Dataset: Latent Style-based Quantum Wasserstein GAN for Drug Design

This is the supporting package to the paper: Latent Style-based Quantum WGAN for Drug Design, arXiv:2603.22399

It contains all the data to generate the tables and graphs of the paper, as well as a python package to automatize their generation.

Instructions:

Unzip the file, the main file is "create_plots_and_tables.py". The data is located in the folder "data".

The main python script is "create_plots_and_tables.py". It accepts command-line arguments to select which data plot or table you want to re-generate. Here is what the help command produces:

usage: create_plots_and_tables.py [-h] [--list-targets] [targets ...]

Create paper figures and tables.

positional arguments:  targets         What to generate: 'all', 'figures', 'tables', or specific                  IDs (e.g. Fig4 Tab6).

options:  -h, --help      show this help message and exit  --list-targets  Show all valid target IDs and exit.

Examples:  python create_plots_and_tables.py all  python create_plots_and_tables.py tables  python create_plots_and_tables.py Fig4 Tab6  python create_plots_and_tables.py --list-targets

In the newly created folders "plots" and "tables" you find the results.

-- Notes -- Not all tables/plots can be reproduced, see below:

Figs 1 - 3 not coming from dataFig 4 - 10 reproducedFig 11 - data exists in the repository (folders with "julien" in the name) but not visualisedFig 12 - 17 reproducedFig 18 - 19 Data (involves active use of the VAE training) is not ready but available upon reasonable requestFig 20 - reproduced, is automatically produced with Fig9Fig 21 - 26 reproduced

Tab 1 - parameter summaryTab 2 - 3 reproducedTab 4 - is the simplified version of Tab 14, reproducedTab 5 - summary table, values available in other tablesTab 6 - 12 reproducedTab 13 - summary link tableTab 14 - 20 reproduced