Tutorial

These tutorials will demonstrate the most foundational usages of XenonPy and XenonPy.MDL and we assume that you can use Jupyter notebook.

• Data access
• Descriptor calculation
• Visualization
• Random NN models
• Pre-trained Model Library
• Transfer Learning
• XenonPy-iQSPR tutorial

Sample data

Before starting these tutorials, you need to prepare some sample data for the benchmark/test.

We selected 1,000 inorganic compounds randomly from the Materials Project database for this using. You can check these MP ids at:

https://github.com/yoshida-lab/XenonPy/blob/master/samples/mp_ids.txt.

To build the sample data, you also have to create an API key at the materials project. Please see The Materials API and follow the official documents.

You can use the following codes to build your sample data. These data will be saved at ~/.xenonpy/userdata.

>>> from xenonpy.datatools import preset
>>> preset.build('mp_samples', api_key='your api key')

If you want to know what exactly the code did, please check:

https://github.com/yoshida-lab/XenonPy/blob/master/samples/sample_data_building.ipynb