Elektra  0.8.19
Plugin: python

The plugin uses Python to do magic things. It basically allows to call plugins written in Python.

What a Python script can do is not really limited by design, so any kind of plugin may be implemented. The python plugin is especially useful to write filter and logging scripts.

Usage

The python plugin requires the configuration parameter script holding the file path to a python script. The mount command would look like

kdb mount file.ini /python python script=/path/to/filter_script.py

if the ini plugin should be used for storage and the python plugin only serves to invoke the filter script.

For a Python script that serves as (json) storage plugin itself, one could also use

kdb mount file.json /python python script=/path/to/json_plugin.py

Plugin Configuration

The python plugin supports following optional configuration values/flags:

Python Scripts

Python scripts must implement a class called ElektraPlugin with one parameter. The class itself can implement the following functions

where config & returned are KeySets and errorKey & parentKey are Keys. For the return codes of the functions, the same rules as for normal plugins apply.

If a function is not available, it simply is not called. A script does not have to implement all functions therefore.

Access to kdb can be retrieved using the Python import

import kdb

Example

An example script that prints some information for each method call would be:

class ElektraPlugin(object):
    def open(self, config, errorKey):
        print("Python script method 'open' called")
        return 0

    def get(self, returned, parentKey):
        print("Python script method 'get' called")
        return 1

    def set(self, returned, parentKey):
        print("Python script method 'set' called")
        return 1

    def error(self, returned, parentKey):
        print("Python script method 'error' called")
        return 1

    def close(self, errorKey):
        print("Python script method 'close' called")
        return 0

Further examples can be found in the python directory.

Disclaimer

Be aware that a Python script will never be as performant as a native C/C++ plugin. Spinning up the interpreter takes additional time and resources. The python plugin does also not clean up resources and is not restartable.