Hi all,

I've come up with a third option that I'm experimenting with at the moment.

I've made a new class (e.g. MyGeometry) that extends MasonGeometry and maintains a reference to an underlying object (e.g. an agent). It seems to be working OK so far, although after reading a shapefile into a GeomVectorField I have to go through and cast all the MasonGeometry objects to MyGeometry objects. This only needs to happen once at the beginning though so doesn't have a large overhead.

Let me know if there are any obvious pitfalls that I'm missing. If anyone is interested I'll post the code.

Cheers,
Nick

On Wed, 23 Mar 2016 04:52:00 -0400, Nick Malleson <[log in to unmask]> wrote:

>Thanks Luís and Ernesto, that's really helpful.
>
>I imagine that extending MasonGeometry will have a lower overhead than maintaining a map of Geometry->Agent.
>
>Ernesto, what are the implications of having to make Agent serializable? MasonGeometry implements Serializable, I can continue to implement Steppable.
>
>Thanks in advance,
>Nick
>
>
>On Tue, 22 Mar 2016 14:57:22 +0000, Ernesto Carrella <[log in to unmask]> wrote:
>
>>The two easiest solutions are probably to either keep a map
>>Geometry-->Agent somewhere  or to subclass MasonGeometry to add a field
>>linking it back to the Agent.
>>First is probably better since you don't have to worry about making Agent
>>serializable as well (although you might not care)
>>
>>On Tue, Mar 22, 2016 at 1:56 PM Nick Malleson <[log in to unmask]>
>>wrote:
>>
>>> Hi all,
>>>
>>> This might have an easy answer, apologies if I have missed something on
>>> the list.
>>>
>>> I am using a GeomVectorField to store my agents. I need to be able to get
>>> back to the original Agent object, not just its associated MasonGeometry. I
>>> create agents and add them to the GeomVectorField a bit like this:
>>>
>>> Agent a = new Agent();
>>> agentGeomVectorField.addGeometry(a.getGeometry());
>>> schedule.scheduleRepeating(a);
>>>
>>> That basically works fine, but now I want to be able to get back to the
>>> underlying agent. I guess I could maintain a separate List of agents as
>>> well as the GeomVectorField, or I could add the agent to its
>>> MasonGeometry's userData field.
>>>
>>> Any opinions on which would be best? How else have people stored their
>>> agents and their geometries using geomason?
>>>
>>> Thanks in advance,
>>> Nick
>>>
>>