numpytimebuilder


Namenumpytimebuilder JSON
Version 0.4.2 PyPI version JSON
download
home_pagehttps://bitbucket.org/nielsenb/numpytimebuilder
SummaryA library for using the NumPy datetime API with aniso8601
upload_time2025-01-10 01:54:16
maintainerNone
docs_urlNone
authorBrandon Nielsen
requires_pythonNone
licenseNone
keywords iso8601 numpy aniso8601 datetime
VCS
bugtrack_url
requirements No requirements were recorded.
Travis-CI No Travis.
coveralls test coverage No coveralls.
            NumPyTimeBuilder
================

aniso8601 builder for NumPy datetimes
-------------------------------------

Features
========
* Provides :code:`NumPyTimeBuilder` compatible with `aniso8601 <https://bitbucket.org/nielsenb/aniso8601>`_
* Returns :code:`datetime64` and :code:`timedelta64` `NumPy types <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_

Installation
============

The recommended installation method is to use pip::

  $ pip install numpytimebuilder

Alternatively, you can download the source (git repository hosted at `Bitbucket <https://bitbucket.org/nielsenb/numpytimebuilder>`_) and install directly::

  $ python setup.py install

Use
===

Parsing datetimes
-----------------

To parse a typical ISO 8601 datetime string::

  >>> import aniso8601
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> aniso8601.parse_datetime('1977-06-10T12:00:00', builder=NumPyTimeBuilder)
  numpy.datetime64('1977-06-10T12:00:00')

Alternative delimiters can be specified, for example, a space::

  >>> aniso8601.parse_datetime('1977-06-10 12:00:00', delimiter=' ', builder=NumPyTimeBuilder)
  numpy.datetime64('1977-06-10T12:00:00')

Since the NumPy :code:`datetime64` implementaton only supports naive datetimes, timezones are explicitly not supported::

  >>> aniso8601.parse_datetime('1977-06-10T12:00:00Z', builder=NumPyTimeBuilder)
  Traceback (most recent call last):
    File "<stdin>", line 1, in <module>
    File "/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py", line 131, in parse_datetime
      return builder.build_datetime(datepart, timepart)
    File "numpytimebuilder/__init__.py", line 37, in build_datetime
      raise NotImplementedError('Timezones are not supported by numpy '
  NotImplementedError: Timezones are not supported by numpy datetime64 type.
  >>> aniso8601.parse_datetime('1979-06-05T08:00:00-08:00', builder=NumPyTimeBuilder)
  Traceback (most recent call last):
    File "<stdin>", line 1, in <module>
    File "/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py", line 131, in parse_datetime
      return builder.build_datetime(datepart, timepart)
    File "numpytimebuilder/__init__.py", line 37, in build_datetime
      raise NotImplementedError('Timezones are not supported by numpy '
  NotImplementedError: Timezones are not supported by numpy datetime64 type

Leap seconds are not currently supported by the NumPy :code:`datetime64` implementation, so leap seconds are explicitly not supported::

  >>> aniso8601.parse_datetime('2018-03-06T23:59:60', builder=NumPyTimeBuilder)
  Traceback (most recent call last):
    File "<stdin>", line 1, in <module>
    File "/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py", line 131, in parse_datetime
      return builder.build_datetime(datepart, timepart)
    File "numpytimebuilder/__init__.py", line 120, in build_datetime
      raise LeapSecondError('Leap seconds are not supported.')
  aniso8601.exceptions.LeapSecondError: Leap seconds are not supported.

Parsing dates
-------------

To parse a date represented in an ISO 8601 string::

  >>> import aniso8601
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> aniso8601.parse_date('1984-04-23', builder=NumPyTimeBuilder)
  numpy.datetime64('1984-04-23')

Basic format is supported as well::

  >>> aniso8601.parse_date('19840423', builder=NumPyTimeBuilder)
  numpy.datetime64('1984-04-23')

To parse a date using the ISO 8601 week date format::

  >>> aniso8601.parse_date('1986-W38-1', builder=NumPyTimeBuilder)
  numpy.datetime64('1986-09-15')

To parse an ISO 8601 ordinal date::

  >>> aniso8601.parse_date('1988-132', builder=NumPyTimeBuilder)
  numpy.datetime64('1988-05-11')

