Dand

<h1 align="left">Dandelion</h1>

<div align="center"> <h4 align="center">Codes for automated and efficient sampling of chemical reaction space for MLIP training</h4> <img src="https://github.com/user-attachments/assets/ed0e5069-6464-485e-b2a5-c3461b8cb528" alt="drawing" width="700"/> </div>

Dandelion is a code for generating datasets that contain both equilibrium and reactive regions of potential energy surfaces, using automated and efficient sampling of chemical reaction space.

Documentation : https://mhyeok1.github.io/dand_docs/

Citation

If you find this work useful for your research, please consider citing:

• Lee et al. Adv. Sci. 12, 2409009 (2025) LINK

This work builds upon pioneering works that should also be cited:

• Grambow et al. Sci. Data 7, 137 (2020) LINK
• Schreiner et al. Sci. Data 9, 779 (2022) LINK

Supporting Information

The datasets used in the paper are available at zenodo.