PlugyPy 1.2.2

A lightweigh Python plugin system

PlugyPy

Installation

You can install the plugin system from PyPi with:

pip install PlugyPy

Or locally by cloning the repository:

git clone https://github.com/not-so-cool-anymore/plugypy.git

and then running the setup file in the main directory with:

pip install .

Writing a configuration file

A PlugyPy configuration (or config) file is a JSON file that contains a deserialized representation of a PlugyPy Configuration object. The Configuration object consists of a boolean named will_load_all and a list named plugins. plugins is a list of PlugyPy Plugin objects will_load_all determines whether all of the plugins in the plugins directory will be loaded or only the enabled ones.

An example of a config file is:

{
    "will_load_all": true,
    "plugins": [
        {
            "name": "example_plugin_name_0",
            "is_enabled": true
        },
        {
            "name": "example_plugin_name_1",
            "is_enabled": false
        },
        {
            "name": "example_plugin_name_2",
            "is_enabled": true
        }
    ] 
}

Where name is the name of the plugin file without the .py file extension, and enabled is the boolean variable that indicates whether a plugin will be loaded (when true) or not (when false).

Usage

Importing the plugin system:

import plugypy

Deserializing JSON configuration file into a Configuration object.

configuration_deserializer = ConfigurationDeserializer('/path/to/configuration/file')
configuration = configuration_deserializer.deserialize_config()

Creating a plugin manager object:

plugin_manager = plugypy.PluginManager('/path/to/plugins/directory', configuration)

The plugin manager object has one extra feature - plugin ownership verification. This feature ensures that the plugin that is being executed belongs to the current user (or if sudo is used to run the program - the sudo caller). This feature can be activated via parsing one extra argument - will_verify_ownership=True, which is set to False by default when not passed.

Discovering plugins:

discovered_plugins = plugin_manager.discover_plugins()

discover_plugins is a list of all the plugins in a given plugins directory.

Importing plugins:

plugins_list = plugin_manager.import_plugins(discovered_plugins)

plugins_list is a list of map objects that has a name key and an object value of the imported executable Python module (plugin executable).

Importing a singe plugin by file name:

single_plugin = plugin_manager.import_plugin('PLUGIN_FILENAME_WITHOUT_PY_EXTENSION')

In this case, the plugin will be imported no matter if a configuration for it exists. This importing method is developed for edge cases in which the imported plugin will be executed only once.

Getting a plugin's information:

plugin_name = plugins_list[n]['name']
plugin_executable_object = plugins_list[n]['object']

n is an index of a plugin.

Executing a plugin's function with no parameters:

plugin = plugins_list[n]
plugin_result = plugin_manager.execute_plugin_function(plugin, 'function_name')

if plugin_result == None:
    print('The plugin returned no result')
else:
    print('The plugin returned: {}'.format(result))

n is an index of a plugin.

Executing a plugin's function with parameters:

plugin = plugins_list[n]

arguments_tuple = ('arg1', 'arg2', 'arg3')
plugin_result = plugin_manager.execute_plugin_function(plugin, 'function_name', args=arguments_tuple)

if plugin_result == None:
    print('The plugin returned no result')
else:
    print('The plugin returned: {}'.format(result))

n is an index of a plugin.

args must be a tuple of argument/s.