ceffyl

Pronounced /ˈkɛfɨ̞l/ ('keff-ill'), meaning 'horse' in Cymraeg/Welsh 🏴󠁧󠁢󠁷󠁬󠁳󠁿🐎

A software package to rapidly and flexibly analyse pulsar timing array (PTA) data via refiting to pulsar timing free spectra.

This can be done by fitting to a free spectrum of the entire PTA or individual pulsars!

Installation

It is highly recommended that you install enterprise-pulsar first via conda-forge before installing ceffyl

To install via `pip'

conda create -n new_env python=3.10
conda activate <new_env>
pip install ceffyl
pip install --no-deps enterprise-pulsar

To install via Anaconda:

conda create -n <new_env> python enterprise-pulsar
conda activate <new_env>
conda install -c conda-forge ceffyl

Then update to the latest version using github and pip!

This is because we use enterprise as a dependency. Enterprise requires tempo2. Tempo2 is notoriously difficult to install directly... We plan to remove this dependency in a future update.

data

Download representations of PTA data to accurately fit spectral models with ceffyl!

• 🆕 NANOGrav 15-year data set

  • PTA free spectrum refit data

• IPTA DR2

  • PTA free spectrum refit data

• NANOGrav 12.5-year data set

  • PTA free spectrum refit data (zipped)

examples

• PTA free spectrum refit example

  • This is the fastest and most accurate refit technique. Fit any GWB spectrum that you'd like in < 5 minutes!

• GFL Lite refit example

  • Fit GWB models quickly and accurately to different combinations of pulsars!

• GFL refit example

  • Fit GWB and custom intrinsic red noise models to different pulsars quickly! Experimental - use with caution!

Do you have your own free spectrum posteriors that you want to work in ceffyl? Learn about making your own KDE posteriors here

Attribution

@article{lamb2023rapid,
  title={Rapid refitting techniques for Bayesian spectral characterization of the gravitational wave background using pulsar timing arrays},
  author={Lamb, William G and Taylor, Stephen R and van Haasteren, Rutger},
  journal={Physical Review D},
  volume={108},
  number={10},
  pages={103019},
  year={2023},
  publisher={APS}
}