cmlenz@2: .. -*- mode: rst; encoding: utf-8 -*-
cmlenz@2: 
cmlenz@124: ===============
cmlenz@124: Date Formatting
cmlenz@124: ===============
cmlenz@2: 
cmlenz@2: 
cmlenz@2: .. contents:: Contents
cmlenz@2:    :depth: 2
cmlenz@2: .. sectnum::
cmlenz@2: 
cmlenz@2: 
cmlenz@18: When working with date and time information in Python, you commonly use the
cmlenz@98: classes ``date``, ``datetime`` and/or ``time`` from the `datetime`_ package.
cmlenz@18: Babel provides functions for locale-specific formatting of those objects in its
cmlenz@40: ``dates`` module:
cmlenz@40: 
cmlenz@98: .. _`datetime`: http://docs.python.org/lib/module-datetime.html
cmlenz@98: 
cmlenz@40: .. code-block:: pycon
cmlenz@18: 
cmlenz@18:     >>> from datetime import date, datetime, time
cmlenz@18:     >>> from babel.dates import format_date, format_datetime, format_time
cmlenz@18: 
cmlenz@18:     >>> d = date(2007, 4, 1)
cmlenz@18:     >>> format_date(d, locale='en')
cmlenz@18:     u'Apr 1, 2007'
cmlenz@18:     >>> format_date(d, locale='de_DE')
cmlenz@18:     u'01.04.2007'
cmlenz@18: 
cmlenz@18: As this example demonstrates, Babel will automatically choose a date format
cmlenz@18: that is appropriate for the requested locale.
cmlenz@18: 
cmlenz@18: The ``format_*()`` functions also accept an optional ``format`` argument, which
cmlenz@18: allows you to choose between one of four format variations:
cmlenz@18: 
cmlenz@18:  * ``short``,
cmlenz@18:  * ``medium`` (the default),
cmlenz@20:  * ``long``, and
cmlenz@20:  * ``full``.
cmlenz@18: 
cmlenz@40: For example:
cmlenz@40: 
cmlenz@40: .. code-block:: pycon
cmlenz@18: 
cmlenz@18:     >>> format_date(d, format='short', locale='en')
cmlenz@18:     u'4/1/07'
cmlenz@18:     >>> format_date(d, format='long', locale='en')
cmlenz@18:     u'April 1, 2007'
cmlenz@18:     >>> format_date(d, format='full', locale='en')
cmlenz@18:     u'Sunday, April 1, 2007'
cmlenz@18: 
cmlenz@18: 
cmlenz@18: Pattern Syntax
cmlenz@124: ==============
cmlenz@18: 
cmlenz@18: While Babel makes it simple to use the appropriate date/time format for a given
cmlenz@18: locale, you can also force it to use custom patterns. Note that Babel uses
cmlenz@18: different patterns for specifying number and date formats compared to the
cmlenz@18: Python equivalents (such as ``time.strftime()``), which have mostly been
cmlenz@18: inherited from C and POSIX. The patterns used in Babel are based on the
cmlenz@18: `Locale Data Markup Language specification`_ (LDML), which defines them as
cmlenz@18: follows:
cmlenz@18: 
cmlenz@18:     A date/time pattern is a string of characters, where specific strings of
cmlenz@18:     characters are replaced with date and time data from a calendar when formatting
cmlenz@18:     or used to generate data for a calendar when parsing. […]
cmlenz@18: 
cmlenz@18:     Characters may be used multiple times. For example, if ``y`` is used for the
cmlenz@18:     year, ``yy`` might produce "99", whereas ``yyyy`` produces "1999". For most
cmlenz@18:     numerical fields, the number of characters specifies the field width. For
cmlenz@18:     example, if ``h`` is the hour, ``h`` might produce "5", but ``hh`` produces
cmlenz@18:     "05". For some characters, the count specifies whether an abbreviated or full
cmlenz@18:     form should be used […]
cmlenz@18: 
cmlenz@18:     Two single quotes represent a literal single quote, either inside or outside
cmlenz@18:     single quotes. Text within single quotes is not interpreted in any way (except
cmlenz@18:     for two adjacent single quotes).
