graphchem 2.3.4

Graph-based machine learning for chemical property prediction

GraphChem: Graph-based machine learning for chemical property prediction

GraphChem is an open source Python package for constructing graph-based machine learning models with a focus on fuel property prediction.

Installation

Prerequisites

• Python 3.10 or newer

Method 1: pip

$ pip install graphchem

Method 2: From source

$ git clone https://github.com/ecrl/graphchem
$ cd graphchem
$ python -m pip install .

If installation fails for RDKit, PyTorch, or torch-geometric, follow the vendor guides: RDKit, PyTorch, PyTorch Geometric.

Documentation

• Read the Docs (Sphinx) — installation, quickstart, API, datasets, stability policy
• API_STABILITY.md — source-tree pointer to the frozen public surface
• CHANGELOG.md

Examples

Notebooks are under examples/ on GitHub. See also the tutorials page in the docs.

For reproducible runs, set seeds before training or evaluation, for example:

import numpy as np
import torch

torch.manual_seed(0)
np.random.seed(0)

Exact numeric results can still vary across platforms and dependency wheel builds (RDKit / PyTorch / PyG).

Contributing and support

• Contribute via pull request; follow CONTRIBUTING.md
• Report bugs and feature requests with GitHub Issues
• Security reports: SECURITY.md
• Code of conduct: CODE_OF_CONDUCT.md

For additional questions, contact Travis Kessler ([e-mail]).