Installing the workbench¶
The 2.x version of the workbench requires Python 3.6 or higher.
A stable version of the workbench can be installed via pip.
pip install ema_workbench
The code is also available from github. The code comes with a requirements.txt file that indicates the key dependencies. Basically, if you have a standard scientific computing distribution for Python such as the Anaconda distribution, most of the dependencies will already be met.
In addition to the libraries available in the default Anaconda distribution, there are various optional dependencies. Please follow the installation instructions for each of these libraries.
From conda or conda forge:
- altair for interactive visualizations
- ipyparallel for support of interactive multiprocessing within the jupyter notebook.
There are also some pip based dependencies:
- SALib , this is a necessary dependency for advanced senstivity analysis.
- platypus-opt , this is an optional dependency for many-objective optimization functionality.
- pydot and Graphviz for some of the visualizations.
The various connectors have their own specific requirements.