cmlenz@18: 
cmlenz@40: For example:
cmlenz@40: 
cmlenz@40: .. code-block:: pycon
cmlenz@18: 
cmlenz@18:     >>> d = date(2007, 4, 1)
cmlenz@18:     >>> format_date(d, "EEE, MMM d, ''yy", locale='en')
cmlenz@18:     u"Sun, Apr 1, '07"
cmlenz@18:     >>> format_date(d, "EEEE, d.M.yyyy", locale='de')
cmlenz@18:     u'Sonntag, 1.4.2007'
cmlenz@18: 
cmlenz@18:     >>> t = time(15, 30)
cmlenz@18:     >>> format_time(t, "hh 'o''clock' a", locale='en')
cmlenz@18:     u"03 o'clock PM"
cmlenz@18:     >>> format_time(t, 'H:mm a', locale='de')
cmlenz@18:     u'15:30 nachm.'
cmlenz@18: 
cmlenz@18:     >>> dt = datetime(2007, 4, 1, 15, 30)
cmlenz@18:     >>> format_datetime(dt, "yyyyy.MMMM.dd GGG hh:mm a", locale='en')
cmlenz@18:     u'02007.April.01 AD 03:30 PM'
cmlenz@18: 
cmlenz@18: The syntax for custom datetime format patterns is described in detail in the
cmlenz@18: the `Locale Data Markup Language specification`_. The following table is just a
cmlenz@18: relatively brief overview.
cmlenz@18: 
cmlenz@18:  .. _`Locale Data Markup Language specification`: http://unicode.org/reports/tr35/#Date_Format_Patterns
cmlenz@18: 
cmlenz@18: Date Fields
cmlenz@18: -----------
cmlenz@18: 
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Field    | Symbol | Description                                            |
cmlenz@18:   +==========+========+========================================================+
cmlenz@18:   | Era      | ``G``  | Replaced with the era string for the current date. One |
cmlenz@18:   |          |        | to three letters for the abbreviated form, four        |
cmlenz@18:   |          |        | lettersfor the long form, five for the narrow form     |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Year     | ``y``  | Replaced by the year. Normally the length specifies    |
cmlenz@18:   |          |        | the padding, but for two letters it also specifies the |
cmlenz@18:   |          |        | maximum length.                                        |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``Y``  | Same as ``y`` but uses the ISO year-week calendar.     |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``u``  | ??                                                     |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Quarter  | ``Q``  | Use one or two for the numerical quarter, three for    |
cmlenz@18:   |          |        | the abbreviation, or four for the full name.           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``q``  | Use one or two for the numerical quarter, three for    |
cmlenz@18:   |          |        | the abbreviation, or four for the full name.           |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Month    | ``M``  | Use one or two for the numerical month, three for the  |
cmlenz@18:   |          |        | abbreviation, or four for the full name, or five for   |
cmlenz@18:   |          |        | the narrow name.                                       |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``L``  | Use one or two for the numerical month, three for the  |
cmlenz@18:   |          |        | abbreviation, or four for the full name, or 5 for the  |
cmlenz@18:   |          |        | narrow name.                                           |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Week     | ``w``  | Week of year.                                          |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``W``  | Week of month.                                         |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Day      | ``d``  | Day of month.                                          |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``D``  | Day of year.                                           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``F``  | Day of week in month.                                  |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``g``  | ??                                                     |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Week day | ``E``  | Day of week. Use one through three letters for the     |
cmlenz@18:   |          |        | short day, or four for the full name, or five for the  |
cmlenz@18:   |          |        | narrow name.                                           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``e``  | Local day of week. Same as E except adds a numeric     |
cmlenz@18:   |          |        | value that will depend on the local starting day of    |
cmlenz@18:   |          |        | the week, using one or two letters.                    |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``c``  | ??                                                     |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18: 
cmlenz@18: Time Fields
cmlenz@18: -----------
cmlenz@18: 
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Field    | Symbol | Description                                            |
