@peterspackman/selfies

TypeScript port of the Python SELFIES library for robust molecular string representation.

Installation

npm install @peterspackman/selfies

Usage

import * as selfies from '@peterspackman/selfies';

// SMILES to SELFIES
const benzene = "c1ccccc1";
const benzeneSelfies = selfies.encoder(benzene);  
// "[C][=C][C][=C][C][=C][Ring1][=Branch1]"

// SELFIES to SMILES
const smiles = selfies.decoder(benzeneSelfies);
// "C1=CC=CC=C1"

// Working with SELFIES
const length = selfies.lenSelfies(benzeneSelfies);  // 8
const symbols = [...selfies.splitSelfies(benzeneSelfies)];
// ['[C]', '[=C]', '[C]', '[=C]', '[C]', '[=C]', '[Ring1]', '[=Branch1]']

API

• encoder(smiles) - Convert SMILES to SELFIES
• decoder(selfies) - Convert SELFIES to SMILES
• setSemanticConstraints() - Configure bonding constraints
• lenSelfies(selfies) - Get symbol count
• splitSelfies(selfies) - Split into symbols
• getAlphabetFromSelfies(selfies[]) - Extract unique symbols
• selfiesToEncoding() / encodingToSelfies() - ML encodings

License

Apache 2.0 (same as original Python implementation)

Citation

If you use this in research, please cite the original SELFIES paper:

@article{krenn2020selfies,
  title={Self-referencing embedded strings (SELFIES): A 100% robust molecular string representation},
  author={Krenn, Mario and H{\"a}se, Florian and Nigam, AkshatKumar and Friederich, Pascal and Aspuru-Guzik, Alan},
  journal={Machine Learning: Science and Technology},
  volume={1},
  number={4},
  pages={045024},
  year={2020},
  publisher={IOP Publishing}
}