Replace black/flake8/isort with ruff for lint + format

[stefanbschneider]

Summary

• Replace black, flake8, and isort with ruff for both linting and formatting.
• ruff.toml: line-length = 100 (matches the previous flake8 setting), select = ["E", "F", "W", "I"] (I replaces isort), and a per-file F401 ignore for __init__.py (previously only passed on the CI flake8 command line via --per-file-ignores, so pre-commit never actually enforced it). *.ipynb and *.md are excluded from ruff format — black never touched notebooks, and ruff's markdown support would otherwise reformat the Python code fences embedded in the docs, which is out of scope here.
• Delete setup.cfg (it only held the flake8 config).
• .pre-commit-config.yaml: drop the black/flake8/isort hooks, add astral-sh/ruff-pre-commit (ruff-check --fix, ruff-format); also bumped the stale pre-commit-hooks pin (v2.3.0 → v6.0.0) while touching this file.
• CI (.github/workflows/python-package.yml): replace the "Lint with flake8" step with ruff check ., and add a new ruff format --check . step. This is a real behavior change, not just a swap — CI previously never checked formatting at all; black/isort only ran locally via pre-commit.
• tests/requirements.txt: flake8 → ruff.
• Reformatted all Python sources with ruff format (mostly re-wrapping lines that black had wrapped at 88 chars, now allowed up to 100 to match the lint config) and ran ruff check --fix (clean — no violations beyond formatting).

README badges

• Swapped the "Code Style: Black" badge for the official Ruff badge.
• Removed the "PyPI" badge — it was actually the python-publish.yml workflow-status badge, which only runs on release tags, so it reads stale/broken most of the time. Replaced with a real PyPI version badge linking to the package's PyPI page.
• Added a License (MIT) badge, since none existed despite the project being MIT-licensed.
• Pointed the Colab links at main instead of master — the repo has no master branch, so those only ever worked via GitHub's redirect.

Dependency updates

Beyond the ruff swap itself: pre-commit-hooks v2.3.0 → v6.0.0. I checked uv pip list --outdated; core runtime deps (gymnasium, numpy, pandas, matplotlib, pygame, shapely) were already bumped and verified in the immediately preceding commit (b07257b), and nothing there is newly outdated except transitive CUDA/nvidia/pydantic-core pins and gymnasium 1.2.2 (pinned exactly by ray[rllib], already inside setup.py's >=0.29.1,<2.0 range) — so those were left alone.

Not included: the stuck-CI ruleset issue

Separately reported (and diagnosed, but not fixed here — it's a repo Settings change, not a code change): the main branch ruleset requires status check build (ubuntu-latest, 3.9), but the CI matrix hasn't run 3.9 since an earlier commit (now 3.10–3.13). That context can never be reported, so every PR — including this one — sits "Expected — Waiting for status to be reported" forever. Fix is to update the ruleset's required checks to swap 3.9 for 3.13 (Settings → Rules → Rulesets → main).

Verification

• pytest — 11 passed
• pytest --nbmake examples/test.ipynb — 1 passed
• pre-commit run --all-files — clean
• ruff check . / ruff format --check . — clean
• Sanity-checked the ruff config actually enforces line length and import order (not just defaults) by injecting a >100-char line and unsorted imports into scratch copies and confirming ruff check flags them (E501, I001) before relying on it.

🤖 Generated with Claude Code

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.