rht

Results 789 comments of rht

ContinuousSpace in the Tornado viz just needs to incorporate https://github.com/projectmesa/mesa-examples/blob/main/examples/boid_flockers/boid_flockers/simple_continuous_canvas.js. ContinuousSpace in JupyterViz has already been implemented. A showcase of it: https://github.com/projectmesa/mesa-examples/pull/81.

Assuming that: 1. mesa-viz-tornado is on feature freeze 2. the JupyterViz, with an API that is similar to mesa-viz-tornado, is the official recommended way to do visualization then this issue...

Because this PR requires mandatory model `super().__init__`, which is an API breaking change. Mesa 2.2.x recommends having model `super().__init__()`, but is not necessary. I forgot that #2026 hasn't been released...

My "ChatGPT" summary of the discussion: IIUC, marking the attributes with `SystemState`, `AgentState`, `ModelState` turns them to be effectively immutable (and thus can be pickled), and in effect, they become...

> SystemState is like a view in NumPy on the state of the simulation. I build it in response to https://github.com/projectmesa/mesa/issues/574. I find this to deviate from the point of...

Revisiting this PR in the light of the measure and data collection divide (point 3 in https://github.com/projectmesa/mesa/discussions/1944#discussioncomment-8364109), to recycle the ideas. I find the API proposed here, in particular ```python...

> There is indeed a family resemblance between Measure and SystemState. That's a concise term for describing the situation. > All are contained in a single, rather complex, data structure....

The animation jitter could be fixed with a `plt.tight_layout()` equivalent of Altair. I just haven't figured it out yet.

Fixed for multigrid, where now you may see overlapping markers: ![2024-01-22T23:22:59,903471196-05:00](https://github.com/projectmesa/mesa/assets/395821/47a18664-3785-420c-ad77-addca1a50587)

> Since altair uses a declarative approach vs an imperative in matplotlib, I think its confusing to have different agent_portrayal functions. But would have to see the agent_portrayal. I need...