neoralab-sascorer

A small, pip-installable wrapper around RDKit's Contrib/SA_Score implementation for computing the synthetic accessibility (SA) score.

📚 Documentation: See the MkDocs site content in docs/index.md.

Features

• Simple Python API for SMILES or RDKit molecule inputs.
• CLI for quick SA score checks from the shell.
• Lightweight packaging with dynamic versioning from git tags.

Requirements

• Python 3.8+
• RDKit (install via conda or ensure a compatible wheel is available for your platform)

Installation

Dependency: This package requires RDKit. pip install neoralab-sascorer expects an RDKit wheel to be available for your platform. If you prefer conda, you can install RDKit via conda and then install this package.

Editable install from this repo:

uv pip install -e .

The fpscores.pkl.gz fragment score file must be present in neoralab-sascorer/src/neoralab_sascorer/ (it is distributed with RDKit's Contrib/SA_Score and should be copied in before building a wheel).

Usage

Python API

from neoralab_sascorer import sa_score

score = sa_score("CC(=O)Oc1ccccc1C(=O)O")
print(score)

from rdkit import Chem
from neoralab_sascorer import sa_score_mol

mol = Chem.MolFromSmiles("CC(=O)Oc1ccccc1C(=O)O")
print(sa_score_mol(mol))

CLI

neoralab-sascorer "CC(=O)Oc1ccccc1C(=O)O"

Output format:

<SMILES>\t<score>

Development

Versioning

This project uses versioningit to derive versions from git tags. For example, a commit after the 0.1.0 tag will produce a version like 0.1.0.dev3.

Pre-commit

Install and run pre-commit hooks locally:

uv pip install pre-commit
pre-commit install
pre-commit run --all-files

Attribution

This package bundles the SA_Score implementation and data originally distributed in RDKit's Contrib/SA_Score, based on the method described by Ertl and Schuffenhauer ("Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions", J. Cheminformatics 1:8, 2009).