fix: migrate codebase to Mesa 4.x API (resolves mesa.space removal)

[abhinavk0220]

Closes #152

Summary

Mesa 4.x removed mesa.space entirely. This PR migrates the full codebase to the new Mesa 4.x APIs so that mesa-llm works with the current version of Mesa.

Changes

• Replace MultiGrid/SingleGrid with OrthogonalMooreGrid from mesa.discrete_space
• Replace old ContinuousSpace(x_max, y_max) with new dimensions-based API
• Fix agent placement to use cell-based API (agent.cell = grid._cells[pos])
• Fix neighbor lookup using get_neighborhood() + coordinate set matching
• Fix model.steps → model.step (renamed in Mesa 4.x)
• Fix Model.__init__(seed=...) kwarg incompatibility with mesa_signals

Testing

All 203 tests passing after migration:

Notes

Tested against Mesa 4.x dev installed from source.

---

GSoC contributor checklist

Context & motivation

Mesa 4.x removed mesa.space entirely with no deprecation warning
any project depending on it simply breaks at import time. I discovered
this while trying to run mesa-llm examples and getting immediate
ImportError. This PR is the full migration of the codebase to the
new Mesa 4.x spatial API.

What I learned

Mesa 4.x's new discrete space architecture is fundamentally different
from the old API not just renamed imports but a completely different
cell-based model where agents live in cells rather than coordinates.
Understanding this required reading Mesa's source code directly, not
just the migration guide. The most subtle change was the coordinate
convention difference between OrthogonalMooreGrid._cells (x,y) and
the internal connections dict (row,col) a distinction that caused
agents to move in the wrong direction and required a separate fix in
PR #195.

Learning repo

🔗 My learning repo: https://github.com/abhinavk0220/GSoC-learning-space
🔗 Relevant model(s): https://github.com/abhinavk0220/GSoC-learning-space/tree/main/models

Readiness checks

• This PR addresses an agreed-upon problem (linked issue or discussion with maintainer approval), or is a small/trivial fix
• I have read the contributing guide and deprecation policy
• I have performed a self-review: I reviewed my own PR as if I were a reviewer and left comments on anything that needs explanation
• Another GSoC contributor has reviewed this PR:
• Tests pass locally (`pytest --cov=mesa tests/)
• Code is formatted (ruff check . --fix)
• If applicable: documentation, examples, and/or migration guide are updated

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 53104782-8bd4-4f75-afc9-c8102aea04d1

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Tip

Flake8 can be used to improve the quality of Python code reviews.

Flake8 is a Python linter that wraps PyFlakes, pycodestyle and Ned Batchelder's McCabe script.

To configure Flake8, add a '.flake8' or 'setup.cfg' file to your project root.

See Flake8 Documentation for more details.

[Jay-Lokhande]

Hey this looks cool,
actually i have cloned your branch but not getting how did you tested against Mesa 4.x would please help me out for this?

[colinfrisch]

I'm not fully up to date with the latest mesa 4.x changes... @wang-boyu what do you think about this ?

[abhinavk0220]

Hey Jay! To test against Mesa 4.x, install Mesa directly from
source:

git clone https://github.com/mesa/mesa.git
cd mesa
pip install -e .

Then install mesa-llm dependencies:

cd ../mesa-llm
pip install -e .

Then run the test suite:

python -m pytest tests/ -q

The key Mesa 4.x changes that affect mesa-llm are:

• mesa.space is removed → use mesa.discrete_space and
  mesa.experimental.continuous_space
• Model(seed=42) constructor changed
• model._time replaces model.steps as step counter

Happy to help if you run into issues! 🙂

[abhinavk0220]

Hi Colin! Happy to provide more context for wang-boyu's review.

The key Mesa 4.x breaking changes this PR addresses:

1. mesa.space removed entirely → migrated to mesa.discrete_space
   (OrthogonalMooreGrid) and mesa.experimental.continuous_space

2. ContinuousSpace API changed → old positional args replaced
   with dimensions=[[0,x],[0,y]] keyword API

3. model.steps → model._time as the internal step counter

4. Model(seed=42) constructor incompatibility with mesa_signals

All 203 tests passing after migration. Happy to answer any
specific questions wang-boyu has!

[Jay-Lokhande]

I was trying to understand changes you made I am not getting that 'hell' named file you have created or might generated.

[abhinavk0220]

Replaced mesa.space imports with mesa.discrete_space and
mesa.experimental.continuous_space throughout. The ContinuousSpace
API also changed old positional args (x_max, y_max) replaced with
dimensions=[[0,x],[0,y]] keyword API. Agent placement migrated from
grid.place_agent() to cell-based API: cell.add_agent(agent) +
agent.cell = cell.

[AdityaChauhanX07]

A few things worth noting for the review:

1. This was tested against Mesa 4.x from source, not the official pre-release (pip install --pre mesa). Worth verifying against the pre-release since that's what most users will actually install.

2. There are merge conflicts that will need resolving before this lands.

[abhinavk0220]

Thanks for the review, @AdityaChauhanX07!

On testing against the pre-release:
Good catch. The tests were run against Mesa 4.x installed from source (pip install -e .). I'll verify against the pip pre-release (pip install --pre mesa) as well — the APIs should be identical but it's worth confirming. If any divergence exists I'll update the PR accordingly.

On merge conflicts:
Agreed, there are conflicts that accumulated while other PRs landed on main. I'll rebase against the current main and resolve them before requesting a final review.

Appreciate you flagging both issues — this kind of review feedback is exactly what helps get a PR merge-ready.