Skip to main content

Python bindings for Linknx

Project description

Copyright (C) 2012-2014 Cyrille Defranoux

Contact me at knx at aminate dot net

Get the source code on Github

Python Version Requirement

Pyknx version 1 was compatible with Python 2 only. It was not compatible with Python 3 at all. Pyknx version 2 is designed to work with a Python 3 environment. The main reason for this requirement is to benefit from unicode string enhancements in Python 3. Do not attempt to use it with Python 2 or you will likely get errors about ascii codec not being able to encode some characters.

Upgrade Notice

Version 2 of Pyknx consists in a rework of all standalone scripts to improve their usability. Pyknx has also been modified to embed acceptance tests. The overall API of the package has been slightly changed. Unless you were using the pyknx*.py standalone scripts of Pyknx version 1, upgrading to version 2 should be seamless. But please note that version 2 is still in beta phase and may thus not completely work as expected. Things should settle down quite soon but in the meantime, you may consider using the latest version of Pyknx 1. As usual, feedback regarding version 2 is greatly appreciated.

What is it?

Pyknx is a package that is aimed at providing basic functionality related to communicating with a Linknx instance. It should help in sending or receiving data to/from Linknx.

Once installed, you should be able to read or write your Linknx objects from your Python environment:

from pyknx import linknx
server = linknx.Linknx() # Connect to localhost:1028

# Print status of all lights. The regex assumes that all lights objects are prefixed with 'Lights'.
for lightObj in server.getObjects('Lights.*'):
        print '{0} is currently {1}'.format(, lightObj.value)

Pyknx allows to execute functions implemented in a python file of your own, with the script running as a daemon. This allows to easily store data from one execution to another. The example below demonstrates how one can implement a function that can be called each time an object changes in Linknx:

counters = {}
def onLightsChanged(context):
        global counters;
        if counters.has_key(context.objectId):
                counter[context.objectId] += 1
                counter[context.objectId] = 1

        print('Lights {0} have been switched {1} times.'.format(context.objectId, counter[context.objectId])

Refer to section How does it work? to know how to configure your environment to make such magic happen.

Why Pyknx?

There is no doubt that Linknx is a very powerful, stable and simple-to-configure solution. It is sufficient for most needs in the frame of home automation. Nevertheless, as a developer, I sometimes find frustrating not having the opportunity to replace a set of XML rules in my Linknx config by a piece of code…

I first wrote a simple python script called lwknxclient whose unique functionality is to read/write Linknx object’s values. This script solved the problem of easily sending data to Linknx from a bash script.

Then, my requirements evolved drastically as I wanted to implement my own alarm system to protect my home. I have a few door switches, cameras and smoke detectors that I wanted to use. I first implemented a simple version in pure XML that worked but it had many drawbacks:

  • the configuration is quite verbose. This is the very nature of XML. Factorization is hardly ever possible.
  • the configuration is tricky to test. I had to test it interactively after each modification. I quickly reached a point from which I was too afraid breaking something to add new functionality to my system.
  • I had to rely on bash scripts for each non-trivial action executed by Linknx, which led to a bunch of scripts disseminated to various places on my server. Difficult to maintain too… And, no offense, but I have to respectfully admit that bash is not the kind of language I am happy to work with.
  • calling external scripts from within shell-cmd actions has a major drawback: the script’s lifetime is equal to the action’s one. If the script has to retain some variables between two executions, it has no solution but polluting Linknx objects pool or storing data to files. None of these are convenient for a non-trivial application.

The answer to those problems was to implement a daemon in Python that would manage my alarm system. This time, I would be able to use the comprehensive Python framework, to factorize my code and to test it automatically!

How does it work?

Pyknx relies on the built-in ioport communication of Linknx. The principle is as following:

  • edit your linknx XML configuration to add a pyknxcallback attribute to each object for which you would like a python callback to be called whenever its value changes. The value of the attribute corresponds to the name of the function to call:

    <object gad="0/1/2" id="KitchenLights" type="1.001" pyknxcallback="onLightsChanged">Kitchen Lights</object>
  • use to automatically append to the linknx XML configuration rules that are required for the communication to work. These rules use ioport actions to send data to the Python daemon but you don’t have to mess with that if you are not willing to:

    $> -i linknx.xml -o patchedlinknx.xml

See –help for a complete list of options. This command will declare the ioport and a few rules that are necessary for the communication to work. In the output xml file, you should read:

<ioport host="" id="pyknxcommunicator" port="1029" type="tcp"/>
  • start Linknx with the above configuration:

    $>linknx --config=patchedlinknx.xml
  • start an instance of the communicator using The name of a file of your own that implements every function declared with pyknxcallback attributes shall be passed to the command line:

    $> -f

And that’s all. Every callback is passed a ‘Context’ instance that implements an ‘object’ property which can be used to identify the object that is the source of the event on Linknx’s side. Simply write ‘context.object.value’ to retrieve or change the value of the object.

