-
I wanted to create a model with a property_layer, and I set agent_portryal to None as I was only interested in seeing how the property layer would change and not in the actual location of the agents themselves. However, the agents were still portrayed but using the default portrayal. _scatter(ax, arguments, **kwargs) After if(agent_portrayal!=None):
_scatter(ax, arguments, **kwargs) I didn't post this in bugs as I wasn't entirely sure if it was one... I think I misunderstood the functionality of None and reverting to the default portrayal is the intended behavior, but if that is the case, I think it would be useful to have the option of not representing the agents at all as well... |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
This is related to #2638 and #2640. I need to think about the best way forward to address this. Yes, we can add a bunch of if statements as per #2640. However, I personally am inclined to see this as an architectural problem. At the moment, we don't distinguish carefully between drawing the space (e.g., the grid structure), the drawing of properties of the space (i.e., property layer portrayal), and the drawing of agents within the space. These are effectively all collapsed into a single |
Beta Was this translation helpful? Give feedback.
This is related to #2638 and #2640.
I need to think about the best way forward to address this. Yes, we can add a bunch of if statements as per #2640. However, I personally am inclined to see this as an architectural problem. At the moment, we don't distinguish carefully between drawing the space (e.g., the grid structure), the drawing of properties of the space (i.e., property layer portrayal), and the drawing of agents within the space. These are effectively all collapsed into a single
draw_space
method. I am inclined to redesign the API and explicitly separate these three things slightly. If you pass the same axes (in the case of matplotlib), you can draw them on top of each other, but…