Terrarium

A living ASCII ecosystem that runs in your terminal. Watch plants spread, herbivores graze and predators hunt—all while you leave your terminal open.

Overview

Terrarium is an idle simulation game. The player watches—and occasionally nudges— a self-sustaining world of plants, herbivores and predators. The core fantasy is:

"I grew something beautiful just by leaving my terminal open."

Requirements

• Python 3.11+
• rich
• A terminal at least 80 × 24 characters (120 × 30+ recommended)

Installation

Install from PyPI

pip install terrarium-game
terrarium-game

Install from source

git clone https://github.com/minoangelo/terrarium
cd terrarium
pip install -r requirements.txt
python main.py

Controls

The game is designed to be satisfying to watch without touching the keyboard.

How It Works

The Grid

The world is a 2-D tile grid (up to 60 × 24, scaled to your terminal). Each tile tracks moisture and nutrients (0–100) that slowly regenerate over time. Static water tiles (~) boost nearby moisture. Dead organic matter (%) fertilises the soil as it decays.

Plants

Plants die when soil nutrients reach zero, leaving organic matter behind.

Animals

Both animals starve if hunger reaches 100—health then decays until death, leaving organic matter.

Milestones

Save Files

The game auto-saves every 30 seconds to ~/.terrarium/save.json. On next launch you will be asked whether to resume. Quitting with q always saves immediately.

Project Structure

terrarium/
├── main.py          # Entry point, game loop, keyboard input
├── world.py         # Grid, Tile and soil simulation
├── entities.py      # Entity classes and per-tick simulation logic
├── renderer.py      # Rich layout, viewport drawing, sidebar
├── events.py        # Event log and milestone tracking
├── persistence.py   # JSON save/load
├── state.py         # Shared runtime/render state dataclasses
└── requirements.txt

Contributing

Contributions are welcome. This project currently uses a lightweight workflow.

1) Set up locally

python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
pip install -r requirements.txt

2) Run and validate

# run the app
python main.py

# basic syntax/import check
python -m compileall .

# lint (primary)
pylint main.py <file>

# lint tests (with project root on PYTHONPATH)
PYTHONPATH=. pylint tests/<file>

# optional fast lint
ruff check .

3) Tests

The repository includes tests under tests/. Use pytest to run them, and keep adding coverage there as features and bug fixes evolve.

# run all tests
pytest

# run a single test file
pytest tests/test_entities.py

# run a single test case
pytest tests/test_entities.py::test_predator_moves_toward_nearest_prey

# run by expression
pytest -k "predator and not slow"

4) Commit and PR style

• Follow Conventional Commits (for example: feat: ..., fix: ..., docs: ...).
• Keep changes focused and small when possible.
• In your PR description, include:
  • what changed,
  • why it changed,
  • which validation commands you ran.

5) Releases and PyPI publishing

• Versioning and release PRs are automated with Release Please.
• Conventional Commit messages are used to determine version bumps and changelog entries.
• Publishing to PyPI runs from GitHub Actions when a GitHub Release is published.
• Configure PyPI trusted publishing for this repository before the first release.

6) Scope guidance

• Keep architecture boundaries intact:
  • world.py for terrain/soil rules,
  • entities.py for simulation behavior,
  • renderer.py for presentation,
  • persistence.py for save/load,
  • main.py for orchestration and input loop.
• Avoid blocking I/O in the main loop.
• Preserve save compatibility where practical.

License

See LICENSE.md.

Contact

• Carmine (Mino) Siena, dev@mino-siena.de