Parsing times
-------------

NumPy offers no :code:`time64` type, so parsing times is explicitly not supported::

  >>> import aniso8601
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> aniso8601.parse_time('11:31:14', builder=NumPyTimeBuilder)
  Traceback (most recent call last):
    File "<stdin>", line 1, in <module>
    File "/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py", line 116, in parse_time
      return _RESOLUTION_MAP[get_time_resolution(timestr)](timestr, tz, builder)
    File "/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py", line 165, in _parse_second_time
      return builder.build_time(hh=hourstr, mm=minutestr, ss=secondstr, tz=tz)
    File "numpytimebuilder/__init__.py", line 32, in build_time
      raise NotImplementedError('No compatible numpy time64 type.')
  NotImplementedError: No compatible numpy time64 type.

Parsing durations
-----------------

The NumPy :code:`timedelta64` type only supports a single component per delta, so durations are returned as a tuple of :code:`timedelta64` objects.

To parse a duration formatted as an ISO 8601 string::

  >>> import aniso8601
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> aniso8601.parse_duration('P1Y2M3DT4H54M6S', builder=NumPyTimeBuilder)
  (numpy.timedelta64(428,'D'), numpy.timedelta64(4,'h'), numpy.timedelta64(54,'m'), numpy.timedelta64(6,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

Reduced accuracy is supported::

  >>> aniso8601.parse_duration('P1Y', builder=NumPyTimeBuilder)
  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

A decimal fraction is allowed on the lowest order element::

  >>> aniso8601.parse_duration('P1YT3.5M', builder=NumPyTimeBuilder)
  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(3,'m'), numpy.timedelta64(30,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

The decimal fraction can be specified with a comma instead of a full-stop::

  >>> aniso8601.parse_duration('P1YT3,5M', builder=NumPyTimeBuilder)
  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(3,'m'), numpy.timedelta64(30,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

Parsing a duration from a combined date and time is supported as well::

  >>> aniso8601.parse_duration('P0001-01-02T01:30:5', builder=NumPyTimeBuilder)
  (numpy.timedelta64(397,'D'), numpy.timedelta64(1,'h'), numpy.timedelta64(30,'m'), numpy.timedelta64(5,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

The above treat years as 365 days and months as 30 days. Calendar level accuracy is not supported. Fractional years and months are supported accordingly::

  >>> aniso8601.parse_duration('P2.1Y', builder=NumPyTimeBuilder)
  (numpy.timedelta64(766,'D'), numpy.timedelta64(12,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))
  >>> aniso8601.parse_duration('P1Y0.5M', builder=NumPyTimeBuilder)
  (numpy.timedelta64(380,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))

Applying durations
^^^^^^^^^^^^^^^^^^

The :code:`apply_duration` helper function is provided for applying duration tuples to a :code:`datetime64` object. It takes a :code:`datetime64` (from :code:`parse_datetime`), a duration tuple (from :code:`parse_duration`), and a `Python operator <https://docs.python.org/3.6/library/operator.html>`_ to be applied::

  >>> import aniso8601
  >>> import operator
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> from numpytimebuilder.util import apply_duration
  >>> datetime = aniso8601.parse_datetime('1977-06-10T12:00:00', builder=NumPyTimeBuilder)
  >>> duration = aniso8601.parse_duration('P3Y2M1DT1H2M3S', builder=NumPyTimeBuilder)
  >>> apply_duration(datetime, duration, operator.add)
  numpy.datetime64('1980-08-09T13:02:03')

**Keep in mind the span of representable datetimes decreases as the resolution increases!** See the `NumPy Datetime Units <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html#datetime-units>`_ documentation for more information.

