[ros-users] rviz is moving to Qt, preview available

Ken Conley kwc at willowgarage.com
Wed Jan 11 19:58:48 UTC 2012


On Wed, Jan 11, 2012 at 8:32 AM, Ye Cheng <cnhzcy14 at gmail.com> wrote:

> Hi guys,
>
> Thank you very much for all the reply! They are very helpful!
>
>  I have taken some time to merge the matplotlib into ros-gui for
>>> real-time  scalar data visualization. This is actually a simple version of
>>> original rxplot. I think the plot quality is better than qwt:
>>>
>>> https://plus.google.com/u/0/photos/110758615329253997840/albums/5687860553056164161/5696030098198932738
>>>
>>
>> The image looks great. Is the repository for this plugin available?
>>
>>
>>
>>   And I'll soon implement rxbag in ros-gui.
>>>
>>
>> This would be a great plugin for ROS GUI.
>> It would be one of the most important rx-tools yet missing in ROS GUI.
>> May be it would be good to put this directly into the "ros_gui" stack?
>>
>
> I'd like to merge all my code in ros_gui stack. But what shall I do?
>
>  But before going further, I have some question:
>>>>
>>>>  1. It looks like the rviz has it's own dock widget plugin
>>>> architecture. Is that an overlap between rviz and ros-gui if I make an rviz
>>>> plugin for ros-gui?
>>>>
>>>
>>> Yes, rviz uses it's own plugin architecture to be independent of ROS GUI.
>>> Actually there is already a rviz plugin for ROS GUI ("rosgui_rviz") -
>>> please take a look at the "ros_gui" stack and its packages:
>>> https://kforge.ros.org/visualization/ros_gui?p=ros_gui;a=tree
>>>
>>
>> In future we may want to make the plugins compatible between ROS_GUI and
>> RViz, I don't know.  For now, if you want to be able to use your code in
>> either, maybe you can keep most of your code in an internal class which is
>> the same for both frameworks and just write small wrapper classes for RViz
>> and for ROS_GUI.
>>
>> RViz and ROS_GUI both use a QMainWindow and QDockWidgets to house their
>> widgets, so I would think that widget code written for one could be used
>> for the other as well.  Though ROS_GUI uses a Python QMainWindow and RViz
>> uses a C++ one.  Interesting.
>>
>>
>>>   2. Is the rviz property display widget available for ros-gui?
>>>>
>>>
>>> I don't know - but the existing plugin "rosgui_rviz" runs a full
>>> instance of rviz inside a ROS GUI dock widget.
>>
>>
>> It is not currently available on its own in ROS_GUI, but I believe it
>> could be integrated.  I plan to publish the API for RViz's property editor
>> widget.  You can see a C++ example of how to use the property editor in
>> visualization/rviz/src/test/property_test.cpp.  You'll need checkouts from
>> trunk of visualization and visualization_common, build both, then "roscd
>> rviz && make test" to build it.
>>
>> I would hope the property editor widget could be used to edit properties
>> of lots of things in ROS_GUI, but I haven't tried it.  Also, I haven't set
>> up the Python bindings to happen for the property editor code yet, I don't
>> know if that will be hard or easy.
>>
>
> This is very useful! But what if both two ros_gui plugins have a same
> widget? Apparently loading a same widget twice is not efficient. At the
> moment I don't think there is a communication scheme between plugins in
> ros_gui. Am I right?
>
>
>  And I'll soon implement rxbag in ros-gui.
>>
>> Good stuff!  I designed rxbag in such a way that it could be included in
>> a larger system, so hopefully that integration should be straightforward.
>>
>> What I really need at the moment is something like rxlab<http://www.ros.org/wiki/rxlab>which is mentioned
>>> here <http://www.ros.org/reps/rep-0101.html>. Is the development of
>>> this package still active?
>>>
>>
>> No, it is not.  rxbag is already implemented as a "unit" in rxlab, so you
>> can look at the source to get a feel for how to integrate the main timeline
>> component.  Another system worth exploring is Ecto<http://ecto.willowgarage.com/>,
>> which is a very well designed tool for quickly and efficiently prototyping
>> and visualizing networks of computation.
>>
>> If you have technical questions regarding either rxbag or rxlab, feel
>> free to email me.
>>
>
> Ecto is new to me, I will take look at it. Actually I'm very interesting
> in the rosh. The ipython integration can help me to analyze the real-time
> data by Python scripting. Maybe a rosh plugin for ros_gui is valuable?
>

Let me know if you need assistance on the rosh side.  rosh development goes
in fits and starts depending on when I have time.  It would help to have
more requirements to drive it.

 - Ken


>
> Cheers,
> Ye
>
>
> On Wed, Jan 11, 2012 at 4:28 AM, Tim Field <tfield at willowgarage.com>wrote:
>
>> On Tue, Jan 10, 2012 at 8:25 AM, Ye Cheng <cnhzcy14 at gmail.com> wrote:
>>
>>>
>>> And I'll soon implement rxbag in ros-gui.
>>>
>>
>> Good stuff!  I designed rxbag in such a way that it could be included in
>> a larger system, so hopefully that integration should be straightforward.
>>
>> What I really need at the moment is something like rxlab<http://www.ros.org/wiki/rxlab>which is mentioned
>>> here <http://www.ros.org/reps/rep-0101.html>. Is the development of
>>> this package still active?
>>>
>>
>> No, it is not.  rxbag is already implemented as a "unit" in rxlab, so you
>> can look at the source to get a feel for how to integrate the main timeline
>> component.  Another system worth exploring is Ecto<http://ecto.willowgarage.com>,
>> which is a very well designed tool for quickly and efficiently prototyping
>> and visualizing networks of computation.
>>
>> If you have technical questions regarding either rxbag or rxlab, feel
>> free to email me.
>>
>> Tim
>>
>>
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
>>
>>
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20120111/8d2c1c47/attachment-0004.html>


More information about the ros-users mailing list