Fix batch_run data collection on final step

[satyamlal]

Summary

This is a bug where the function of batch_run, in the Mesa framework, is not collecting data on the last step of the simulation. Such an issue is affecting users because they need accurate data collection at the end of the simulation step.

Bug / Issue

The issue is tracked under#2514. Within the batch_run function, DataCollector wasn't collecting information on final step because of incorrect step range logic. The missing data had an appearance in simulations that were more than one step.

Implementation

To resolve this issue, the logic for generating the list of steps for data collection was modified. Specifically:

The line generating steps was changed from: file path: (..mesa\batchrunner.py)

    steps = list(range(0, model.steps, data_collection_period))
    if not steps or steps[-1] != model.steps - 1:
        steps.append(model.steps - 1)

to:

    # steps for data collection
    steps = list(range(0, model.steps, data_collection_period))
    
    # final step for data collection
    if steps and steps[-1] != model.steps:
        steps.append(model.steps)

Testing

1. I created a simple model with only a few steps to make it easier to track the data.
2. Ran it using the original code and saw that the last step’s data was missing.
3. Ran it again with the fixed version and confirmed that the data for the last step was included this time.
4. Compared the outputs to ensure the fix didn’t break anything else.
   Everything looks good, and the results are consistent now!

Additional Notes

While working on this issue, I consistently encountered a warning related to the following line:

from tqdm.auto import tqdm

Despite my efforts, I wasn’t able to identify a solution for this warning. If anyone has insights or suggestions on resolving this, I would greatly appreciate your guidance.

[github-actions]

Performance benchmarks:

Model	Size	Init time [95% CI]	Run time [95% CI]
BoltzmannWealth	small	🔴 +16.5% [+15.1%, +17.8%]	🔵 +2.2% [+1.6%, +2.7%]
BoltzmannWealth	large	🔵 +2.2% [+0.9%, +4.3%]	🔴 +4.7% [+3.3%, +6.1%]
Schelling	small	🔵 -0.1% [-0.5%, +0.3%]	🔵 +0.4% [+0.3%, +0.6%]
Schelling	large	🔵 +0.3% [-0.1%, +0.9%]	🔵 +0.5% [-1.3%, +2.3%]
WolfSheep	small	🔵 -1.1% [-1.8%, -0.3%]	🔵 +1.7% [-2.9%, +6.5%]
WolfSheep	large	🔵 -1.1% [-2.1%, -0.2%]	🔵 -2.0% [-4.7%, +0.7%]
BoidFlockers	small	🔵 +0.9% [+0.2%, +1.7%]	🔵 +0.0% [-0.7%, +0.9%]
BoidFlockers	large	🔵 +0.9% [+0.3%, +1.6%]	🔵 -0.1% [-0.7%, +0.5%]

[satyamlal]

I noticed that some checks are failing in the CI pipeline. I would appreciate it if you could review my PR and guide me on resolving these issues. I’m eager to learn and improve based on your feedback.

[quaquel]

As can be seen by clicking on details for any of the failed tests, the batch run related tests are failing. This makes sense because this PR changes the behavior. Figure out where in the tests it fails and update it for the correct behavior. Probably there is a check (i.e., some assert statement) somewhere on the number of steps for which data is collected.

[satyamlal]

Thanks for guiding @quaquel . I am working on it to fix the issue.

[EwoutH]

Let us know if you need help or are stuck on something!

[tpike3]

@satyamlal Where you still planning on completing this?

[satyamlal]

@tpike3 I did try to solve the issue sir. But I am unable to fix the and commit my changes, as my code is facing test case issues. So for now I won't be able to work on this.

[quaquel]

I have looked into this and tried to resolve the issues, but I am failing. Part of the problem seems to be that #2514 is not fully diagnosing the issue. Some of it might be due to a mismatch between 0 and 1 based indexing.

I am closing this PR.