pespace.inference.interface_bilby

Interface for using bilby to perform parameter estimation.

This module provides a bilby-compatible likelihood class enabling the use of various external samplers supported in bilby.

Classes

LikelihoodBilbyInterface(waveform, detector, ...)

Bilby-compatible likelihood class for sampling.
class pespace.inference.interface_bilby.LikelihoodBilbyInterface(waveform, detector, channels)[source]

Bases: Likelihood

Bilby-compatible likelihood class for sampling.

TODO:

  • Add support for multiple detectors with shared or different frequency samples, observation durations, or cadences

  • Implement phase, time, and distance marginalization

Parameters:
__init__(waveform, detector, channels)[source]

Initialize the Bilby likelihood interface.

Parameters:

  • waveform (BaseWaveform or WaveformLALSimulationInterface) – Waveform model instance that provides the waveform_container attribute and update_waveform method for generating gravitational wave signals.

  • detector (InterferometerAntenna or tuple of InterferometerAntenna) – Detector instance(s) for computing detector responses. Each detector must have tdi_data and update_detector_response attributes/methods.

  • channels (tuple of str) – TDI channels to use for likelihood computation. Must be a subset of (‘A’, ‘E’, ‘T’). Common choices are (‘A’, ‘E’, ‘T’) or (‘A’, ‘E’).

Raises:

ValueError – If any channel in channels is not one of ‘A’, ‘E’, or ‘T’.

Notes

The likelihood computation uses the Whittle approximation, which assumes Gaussian and stationary noise.

log_likelihood()[source]

Calculate the log-likelihood value for the current parameters.

This method updates the waveform model with the current parameter values, computes the detector response for each detector, and evaluates the Whittle likelihood across all specified TDI channels.

Returns:

The log-likelihood value summed over all detectors and channels.

Return type:

float