Unable to run visualization properly!

[PreetamKulkarni]

I had written my code a month ago and everything worked fine. I installed the latest version of Mesa 2.1.1 today and I have been facing issues with the visualization. I am not sure if installing the upgrade is what caused the issue. I tried to run the boid flockers example and it worked fine.

The error log I see in "spyder" is as follows:

WARNING:tornado.general:403 GET /local/SimpleCanvas//Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF%20CAREER%20-%20Project/Research%20Paper%20Submissions/WSC%202023/Code/Python%20MESA%20code/simple_continuous_canvas.js (127.0.0.1): /Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF CAREER - Project/Research Paper Submissions/WSC 2023/Code/Python MESA code/simple_continuous_canvas.js is not in root static directory
WARNING:tornado.access:403 GET /local/SimpleCanvas//Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF%20CAREER%20-%20Project/Research%20Paper%20Submissions/WSC%202023/Code/Python%20MESA%20code/simple_continuous_canvas.js (127.0.0.1) 15.32ms
Socket opened!
{"type":"reset"}
100

WARNING:tornado.general:403 GET /local/SimpleCanvas//Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF%20CAREER%20-%20Project/Research%20Paper%20Submissions/WSC%202023/Code/Python%20MESA%20code/simple_continuous_canvas.js (127.0.0.1): /Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF CAREER - Project/Research Paper Submissions/WSC 2023/Code/Python MESA code/simple_continuous_canvas.js is not in root static directory
WARNING:tornado.access:403 GET /local/SimpleCanvas//Users/preetamkulkarni/Library/CloudStorage/OneDrive-UniversityofTexasatArlington/NSF%20CAREER%20-%20Project/Research%20Paper%20Submissions/WSC%202023/Code/Python%20MESA%20code/simple_continuous_canvas.js (127.0.0.1) 15.32ms

[PreetamKulkarni]

I was able to resolve this by adding the proper address of the file "simple_continuous_canvas.js" in "SimpleContinuousModule.py". This was my mistake because I had moved files around and I forgot to update.

However, I do have this issue with the legends and colors not showing up on the screen correctly after I updated to Mesa 2.1.1.

Please help me resolve this.

Thanks in advance :)

[rht]

• Have you tried downgrading to 1.x, and confirmed the problem disappears?
• The error message indicates there is already an existing Tornado server running. Have you made sure your session is fresh and clean before retrying?

[PreetamKulkarni]

Hi,
I used "ctrl + ." before running code and this restarts the kernal. I also tried closing spyder IDE completely and restarted it. I also removed mesa 2.1.1 and installed 1.2.1 but it is still not showing the legend and the color on the graphs. I am not sure what is causing the issue. I checked the version and it shows I have downgraded to a lower version:

pip show mesa
Name: Mesa
Version: 1.2.1
Summary: Agent-based modeling (ABM) in Python 3+
Home-page: https://github.com/projectmesa/mesa
Author: Project Mesa Team
Author-email: [email protected]
License: Apache 2.0
Location: /Users/preetamkulkarni/opt/anaconda3/lib/python3.9/site-packages
Requires: pandas, tornado, tqdm, cookiecutter, networkx, click, numpy
Required-by: Mesa-Geo
Note: you may need to restart the kernel to use updated packages.

[rht]

Can you try running the server from a terminal instead of via Spyder? The Tornado viz framework is known to not work well with Jupyter, because Jupyter has Tornado. I suspect that Spyder has some sort of Jupyter stuff under the hood.

There is also the new Solara viz framework, but there are no visualizers yet for continuous space and geo stuff.

[PreetamKulkarni]

I tried to update Anaconda and Spyder and it fixed the issue. I am able to recreate the issue by upgrading to Mesa 2.1.1.
Everything works fine if I am on Mesa 1.2.1 and for now, I will continue to use Mesa 1.2.1.

[NickGotts]

I have been struggling with a very similar problem. I have been trying to port a continuous space model written under Mesa 0.8.8.1 to Mesa 2.1.1. The model runs (after small changes such as to account for an updated version of shapely), but the visualization does not work. Moreover, after an unsuccessful run, the boid_flockers example sometimes (but not always) behaves in the same way, and I have to close and reopen the browser (Firefox) to get it to do so. Other examples, using grid spaces, are unaffected even when boid_flockers is. I'm running the model from a powershell terminal of an anaconda environment using Python 3.8,, which is the same way I've been running the model throughout its development. Any suggestions, please?

[rht]

1. Do you have a minimal reproducer?
2. Can you post error messages in the console that you have encountered?

[rht]

Moreover, after an unsuccessful run, the boid_flockers example sometimes (but not always) behaves in the same way

If I run boid_flockers I should see the problem?

[NickGotts]

Thanks for getting back to me so promptly!

My "minimal reproducer" is an early version of my model, which is still not very minimal; I will try cutting it down further. The skeleton of it is based on boid_flockers itself, particularly with regard to the interaction with JavaScript (a language of which I know very little). No error messages appear on the console. However, if I start a Jupyter notebook from within anaconda, navigate to the folder with "run.py" in it and input "import run", I get the following:

Interface starting at http://127.0.0.1:8521/

---

RuntimeError Traceback (most recent call last)
Cell In[1], line 1
----> 1 import run

File ~\anaconda3\envs\mesaenv38new\mesa-examples\examples\distancing2csimp\run.py:3
1 from distancing2c.server import server # noqa
----> 3 server.launch(open_browser=True)

File ~\anaconda3\lib\site-packages\mesa\visualization\ModularVisualization.py:407, in ModularServer.launch(self, port, open_browser)
405 tornado.autoreload.start()
406 try:
--> 407 tornado.ioloop.IOLoop.current().start()
408 except KeyboardInterrupt:
409 tornado.ioloop.IOLoop.current().stop()

File ~\anaconda3\lib\site-packages\tornado\platform\asyncio.py:195, in BaseAsyncIOLoop.start(self)
194 def start(self) -> None:
--> 195 self.asyncio_loop.run_forever()

File ~\anaconda3\lib\asyncio\base_events.py:560, in BaseEventLoop.run_forever(self)
558 """Run until stop() is called."""
559 self._check_closed()
--> 560 self._check_running()
561 self._set_coroutine_origin_tracking(self._debug)
562 self._thread_id = threading.get_ident()

File ~\anaconda3\lib\asyncio\base_events.py:552, in BaseEventLoop._check_running(self)
550 def _check_running(self):
551 if self.is_running():
--> 552 raise RuntimeError('This event loop is already running')
553 if events._get_running_loop() is not None:
554 raise RuntimeError(
555 'Cannot run the event loop while another loop is running')

RuntimeError: This event loop is already running

Socket opened!
{"type":"reset"}

---

The model runs when started, so the problem seems to be closely tied to the visualization.

As for your query about boid_flockers, I don't know, but I would guess not, since boid_flockers usually works. I am guessing there is something about my model which causes the subsequent failure of the boid_flockers visualization, but what that could be, I don't know, and it does seem to be specific to continuous space, as other example models run correctly even after visualization fails for both my model and boid_flockers. I will see if I can reproduce the problem by running boid_flockers repeatedly, interrupting it, trying to start up a new run while the old one is paused, and so on.