Worm2D

Worm2D is a neuromechanical simulator of C. elegans locomotion. This is an initiative of the OpenWorm project. It is based on earlier work on development of 2D worm neuromechanical models of Netta Cohen and team, as well as Dr. Eduardo J. Izquiredo, Dr. Erick Olivares and Prof. Randall Beer (See HISTORY.md).

Worm2D is a hybrid C++/Python package: C++ binaries implement the physics, nervous system, and evolutionary search, while Python handles configuration, binary dispatch, and post-run analysis/plotting.

Build & Install

The package uses scikit-build-core with CMake. nlohmann-json is bundled as a header at src/cpp/nlohmann/json.hpp — no system package needed.

Install in editable mode (builds C++ and installs Python package):

pip install -e .[all]

The [all] extra pulls in [neuroml] (pyNeuroML, NEURON<9.0) and [dev] (ruff, pytest, ipywidgets, OSBModelValidation, modelspec).

Running Tests

Quick test (runs after pip install -e .):

python tests/test.py # runs a single optimisation run & saves output to exampleRun folder, see below

Full test suite (cleans prior run folders, then runs all model variants and OMV tests):

./test_all.sh        # full
./test_all.sh -q     # quick (skips OMV and most model-specific tests)

Individual test scripts in tests/ and src/worm2d/ follow the naming pattern test<ModelVariant>.py. Run any directly:

cd tests
python testW2DCE.py

CLI

After install, the worm2d command is available:

worm2d -f <outputFolder> [-E] [-O <modelName>] [-T <modelFolder>] [...]

Key flags: -E / --doEvol runs the evolutionary search; -f specifies the output folder (required); -O sets the model name when using the Worm2D binary family.