Python environment

In OGS we make use of Python packages at different stages, e.g.:

• TESPy for simulating thermal engineering plants in a benchmark
• pvpython for pre- and post-processing
• nbconvert for testing Jupyter Notebooks
• …

Python packages are usually installed via pip inside an isolated environment (a virtual environment).

Pip

When configuring OGS with OGS_USE_PIP=ON Python creates a new virtual environment in the .venv-directory inside your build directory. It will also install required Python packages into this environment. You can see the current environment definition in the file requirements.txt inside your build-directory.

To manually add Python packages run the following inside your build-directory:

.venv/bin/pip install python-package-name

To activate the environment run source .envrc inside your build directory. If you have the direnv-tool installed and setup the virtual environment will be activated automatically upon changing into the build directory.

Pip & Benchmarks

You can use the argument PYTHON_PACKAGES on AddTest() to specify additional Python package dependencies.

The following example would install the latest version of numpy and pandas version 0.1.2:

AddTest(
    ...
    PYTHON_PACKAGES numpy pandas==0.1.2
    ...
)