Hi all,
We’ve recently released RAFFLE, an open source Fortran library (with an optional Python wrapper) for predicting interface structures.
RAFFLE employs an iterative approach to structure search, generating candidate interfaces, learning their energetics, and refining its atom placement rules with each iteration. This active learning process improves the efficiency of sampling interface energy landscapes, directing the search towards low-energy configurations.
I thought I’d share it here in case it’s useful to anyone. We are actively developing the library and exploring new features, so feedback and contributions are very welcome!
The software is developed in Fortran with a Python wrapper. It can be used in one of three ways:
- Fortran library (using fpm or CMake), run by calling procedures programs
- Fortran executable (using fpm or CMake), run using a custom format input file
- Python library (using fpm or CMake), run by calling procedures in script
For the Fortran side of it, it uses a set of derived types to hand atomic structure data (might be turned into a separate, dependent library in the future).
RAFFLE is developed by Ned Thaddeus Taylor, Joe Pitfield, and Steve Hepplestone.
Feel free to ask any questions or share any suggestions!
