July 2015


Options: Use Monospaced Font
Show HTML Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Ernesto Carrella <[log in to unmask]>
Reply To:
MASON Multiagent Simulation Toolkit <[log in to unmask]>
Wed, 22 Jul 2015 08:43:25 +0000
text/plain (4036 bytes) , text/html (5 kB)
Depending on how much work you need MASON GUI to do, you could also ditch
display2D and draw using Jung directly. I always felt Jung 2.0 have some of
the most elegant API for drawing simple graphs (aging a bit now that Java 8
replaced most of the apache commons used by Jung, but still great).
Nice tutorial here:

If you have a clear sense of where each vertex lies in a 2d space, use a
StaticLayout in Jung where you pass a function Vertex ---> Point2D.
You can also mod the vertices to show what your portrayals are currently
You can then use pass another function Edge--->Stroke to color edges when
you want them highlighted or not.

What is left then is to integrate MASON and Jung and that's not too hard
either, you need to
1) Make Jung set inspectors to MASON controller [page 243 of MASON manual]
whenever something is clicked (through mouse plugins)
2) Make Jung graph update in step (calling repaint) by scheduling the
update function in the GUIState (calling scheduleImmediatelyAfter)

You could try something goulish where you paint jung graph through jung on
a transparent pane and then paste it on top of the display2D, but my guess
is that resizes and zooms are going to be a nightmare.

On Tue, Jul 21, 2015 at 5:25 PM Joey Harrison <[log in to unmask]> wrote:

> Hi Miriam,
> Here are a couple suggestions:
> 1) Since your graph doesn't change, store a JUNG version and a MASON
> Network version. Use the NetworkPortrayal2D to visualize it and do your
> shortest path calls on the JUNG version.
> 2) To visualize the shortest path, create a separate Network on a separate
> layer with all the same nodes, but only the edges contained in the shortest
> path. When you draw the edges for that layer, make them orange, wide, and
> mostly transparent and it will look like highlighting.
> Hope that helps,
> Joey
> On Tue, Jul 21, 2015 at 9:44 AM, Tschanen Miriam <[log in to unmask]>
> wrote:
>>  Dear all,
>> I am currently in the process of writing a custom portrayal class that
>> can display JUNG graphs and I have a couple of questions.
>> First off, I realise that there is an example of how to do this linked on
>> the MASON homepage, but that display class adds a lot of fluff that I do
>> not need. For example, I will not need an algorithm to plan the graph
>> layout, because all my nodes are already embedded in a 2D field before
>> creating the graph. I use the graph only to store weighted connections
>> between the nodes, and just need a portrayal class to draw those
>> connections. As such this is very similar to how the native Network class
>> works with NetwortPortrayal2D, but I would still like to use some of JUNG's
>> features, like its shortes path algorithms. So what I'm trying to do is to
>> write my own adaption of the NetwortPortrayal2D class that displays a JUNG
>> Graph instead of a Network.
>> While digging through the NetwortPortrayal2D code I noticed something
>> strange. I found this line in the hitOrDraw method:
>> if (edgemap.containsKey(edge)) continue;
>>                     edgemap.put(edge, edge);
>> I cannot for the live of me imagine why anyone would add an object to a
>> HashMap as both key *and *value. Why not just use a set? Is this an
>> oversight, or am I missing something crucial here?
>> Also, my graph is static and does not change at all during the
>> simulation, so there would be no reason to redraw it at every time step. I
>> found a flag called immutableField in the FieldPortrayal class, is setting
>> this value to true enough to prevent redrawing? Or is there something else
>> I need to do?
>> And finally, I would like to highlight certain edges during the
>> simulation, e.g. the shortest path between two nodes for any pair of nodes
>> selected by the user. What would be the easiest way to implement the
>> highlighting? I'm not looking for a complete example, just some pointers to
>> get me started. Can this be done without redrawing the entire graph?
>> Thanks in advance for any help.
>> Sincerely,
>> Miriam Tschanen