python wrapper for RTI DDS

pip install dds==1.0


pyDDS (a Python 2.7 wrapper for RTI DDS.)


To use this wrapper, you must compile your generated topics (the output of rtiddsgen) into a shared library. This will be a .so on Linux, .DLL on windows and .dylib on OS X. The location of this library and the RTI libraries nddsc and nddscore must be in the library search path (LD_LIBRARY_PATH on Linux, PATH on windows, and DYLD_LIBRARY_PATH on OS X).


Lets say you have a topic that looks like:

#ifndef MY_DDS
#define MY_DDS

module my {
    module dds {
        enum my_enum {

        struct my_custom_topic
            string   name; //@key
            string   value;
            my_enum  mode;

#endif // MY_DDS

Once this topic is generated (rtiddsgen), compiled, and packaged as a shared library, you can use it with this library.


Lets say you just want to print out a representation of the topic data. Lets also say that your topic library is called and the location of this library is in LD_LIBRARY_PATH. To subscribe:

import dds
import json
import time

def print_repr(data):
    print json.dumps(data, indent=4)

dds_instance = dds.DDS('my_topics')
topic = dds_instance.get_topic('')

while True:

That's all there is to basic topic subscription! All data samples are delivered to the callback as a python dictionary. Also of note: the callback will occur in a separate thread, so you must take that into consideration for any non thread-safe operations.

If desired, you can also specify a few other options:

  • instance revoked A publisher can revoke a topic instance. To be notified of these events, specify a callback with the keyword argument instance_revoked_cb
  • liveliness lost If a publisher goes down, or the network connection fails, DDS can notify subscribers that a publisher seems to have gone down. To be notified of this condition, specify the keyword argument liveliness_lost_cb
  • content filtering Sometimes you are only interested in topic data that matches some condition. Instead of examining all the topic data samples and throwing out the ones you don't want, you can tell DDS to only send you samples that match a specified filter. For example, say you only want samples that have mode == mode_2. This can be accomplished by specifying the keyword argument filter_expression="mode MATCH 'mode_3'". See the docs for more details.

Subscriptions can also be canceled by calling topic.unsubscribe()


To publish a data sample, you simply construct a python dictionary that matches the structure of the topic. e.g.:

import dds

dds_instance = dds.DDS('my_topics')
topic = dds_instance.get_topic('')

sample = {
    'name': 'my key name',
    'value': 'whatever value you want',
    'mode': 'mode_2'


The dictionary does not have to be complete. i.e. it can be sparsely specified.


sample = {
    'name': 'my key name',
    'mode': 'mode_2'

is a valid topic sample that can be successfully published. If topic fields are not specified, they will be filled in with default values depending on the type. Strings will be populated with the empty string, number types will get zero, enums will get the first enum value, etc.

A publisher can also 'revoke' a topic. If a topic has keyed fields (the // @key decoration in the IDL) then there can be multiple instances of the topic on the DDS but simultaneously. To revoke a particular instance, call topic.dispose(sample) where sample has the keyed fields specified to match the topic instance you wish to revoke.

For more detailed documentation, see the inline docs in