Parsing intervals
-----------------

To parse an interval specified by a start and end::

  >>> import aniso8601
  >>> from numpytimebuilder import NumPyTimeBuilder
  >>> aniso8601.parse_interval('2007-03-01T13:00:00/2008-05-11T15:30:00')
  (numpy.datetime64('2007-03-01T13:00:00'), numpy.datetime64('2008-05-11T15:30:00'))

Intervals specified by a start time and a duration are supported::

  >>> aniso8601.parse_interval('2007-03-01T13:00:00/P1Y2M10DT2H30M', builder=NumPyTimeBuilder)
  (numpy.datetime64('2007-03-01T13:00:00'), numpy.datetime64('2008-05-09T15:30:00'))

A duration can also be specified by a duration and end time::

  >>> aniso8601.parse_interval('P1M/1981-04-05', builder=NumPyTimeBuilder)
  (numpy.datetime64('1981-04-05'), numpy.datetime64('1981-03-06'))

Notice that the result of the above parse is not in order from earliest to latest. If sorted intervals are required, simply use the :code:`sorted` keyword as shown below::

  >>> sorted(aniso8601.parse_interval('P1M/1981-04-05', builder=NumPyTimeBuilder))
  [numpy.datetime64('1981-03-06'), numpy.datetime64('1981-04-05')]

The end of an interval is returned as a datetime when required to maintain the resolution specified by a duration, even if the duration start is given as a date::

  >>> aniso8601.parse_interval('2014-11-12/PT4H54M6.5S', builder=NumPyTimeBuilder)
  (numpy.datetime64('2014-11-12'), numpy.datetime64('2014-11-12T04:54:06.500'))
  >>> aniso8601.parse_interval('2007-03-01/P1.5D', builder=NumPyTimeBuilder)
  (numpy.datetime64('2007-03-01'), numpy.datetime64('2007-03-02T12:00:00'))

Repeating intervals are supported as well, and return a generator::

  >>> aniso8601.parse_repeating_interval('R3/1981-04-05/P1D', builder=NumPyTimeBuilder)
  <generator object _date_generator at 0x7fd76fe9abe0>
  >>> list(aniso8601.parse_repeating_interval('R3/1981-04-05/P1D', builder=NumPyTimeBuilder))
  [numpy.datetime64('1981-04-05'), numpy.datetime64('1981-04-06'), numpy.datetime64('1981-04-07')]

Repeating intervals are allowed to go in the reverse direction::

  >>> list(aniso8601.parse_repeating_interval('R2/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder))
  [numpy.datetime64('1980-03-05T01:01:00'), numpy.datetime64('1980-03-04T23:59:00')]

Unbounded intervals are also allowed (Python 2)::

  >>> result = aniso8601.parse_repeating_interval('R/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder)
  >>> result.next()
  numpy.datetime64('1980-03-05T01:01:00')
  >>> result.next()
  numpy.datetime64('1980-03-04T23:59:00')

or for Python 3::

  >>> result = aniso8601.parse_repeating_interval('R/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder)
  >>> next(result)
  numpy.datetime64('1980-03-05T01:01:00')
  >>> next(result)
  numpy.datetime64('1980-03-04T23:59:00')

The above treat years as 365 days and months as 30 days. Calendar level accuracy is not supported. Fractional months and years are supported accordingly::

  >>> aniso8601.parse_interval('P1.1Y/2001-02-28', builder=NumPyTimeBuilder)
  (numpy.datetime64('2001-02-28'), numpy.datetime64('2000-01-24'))
  >>> aniso8601.parse_interval('2001-02-28/P1Y2.5M', builder=NumPyTimeBuilder)
  (numpy.datetime64('2001-02-28'), numpy.datetime64('2002-05-14'))

Development
===========

Setup
-----

It is recommended to develop using a `virtualenv <https://virtualenv.pypa.io/en/stable/>`_.

Inside a virtualenv, development dependencies can be installed automatically::

  $ pip install -e .[dev]

`pre-commit <https://pre-commit.com/>`_ is used for managing pre-commit hooks::

  $ pre-commit install

To run the pre-commit hooks manually::

  $ pre-commit run --all-files

Tests
-----

Tests can be run using the `unittest testing framework <https://docs.python.org/3/library/unittest.html>`_::

   $ python -m unittest discover numpytimebuilder

Contributing
============

numpytimebuilder is an open source project hosted on `Bitbucket <https://bitbucket.org/nielsenb/numpytimebuilder>`_.

Any and all bugs are welcome on our `issue tracker <https://bitbucket.org/nielsenb/numpytimebuilder/issues>`_.