cmlenz@18:   +==========+========+========================================================+
cmlenz@18:   | Period   | ``a``  | AM or PM                                               |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Hour     | ``h``  | Hour [1-12].                                           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``H``  | Hour [0-23].                                           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``K``  | Hour [0-11].                                           |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``k``  | Hour [1-24].                                           |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Minute   | ``m``  | Use one or two for zero places padding.                |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18:   | Second   | ``s``  | Use one or two for zero places padding.                |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``S``  | Fractional second, rounds to the count of letters.     |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``A``  | Milliseconds in day.                                   |
cmlenz@18:   +----------+--------+--------------------------------------------------------+ 
cmlenz@18:   | Timezone | ``z``  | Use one to three letters for the short timezone or     |
cmlenz@18:   |          |        | four for the full name.                                |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``Z``  | Use one to three letters for RFC 822, four letters for |
cmlenz@18:   |          |        | GMT format.                                            |
cmlenz@18:   |          +--------+--------------------------------------------------------+
cmlenz@18:   |          | ``v``  | Use one letter for short wall (generic) time, four for |
cmlenz@18:   |          |        | long wall time.                                        |
cmlenz@233:   |          +--------+--------------------------------------------------------+
cmlenz@233:   |          | ``V``  | Same as ``z``, except that timezone abbreviations      |
cmlenz@233:   |          |        | should be used regardless of whether they are in       |
cmlenz@233:   |          |        | common use by the locale.                              |
cmlenz@18:   +----------+--------+--------------------------------------------------------+
cmlenz@18: 
cmlenz@18: 
cmlenz@394: Time Delta Formatting
cmlenz@394: =====================
cmlenz@394: 
cmlenz@394: In addition to providing functions for formatting localized dates and times,
cmlenz@394: the ``babel.dates`` module also provides a function to format the difference
cmlenz@394: between two times, called a ''time delta''. These are usually represented as
cmlenz@394: ``datetime.timedelta`` objects in Python, and it's also what you get when you
cmlenz@394: subtract one ``datetime`` object from an other.
cmlenz@394: 
cmlenz@394: The ``format_timedelta`` function takes a ``timedelta`` object and returns a
cmlenz@394: human-readable representation. This happens at the cost of precision, as it
cmlenz@394: chooses only the most significant unit (such as year, week, or hour) of the
cmlenz@394: difference, and displays that:
cmlenz@394: 
cmlenz@394: .. code-block:: pycon
cmlenz@394: 
cmlenz@394:     >>> from datetime import timedelta
cmlenz@394:     >>> from babel.dates import format_timedelta
cmlenz@394:     >>> delta = timedelta(days=6)
cmlenz@394:     >>> format_timedelta(delta, locale='en_US')
cmlenz@394:     u'1 week'
cmlenz@394: 
cmlenz@394: The resulting strings are based from the CLDR data, and are properly
cmlenz@394: pluralized depending on the plural rules of the locale and the calculated
cmlenz@394: number of units.
cmlenz@394: 
cmlenz@394: The function provides parameters for you to influence how this most significant
cmlenz@394: unit is chosen: with ``threshold`` you set the value after which the
cmlenz@394: presentation switches to the next larger unit, and with ``granularity`` you
cmlenz@394: can limit the smallest unit to display:
cmlenz@394: 
cmlenz@394: .. code-block:: pycon
cmlenz@394: 
cmlenz@394:     >>> delta = timedelta(days=6)
cmlenz@394:     >>> format_timedelta(delta, threshold=1.2, locale='en_US')
cmlenz@394:     u'6 days'
cmlenz@394:     >>> format_timedelta(delta, granularity='month', locale='en_US')
cmlenz@395:     u'1 month'
cmlenz@394: 
cmlenz@394: 
cmlenz@255: Time-zone Support
cmlenz@124: =================
cmlenz@29: 
cmlenz@31: Many of the verbose time formats include the time-zone, but time-zone
cmlenz@31: information is not by default available for the Python ``datetime`` and
cmlenz@31: ``time`` objects. The standard library includes only the abstract ``tzinfo``
cmlenz@31: class, which you need appropriate implementations for to actually use in your
cmlenz@29: application. Babel includes a ``tzinfo`` implementation for UTC (Universal
cmlenz@31: Time).
