Molecular Generation With Variational Auto Encoding and Generative Adversarial Neworks

Developer Notes :

This Project Is a work in progress
Data has been preprocessed for use with machine learning. It is not included in this repo as it is 500mb+

Features

The project consists of the following main components :

Preprocessing - A dataset of smiles is sorted and processed by accessing the dataset of smiles, rescaling and normalising the molecules so trained models has less representations and relationships to learn. This will improve the accuracy.

----- Work In Progress -----

Training - Training is currently done using either a Vartiational Auto Encoder or Generative Adversaral Network.
New Molecules Generation - New molecules are generated.

A more detailed documentation of features will be updated soon.

Installation

Clone the repository:

git clone https://github.com/DanielFlockhart/Molecular-Generation.git

Navigate to the project directory:
```
cd Molecular-Generation
```
Install the required dependencies:
```
pip install -r requirements.txt
```

Usage

Code is not yet set up with User Interface for usage - Will update README.md at a later stage

To Preprocess data

   cd src
   python main.py

   ------------------------- Initialising -------------------------
   ------------------------- Preprocessing Data -------------------------
   ------------------------- Downloading Images -------------------------
   100%|████████████████████████████████████| 48167/48167 [06:09<00:00, 130.33it/s]
   ------------------------- Scaling Images -------------------------
   100%|████████████████████████████████████| 47471/47471 [04:02<00:00, 195.92it/s]
   ------------------------- Preprocessing Summary -------------------------
   Generated From Smiles        : 98.55502730084913%
   Bound Used                   : 617.0
   Amount of Images Above Bound : 902/47471

Contribution

Contributions to this project are welcome! If you have any suggestions, improvements, or new features to propose, please submit a pull request. You can also report any issues or bugs by opening an issue on the project's GitHub repository.

When contributing, please follow the existing code style, write clear and concise commit messages, and provide appropriate documentation.

License

This project is licensed under the MIT License. Feel free to use, modify, and distribute it as per the terms of the license.

Credits

The project acknowledges the following resources for their contributions:

RDKit - Converting SMILEs to 2D Structures
LiverpoolUniversity - Dataset for theoretical predictions of new applications for existing compounds

Thank you for using the Molecular Generation project! We hope it proves to be useful for your chemical analysis and research.

To Do

Reverse image to smiles - Molecular Profiles
Train a full Model
Write Paper 2
Under the other conditions version