Example models should call super in init

[drewrey]

Some example models are not calling super() in their respective __init__ methods.

[njvrzm]

One detriment of this is that random.seed() isn't getting called for these examples, so they're not deterministic. Another is that each subclass must remember to set self.running = True. On the other hand, it's not clear that Model.__init__ is the best place for these things to be happening - maybe it would be better for the base class to have no __init__ and to add, for instance, set_seed and set_running methods.

[dmasad]

The base classes are more there as placeholders, and the constructors aren't necessarily intended to be initialized via super() -- especially when in some cases (e.g. WolfSheep) there are multiple inheritances between the base class and the actual agent class. If we want to require agents to call super().__init__(), we need to make sure that it actually contains something useful.

At the moment, I also don't think the examples need to be deterministic.

[jeetu7]

This message is just to document the error message that the system will print once I run the web visualization and hit reset. (if i miss setting self.running = True in my model subclass)

  File "/home/user/learn/venvs/mesa3_venv/lib/python3.5/site-packages/tornado/gen.py", line 282, in wrapper
    yielded = next(result)
  File "/home/user/learn/venvs/mesa3_venv/lib/python3.5/site-packages/mesa/visualization/ModularVisualization.py", line 270, in run_model
    while self.model.schedule.steps < self.max_steps and self.model.running:
AttributeError: 'RoboModel' object has no attribute 'running'

[dmasad]

@jeetu7 That issue is fixed in master, so the fix will be included in the next release.

[jackiekazil]

And that release should have been released a few days ago.
Let me know if it still needs to be addressed.

On Tue, Jun 21, 2016 at 2:39 PM, David Masad notifications@github.com
wrote:

@jeetu7 https://github.com/jeetu7 That issue is fixed in master, so the
fix will be included in the next release.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#248 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAKLTr6WrIgEEbStYoTlztU7TKb1D15Lks5qOFoQgaJpZM4IuAh0
.

Jacqueline Kazil | @jackiekazil

[jeetu7]

Still one small confusion. Which one is the recommended thing to do now?

1. Call super in the subclassed model class
2. specify self.running = True (as evident from latest examples in repo)

[dmasad]

At the moment, 2.

On Thu, Jun 23, 2016, 08:16 jeetu7 notifications@github.com wrote:

Still one small confusion. Which one is the recommended thing to do now?

1. Call super in the subclassed model class
2. specify self.running = True (as evident from latest examples in repo)

—
You are receiving this because you commented.

Reply to this email directly, view it on GitHub
#248 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAkb8j1J2apiOQoUc6TwaYYOgSPtYWYQks5qOnkKgaJpZM4IuAh0
.

[joelostblom]

I just ran into the error message @jeetu7 reported when using the MoneyModel example from the tutorial on ReadTheDocs. I instlled the latest version of mesa from Github, but I still run into the same issue. Adding self.running = True under the MoneyModel __init__ method as in the updated Github MoneyModel example (and as mentioned above) solved the problem.

[jackiekazil]

Newer discussion is here-- #443. I am closing this convo.