Python for Dymola

Sometimes it is required to sweep over design parameters, simulate, and gather data over and over again.

The use of Python with the Dymola API allows us to automate this process.

If using VSCode, you can visualize the data within VSCode by first installing the Jupyter extension. You then need to add a breakpoint somewhere after the data you are interested in visualizing is processed by the Python interpreter. Now run the script in Debug mode and go to the debug menu, find the variable you are interested in visualizing and right-click on it, then choose 'View value in Data Viewer'.

You can read more on that here.

Be a good person and create a virtual environment¹ and then activate it. People who don't, only want to watch the world burn.

$ python3 -m venv /path/to/new/virtual/environment
$ source /path/to/venv/bin/activate

Update pip and install pip-tools². I only say this because I care about you, complete random stranger.

(venv) $ pip install -U pip
(venv) $ pip install pip-tools
(venv) $ pip-sync

If you use Windows, first of all ew and shame on you, then you need to prefix your commands with python3 -m as follows,

(venv) PS> python3 -m pip install -U pip

etc...

Make sure to add conda to the PATH variable. Open the Anaconda3 command prompt and type

where conda

Add the directories listed to PATH. Then in VSCode, open a terminal prompt and type conda init. Once you create your conda environment, you should now be able to activate it.

If using macOS, don't bother with any of that. Simply use

brew install --cask anaconda

If you don't want the bloated Anaconda installation and don't mind being behind the driver seat, you can use micromamba instead

brew install micromamba

If using micromamba, it would be a good idea to add an alias to make your life easier. An example on how to do that if you are using macOS/Linux is as follows:

echo "alias conda='micromamba'" >> ~/.zshrc

Don't forget to create an environment and activate it.

conda create -n <env_name>
conda activate <env_name>
conda install -c conda-forge python=<MAJOR.MINOR.PATCH>

Using conda/micromamba allows you to easily install and use PyFMI from Modelon. To install

conda install -c conda-forge pyfmi

In addition, it makes it easier to install FMPy from CATIA (© Dassault Systèmes)

conda install -c conda-forge fmpy

You can start the FMPy GUI with

python -m fmpy.gui

The FMPy Web App is built with Dash and a great way to share your FMUs with anyone that has a web browser. To start it, run

python -m fmpy.webapp <model_name>.fmu