GEMSEO depends on the packages listed below, some of them are optional.

You may use more recent versions of these packages, but we cannot guarantee the backward compatibility. However, we provide a large set of tests with a high code coverage so that you can fully check your configuration.

See also

Fully check your configuration with Test with unit tests.

Core features

The required dependencies provide the core features of GEMSEO, these are (shown here for Python 3):

  • custom_inherit ==2.4.0

  • fastjsonschema <=2.15.1

  • future

  • genson ==1.2.2

  • h5py >=2.3,<=3.2.1

  • jinja2 <=3.0.1

  • matplotlib >=2,<=3.4.3

  • networkx >=2.2,<=2.5

  • numpy >=1.10,<=1.20.3

  • pyxdsm <=2.2.0

  • requests

  • scipy >=1.1,<=1.7.1

  • six

  • tqdm >=4,<=4.61.0

  • xdsmjs >=1.0.0,<=1.0.1

The minimal dependencies will allow to execute MDO processes but not all post processing tools will be available.

Full features

Some packages are not required to execute basic scenarios, but provide additional features, they are listed below. The dependencies are independent, and can be installed one by one to activate the dependent features of listed in the same table. Installing all those dependencies will provide the full features set of GEMSEO. All these tools are open source with non-viral licenses (see Credits):

  • graphviz ==0.16: coupling graph generation

  • nlopt >=2.4.2,<=2.7.0: optimization library

  • openturns >=1.13,<=1.18: designs of experiments, machine learning, uncertainty quantification

  • pandas >=0.16,<=1.3.4: scatterplot matrix

  • pdfo ==1.0.0: derivative-free optimization algorithms

  • pydoe2 >=0.3.8,<=1.3.0: design of experiments

  • pyside2 <=5.15.2: grammar editor GUI

  • scikit-learn >=0.18,<=1.0.1: gaussian process surrogate model and SOM, kmeans

  • sympy >=0.7,<=1.9: symbolic calculations for analytic disciplines

  • openpyxl <=3.0.7: excel reading with pandas

  • xlwings <=0.21.4: excel reading with pandas