Welcome to PEREGRINE! A public, open-source pipeline designed for targeted inference of gravitational wave signals from multiple source types.

Currently, PEREGRINE has two public branches:

• cbc: Inferring source parameters of standard LVKC CBC sources.
• overlapping: Inferring source parameters of overlapping or coincident gravitational wave detections in a 2G network.

• Sequential simulation-based inference for gravitational wave signals (Bhardwaj, Alvey, Miller, Nissanke, Weniger)
• What to do when things get crowded? Scalable joint analysis of overlapping gravitational wave signals (Alvey, Bhardwaj, Nissanke, Weniger)

PEREGRINE leverages the power of (Truncated Marginal) Neural Ratio Estimation (TMNRE) via the open-source package swyft.

• For details on (TM)NRE, swyft and more, please visit The Undark Lab.
• Documentation for swyft is available at: swyft documentation.

For any queries related to PEREGRINE, please feel free to contact:

• Uddipta Bhardwaj
• James Alvey
• Alternatively, please feel free to open a github issue or pull request.

If you use PEREGRINE in your analysis, or find it useful, we would ask that you please use the following citations:

@article{Bhardwaj:2023xph,
    author = "Bhardwaj, Uddipta and Alvey, James and Miller, Benjamin Kurt and Nissanke, Samaya and Weniger, Christoph",
    title = "{Peregrine: Sequential simulation-based inference for gravitational wave signals}",
    eprint = "2304.02035",
    archivePrefix = "arXiv",
    primaryClass = "gr-qc",
    month = "4",
    year = "2023"
}

@article{Miller:2022shs,
    author = "Miller, Benjamin Kurt and Cole, Alex and Weniger, Christoph and Nattino, Francesco and Ku, Ou and Grootes, Meiert W.",
    title = "{swyft: Truncated Marginal Neural Ratio Estimation in
Python}",
    doi = "10.21105/joss.04205",
    journal = "J. Open Source Softw.",
    volume = "7",
    number = "75",
    pages = "4205",
    year = "2022"
}