Skip to content

Installation

AgaveChem supports Python versions >= 3.10.

Requirements

  • Python (version >= 3.10)
  • RDKit — cheminformatics toolkit
  • rdchiral-plus — improved template extraction and application library (installed automatically)
  • PyTorch — required for the neural mapper; a CPU-only build is sufficient for inference
  • Transformers (Hugging Face) — required for the neural mapper

All dependencies are declared in requirements.txt and are installed automatically by pip.

Install AgaveChem from PyPi:

pip install agave_chem

Or install AgaveChem with pip directly from the GitHub repo:

pip install git+https://github.com/denovochem/agave_chem.git

Installation Option #2: Clone the repository

Install the latest version of AgaveChem from GitHub. The version on GitHub is not guaranteed to be stable but may include new features not available via other install options.

git clone https://github.com/denovochem/agave_chem.git
cd agave_chem
pip install .

Verifying the installation

After installation, verify that AgaveChem is importable and the neural mapper loads correctly:

from agave_chem import MCSReactionMapper

mapper = MCSReactionMapper("test")
result = mapper.map_reaction("CC(=O)O.OCC>>CC(=O)OCC")
print(result["selected_mapping"])

The neural mapper (NeuralReactionMapper) will download model weights on first use if they are not already present.