Skip to main content


Project description


Pragmatic & flexible configuration loader that makes your app settings clean and sexy.

.. image::

.. image::

.. image::

.. image::

You can find documentation here:

Supported platforms

* Python 2.6
* Python 2.7
* Python 3.2
* Python 3.3


Idea is simple and has been developed by (probably not only) django
community, but has been never standardized. Here it is:

1) keep your configuration in separate python modules inside your
settings/conf/config/<however-you-call-it> *module* like this::

|-- => init used for loading code
|-- => this should be common stuff used in all environments
|-- => change/add all you need to run the app in development
|-- => change/add all you need to run the app in production
`-- => everything that depends on your machine you currently
working on for your own (and everybody in a team)
safety - remove it from version control

2) load those files in the order you want to get the
configuration. Use environment variables to decide which settings
should be loaded. Assuming the variable is called ``ENV``, here's
the code that loads the configuration inside settings/

.. code-block:: python

import os
import confy

confy = confy.loader(__file__)

config = confy.merge(
confy.from_modules('base', os.environ.get('ENV', 'development')),
confy.from_modules('local', silent=True),

3) so in the end of the day you can simply import it easily and
everythink is setup for you

.. code-block:: python

>>> from settings import config
>>> # DONE!
>>> config.DB_USER # etc...

This is the basic stuff but "confy" has a lot more to offer.


1) I believe in clean and elegant solutions.

2) *.ini and "configparser"-like parsers (eg from pylons/pyramid) are
great but there are few issues with them::

=> Each setting has a type. And you have to write code that changes text into other types.

=> __getitem__ syntax is verbose, actually it's too verbose in some cases

You don't have this problem when you keep your config/settings
inside simple python files, however there is no standardized
one-and-only-one-good-way how to keep and load those kind of files.

3) Recently I've found interesting presentation
about configuration injected from environment variables. And I love
the idea (untill slide No 12), but it has the same issues as *ini

Configuration sources

With "confy" you can simply load the configuration from any source you
want, from any place of your system. That gives you flexibility on how
and where you want to keep configuration.

.. code-block:: python

import os
import confy

confy = confy.loader(__file__)

config = confy.merge(

# primarily read modules
confy.from_modules('base', os.environ.get('ENV', 'development')),
confy.from_modules('local', silent=True), # no complains if no ''

# if you'd like to setup things directly through your
# environment variables, here's how you can do that
], silent=True), # if environ vars are not set it's not a problem

# if need additional data from ini files - no problem
confy.from_ini('~/.project_sensitive_data.ini', silent=True),

# additional configuration might be kept under given directory
# in envdir-like manner (
# If settings/ directory contains "envvars" dir all the variables
# will be loaded into a configuration
confy.from_dirs('envvars', silent=True),

Configuration are loaded one after another, so please keep in mind
that variables might override each other.


Simple and easy::

$ pip install confy


* Jakub Janoszek (

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

confy-0.3.8.tar.gz (13.1 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page