Contents of the package

The archive comes with a package named pyknx that offers the following pure-python modules:

  • common module that implements the communication with a linknx server. With this module, one can retrieve linknx objects, read or write their value, read linknx configuration, …
  • this module contains the Communicator daemon, whose purpose is to receive events from linknx, through ioports. It is then easy to write callbacks to react to object modifications. Additional scripts based on pyknx are provided (see below) in order to make this bidirectional communication with linknx just a few keystrokes away from now!
  • internal module that provides logging functionality for the package.
  • an internal module that implements common functionality related to socket communication. The end-user is unlikely to use this module directly.

This package also provides additional python scripts that are intended to run as standalone executables. They are briefly explained in the sections above but the –help argument of each script should be enough to help you understand how it really works.

  • is used to automatically patch your linknx XML configuration in order to generate the ioport service and the rules necessary for the communication between Linknx and the Python daemon to work.
  • is the script that represents the daemon itself. Simply tell it where to find your user-defined python file with your implementation and it should work.
  • can be used to ask the daemon to perform a function call. For instance ‘ -amyArgument=2 myCallback’ should call the function myCallback(context) in your user-defined file. The passed context will contain a property named myArgument whose value is 2. This utility script is useful to help making external applications pass data to your daemon.
  • is the replacement for the former I initially wrote a few months ago. It is a lightweight client for linknx that can retrieve or change the value of one or multiple objects at once (multiple-object requests are available for read requests only).

How to install

Two standard ways: using pip or calling manually. With pip for Python 3 (, simply do:

pip3 install pyknx

You can optionally add –install-option=”–user” to tell to install in your home rather than one the system-wide locations.

The other way: uncompress archive. Then calling directly boils down to:

python install

You can optionally add –user to install in your home. Please refer to distutils documentation for further details.


Pyknx is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Pyknx is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

For the full version of the GNU General Public License, please refer to

Feel free to contact me by email (knx at aminate dot net) for feature requests or bug reports or simply to let me know whether you find this package useful or terribly useless! User feedback is never a waste of time.

Change Log


First version. This corresponds to a version that has been thoroughly tested with my own alarm system implemented in Python and with unit tests.

Until 1.0.8

Minor changes, mainly in documentation. A few bugfixes.


Added a “flags” member of pyknx.linknx.ObjectConfig, that contains the actual flags of the object as defined in the XML. If no attribute is set, this member is equal to ‘cwtu’. Changed behaviour for the “init” member: is now set to ‘request’ whenever the attribute is absent from XML. This is for consistency with respect to “flags”.


Fixed a bug in an import of pyknx.logger was missing and caused a crash when attempting to handle the exception thrown while waiting for incoming data.


Fixed a bug in Object.value for objects of type ‘float’. The returned value was a string (the raw value sent by linknx). It is now a float as expected.


Changed the encoding of the output file generated by to ISO-8859-1 (aka latin-1) because other encodings are not well supported by linknx. (WARNING: this change has been reverted in next version) Added a XML header that specifies the encoding used when sending a request to linknx. This encoding is <?xml version=”1.0” encoding=”ISO-8859-1”?>. If you encounter encoding problems, make sure that your linknx XML config file is encoded in either ascii or latin-1, not in utf-8 (for instance). I am currently working on a version 2 of pyknx that will rely on Python 3, which version of Python does no more suffer from the encoding issues of older versions.


Reverted changes regarding encoding implemented in 1.0.12. It appeared that it was not a good fix. Linknx had to handle utf8 request properly, so that clients can communicate with it in UTF8 too. Knxweb uses UTF8 and it was not an option to change that. A patch has been submitted to linknx’s maintainer to fix that. Before that patch is integrated, it is very likely that v1.0.13 of pyknx will work nicely with most configurations, because encoding issues appear in rare cases.


Fixed a minor bug in (which impacts passing a communicator address that specifies the host by a name rather than by IP address was leading to a configuration incompatible with linknx. Linknx uses pth_connect and this function does not support named hosts. Added a conversion with socket.gethostbyname() in order to be sure to write the host’s IP into the patched configuration.


Reworked standalone scripts (pyknxconf, pyknxclient, pyknxcall) to increase ease of use and consistency: - for instance, Pyknxclient can now read several object at once - replaced argument parsing previously done with getopt by argparse that appears to be more efficient. These breaking changes cause backward incompatibility for clients of these scripts. Please refer to the documentation of these script to learn more about their updated usage.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
pyknx-2.0.0b4.tar.gz (20.4 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page