VERSION | DOWNLOADS | TESTS | COVERAGE |
---|---|---|---|
Pluginloader is a library to allow an easy way to load plugins. They can be managed by interfaces or just method validators.
- Each plugin can be instanciated several times.
- Customizable filter to select if a class should be loaded as a plugin.
- Sandboxed: plugins cannot access the main program or other plugins by default, because they are loaded in isolated environments.
- Parameterizable context: Sometimes you need to pass some classes, functions or variables preloaded to the plugins. It is possible and easy.
Two options: to install it in your system/project:
pip install pluginloader
You can load all plugins in a file, just doing:
loader = PluginLoader()
plugins = loader.load_file('plugins.py')
With those simple lines you will have in the variable plugins
a dictionary with each class inside the plugins.py
file as key and a factory as value.
Let's see an example. Using the plugins.py
file:
class Foo(object):
pass
We can have an object of that class just with:
loader = PluginLoader()
plugins = loader.load_file('plugins.py')
instance1 = plugins['Foo']()
instance2 = plugins['Foo']()
Simple and easy.
This is a simple module with a simple API. It just contains one class, PluginLoader
, with these public methods:
Loads all plugins in a file.
Parameters:
filename
: File name to be loaded.onlyif
: Value or function that will be called with each class found. It will skip the plugin if this function returnsFalse
.context
: Dict with the context where the method should be loaded in. It usually will map a class, function or variable name to the class, function or value in the main program, so it can be used within the plugin.
Loads all plugins in a directory.
Parameters:
path
: Path where plugins are in.onlyif
: Value or function that will be called with each class found. It will skip the plugin if this function returnsFalse
.recursive
: Boolean value to allow recursive read of directories.context
: Dict with the context where the method should be loaded in.
Links will always be ignored.
The onlyif
functions have this format:
def condition(obj_name, class_name, file_name)
where:
- obj_name
is the name of the object. It can be a variable, function, class or instance.
- class_name
is the class of the object.
- file_name
is the file where the object has been declared.
Copyright (c) 2014 Miguel Ángel García (@magmax9).
Licensed under the MIT license.