cmlenz@31: 
cmlenz@31: For real time-zone support, it is strongly recommended that you use the
cmlenz@29: third-party package `pytz`_, which includes the definitions of practically all
cmlenz@29: of the time-zones used on the world, as well as important functions for
cmlenz@40: reliably converting from UTC to local time, and vice versa:
cmlenz@40: 
cmlenz@40: .. code-block:: pycon
cmlenz@29: 
cmlenz@29:     >>> from datetime import time
cmlenz@34:     >>> from pytz import timezone, utc
cmlenz@34:     >>> dt = datetime(2007, 04, 01, 15, 30, tzinfo=utc)
cmlenz@34:     >>> eastern = timezone('US/Eastern')
cmlenz@34:     >>> format_datetime(dt, 'H:mm Z', tzinfo=eastern, locale='en_US')
cmlenz@34:     u'11:30 -0400'
cmlenz@29: 
cmlenz@29: The recommended approach to deal with different time-zones in a Python
cmlenz@29: application is to always use UTC internally, and only convert from/to the users
cmlenz@29: time-zone when accepting user input and displaying date/time data, respectively.
cmlenz@34: You can use Babel together with ``pytz`` to apply a time-zone to any
cmlenz@34: ``datetime`` or ``time`` object for display, leaving the original information
cmlenz@40: unchanged:
cmlenz@40: 
cmlenz@40: .. code-block:: pycon
cmlenz@34: 
cmlenz@34:     >>> british = timezone('Europe/London')
cmlenz@34:     >>> format_datetime(dt, 'H:mm zzzz', tzinfo=british, locale='en_US')
cmlenz@34:     u'16:30 British Summer Time'
cmlenz@29: 
cmlenz@35: Here, the given UTC time is adjusted to the "Europe/London" time-zone, and
cmlenz@35: daylight savings time is taken into account. Daylight savings time is also
cmlenz@35: applied to ``format_time``, but because the actual date is unknown in that
cmlenz@35: case, the current day is assumed to determine whether DST or standard time
cmlenz@35: should be used.
cmlenz@35: 
cmlenz@29:  .. _`pytz`: http://pytz.sourceforge.net/
cmlenz@29: 
cmlenz@29: 
cmlenz@255: Localized Time-zone Names
cmlenz@255: -------------------------
cmlenz@255: 
cmlenz@255: While the ``Locale`` class provides access to various locale display names
cmlenz@255: related to time-zones, the process of building a localized name of a time-zone
cmlenz@255: is actually quite complicated. Babel implements it in separately usable
cmlenz@255: functions in the ``babel.dates`` module, most importantly the
cmlenz@255: ``get_timezone_name`` function:
cmlenz@255: 
cmlenz@255: .. code-block:: pycon
cmlenz@255: 
cmlenz@255:     >>> from pytz import timezone
cmlenz@255:     >>> from babel import Locale
cmlenz@255:     >>> from babel.dates import get_timezone_name
cmlenz@255: 
cmlenz@255:     >>> tz = timezone('Europe/Berlin')
cmlenz@255:     >>> get_timezone_name(tz, locale=Locale.parse('pt_PT'))
cmlenz@255:     u'Hor\xe1rio Alemanha'
cmlenz@255: 
cmlenz@255: You can pass the function either a ``datetime.tzinfo`` object, or a
cmlenz@255: ``datetime.date`` or ``datetime.datetime`` object. If you pass an actual date,
cmlenz@255: the function will be able to take daylight savings time into account. If you
cmlenz@255: pass just the time-zone, Babel does not know whether daylight savings time is
cmlenz@255: in effect, so it uses a generic representation, which is useful for example to
cmlenz@255: display a list of time-zones to the user.
cmlenz@255: 
cmlenz@255: .. code-block:: pycon
cmlenz@255: 
cmlenz@255:     >>> from datetime import datetime
cmlenz@255: 
cmlenz@255:     >>> dt = tz.localize(datetime(2007, 8, 15))
cmlenz@255:     >>> get_timezone_name(dt, locale=Locale.parse('de_DE'))
cmlenz@255:     u'Mitteleurop\xe4ische Sommerzeit'
cmlenz@255:     >>> get_timezone_name(tz, locale=Locale.parse('de_DE'))
cmlenz@255:     u'Deutschland'
cmlenz@255: 
cmlenz@255: 
cmlenz@18: Parsing Dates
cmlenz@124: =============
cmlenz@18: 
cmlenz@40: Babel can also parse date and time information in a locale-sensitive manner:
cmlenz@40: 
cmlenz@40: .. code-block:: pycon
cmlenz@18: 
cmlenz@18:     >>> from babel.dates import parse_date, parse_datetime, parse_time
cmlenz@255: 
cmlenz@255: .. note:: Date/time parsing is not properly implemented yet