References
==========

* `NumPy datetimes and timedeltas <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_
* `aniso8601 and sub-microsecond precision <https://bitbucket.org/nielsenb/aniso8601/issues/10/sub-microsecond-precision-in-durations-is>`_

            

Raw data

            {
    "_id": null,
    "home_page": "https://bitbucket.org/nielsenb/numpytimebuilder",
    "name": "numpytimebuilder",
    "maintainer": null,
    "docs_url": null,
    "requires_python": null,
    "maintainer_email": null,
    "keywords": "iso8601 numpy aniso8601 datetime",
    "author": "Brandon Nielsen",
    "author_email": "nielsenb@jetfuse.net",
    "download_url": "https://files.pythonhosted.org/packages/c4/64/930d4566de0ff32809b8880cac25eb61abdf917dc675d4f469ae54e60e0a/numpytimebuilder-0.4.2.tar.gz",
    "platform": null,
    "description": "NumPyTimeBuilder\n================\n\naniso8601 builder for NumPy datetimes\n-------------------------------------\n\nFeatures\n========\n* Provides :code:`NumPyTimeBuilder` compatible with `aniso8601 <https://bitbucket.org/nielsenb/aniso8601>`_\n* Returns :code:`datetime64` and :code:`timedelta64` `NumPy types <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_\n\nInstallation\n============\n\nThe recommended installation method is to use pip::\n\n  $ pip install numpytimebuilder\n\nAlternatively, you can download the source (git repository hosted at `Bitbucket <https://bitbucket.org/nielsenb/numpytimebuilder>`_) and install directly::\n\n  $ python setup.py install\n\nUse\n===\n\nParsing datetimes\n-----------------\n\nTo parse a typical ISO 8601 datetime string::\n\n  >>> import aniso8601\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> aniso8601.parse_datetime('1977-06-10T12:00:00', builder=NumPyTimeBuilder)\n  numpy.datetime64('1977-06-10T12:00:00')\n\nAlternative delimiters can be specified, for example, a space::\n\n  >>> aniso8601.parse_datetime('1977-06-10 12:00:00', delimiter=' ', builder=NumPyTimeBuilder)\n  numpy.datetime64('1977-06-10T12:00:00')\n\nSince the NumPy :code:`datetime64` implementaton only supports naive datetimes, timezones are explicitly not supported::\n\n  >>> aniso8601.parse_datetime('1977-06-10T12:00:00Z', builder=NumPyTimeBuilder)\n  Traceback (most recent call last):\n    File \"<stdin>\", line 1, in <module>\n    File \"/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py\", line 131, in parse_datetime\n      return builder.build_datetime(datepart, timepart)\n    File \"numpytimebuilder/__init__.py\", line 37, in build_datetime\n      raise NotImplementedError('Timezones are not supported by numpy '\n  NotImplementedError: Timezones are not supported by numpy datetime64 type.\n  >>> aniso8601.parse_datetime('1979-06-05T08:00:00-08:00', builder=NumPyTimeBuilder)\n  Traceback (most recent call last):\n    File \"<stdin>\", line 1, in <module>\n    File \"/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py\", line 131, in parse_datetime\n      return builder.build_datetime(datepart, timepart)\n    File \"numpytimebuilder/__init__.py\", line 37, in build_datetime\n      raise NotImplementedError('Timezones are not supported by numpy '\n  NotImplementedError: Timezones are not supported by numpy datetime64 type\n\nLeap seconds are not currently supported by the NumPy :code:`datetime64` implementation, so leap seconds are explicitly not supported::\n\n  >>> aniso8601.parse_datetime('2018-03-06T23:59:60', builder=NumPyTimeBuilder)\n  Traceback (most recent call last):\n    File \"<stdin>\", line 1, in <module>\n    File \"/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py\", line 131, in parse_datetime\n      return builder.build_datetime(datepart, timepart)\n    File \"numpytimebuilder/__init__.py\", line 120, in build_datetime\n      raise LeapSecondError('Leap seconds are not supported.')\n  aniso8601.exceptions.LeapSecondError: Leap seconds are not supported.