# ProteinSolver: Solving the Inverse Protein Folding Problem with Graph Neural Networks

[![HTML Manuscript](https://img.shields.io/badge/manuscript-HTML-orange.svg)](https://ostrokach.gitlab.io/proteinsolver/FGIDVYN9fMx8qvsyPSKtoLaEucROasB/report/)
[![docs](https://img.shields.io/badge/docs-v0.1.1-blue.svg)](https://ostrokach.gitlab.io/proteinsolver/v0.1.1/FGIDVYN9fMx8qvsyPSKtoLaEucROasB/)
[![conda](https://img.shields.io/conda/dn/ostrokach/proteinsolver.svg)](https://anaconda.org/ostrokach/proteinsolver/)
[![build status](https://gitlab.com/ostrokach/proteinsolver/badges/v0.1.1/build.svg)](https://gitlab.com/ostrokach/proteinsolver/commits/v0.1.1/)
[![coverage report](https://gitlab.com/ostrokach/proteinsolver/badges/v0.1.1/coverage.svg)](https://ostrokach.gitlab.io/proteinsolver/v0.1.1/FGIDVYN9fMx8qvsyPSKtoLaEucROasB/htmlcov/)

## Introduction

`proteinsolver` is a deep neural network which learns to solve (ill-defined) constraint satisfaction problems (CSPs) from training data.

### Environment variables

- `DATAPKG_DATA_DIR` - Location of training and validation data.