A convenience class for providing default values for a django app setting.
This app provides a baseclass to easily realize AppSettings for any django project. The advantage of using an AppSettings class lies in the possibility for the programmer to assign default values for settings if the setting is not present in the main settings.py
- Django >= 1.6
- From the pip repository: pip install django-pluggableappsettings
- or directly from github: `pip install git+git://github.com/NB-Dev/django-pluggableappsettings.git
Create your AppSettings class in any of your project’s files. E.g. in ‘app_settings.py’.
Define your settings by setting the class attributes as one of the provided settings types
from django_pluggableappsettings import AppSettings, Setting class MyAppSettings(AppSettings): MY_SETTING = Setting('DEFAULT_VALUE')
Access the setting from anywhere:
from app_settings import MyAppSettings setting = MyAppSettings.MY_SETTING
Provided Setting Types
Different setting types are provided with the package:
Setting(default_value, setting_name, aliases)
The most basic setting that looks up the setting’s value from the settings.py usually the attribute name is used for the detection. If, however, the settings_name parameter is given, this name is used instead for the lookup in the settings.py. It simply returns the value from the settings.py or, if that is not set, the default value. If no default value is provided and the setting is not set in your settings.py, an AttributeError is thrown. Also a list of aliases can be passed to allow for multiple names of one setting (e.g. for backwards compatibility)
CalledOnceSetting(default_value, setting_name, aliases, force_callable=False)
Checks whether the value is callable and calls it once before returning. Subsequent accesses to this setting return the cached return value of the first call. If force_callable is True, the setting throws a ValueError if the value of the setting is not callable.
CalledEachTimeSetting(default_value, setting_name, aliases, force_callable=False)
Checks whether the value is callable. If so, the callable is called each time when the setting is accessed. If force_callable is True, the setting throws a ValueError if the value of the setting is not callable.
ClassSetting(default_value, setting_name, aliases)
Behaves as a Setting but accepts only Classes or dotted paths to classes as values. If the value is a dotted path, the path is translated to a class before returning, so the returned value is always a class.
IntSetting(default_value, setting_name, aliases)
Accepts only values that are of type int or can be casted to type int
FloatSetting(default_value, setting_name, aliases)
Accepts only values of type float of values that can be casted to type float
StringSetting(default_value, setting_name, aliases)
Accepts only strings as value
IterableSetting(default_value, setting_name, aliases)
Makes sure that the value is an iterable
TypedSetting(default_value, setting_name, aliases)
A class that checks whether the given value is of a certain type and optionally allows casting the value to that type. Used as a base class for all type checking classes and can be easily subclassed to allow checking of various value types.
To create your own type checking setting simply subclass this type and set the class attributes _setting_type and _cast_value according to your needs. The _setting_type attribute specifies the desired type while the _cast_value attribute specifies whether the value should be casted to the _setting_type. A _cast_value of True essentially results in a call of value = _setting_type(value).
E.g. The IntSetting is defined as follows:
class IntSetting(TypedSetting): """ An integer setting """ _setting_type = int _cast_value = True
If you need more elaborate casting functions, you can overwrite the cast_value(self, value) function of your type which should return the casted value.
You can access any setting by simply importing your AppSettings class and accessing the corresponding attribute.
Tests with AppSettings
The package provides a convenient override_appsettings decorator / context manager to allow for the temporary override of AppSettings values. It is used just like Django’s override_settings decorator but with an extra argument: The AppSettings-Class that is to be altered has to be passed in as first argument. Following should be keyword, value arguments where the keyword is the name of the setting to be overridden and the value is the desired return value.
from django_pluggableappsettings.test.utils import override_appsettings from myapp.appsettings import MyAppSettings class SomeTestCase(TestCase): @override_appsettings(MyAppSettings, SETTING='new_value') def test_decorated(self): MyAppSettings.SETTING # This returns 'new_value' def test_context_manager(self): with override_appsettings(MyAppSettings, SETTING='new_value'): MyAppSettings.SETTING # This returns 'new_value'
Running the tests
The included tests can be run standalone by running the tests/runtests.py script. You need to have Django and mock installed for them to run. If you also want to run coverage, you need to install it before running the tests
v. 1.1.6 (2017-05-19)
- Fixing the README to be correctly displayed on pypi
v. 1.1.5 (2017-05-19)
- Version bump as I forgot to convert the readme. Added a publish.py to automate publishing in future.
v. 1.1.4 (2017-05-19)
- Adding tests for Django 1.10 and 1.11.
v. 1.1.3 (2016-01-27)
- Adding the possibility to look for a settings value under a different name in the settings.py by usage of the settings_name parameter
- Fixing a bug that caused all AppSettings instances to share the same cache of loaded settings. This could cause the settings to be overridden by other settings
v. 1.1.2 (2016-01-15)
- Adding an override_appsettings decorator / context manager to allow the overriding of AppSettings values in test
- Added the possibility to retrieve non-Setting attributes from the AppSettings class to allow for custom attributes or custom functions.
- I screwed up with pypi and need to bump the version number - Sorry
- Changing structure of Setting class to being able to add repeatedly called functions as setting.
Warning: This breaks compatibility of custom settings classes. To fix this, simply rename the get method of your custom classes to _get - Added a CalledEachTimeSetting that takes a callable that is called each time the setting’s value is accessed - Renamed the CallableSetting to CalledOnceSetting to make the differentiation to the CalledEachTimeSetting clearer. The old name will stay as an alias for now. - The CalledEachTimeSetting and the CalledOnceSetting take an force_callable kwarg to set whether the value of the setting is enforced to be callable or not.
- Releasing first stable version
- Added ‘aliases’ parameter to Setting to allow multiple names for one setting (e.g. for backwards compatibility)
- Extended code to also work with Python 3
- Added TypedSetting Setting type which allows for the setting to be typechecked
- Added IntSetting, FloatSetting, StringSetting and ÌterableSetting` as subtypes of `TypedSetting
- Added the changelog
- Redesign of settings to allow different types of settings that can now also provide type checking.
- Settings are now explicitly defined and no _DEFAULT_ prefix is needed anymore
- Also no staticmethod decorator is needed anymore
- Allow the easy definition of multiple allowed setting types so that a setting could e.g. accept either string or an Integer
- Allow the chaining of callables with typed settings to check that the return value of a callable is of the correct type
This Project is maintaned by Northbridge Development Konrad & Schneider GbR Softwareentwicklung