\n\nParsing dates\n-------------\n\nTo parse a date represented in an ISO 8601 string::\n\n  >>> import aniso8601\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> aniso8601.parse_date('1984-04-23', builder=NumPyTimeBuilder)\n  numpy.datetime64('1984-04-23')\n\nBasic format is supported as well::\n\n  >>> aniso8601.parse_date('19840423', builder=NumPyTimeBuilder)\n  numpy.datetime64('1984-04-23')\n\nTo parse a date using the ISO 8601 week date format::\n\n  >>> aniso8601.parse_date('1986-W38-1', builder=NumPyTimeBuilder)\n  numpy.datetime64('1986-09-15')\n\nTo parse an ISO 8601 ordinal date::\n\n  >>> aniso8601.parse_date('1988-132', builder=NumPyTimeBuilder)\n  numpy.datetime64('1988-05-11')\n\nParsing times\n-------------\n\nNumPy offers no :code:`time64` type, so parsing times is explicitly not supported::\n\n  >>> import aniso8601\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> aniso8601.parse_time('11:31:14', builder=NumPyTimeBuilder)\n  Traceback (most recent call last):\n    File \"<stdin>\", line 1, in <module>\n    File \"/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py\", line 116, in parse_time\n      return _RESOLUTION_MAP[get_time_resolution(timestr)](timestr, tz, builder)\n    File \"/home/nielsenb/Jetfuse/numpytimebuilder/python2/lib/python2.7/site-packages/aniso8601/time.py\", line 165, in _parse_second_time\n      return builder.build_time(hh=hourstr, mm=minutestr, ss=secondstr, tz=tz)\n    File \"numpytimebuilder/__init__.py\", line 32, in build_time\n      raise NotImplementedError('No compatible numpy time64 type.')\n  NotImplementedError: No compatible numpy time64 type.\n\nParsing durations\n-----------------\n\nThe NumPy :code:`timedelta64` type only supports a single component per delta, so durations are returned as a tuple of :code:`timedelta64` objects.\n\nTo parse a duration formatted as an ISO 8601 string::\n\n  >>> import aniso8601\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> aniso8601.parse_duration('P1Y2M3DT4H54M6S', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(428,'D'), numpy.timedelta64(4,'h'), numpy.timedelta64(54,'m'), numpy.timedelta64(6,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nReduced accuracy is supported::\n\n  >>> aniso8601.parse_duration('P1Y', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nA decimal fraction is allowed on the lowest order element::\n\n  >>> aniso8601.parse_duration('P1YT3.5M', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(3,'m'), numpy.timedelta64(30,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nThe decimal fraction can be specified with a comma instead of a full-stop::\n\n  >>> aniso8601.parse_duration('P1YT3,5M', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(365,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(3,'m'), numpy.timedelta64(30,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nParsing a duration from a combined date and time is supported as well::\n\n  >>> aniso8601.parse_duration('P0001-01-02T01:30:5', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(397,'D'), numpy.timedelta64(1,'h'), numpy.timedelta64(30,'m'), numpy.timedelta64(5,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nThe above treat years as 365 days and months as 30 days. Calendar level accuracy is not supported. Fractional years and months are supported accordingly::\n\n  >>> aniso8601.parse_duration('P2.1Y', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(766,'D'), numpy.timedelta64(12,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n  >>> aniso8601.parse_duration('P1Y0.5M', builder=NumPyTimeBuilder)\n  (numpy.timedelta64(380,'D'), numpy.timedelta64(0,'h'), numpy.timedelta64(0,'m'), numpy.timedelta64(0,'s'), numpy.timedelta64(0,'ms'), numpy.timedelta64(0,'us'), numpy.timedelta64(0,'ns'), numpy.timedelta64(0,'ps'), numpy.timedelta64(0,'fs'), numpy.timedelta64(0,'as'))\n\nApplying durations\n^^^^^^^^^^^^^^^^^^\n\nThe :code:`apply_duration` helper function is provided for applying duration tuples to a :code:`datetime64` object. It takes a :code:`datetime64` (from :code:`parse_datetime`), a duration tuple (from :code:`parse_duration`), and a `Python operator <https://docs.python.org/3.6/library/operator.html>`_ to be applied::\n\n  >>> import aniso8601\n  >>> import operator\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> from numpytimebuilder.util import apply_duration\n  >>> datetime = aniso8601.parse_datetime('1977-06-10T12:00:00', builder=NumPyTimeBuilder)\n  >>> duration = aniso8601.parse_duration('P3Y2M1DT1H2M3S', builder=NumPyTimeBuilder)\n  >>> apply_duration(datetime, duration, operator.add)\n  numpy.datetime64('1980-08-09T13:02:03')\n\n**Keep in mind the span of representable datetimes decreases as the resolution increases!** See the `NumPy Datetime Units <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html#datetime-units>`_ documentation for more information.\n\nParsing intervals\n-----------------\n\nTo parse an interval specified by a start and end::\n\n  >>> import aniso8601\n  >>> from numpytimebuilder import NumPyTimeBuilder\n  >>> aniso8601.parse_interval('2007-03-01T13:00:00/2008-05-11T15:30:00')\n  (numpy.datetime64('2007-03-01T13:00:00'), numpy.datetime64('2008-05-11T15:30:00'))\n\nIntervals specified by a start time and a duration are supported::\n\n  >>> aniso8601.parse_interval('2007-03-01T13:00:00/P1Y2M10DT2H30M', builder=NumPyTimeBuilder)\n  (numpy.datetime64('2007-03-01T13:00:00'), numpy.datetime64('2008-05-09T15:30:00'))\n\nA duration can also be specified by a duration and end time::\n\n  >>> aniso8601.parse_interval('P1M/1981-04-05', builder=NumPyTimeBuilder)\n  (numpy.datetime64('1981-04-05'), numpy.datetime64('1981-03-06'))\n\nNotice that the result of the above parse is not in order from earliest to latest. If sorted intervals are required, simply use the :code:`sorted` keyword as shown below::\n\n  >>> sorted(aniso8601.parse_interval('P1M/1981-04-05', builder=NumPyTimeBuilder))\n  [numpy.datetime64('1981-03-06'), numpy.datetime64('1981-04-05')]\n\nThe end of an interval is returned as a datetime when required to maintain the resolution specified by a duration, even if the duration start is given as a date::\n\n  >>> aniso8601.parse_interval('2014-11-12/PT4H54M6.5S', builder=NumPyTimeBuilder)\n  (numpy.datetime64('2014-11-12'), numpy.datetime64('2014-11-12T04:54:06.500'))\n  >>> aniso8601.parse_interval('2007-03-01/P1.5D', builder=NumPyTimeBuilder)\n  (numpy.datetime64('2007-03-01'), numpy.datetime64('2007-03-02T12:00:00'))\n\nRepeating intervals are supported as well, and return a generator::\n\n  >>> aniso8601.parse_repeating_interval('R3/1981-04-05/P1D', builder=NumPyTimeBuilder)\n  <generator object _date_generator at 0x7fd76fe9abe0>\n  >>> list(aniso8601.parse_repeating_interval('R3/1981-04-05/P1D', builder=NumPyTimeBuilder))\n  [numpy.datetime64('1981-04-05'), numpy.datetime64('1981-04-06'), numpy.datetime64('1981-04-07')]\n\nRepeating intervals are allowed to go in the reverse direction::\n\n  >>> list(aniso8601.parse_repeating_interval('R2/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder))\n  [numpy.datetime64('1980-03-05T01:01:00'), numpy.datetime64('1980-03-04T23:59:00')]\n\nUnbounded intervals are also allowed (Python 2)::\n\n  >>> result = aniso8601.parse_repeating_interval('R/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder)\n  >>> result.next()\n  numpy.datetime64('1980-03-05T01:01:00')\n  >>> result.next()\n  numpy.datetime64('1980-03-04T23:59:00')\n\nor for Python 3::\n\n  >>> result = aniso8601.parse_repeating_interval('R/PT1H2M/1980-03-05T01:01:00', builder=NumPyTimeBuilder)\n  >>> next(result)\n  numpy.datetime64('1980-03-05T01:01:00')\n  >>> next(result)\n  numpy.datetime64('1980-03-04T23:59:00')\n\nThe above treat years as 365 days and months as 30 days. Calendar level accuracy is not supported. Fractional months and years are supported accordingly::\n\n  >>> aniso8601.parse_interval('P1.1Y/2001-02-28', builder=NumPyTimeBuilder)\n  (numpy.datetime64('2001-02-28'), numpy.datetime64('2000-01-24'))\n  >>> aniso8601.parse_interval('2001-02-28/P1Y2.5M', builder=NumPyTimeBuilder)\n  (numpy.datetime64('2001-02-28'), numpy.datetime64('2002-05-14'))\n\nDevelopment\n===========\n\nSetup\n-----\n\nIt is recommended to develop using a `virtualenv <https://virtualenv.pypa.io/en/stable/>`_.