Skip to main content

Managing a config globally throughout a Python application

Project description


Managing a config globally throughout a Python application


With gconf, yaml configuration files can be loaded on module-scope such that its values can be accessed from anywhere in a running application.


Loading yaml files

One or more yaml configuration files are loaded globally from path-like objects, usually in a program's main function. The resulting configuration is an overlay of all loaded configurations with those that were loaded later taking precedence.

Loading a single config file:


Loading multiple config files at once:

gconf.load('conf1.yml', 'conf2.yml')

Loading multiple config files one after the other:


The two examples above produce an identical result. In both cases config2 takes precedence over config1 and overrides its values.

Each invocation of load returns a list of all paths that have actually been loaded.

Loading the first config from a list of paths:

gconf.load_first('conf1.yml', 'conf2.yml')

If conf1.yml exists, it is loaded and conf2.yml is ignored. if conf1.yml does not exist, conf2.yml is loaded.

load_first returns the path that has actually been loaded.

All loading functions raise a FileNotFoundError if no file is found. This can be prevented by setting the keyword-argument required=False.

Manually adding values

A dict can be added to the config from within the application. This is equivalent to loading a config file with the same content and overrides values if they are already present or adds them if not.

gconf.add({'parent': {'child': 'new child'}})

Warning: using this functionality, it is possible to the gconf module as a store for global variables. Global variables are a code smell and should not be used! Please use the gconf.add() function only if you know exactly what you are doing.

Accessing config values

There are several ways of addressing the values that are stored in the config, all of them using the module-level get method.


gconf.get('parent.child.some value')

String arguments

gconf.get('parent', 'child', 'some value')

A mix of both

gconf.get('parent.child', 'some value')

Top-level item as dictionary

gconf.get()['parent']['child']['some value']

Some intermediate item as dictionary

gconf.get('parent')['child']['some value']

To access list items, simply use their index in the path


Default value

The get method accepts a default argument, which is returned if the specified item does not exist:

gconf.get('non-existing', default=default_value)

This even works for falsey values like False, None or the empty string.

Environment Variables Override

Config values are overridden by environment variables. As keys of environment variables are uppercase with underscores by convention, gconf expects them in that way. The conversion rules are:

  1. GCONF_ prefix is added to avoid conflicts
  2. hierarchy levels are separated by underscore
  3. spaces are replaced by underscore
  4. all text is in all-caps


gconf.get environment var
parent.some child GCONF_PARENT_SOME_CHILD
list.0.entry GCONF_LIST_0_ENTRY


If an attempt is made at loading non-existing files and required=True (the default), an FileNotFoundError is raised.

If no default is provided, an attempt to access a non-existing item raises a KeyError.

Temporary override

Parts of the config can be temporarily overridden through a context manager. Pass it a dictionary that overlays the existing one:

with gconf.override_conf({'parent': {'child': 'override_value'}}):
    gconf.get('parent.child')  # => 'override_value'
gconf.get('parent.child')  # => 'original_value'

To temporarily remove parts of the config the DELETED constant can be used:

import gconf

with gconf.override_conf({'parent': {'child': gconf.DELETED}}):
    gconf.get('parent.child')  # => KeyError
gconf.get('parent.child')  # => 'original_value'


The global gconf dict can be completely reset.


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

gconf-0.8.2.tar.gz (17.9 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