EXA4MIND's Toolboxes

EXA4MIND's Toolboxes offer simple tools and examples to

• preprocess data (Preprocessing Toolbox),
• validate data (Validation Toolbox), and
• process data (Analytics Toolbox, with a focus on AI-related methods).

The tools are mostly written in Python and work from the command line with an interface based on the Click library. The portfolio of tools, which shall also serve as examples and boilerplates for creating custom preprocessors, is still being expanded.

Preprocessing Toolbox

The preprocessing toolbox is available under https://opencode.it4i.eu/exa4mind/platform/toolboxes/preprocessing-toolbox, with further documentation in README.md and LICENSE files.

The public version currently offers preprocessors to ingest data in files and directories into PostgreSQL, to convert and cloud-mask LANDSAT images, and to preprocess and re-format simulation, force-field and experimental data for Molecular Dynamics.

Validation Toolbox

The validation toolbox is available under https://opencode.it4i.eu/exa4mind/platform/toolboxes/validation-toolbox.

We have published tools for validating and preprocessing public health data from the Robert-Koch-Institut (RKI, Germany) and GPS & Odometry data as generated in autonomous-driving experiments.

Analytics and AI Toolbox

The analytics toolbox is available under https://opencode.it4i.eu/exa4mind/platform/toolboxes/analytics-and-ai-toolbox.

Currently, we have a tool on checking the suitability of training datasets for fair and unbiased AI in this toolbox.