parsedatetime 2.2

Installing

You can install parsedatetime using:

pip install parsedatetime

Running Tests

From the source directory:

make test

To run tests on several python versions, type make tox:

$ make tox
[... tox creates a virtualenv for every python version and runs tests inside of each]
py27: commands succeeded
py35: commands succeeded

This assumes that you have the versions you want to test under installed as part of your PyEnv environment:

pyenv install -s 2.6.9
pyenv install -s 2.7.11
pyenv install -s 3.5.2
pyenv install -s pypy-5.3
pyenv global 2.7.11 3.5.2 2.6.9 pypy-5.3

The tests depend on PyICU being installed. PyICU depends on icu4c which on OS X requires homebrew:

brew install icu4c

and then will require that you pass in the LDFLAGS and CPPFLAGS that homebrew will display after the install:

LDFLAGS:  -L/usr/local/opt/icu4c/lib
CPPFLAGS: -I/usr/local/opt/icu4c/include

The Makefile contains the OS X default values for them so you may need to tweak them.

Using parsedatetime

An example of how to use parsedatetime:

import parsedatetime

cal = parsedatetime.Calendar()

cal.parse("tomorrow")

To get it to a Python datetime object:

from datetime import datetime

time_struct, parse_status = cal.parse("tomorrow")
datetime(*time_struct[:6])

Parse datetime with timezone support (using pytz package):

import parsedatetime
import pytz
from pytz import timezone

cal = parsedatetime.Calendar()

datetime_obj, _ = cal.parseDT(datetimeString="tomorrow", tzinfo=timezone("US/Pacific"))

More detailed examples can be found in the examples directory.

Documentation

The generated documentation is included by default in the docs directory and can also be viewed online at https://bear.im/code/parsedatetime/docs/index.html

The docs can be generated by running:

make docs

Notes

The Calendar class has a member property named ptc which is created during the class init method to be an instance of parsedatetime_consts.CalendarConstants().

History

The code in parsedatetime has been implemented over the years in many different languages (C, Clipper, Delphi) as part of different custom/proprietary systems I’ve worked on. Sadly the previous code is not “open” in any sense of that word.

When I went to work for Open Source Applications Foundation and realized that the Chandler project could benefit from my experience with parsing of date/time text I decided to start from scratch and implement the code using Python and make it truly open.

After working on the initial concept and creating something that could be shown to the Chandler folks, the code has now evolved to its current state with the help of the Chandler folks, most especially Darshana.