Unstandard Python library of useful and highly-reusable functions.
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
test Start of Python 3 compatibility Nov 22, 2013
.gitignore Trying to add tests, something is borken though. Sep 21, 2011
.travis.yml Shhh, travis. Nov 25, 2013
LICENSE Add LICENSE file. Jan 3, 2015
MANIFEST.in Include README, fixes #24 Oct 20, 2013
Makefile Convenience Makefile. Mar 10, 2013
README.md Add LICENSE file. Jan 3, 2015
setup.cfg Prettier. Sep 21, 2011
setup.py Adding tagbuilder. Jun 21, 2014
tox.ini Add tox. Mar 23, 2013


Unstandard Library for Python

Have you ever written code that you used in more than one project? Me too.

This is a compilation of highly-reusable code for Python.

Getting started

Install the latest release of unstdlib:

$ pip install unstdlib

Use it in your code:

from unstdlib import get_many, groupby_dict

# ...

user_id, sort_by, page_num = get_many(request.params, ['user_id'], optional=['sort_by', 'page_num'])

# ...
data_by_tags = groupby_dict(data, keyfunc=lambda o: o.tag)

for n in data_by_tags['news']:
    print n


(TODO: Format this better and select the most useful ones.)


  • datetime_.iterate_date(start, stop=None, step=datetime.timedelta(days=1))
  • datetime_.iterate_date_values(d, start_date=None, stop_date=None, default=0)
  • datetime_.isoformat_as_datetime(s)
  • datetime_.truncate_datetime(t, resolution)
  • datetime_.now(timezone=None)
  • dict_.get_many(d, required=[], optional=[], one_of=[])
  • dict_.pop_many(d, keys, default=None)
  • @exception_.convert_exception(from_exception, to_exception, *to_args, **to_kw)
  • functools_.assert_hashable(*args, **kw)
  • @functools_.memoized(fn=None, cache=None)
  • @functools_.memoized_property(object)
  • @functools_.deprecated(message, exception=PendingDeprecationWarning)
  • list_.groupby_count(i, key=None, force_keys=None)
  • list_.iterate(maybe_iter, unless=(basestring, dict))
  • list_.is_iterable(maybe_iter, unless=(basestring, dict))
  • list_.iterate_chunks(i, size=10)
  • list_.iterate_items(dictish)
  • list_.iterate_flatten(q)
  • @list_.listify(fn=None, wrapper=list)
  • string_.random_string(length=6, alphabet=string.letters+string.digits)
  • string_.number_to_string(n, alphabet)
  • string_.string_to_number(s, alphabet)
  • string_.dollars_to_cents(s, allow_negative=False)
  • string_.to_str(obj, encoding='utf-8', **encode_args)
  • string_.to_unicode(obj, encoding='utf-8', fallback='latin1', **decode_args)
  • string_.to_int(s, default=0)
  • string_.format_int(n, singular=_Default, plural=_Default)
  • string_.slugify(s, delimiter='-')
  • type_.is_subclass(o, bases)
  • os_.chdir(new_dir): like os.chdir, but also a context manager: with chdir("/tmp/"): pass


  • validate(d, key, validator)
  • validate_many(d, schema)


  • enumerate_query_by_limit(q, limit=1000)


  • get_cache_buster(src_path, method='importtime')
  • literal(s)
  • tag(tagname, content='', attrs=None)
  • javascript_link(src_url, src_path=None, cache_bust=None, content='', extra_attrs=None)
  • stylesheet_link(src_url, src_path=None, cache_bust=None, content='', extra_attrs=None)

Organization & Philosophy

This library includes code with no dependencies and code with dependencies. Only the no-dependency code will be imported by default. Each collection of code with a specific dependency is bundled under its own module (such as Django-specific code).

  • unstdlib.standard contains code that does not require additional dependencies on top of Python 2.5.
  • unstdlib.sqlalchemy contains code that is SQLAlchemy-specific.
  • unstdlib.django (someday) contains code that is Django-specific.

We value simplicity and elegance over robustness and optimization. This library should serve as a good foundation for your own application-specific code instead of a complete framework. In other words, it is preferred to have a 2-line function that covers 80% of use cases than a 20-line function that covers 100% of the use cases. Unexpected behaviour on bad input is fine as long as expected usage is well documented.


Forks are highly encouraged. Everyone should have a collection of code they commonly reuse. If you feel your code will be useful to others, make sure that it is conforming to the spirit of the library outlined in the Organization & Philosophy section and send over a pull request.


MIT (See LICENSE file).