For the installation and usage of the framework Python 3.6 or higher is needed. OSP-core is available on PyPI, so it can be installed using pip:

~/test$ pip install osp-core

However, we highly encourage the use of a virtual environment or a conda environment.

# virtual environment 
~/test$ python3 -m venv SimPhoNy
~/test$ source SimPhoNy/bin/activate
(SimPhoNy) ~/test$ 
# conda
~/test$ conda create -n <env name>
~/test$ conda activate <env name>

Unfortunately, OSP-core is not available on the Anaconda Repository, so in both cases, after setting up the environment, it must be installed with pip install osp-core.

After installing OSP-core, you can install your ontology namespaces. We provide the tool pico (pico installs cuds ontologies) for that purpose.

pico install <path/to/ontology.yml>

# If you have issues using pico directly, you can use
python -m osp.core.pico install <path/to/ontology.yml>

Wrapper installation

Wrappers are currently not available on PyPI, so they must be installed from source. First, the repository is cloned:

git clone<some-wrapper>.git
cd some-wrapper

Local wrapper installation

With OSP-core installed, if the wrapper has its own ontology, it must be installed:

pico install <path/to/ontology.yml>

For the wrappers that require the installation of a backend, a script is provided. It will automatically call, where the engine specific requirements are installed.


Now, the wrapper can be installed:

python3 install

Wrapper Docker image

Some wrappers also provided a Dockerfile for an automatic installation in a container. Simply run the script. There is no need to install OSP-core either.


Installing OSP-core from source

If you are a developer or an advanced user, you might be interested in installing OSP-core from source.

To do so, first the repository must be cloned:

git clone
cd osp-core

The installation is based on setuptools:

# build and install (recommended)
pip install .

# alternative
python3 install


# build for in-place development (recommended)
pip install -e .

# alternative
python3 develop