\n\nInside a virtualenv, development dependencies can be installed automatically::\n\n  $ pip install -e .[dev]\n\n`pre-commit <https://pre-commit.com/>`_ is used for managing pre-commit hooks::\n\n  $ pre-commit install\n\nTo run the pre-commit hooks manually::\n\n  $ pre-commit run --all-files\n\nTests\n-----\n\nTests can be run using the `unittest testing framework <https://docs.python.org/3/library/unittest.html>`_::\n\n   $ python -m unittest discover numpytimebuilder\n\nContributing\n============\n\nnumpytimebuilder is an open source project hosted on `Bitbucket <https://bitbucket.org/nielsenb/numpytimebuilder>`_.\n\nAny and all bugs are welcome on our `issue tracker <https://bitbucket.org/nielsenb/numpytimebuilder/issues>`_.\n\nReferences\n==========\n\n* `NumPy datetimes and timedeltas <https://docs.scipy.org/doc/numpy/reference/arrays.datetime.html>`_\n* `aniso8601 and sub-microsecond precision <https://bitbucket.org/nielsenb/aniso8601/issues/10/sub-microsecond-precision-in-durations-is>`_\n",
    "bugtrack_url": null,
    "license": null,
    "summary": "A library for using the NumPy datetime API with aniso8601",
    "version": "0.4.2",
    "project_urls": {
        "Documentation": "https://numpytimebuilder.readthedocs.io/",
        "Homepage": "https://bitbucket.org/nielsenb/numpytimebuilder",
        "Source": "https://bitbucket.org/nielsenb/numpytimebuilder",
        "Tracker": "https://bitbucket.org/nielsenb/numpytimebuilder/issues"
    },
    "split_keywords": [
        "iso8601",
        "numpy",
        "aniso8601",
        "datetime"
    ],
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "188a40cc8ed4d408a3d860c2a67093a93e46a965955a5409f2f81fde65174aa8",
                "md5": "67765593435d63d1b682f1be146de00e",
                "sha256": "6909dd8c818b774d16086a1a66730498924cf41a8044c8fbb58865c841bc3715"
            },
            "downloads": -1,
            "filename": "numpytimebuilder-0.4.2-py2.py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "67765593435d63d1b682f1be146de00e",
            "packagetype": "bdist_wheel",
            "python_version": "py2.py3",
            "requires_python": null,
            "size": 16650,
            "upload_time": "2025-01-10T01:54:13",
            "upload_time_iso_8601": "2025-01-10T01:54:13.898193Z",
            "url": "https://files.pythonhosted.org/packages/18/8a/40cc8ed4d408a3d860c2a67093a93e46a965955a5409f2f81fde65174aa8/numpytimebuilder-0.4.2-py2.py3-none-any.whl",
            "yanked": false,
            "yanked_reason": null
        },
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "c464930d4566de0ff32809b8880cac25eb61abdf917dc675d4f469ae54e60e0a",
                "md5": "af553189938763c9f51b3a0679cde8f4",
                "sha256": "c29a9aeb64aa9dca656003815e408695a968e3f4789f4ec43686c21280089a81"
            },
            "downloads": -1,
            "filename": "numpytimebuilder-0.4.2.tar.gz",
            "has_sig": false,
            "md5_digest": "af553189938763c9f51b3a0679cde8f4",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 17802,
            "upload_time": "2025-01-10T01:54:16",
            "upload_time_iso_8601": "2025-01-10T01:54:16.256396Z",
            "url": "https://files.pythonhosted.org/packages/c4/64/930d4566de0ff32809b8880cac25eb61abdf917dc675d4f469ae54e60e0a/numpytimebuilder-0.4.2.tar.gz",
            "yanked": false,
            "yanked_reason": null
        }
    ],
    "upload_time": "2025-01-10 01:54:16",
    "github": false,
    "gitlab": false,
    "bitbucket": true,
    "codeberg": false,
    "bitbucket_user": "nielsenb",
    "bitbucket_project": "numpytimebuilder",
    "lcname": "numpytimebuilder"
}
        
Elapsed time: 0.59275s