Re: Probes

[Paul E Johnson]

"Marcus G. Daniels" wrote:
> 
> In Java, if you want to save a whole object at a timestep, use a
> scheduled "Globals.env.hdf5AppArchiver.putShallow$object (name,
obj);"
> call.  That will write entires to APP.hdf where APP is the name of
> your application.  Then you can get them back on another run with
> "obj = Globals.env.hdf5AppArchiver.getObject (name);", or load the
object
> value into the R statistical package.  Use putDeep instead of
putShallow
> if the object you're saving has other objects in it.  putShallow
has
> the advantage of saving in a more efficient format that loads in an
> intuitive table format in R.
> 

I've been meaning to work out an example of an app that outputs data
that I would then analyze in R.  I'm considering doing it to JHeatbugs. 
If I understand the drift of our conversations about ways to save data
in Swarm and hdf5, my plan should be to create a new class "Output" that
would be full of public instance variables that have the data I want
saved.   Then putShallow would write the numbers.  In the case of
Heatbugs, I think this means changing a bit about the way the Observer's
graph is created, and a "stand alone" averager object will be created,
which can both give data to the output class and also to the graph.  

If anybody has advice about how to pursue this, I will wait with open
ears.  If anybody  has a small working example, fork it over.
"Marcus G. Daniels" wrote:
> 
> >>>>> "FW" == Fred Wan <address@hidden> writes:
> 
> FW> How can this best be done with output from probes, i.e., instead
> FW> of feeding them into display objects, feeding them into other
> FW> (data-collection) objects or files?
> 
> In Java, if you want to save a whole object at a timestep, use a
> scheduled "Globals.env.hdf5AppArchiver.putShallow$object (name, obj);"
> call.  That will write entires to APP.hdf where APP is the name of
> your application.  Then you can get them back on another run with
> "obj = Globals.env.hdf5AppArchiver.getObject (name);", or load the object
> value into the R statistical package.  Use putDeep instead of putShallow
> if the object you're saving has other objects in it.  putShallow has
> the advantage of saving in a more efficient format that loads in an
> intuitive table format in R.
> 
> If you just want to save a single field of an object with a probe,
> then you can make a "value = messageProbe.doubleDynamicCallOn
> (target);" call where you get messageProbe from a
> "probeMap.getProbeForMessage (messageName);".  You'll have to save
> `value' using Java file interfaces, or some ad-hoc way.
> 
>                   ==================================
>    Swarm-Support is for discussion of the technical details of the day
>    to day usage of Swarm.  For list administration needs (esp.
>    [un]subscribing), please send a message to <address@hidden>
>    with "help" in the body of the message.

-- 
Paul E. Johnson                       email: address@hidden
Dept. of Political Science            http://lark.cc.ukans.edu/~pauljohn
University of Kansas                  Office: (785) 864-9086
Lawrence, Kansas 66045                FAX: (785) 864-5700


                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================