Getting Started

Welcome to ProQSAR — an opinionated, reproducible pipeline for QSAR modelling and small-molecule featurisation.

This quickstart shows how to install ProQSAR, verify the installation, and run a minimal pipeline. For full reference documentation, see the project docs and the API reference in the repository.

Introduction

ProQSAR is a lightweight toolkit for end-to-end QSAR workflows: data standardization, featurization, dataset splitting, model training, uncertainty estimation, and evaluation. It provides a simple CLI and a modular Python API so you can run quick experiments or embed ProQSAR components into larger CADD pipelines.

Requirements

Python >= 3.11
Recommended: an isolated virtual environment (venv/virtualenv or Conda)

Virtual environment (recommended)

Using a virtual environment prevents dependency conflicts.

Using venv (cross-platform)

python3 -m venv proqsar-env
source proqsar-env/bin/activate    # Linux / macOS
proqsar-env\Scripts\activate     # Windows (PowerShell)

Using Conda

conda create -n proqsar-env python=3.11 -y
conda activate proqsar-env

Installation

Install the package from PyPI, Conda (channel: tieulongphan), or using the official Docker image.

From PyPI:

pip install proqsar

From conda:

conda install -c tieulongphan proqsar

Docker:

docker pull tieulongphan/proqsar:latest
docker run --rm tieulongphan/proqsar:latest proqsar --help

Quick verification

Verify the installed package and check the version:

python -c "import importlib.metadata as m; print(m.version('proqsar'))"

Development & contributing

Fork the repository and work on a feature branch.
Add unit tests and run the test-suite locally.
Follow code style (PEP8, type hints) and run pre-commit hooks.
Open a PR with a clear description and tests.

Support

Report bugs or request features on GitHub:

ProQSAR Issues