sfts 0.2.0

Scalable data-analysis framework for long-duration gravitational-wave signals

sfts

Short Fourier Transforms for Fresnel-weighted Template Summation.

Implementation of gravitational-wave data-analysis tools described in Tenorio & Gerosa (2025) to operate using Short Fourier Transforms (SFTs).

See the examples for a quick-start on using SFTs and iphenot (/ˈaɪv ˈnɒt/).

sfts contains two main modules:

1. kernels.py: Fresnel and Dirichlet kernels to compute scalar products using SFTs.
2. iphenot.py: jaxified re-implementation of the inspiral part of the IMRPhenomT waveform approximant.

The SFT convention in this package is compatible with that in the LVK .sft file format. Checkout fasttracks' search example to learn about reading .sft files into jax arrays.

How to install

sfts can be pulled in from PyPI:

$ pip install sfts

To pull in jax's GPU capabilities, use:

$ pip install sfts[cuda]

Alternatively, this repository itself is pip-installable.

Cite

If the tools provided by sfts were useful to you, we would appreciate a citation of the accompanying paper:

@article{Tenorio:2025yca,
    author = "Tenorio, Rodrigo and Gerosa, Davide",
    title = "{SFTs: a scalable data-analysis framework for long-duration gravitational-wave signals}",
    eprint = "2502.11823",
    archivePrefix = "arXiv",
    primaryClass = "gr-qc",
    month = "2",
    year = "2025"
}

Whenever applicable, please consider also citing the IMRPhenomT papers listed here.