Feel++ mo2fmu converter

Modelica to FMU converter based on dymola

Installation

From PyPI (Recommended)

Install the latest stable release from PyPI:

pip install feelpp-mo2fmu

Or using uv (faster):

uv pip install feelpp-mo2fmu

From Source

For development or to use the latest unreleased features:

git clone https://github.com/feelpp/mo2fmu.git
cd mo2fmu
uv venv .venv-mo2fmu
source .venv-mo2fmu/bin/activate  # On Windows: .venv-mo2fmu\Scripts\activate
uv pip install -e ".[all]"

Configuration

Dymola Location

The Dymola installation location can be configured via environment variables:

export DYMOLA_ROOT=/opt/dymola-2025xRefresh1-x86_64/
export DYMOLA_EXECUTABLE=/usr/local/bin/dymola
export DYMOLA_WHL=Modelica/Library/python_interface/dymola-2025.1-py3-none-any.whl

These environment variables are used by:

• Tests (avoiding hardcoded paths)
• CI/CD workflows
• Command-line interface defaults

Environment Variables:

• DYMOLA_ROOT: Path to Dymola installation root directory (default: /opt/dymola-2025xRefresh1-x86_64/)
• DYMOLA_EXECUTABLE: Path to Dymola executable binary (default: /usr/local/bin/dymola)
• DYMOLA_WHL: Relative path to Dymola Python wheel from DYMOLA_ROOT (default: Modelica/Library/python_interface/dymola-2025.1-py3-none-any.whl)

Usage in command line

$ mo2fmu --help
Usage: mo2fmu [OPTIONS] MO OUTDIR

Options:
  --fmumodelname TEXT          change the model name of the FMU (default: .mo
                               file stem)
  --load TEXT                  load one or more Modelica packages.
  --flags TEXT                 one or more Dymola flags for FMU translation.
  --type [all|cs|me|csSolver]  the FMI type: cs, me, all, or csSolver.
  --version TEXT               the FMI version.
  --dymola PATH                path to Dymola root.
  --dymolapath PATH            path to Dymola executable.
  --dymolawhl PATH             path to Dymola whl file, relative to Dymola
                               root.
  -v, --verbose                verbose mode.
  -f, --force                  force FMU generation even if file exists.
  --help                       Show this message and exit.----

Usage in Python

Here is an example of how to use the mo2fmu function in Python that would convert a Modelica file to an FMU:

from feelpp.mo2fmu import mo2fmu
mo2fmu(
    mo_file="path/to/model.mo",
    outdir="path/to/output/dir",
    fmumodelname="MyFMUModel",
    load=["Modelica", "SomePackage"],
    flags=["-d=initialization"],
    fmi_type="cs",
    fmi_version="2.0",
    dymola_root="/path/to/dymola/root",
    dymola_executable="/path/to/dymola/executable",
    dymola_whl="/path/to/dymola.whl",
    verbose=True,
    force=False
)

Continuous Integration

Our GitHub Actions workflow (.github/workflows/ci.yml) includes:

• build_wheel: Python wheel compilation and artifact upload.
• docs: Builds the Antora site, deploys to GitHub Pages on master.
• deliver: Docker image build & push to GHCR.
• release: On tags vX.Y.Z, publishes binaries, wheels, datasets, and creates a GitHub release.

Versioning & Release

Project version is centrally defined in:

• docs/antora.yml
• docs/package.json

Contributing

We welcome contributions! Please:

• Fork the repository and create a feature branch.
• Adhere to existing coding conventions; add python tests where appropriate.
• Update documentation (docs/) for any new features.
• Submit a pull request with a clear description of your changes.

License

This project is licensed under the MIT License. See LICENSE for full details.