libfp is a library for calculating crystalline fingerprints and measuring similarities of materials (e.g., crystals, clusters, and molecules). The library is written in C with a Python interface.
Research papers where this code has been used include:
- Accelerating Structural Optimization through Fingerprinting Space Integration on the Potential Energy Surface, J. Phys. Chem. Lett. 15, 3185 (2024)
- Quantum structural fluxion in superconducting lanthanum polyhydride, Nature Commun. 14, 1674 (2023)
- Phase Transition Pathway Sampling via Swarm Intelligence and Graph Theory, J. Phys. Chem. Lett. 10, 5019 (2019)
Before installing libfp, ensure you have the following:
- Python header files (python-dev)
- C compiler (e.g., gcc, clang)
- NumPy
- OpenBLAS, LAPACK, or MKL
You can install libfp using one of the following methods:
To install the latest stable version from PyPI, simply run:
pip install libfp
-
Clone the repository:
git clone https://github.com/Rutgers-ZRG/libfp.git cd libfp -
Install using pip:
pip install .
To use libfp in your Python project:
import libfp
# Your code hereExample
Examples are found in examples directory.
The detailed algorithm of libfp is described in the following paper:
- "A fingerprint based metric for measuring similarities of crystalline structures", Li Zhu, Maximilian Amsler, Tobias Fuhrer, Bastian Schaefer, Somayeh Fareji, Alireza Ghasemi, Migle Grauzinyte, Chris Wolverton, and Stefan Goedecker J. Chem. Phys. 144, 034203 (2016)
This library is derived from an original Fortran version, which I wrote during my postdoctoral work in Prof. Goedecker's research group at the University of Basel. I am grateful to Prof. Goedecker for his guidance during my postdoc period. I also acknowledge support from Rutgers startup funding and the National Science Foundation Division of Materials Research (NSF-DMR), under Grant No. 2226700.