Conditional Constrained Graph Variational Autoencoders (CCGVAE) for Molecule Design

This repository contains the code used to generate the results reported in the paper: Conditional Constrained Graph Variational Autoencoders for Molecule Design.

@article{rigoni2020conditional,
  title={Conditional Constrained Graph Variational Autoencoders for Molecule Design},
  author={Rigoni, Davide and Navarin, Nicol{\`o} and Sperduti, Alessandro},
  journal={arXiv preprint arXiv:2009.00725},
  year={2020}
}

All the files related to the CCGVAE model will be uploaded soon.

Dependencies

This project uses the conda environment. In the root folder you can find, for each model, the .yml file for the configuration of the conda environment and also the .txt files for the pip environment. Note that some versions of the dependencies can generate problems in the configuration of the environment. For this reason, although the setup.bash file is present for the configuration of each project, it is better to configure them manually.

Structure

The project is structured as follows:

• data: contains the code to execute to make the dataset;
• results: contains the checkpoints and the results;
• model: contains the code about the model;
• utils: contains all the utility code;
• histogramAnalysis: contains all the code necessary to print the images about the histogram distribution.

Usage

Data Download

First you need to download the necessary files and configuring the environment by running the following commands:

sh setup.bash install
conda activate givae

Data Pre-processing

In order to make de datasets type the following commands:

cd data
python make_dataset.py --dataset [dataset]

Where dataset can be:

• qm9
• zinc

Model Training

In order to train the model use:

python CCGVAE.py --dataset [dataset] --config '{"generation":0, "log_dir":"./results", "use_mask":false}'

Model Test

In order to generate new molecules:

python CCGVAE.py --dataset [dataset] --restore results/[checkpoint].pickle --config '{"generation":1, "log_dir":"./results", "use_mask":false}'

While, in order to reconstruct the molecules:

python CCGVAE.py --dataset [dataset] --restore results/[checkpoint].pickle --config '{"generation":2, "log_dir":"./results", "use_mask":true}'

In order to analyze the results, we used the following environmet: ComparisonsDGM.

Information

For any questions and comments, contact Davide Rigoni.

NOTE: Some functions are extracted from the following source code.

Licenze

MIT