Arrow

-------------------

- [NEW] ``factory.get()`` can now create arrow objects from a ISO calendar tuple, for example:

.. code-block:: python

>>> arrow.get((2013, 18, 7))
<Arrow [2013-05-05T00:00:00+00:00]>

- [NEW] Added a new token ``x`` to allow parsing of integer timestamps with milliseconds and microseconds.
- [NEW] Formatting now supports escaping of characters using the same syntax as parsing, for example:

.. code-block:: python

>>> arw = arrow.now()
>>> fmt = "YYYY-MM-DD h [h] m"
>>> arw.format(fmt)
'2019-11-02 3 h 32'

- [NEW] Added ``humanize`` week granularity translations for Chinese, Spanish and Vietnamese.
- [CHANGE] Added ``ParserError`` to module exports.
- [FIX] Added support for midnight at end of day. See `703 <https://github.com/arrow-py/arrow/issues/703>`_ for details.
- [INTERNAL] Created Travis build for macOS.
- [INTERNAL] Test parsing and formatting against full timezone database.

-------------------

- [NEW] Added ``humanize`` week granularity translations for Portuguese and Brazilian Portuguese.
- [NEW] Embedded changelog within docs and added release dates to versions.
- [FIX] Fixed a bug that caused test failures on Windows only, see `668 <https://github.com/arrow-py/arrow/issues/668>`_ for details.

-------------------

- [NEW] Added ``humanize`` week granularity translations for Japanese.
- [FIX] Fixed a bug that caused Arrow to fail when passed a negative timestamp string.
- [FIX] Fixed a bug that caused Arrow to fail when passed a datetime object with ``tzinfo`` of type ``StaticTzInfo``.

.. code-block:: python

>>> arrow.get("1565358758", "X")
>>> arrow.get("1565358758.123413", "X")
>>> arrow.get(1565358758)
>>> arrow.get(1565358758.123413)

- [CHANGE] When a meridian token (a|A) is passed and no meridians are available for the specified locale (e.g. unsupported or untranslated) a ``ParserError`` is raised.
- [CHANGE] The timestamp token (``X``) will now match float timestamps of type ``str``: ``arrow.get(“1565358758.123415”, “X”)``.
- [CHANGE] Strings with leading and/or trailing whitespace will no longer be parsed without a format string. Please see `the docs <https://arrow.readthedocs.io/#regular-expressions>`_ for ways to handle this.
- [FIX] The timestamp token (``X``) will now only match on strings that **strictly contain integers and floats**, preventing incorrect matches.
- [FIX] Most instances of ``arrow.get()`` returning an incorrect ``Arrow`` object from a partial parsing match have been eliminated. The following issue have been addressed: `91 <https://github.com/arrow-py/arrow/issues/91>`_, `#196 <https://github.com/arrow-py/arrow/issues/196>`_, `#396 <https://github.com/arrow-py/arrow/issues/396>`_, `#434 <https://github.com/arrow-py/arrow/issues/434>`_, `#447 <https://github.com/arrow-py/arrow/issues/447>`_, `#456 <https://github.com/arrow-py/arrow/issues/456>`_, `#519 <https://github.com/arrow-py/arrow/issues/519>`_, `#538 <https://github.com/arrow-py/arrow/issues/538>`_, `#560 <https://github.com/arrow-py/arrow/issues/560>`_.

-------------------

- [CHANGE] ``ArrowParseWarning`` will no longer be printed on every call to ``arrow.get()`` with a datetime string. The purpose of the warning was to start a conversation about the upcoming 0.15.0 changes and we appreciate all the feedback that the community has given us!

-------------------

- [NEW] Added support for ``week`` granularity in ``Arrow.humanize()``. For example, ``arrow.utcnow().shift(weeks=-1).humanize(granularity="week")`` outputs "a week ago". This change introduced two new untranslated words, ``week`` and ``weeks``, to all locale dictionaries, so locale contributions are welcome!
- [NEW] Fully translated the Brazilian Portuguese locale.
- [CHANGE] Updated the Macedonian locale to inherit from a Slavic base.
- [FIX] Fixed a bug that caused ``arrow.get()`` to ignore tzinfo arguments of type string (e.g. ``arrow.get(tzinfo="Europe/Paris")``).
- [FIX] Fixed a bug that occurred when ``arrow.Arrow()`` was instantiated with a ``pytz`` tzinfo object.
- [FIX] Fixed a bug that caused Arrow to fail when passed a sub-second token, that when rounded, had a value greater than 999999 (e.g. ``arrow.get("2015-01-12T01:13:15.9999995")``). Arrow should now accurately propagate the rounding for large sub-second tokens.