615,54 @@ def format_milliseconds_in_day(self, num):
 
     def format_timezone(self, char: str, num: int) -> str:
         width = {3: 'short', 4: 'long', 5: 'iso8601'}[max(3, num)]
+
+        # It could be that we only receive a time to format, but also have a
+        # reference date which is important to distinguish between timezone
+        # variants (summer/standard time)
+        value = self.value
+        if self.reference_date:
+            value = datetime.combine(self.reference_date, self.value)
+
         if char == 'z':
-            return get_timezone_name(self.value, width, locale=self.locale)
+            return get_timezone_name(value, width, locale=self.locale)
         elif char == 'Z':
             if num == 5:
-                return get_timezone_gmt(self.value, width, locale=self.locale, return_z=True)
-            return get_timezone_gmt(self.value, width, locale=self.locale)
+                return get_timezone_gmt(value, width, locale=self.locale, return_z=True)
+            return get_timezone_gmt(value, width, locale=self.locale)
         elif char == 'O':
             if num == 4:
-                return get_timezone_gmt(self.value, width, locale=self.locale)
+                return get_timezone_gmt(value, width, locale=self.locale)
         # TODO: To add support for O:1
         elif char == 'v':
-            return get_timezone_name(self.value.tzinfo, width,
+            return get_timezone_name(value.tzinfo, width,
                                      locale=self.locale)
         elif char == 'V':
             if num == 1:
-                return get_timezone_name(self.value.tzinfo, width,
+                return get_timezone_name(value.tzinfo, width,
                                          uncommon=True, locale=self.locale)
             elif num == 2:
-                return get_timezone_name(self.value.tzinfo, locale=self.locale, return_zone=True)
+                return get_timezone_name(value.tzinfo, locale=self.locale, return_zone=True)
             elif num == 3:
-                return get_timezone_location(self.value.tzinfo, locale=self.locale, return_city=True)
-            return get_timezone_location(self.value.tzinfo, locale=self.locale)
+                return get_timezone_location(value.tzinfo, locale=self.locale, return_city=True)
+            return get_timezone_location(value.tzinfo, locale=self.locale)
         # Included additional elif condition to add support for 'Xx' in timezone format
         elif char == 'X':
             if num == 1:
-                return get_timezone_gmt(self.value, width='iso8601_short', locale=self.locale,
+                return get_timezone_gmt(value, width='iso8601_short', locale=self.locale,
                                         return_z=True)
             elif num in (2, 4):
-                return get_timezone_gmt(self.value, width='short', locale=self.locale,
+                return get_timezone_gmt(value, width='short', locale=self.locale,
                                         return_z=True)
             elif num in (3, 5):
-                return get_timezone_gmt(self.value, width='iso8601', locale=self.locale,
+                return get_timezone_gmt(value, width='iso8601', locale=self.locale,
                                         return_z=True)
         elif char == 'x':
             if num == 1:
-                return get_timezone_gmt(self.value, width='iso8601_short', locale=self.locale)
+                return get_timezone_gmt(value, width='iso8601_short', locale=self.locale)
             elif num in (2, 4):
-                return get_timezone_gmt(self.value, width='short', locale=self.locale)
+                return get_timezone_gmt(value, width='short', locale=self.locale)
             elif num in (3, 5):
-                return get_timezone_gmt(self.value, width='iso8601', locale=self.locale)
+                return get_timezone_gmt(value, width='iso8601', locale=self.locale)
 
     def format(self, value: SupportsInt, length: int) -> str:
         return '%0*d' % (length, value)

babel/localtime/__init__.py

@@ -14,8 +14,6 @@
 from datetime import datetime, timedelta, tzinfo
 from threading import RLock
 
-import pytz
-
 if sys.platform == 'win32':
     from babel.localtime._win32 import _get_localzone
 else:
@@ -61,7 +59,7 @@ def _isdst(self, dt: datetime) -> bool:
         return tt.tm_isdst > 0
 
 
-def get_localzone() -> pytz.BaseTzInfo:
+def get_localzone() -> tzinfo:
     """Returns the current underlying local timezone object.
     Generally this function does not need to be used, it's a
     better idea to use the :data:`LOCALTZ` singleton instead.
@@ -71,5 +69,5 @@ def get_localzone() -> pytz.BaseTzInfo:
 
 try:
     LOCALTZ = get_localzone()
-except pytz.UnknownTimeZoneError:
+except LookupError:
     LOCALTZ = _FallbackLocalTimezone()